ABSTRACT Title of Dissertation: DUAL-CHAMBERED MEMBRANE BIOREACTOR FOR THE DYNAMIC CO- CULTURE OF DERMAL STRATIFIED TISSUES Javier Navarro Rueda, Doctor of Philosophy, 2019 Dissertation directed by: Fischell Family Distinguished Professor & Department Chair John P. Fisher, Fischell Department of Bioengineering Every year over 11 million patients suffer severe burns worldwide. Facial burn statistics include victims of violence (warfare, acid attacks, scalding) and trauma (flame, electrical, chemical). Skin is the first barrier against external mechanical and biochemical factors, such as burning agents, and is composed of the epidermis, dermis, and hypodermis layers. When burned, skin cannot regulate temperature or fluid transport, or stop bacterial infection. Due to the importance of the skin barrier, natural healing and grafting treatments aim to quickly close the wounds with fast proliferation of fibroblasts and collagen deposition, a process that results in scarring, loss of function, and disfigurement. Tissue engineering has produced epidermis- dermis skin scaffolds for clinical use and in vitro dermal models. Throughout this work we studied 3D printing and bioreactor strategies for the simultaneous physiologic and topographic reconstruction of burnt facial skin tissues. First, we formulated a keratin-based bioink that can be used for 3D printing on a lithography- based 3D printer. Second, we implemented the keratin bioink in the production of Halofuginone-laden face masks for the improvement of contracture, scarring, and aesthetics in severe skin wound healing in an animal model. Due to lack of collagen organization and microstructural development, we introduced a novel dual- chambered (DCB) bioreactor system to study stratified tissues. For this, crosslinking density of the keratin-based hydrogels was used to fine tune the transport properties of membranes for potential use in guided tissue regeneration applications. Then, we assessed the viability of our novel DCB for co-culturing adjacent cell populations with the inclusion of a regulatory keratin membrane. Last, having studied the DCB with flat interfaces, we assessed its viability for perfusing curved interfaces. The integration of both curvature and cell populations allowed to assess the synergistic development of adjacent dermis fibroblasts and hypodermis stem-cell-derived adipocytes and evaluate whether including topography parameters would alter cell viability in the DCB. The strategies developed here elucidate on tissue stratification and aesthetic reconstruction. Furthermore, the keratin-based bioink, the engineered membranes, and the DCBs can be extended to study other stratified or gradient tissues and to fine-tune communication between cell populations in complex 3D constructs. DUAL-CHAMBERED MEMBRANE BIOREACTOR FOR THE DYNAMIC CO- CULTURE OF DERMAL STRATIFIED TISSUES by Javier Navarro Rueda Dissertation submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Doctor of Philosophy 2019 Advisory Committee: Professor John P. Fisher, Chair Professor Don L. DeVoe Associate Professor Helim Aranda-Espinoza Associate Professor Keith E. Herold Assistant Professor Kimberly Stroka ? Copyright by Javier Navarro Rueda 2019 Dedication This work is dedicated to my wife, Ana Catalina, and our daughter, Ana Sofia, whose love and sacrifice have carried us to this point and will guide us forward. ii Acknowledgements I would like to thank my advisor Dr. John Fisher for his support and guidance in the laboratory and the classroom; his dedication to his work, his group, and his family has further inspired me to continue pursuing my path in academia. I would like to thank Dr. Marco Santoro and his family for their guidance, support, and friendship both inside and outside the scope of this work. Next, I thank my undergraduate researchers that stuck throughout all the stages of this work and contributed significantly to its completion, Jay Swayambunathan and Morgan Janes. I thank the members of my committee for their input throughout completion of my proposal and dissertation. I thank all past and present members of the Tissue Engineering and Biomaterials Laboratory, and my fellow class companions who started this journey with me, for listening to my questions and patiently going out of their way to help me at every turn; I am grateful for the bonds of friendship that we have formed. And most importantly, I would like to thank my family as this is the result of their encouragement and love. iii Table of Contents Dedication ..................................................................................................................... ii Acknowledgements ...................................................................................................... iii Table of Contents ......................................................................................................... iv List of Tables ............................................................................................................. viii List of Figures .............................................................................................................. ix Chapter 1: Introduction ................................................................................................. 1 Objectives ................................................................................................................. 3 Chapter 2: Current and Future Perspectives on Skin Tissue Engineering: Key Features of Biomedical Research, Translational Assessment, and Clinical Application ............ 4 2.1. Introduction ........................................................................................................ 4 2.1.1. The premise of tissue engineering............................................................... 4 2.1.2. Normal skin structure and function ............................................................ 5 2.2. Relevant applications of tissue engineering in the development of skin substitutes .................................................................................................................. 7 2.2.1. Use and development of skin models for industry ...................................... 7 2.2.2. Wound healing .......................................................................................... 10 2.3. Advanced tissue engineering approaches to regenerate skin ........................... 13 2.3.1. Scaffolds to guide regeneration ................................................................ 14 2.3.2. Cell therapies ............................................................................................ 18 2.3.3. 3D Printing and biofabrication of skin tissue constructs ......................... 22 2.3.4. Delivery of immunomodulatory cues ........................................................ 32 2.4. Current perspectives and future directions of skin tissue engineering ............ 36 2.4.1. The biomedical research community ........................................................ 37 2.4.2. Regulatory agencies .................................................................................. 39 2.4.3. Clinicians and patients ............................................................................. 41 2.5. Conclusions ...................................................................................................... 45 Chapter 3: Bioinks for Three-Dimensional Printing in Regenerative Medicine ........ 46 3.1. Introduction ...................................................................................................... 46 3.2. Fundamentals of 3D printing ........................................................................... 47 3.2.1. Extrusion based printing ........................................................................... 49 3.2.2. Selective laser sintering ............................................................................ 51 3.2.3. Inkjet bioprinting ...................................................................................... 51 3.2.4. Stereolithography ...................................................................................... 53 3.3. Bioinks ............................................................................................................. 54 3.3.1. Matrix or matrix-mimicking bioinks ......................................................... 58 3.3.2. Sacrificial bioinks ..................................................................................... 84 3.3.3. Supporting bioinks and supporting baths ................................................. 88 3.3.4. Current translation of 3D bioprinting ...................................................... 90 3.4. Conclusion and future directions ..................................................................... 94 Chapter 4: Modeling Skin: Epithelial Barrier Models and Bioreactors ...................... 96 3 4.1. Skin as an epithelial barrier model ................................................................. 96 4.1.1. Introduction............................................................................................... 96 iv 4.1.2. Modeling skin epithelium .......................................................................... 99 4 4.2. Bioreactor systems for skin .......................................................................... 106 4.2.1. Introduction............................................................................................. 106 4.2.2. Complex and multi-chambered bioreactors............................................ 107 4.2.3. Bioreactor systems for the study of skin. ................................................ 109 Chapter 5: Development and Characterization of a 3D Printed, Keratin-Based Hydrogel ................................................................................................................... 112 5.1. Introduction .................................................................................................... 112 5.2. Methods.......................................................................................................... 115 5.2.1. Keratin preparation ................................................................................ 115 5.2.2. Resin formulation .................................................................................... 115 5.2.3. Keratin printing ...................................................................................... 116 5.2.4. Keratin casting ........................................................................................ 116 5.2.5. Resolution assessment ............................................................................. 117 5.2.6. Mechanical testing .................................................................................. 117 5.2.7. Swelling degree and uptake capacity ...................................................... 118 5.2.8. Crosslinking density approximation ....................................................... 119 5.2.9. Cytotoxicity analysis ............................................................................... 119 5.2.10. Live/Dead staining ................................................................................ 120 5.2.11. Microstructure observations ................................................................. 121 5.2.12. Statistical analysis ................................................................................ 121 5.3. Results ............................................................................................................ 121 5.4. Discussion ...................................................................................................... 130 Chapter 6: In Vivo Evaluation of a 3D Printed, Keratin-Based Hydrogels in a Porcine Thermal Burn Model................................................................................................. 137 6.1. Introduction .................................................................................................... 137 6.2. Methods.......................................................................................................... 140 6.2.1. Keratin extraction and bioink preparation ............................................. 140 6.2.2. 3D printing ? continuous Digital Light Processing (cDLP) ................... 141 6.2.3. Rehydration and sol fraction of keratin hydrogels ................................. 141 6.2.4. In vitro collagen gel contracture assay .................................................. 142 6.2.5. In vivo porcine burn model ..................................................................... 143 6.2.6. Wound evaluation ................................................................................... 144 6.2.7. Histology ................................................................................................. 146 6.2.8. Statistics .................................................................................................. 146 6.3. Results and Discussion .................................................................................. 146 6.3.1. Effects of sterilization and rehydration on keratin hydrogels ................ 149 6.3.2. In vitro contracture of keratin-based hydrogels ..................................... 153 6.3.3. In vivo assessments of keratin-based hydrogels ..................................... 156 6.4. Conclusions .................................................................................................... 164 Chapter 7: Development of Keratin-Based Membranes for Potential Use in Skin Repair * ..................................................................................................................... 166 7.1. Introduction .................................................................................................... 166 7.2. Materials and methods ................................................................................... 168 7.2.1. Keratin preparation ................................................................................ 168 7.2.2. Keratin resin and curing ......................................................................... 169 v 7.2.3. Initiator consumption .............................................................................. 171 7.2.4. Sol fraction .............................................................................................. 171 7.2.5. Fourier-transform infrared spectroscopy (FTIR) ................................... 172 7.2.6. Thermogravimetric analysis and differential scanning calorimetry ...... 172 7.2.7. Swelling ................................................................................................... 174 7.2.8. Mechanical properties ............................................................................ 174 7.2.9. Partition coefficient ................................................................................ 175 7.2.10. Permeability .......................................................................................... 176 7.2.11. Transport of adipogenic molecules across keratin membranes ........... 177 7.2.12. Statistics ................................................................................................ 178 7.3. Results ............................................................................................................ 178 7.4. Discussion ...................................................................................................... 186 7.4.1. The effects of energy density (ED) on the crosslinking density (CD) of keratin membranes ............................................................................................ 187 7.4.2. Characterization of membrane properties dependent on the network microstructure ................................................................................................... 190 7.4.3. The effects of crosslink degree (CD) on transport phenomena across keratin membranes ............................................................................................ 195 7.5. Conclusions .................................................................................................... 198 7.6. Supplementary Materials and Methods ......................................................... 199 7.6.1. Rinsing of the leachable components over time ...................................... 199 7.6.2. Effect of lyophilization on swelling properties ....................................... 200 7.7. Supplementary Figures .................................................................................. 201 Chapter 8: Dual-Chambered Membrane Bioreactor for Co-Culture of Stratified Cell Populations ................................................................................................................ 203 8.1. Introduction .................................................................................................... 203 8.2. Methods.......................................................................................................... 206 8.2.1. Bioreactor design and 3D printing ......................................................... 206 8.2.2. Computational modeling ......................................................................... 208 8.2.3. Preparation of keratin membranes ......................................................... 209 8.2.4. Membrane degradation assessment ........................................................ 210 8.2.5. Diffusion in the DCB ............................................................................... 210 8.2.6. Bulk convection in the DCB .................................................................... 212 8.2.7. Dynamic assessment of transport in the DCB (convection and diffusion) ........................................................................................................................... 212 8.2.8. Cell seeding and viability in the DCB..................................................... 213 8.2.9. Adipogenic differentiation in the DCB across the keratin membrane .... 214 8.2.10. Statistics ................................................................................................ 215 8.3. Results and Discussion .................................................................................. 215 8.3.1. CFD modeling of the DCB membrane system ........................................ 216 8.3.2. Assessment of diffusive and convective transport within the DCB ......... 220 8.3.3. Cell cultures, growth, and differentiation within the DCB ..................... 230 8.4. Conclusions .................................................................................................... 235 8.5. Supplementary Figures .................................................................................. 236 Chapter 9: Studies on Complex Topographical and Physiological Reconstruction of Skin in a Dual-Chambered Bioreactor ...................................................................... 238 vi 9.1. Introduction .................................................................................................... 238 9.2. Methods.......................................................................................................... 241 9.2.1. Preparation of keratin membranes ......................................................... 241 9.2.2. Dual-Chambered Bioreactor (DCB) setup and its variation for imaging ........................................................................................................................... 241 9.2.3. Porous scaffolds for the DCB ................................................................. 242 9.2.4. Flat and curved interfaces in the imaging DCB ..................................... 243 9.2.5. Dermal-hypodermal constructs with curved interfaces in the DCB ....... 243 9.2.6. Staining and Imaging .............................................................................. 244 9.2.7. ELISA quantifications ............................................................................. 245 9.2.8. Statistics .................................................................................................. 245 9.3. Results and Discussion .................................................................................. 246 9.3.1. Imaging and tracking flow inside the DCB for curved interfaces .......... 247 9.3.2. Combination of topography and physiology in the DCB ........................ 251 9.4. Conclusions .................................................................................................... 259 Chapter 10: Summary .............................................................................................. 261 Chapter 11: Future Directions .................................................................................. 264 References ................................................................................................................. 267 vii List of Tables Table 2.1. Major structures, cell types, and ECM components present in each layer of normal skin tissue????????????????? 6 Table 2.2. Select listing of tissue engineered skin models developed or used in industry??????????????????????...? 10 Table 3.1. Bioink materials compatible with associated printing techniques?.. 56 Table 5.1. Formulation design of the keratin resin??????????? 122 Table 5.2. Comparison between printed keratin samples and design dimensions???????????????.?????..?. 125 Table 5.3. Compressive modulus and swelling degree of printed and casted keratin???????????????????????... 126 Table 5.4. Dynamic mechanical analysis of printed keratin???????.. 127 Table 6.1. Scoring parameters for wound healing in vivo???????.? 145 Table 6.2. Incidence of histological observation after 30d???????? 157 Table 6.3. Incidence of histological observation after 70d???????? 158 Table 8.1. Qualitative assessment of keratin-based membranes recovered from DCB runs?????????????????????.?. 210 viii List of Figures Figure 2.1. Schematic representation of the major structures and layers of skin tissue that are necessary for normal skin function???????? 6 Figure 2.2. Mesenchymal stem cells (MSCs) participate in many aspects of wound healing. They directly and indirectly promote several cellular functions by: (Clockwise from top) releasing pro-angiogenic cytokines; recruiting macrophages and producing immunomodulatory cytokines; releasing chemokines as well as factors for cell proliferation and remodeling; differentiating into fibroblasts and even skin appendage cells???????????????????????.?? 20 Figure 2.3. Common techniques of 3D printing for skin tissue engineering. a) In electrospinning, the extruded polymer solution is subjected to voltage differences that generate filaments at the micro- or nano-scale, depending on the printing conditions. The fibers can be deposited into a planar surface or woven onto non-planar structures. b) In microextrusion printing, the polymer solution containing cells and other biologics is extruded through a needle and deposited layer-by- layer on the platform. Multiple layers can be assembled by controlling the needle movement. c) Ink-jet printing enables the dropwise deposition of the bioink. Typically, low viscosity solutions are used and the droplets can be generated via localized temperature or pressure variations. d) In Laser-Induced Forward Transfer (LIFT), also called Laser-Assisted Bioprinting (LAB), a focused laser beam is pulsed on top of a donor layer containing the desired bioink formulation. Energy is transferred from the laser and through the energy absorbing layer (typically metal-coated glass) to create localized vapor pockets that dislodge the donor layer bionk in the form of droplets?..????. 29 Figure 2.4. Major considerations for a successful tissue engineered skin product from the perspectives of biomedical research organizations (e.g. academia and industry), regulatory agencies, and the clinic??..? 37 Figure 3.1. Bioink categories: From left to right, a desired final geometry volume can be fabricated with three different bioink approaches to result in the ultimate final print. The print setup shows how different bioinks would be incorporated during the print fabrication?????????... 58 Figure 4.1. The four types of interfacial barriers discussed in this review. (A) The air-tissue interface shows molecular transport (purple diamonds) from air into the tissue space (i.e. skin or lungs). (B) The air-liquid interface shows molecular transport (yellow diamonds) from air into liquid, which is often the vasculature of the body, represented by blue/red ix vessels (i.e. skin). (C) The liquid-tissue interface shows molecular transport (green diamonds) from liquid into the tissue space (i.e. the brain). (D) The liquid-liquid interface shows molecular transport (orange diamonds) from liquid into liquid (i.e. the placenta)???. 98 Figure 4.2. Models of transport through skin. Traditional models for skin transport include (A) liquid-to-liquid interface systems, usually Franz diffusion cells where the donor chamber acts as the apical side of the membrane and the receiver side acts as the basolateral side. Skin models still widely rely on variation of the transwell inserts (B-C), where the exposure of the barrier to air allows assessing the physiological behavior of keratinocytes and the external role of skin. (B) Topical application is modeled by applying the treatment on the air-exposed (apical) side of the barrier, while (C) systemic application is modeled by delivering the treatment via the liquid-exposed (basolateral) side????????????????????????.... 104 Figure 5.1. Printed and casted keratin hydrogels. Representative images of casted and printed keratin hydrogels (5 mm diameter and 5 mm height) using: A) 4, B) 5, and C) 6% wt/vol keratin resin. Printed samples show defined flat surfaces, while casted samples present concave upper surfaces due to the meniscus of the liquid resin in the mold during crosslinking. D) Comparison of height of printed versus casted hydrogels (n=6). Measurements were taken to the bulk of the hydrogel due to issues with the formation of a meniscus on the casted samples. Statistical significance (p<0.05) between fabrication methods is denoted with an *???????????????????.. 124 Figure 5.2. Resolution assessment of printed 4% and 6% keratin. Typical images of the different geometries printed with 4% and 6% keratin resin to assess resolution. (A-D) Cube sample dimension are noted as length x width x height (mm). (E-J) Cylinder sample dimensions are noted as diameter x height (mm). There was low success obtaining squared geometries with right angles using 6% keratin; there was no discernable difference between cubes and cylinders of the same dimensions, thus the lack of larger cube or pyramid samples using this resin. (K-L) Pyramid samples with cube and cylinder steps; the red boxes highlight the steps obtained, indicating proper printing of multiple size features within the same samples. All images show the cross-section (XY plane) of samples under 2.5x magnification?... 125 Figure 5.3. Media uptake capacity and cytotoxicity of printed keratin. A) Uptake capacity of the 4% printed keratin hydrogels. Samples were lyophilized and then rehydrated in PBS or MEM (n=5 each) over a period of 5 d. During this period, the increasing mass of each sample was measured and normalized against the corresponding initial dried x mass to obtain the relative uptake. The main graph shows the detailed uptake during the first hour, the insert graph shows the extended uptake over 5 d. B) Cytotoxic evaluation of the 3D printed keratin hydrogels by direct contact (DC) and conditioned media (CM) tests. L929 cells were cultured accordingly with keratin and a HDPE non- cytotoxic control and compared to those cultured under normal conditions in growth media (Live) and those treated with ethanol (Dead) (n=9, for all study and control groups). No statistical differences were found between Live, Keratin, and HDPE groups, but all three were different to the Dead group (Denoted *, p < 0.05)?. 128 Figure 5.4. Cytotoxicity and microstructural imaging of printed 4% keratin. (A-D) Live/Dead staining of L929 cells on: A) Positive control on TCPS. B) Negative control on TCPS. C) Cells with conditioned media. D) Cells seeded directly onto keratin scaffold. Dotted line represents edge of the scaffold. (E-G) SEM micrographs of the scaffolds. Hydrogels were dialyzed for 48 h with excess PBS, followed by lyophilization, sputter- coating with Au-Pd, and imaged to reveal a porous honeycomb-like structure??????????????????????.? 130 Figure 6.1. Keratin-based photosensible bioink was used to produce standardized scaffolds for in vitro and in vivo assessment of their viability as drug- delivery systems for the treatment of dermal burn wounds. A) 4% wt/vol keratin hydrogels were 3D printed on a lithography-based EnvisionTEC Perfactory 4 DLP printer. B) Crosslinked samples are stable and retain their printed dimensions while stored and rinsed in PBS. C) The 3D printing protocol yields consistent, large-quantity batches of hydrogels with reproducible dimensions for standardized in vitro and in vivo testing. D) Custom 3D printed cases were produced to hold and transport the hydrogels throughout the manufacturing process from laboratory to surgery. The cases successfully supported i) the printed samples that underwent rinsing; ii) freezing at -80 ?C and lyophilization; and iii) sterilization with gamma irradiation. E) Sol fraction at the end of the manufacturing process reveals significant effects of gamma irradiation on mass loss of thinner constructs. The 2- and 3-mm thick samples have less sol fraction before sterilization than 4- and 5-mm thick constructs, but gamma irradiation alters their crosslinked networks and results in higher soluble components at the end-point; statistical significance: * difference between non-sterilized and gamma irradiated sample (mean comparison t-Test, p?0.05). F) Sterilization and rehydration media also have significant effects on the thickness of the end-product hydrogels. All printed scaffolds differ from the designed constructs and lyophilization does not significantly further reduce the thickness, but gamma irradiation does. Rehydration in HH or PBS cannot fully restore the dimensions, resulting in thickness loss of over 2 mm in the most extreme cases. Statistical xi significance: * mean thickness differs from the Designed value (t-Test, p?0.05); ? thickness differs from the Printed value (ANOVA per thickness group, p?0.05); # thickness differs from the ReHyd. PBS (No Ster.) value (ANOVA per thickness group, p?0.05)?????.... 148 Figure 6.2. A) In vitro contracture assay using a cell-laden collagen gel treated with loaded or unloaded keratin hydrogels. The use of 3D printed keratin constructs was generally positive to reduce relative contracture compared to no treatment group; the inclusion of Halofuginone further reduces contracture. Statistical significance: * difference when compared to HDF cells only group, p?0.01; # difference when compared to 3D printed keratin group, p?0.01. B) Combined histomorphologic scores at 30 and 70d post-implantation, expressed as average mean ? SEM, best possible outcome = 10, worst = 0. Statistical significance: + mean histomorphologic score is different at days 30 and 70, p?0.001. C) Representative macroscopic images showing treatment groups at day 0 (burn induction), day 3 (initial treatment applications), day 30 and day 70 (biopsy collection, end points)??????????????????????...? 154 Figure 6.3. Representative histological sections (Masson?s trichrome stain, 20x magnification) of all treatment groups, showing prominent epidermal and dermal changes. At 30 days post-procedure there is marked epidermal hyperplasia and marked collagen degeneration (staining red and/or pale blue by Masson?s) predominantly involving papillary dermis and upper reticular dermis. After 70 days, there is considerably less apparent dermal changes characterized by intensity (blue staining) and uniformity of newly deposited collagen within the dermis but without significant improvement in the epidermis. Scale shown: 2 mm????????????????????????.... 160 Figure 6.4. Representative (low-power, x20) histological sections (Picro Sirius Red stained) of all treatment groups. Collagen content and order is characterized by the intensity and pattern of polarized light induced birefringence of collagen fibers. Intense red-orange-yellowish birefringence indicates mature and organized collagen fibers (control, normal tissue, far left). Green birefringence has generally been correlated to immature, fine and less organized collagen, although color change can be correlated to rotations of the samples and light 1 incidence . Here, color was not considered to assess quality of the collagen but merely to identify it and its general orientation trends. At 30-days post-procedure, similar dermal changes are demonstrated by degree of collagen alterations in all samples; collagen is only present but disorganized in the lower reticular dermis. At 70 days, images show similar dermal changes in all groups, with collagen present xii throughout the dermal layers although disorganized. Scale shown: 1 mm????????????????????????.... 161 Figure 7.1. Photocrosslinking of keratin membranes. A) Chemical mechanism for dityrosine bonding in oxidized keratin: i) keratose contains tyrosine within their peptide chains; ii) tyrosine has a susceptible hydroxyl group that be deprotonated upon interaction with free radicals formed between SPS and riboflavin upon UV exposure; iii) unbalanced terminals bond and recombine with unbalanced terminals in adjacent chains; iv) final chemical equilibrium is reached via keto-enol tautomerism when dityrosine bonds between keratin chains are stable; the crosslinking reaction is terminated once removed from a UV source due to the reducing effect of the hydroquinone photoinhibitor. B) Dityrosine chemistry was used to cast keratin membranes: i) a translucent sheet was covered with a thin layer of resin; ii) the casting mold was then laid over the flat sheet and the two were clamped together. The clamped setup was then crosslinked under UV to seal them together; iii) the sealed molds were filled with keratin resin as needed and exposed to UV; iv) the cured mold can be cut from the base sheet using fine thread. C) Examples of crosslinked circular samples 13mm in diameter with different thicknesses (1, 1.5, or 2 mm). D) Final crosslinked samples 12-1.5 (exposure time 12 min, thickness 1.5 mm) stored in PBS?????????????.. 170 Figure 7.2. The effects of energy density (ED) on the crosslinking density (CD) of keratin membranes. A) The ED parameter was used to produce a wide range of samples by combining thicknesses and exposure times at a 2 fixed UV intensity of 350 mW/dm . Bars that share the same color were produced with different parameter combinations but have the same ED (listed on top on each bar). The consumption of riboflavin (B) and the sol fraction of the hydrogels (C) confirmed that the crosslinking reaction (Figure 7.1A) occurs and is ED-dependent. D) -1 FTIR spectra shows four peaks between 1400 and 1650 cm , indicative of changes in aromatic structures, that increasing for the higher ED samples. E) The area under the FITR peaks for the samples was normalized against the area under the peaks for unreacted keratin; this quantification shows the change in aromatic C=C stretch between samples and its relation to ED. F) TGA profiles for comparison of crosslinked and uncrosslinked keratin, indicating the formation of bonds and changes in the profiles between 150 and 275 ?C. G) DSC thermograms for keratin presents a characteristic peak at 210?C which changes in magnitude in crosslinked samples (blue region). Organized curves can be further compared to show how the peak shifts to 220- 230?C and varies depending on the ED of the sample; low energy samples present a defined peak, which progressively decreases as the energy density increases. H) Quantification of these peaks can be xiii related to changes in the enthalpy of the system, indicative of the formation of bonds and the crosslinking degree of the network dependent on the ED. For all plots, samples that do not share the same letter are significantly different (p<0.05)?????????..? 181 Figure 7.3. Characterization of membrane properties dependent on the network microstructure. A) Swelling profiles of keratin hydrogels in MEM as a function of ED, showing how low ED samples reach their swelling saturation capacity faster than high ED groups yet they also start a process of degradation at earlier timepoints. B) The drop in swelling capacity between days 3 and 28 are indicative of degradation for low ED groups; high ED samples were proven to be highly stable when 3 crosslinked with at least 67.2 mJ/mm . C) Different combinations of thickness and exposure time were used to produce duplicated ED values to assess the viability of the ED parameter as proposed in Equation 7.1. Even if the duplicates follow close trends, statistical differences indicate that higher complexity in the equation is required (statistical significance p<0.01 (+) or p<0.05 (*)). D-E) Mechanical characterization of the hydrogels shows the relation between ED and elastic modulus, ultimate stress, and ultimate strain., Higher ED samples have higher mechanical properties, following non-linear trends that show a saturation profile that further elucidates on the crosslinking limitations and maximum capacities. For all plots, samples that do not share the same letter are significantly different (p?0.05)??????????????????????.? 183 Figure 7.4. The effects of crosslink degree (CD) on transport phenomena across keratin membranes. A) Transport phenomena across the keratin membranes were studied using simplified models of the partition coefficient, permeability, and diffusion transport of adipogenic molecules. B) The partition coefficient of keratin hydrogels in equilibrium state with 10, 150, and 2000 kDa FITCd solutions; ED had greater significant effect on the coefficient for smaller molecules. C) Permeability, on the other hand, is the dynamic interaction of the gel with the solute and was affected by ED for the high MW molecules. D-E) Adipogenic differentiation of hMSCs was significantly affected by the variation of the CD of the membranes used. Imaging and quantification of intracellular lipids determined that membranes with higher CD allow better transport of nutrients and adipogenic molecules after 28 d. For all plots, samples that do not share the same letter are significantly different (p?0.05)?????????????.? 185 Figure 7.S1. Removal of all soluble mass from the crosslinked membranes. Low CD samples present a high loss of mass after the initial, but stabilize after the third or fifth rinse; on the other hand, high CD samples, have a lower mass loss and are stable after the first rinse. Measurement of xiv these mass changes on the microbalance, and the low standard deviation errors, allow us to conclude that no mass changes are occurring once the rinsing protocol is complete, indicative of the fact that all leachable products left after the UV crosslinking reaction have been removed. For all plots, samples that do not share the same letter are significantly different (ANOVA with Tukey?s comparison, p<0.05)??????????????????????..? 201 Figure 7.S2. The effects of lyophilization on swelling of the crosslinked membranes. Lyophilized and non-lyophilized keratin membranes were rehydrated to assess any additional entanglement caused in the freezing and drying processes. A) During rehydration, there are significant differences due to lyophilization at all time points for both 3 3 low (12-1.5, ED 16.8 mJ/mm ) and high (96-1.5, ED 134.4 mJ/mm ) CD membranes. On the other hand, diameter (B) and thickness (C) are recovered over time and are comparable to non-lyophilized samples. Overall, dimensions can be broadly restored with rehydration, but the mass uptake can be irreversibly decreased; this is indicative that swelling assessments, which require a lyophilization step, can be offset due to additional entanglement. ANOVA with Tukey?s comparison was performed on each group independently, samples that do not share the same letter are significantly different within that group as denoted by color (p<0.05). Comparison between lyophilized and non- lyophilized samples was assessed at each time point using two-sample t-test for the mean (p<0.01 (+) or p<0.05 (*))??????..?? 202 Figure 8.1. DCB design and 3D printing. A) Layered microstructure of human skin and our approach to the development of a simplified layered tissue engineering scaffold. B) Approach to the formation of strata and gradients in a DCB. i) When cultured independently, scaffolds have a single, homogeneous cell population; layering the scaffolds in vivo would result in poor integration and uncontrolled gradient interface. ii) The DCB allows two flow lines of media into a common chamber; without a barrier, the two flows would mix producing a gradient. iii) Inclusion of a membrane in the DCB would keep the flow lines separated producing a stratified construct; permeability of the membrane would regulate transport between the two sides. C) i) The DCB is composed of two equal master pieces that can be sealed together; ii) the master piece is composed of an inlet, an outlet, and an open centerpiece, as well as including guide channels for sealing gaskets; iii-iv) when combined, the cross-section reveals a common chamber that provides an interface for communication between the two flow lines. D-E) The master pieces and custom porous scaffolds for cell culturing were 3D printed on a Perfactory 4 DLP printer (EnvisionTEC) using EShell 300 resin. F) The final DCB system assembled has been optimized to have maximum external dimensions xv of 64 x 40 x 23 mm (length, width, height), an effective internal volume of 3.8 ml (total volume as seen in Fig. Civ in green, shades distinguish the two halves that compose it), and inlet and outlet connections to standard 1/8? (3.175 mm) inner diameter tubing?.. 207 Figure 8.2. CFD modeling of the DCB membrane system. A) CAD model for the proposed DCB system with porous scaffolds inserted in the line chambers, detailing the position of the inlets (1 and 2) and the outlets (1 and) and the orientation with respect to gravity; the membrane cavity houses the barrier between the two lines. B) Resulting heat maps of the volume fraction of Donor Fluid (red) and Receiver Fluid (blue), both with the identical properties of water, within the DCB for combinations of membrane porosity (55 or 99%) and inlet flow rate (1, 4, or 10 ml/min). Once stable, the simulations indicate different profiles of mixing at the interface of the two flow lines and the membrane. C) The volume fractions for Donor and Receiver Fluids at every normalized iteration point (proportional to time) were plotted to create mixing profiles as functions of porosity and flow rate; these profiles indicate the point in time, relative to the other cases, were the DCB systems reach concentration equilibrium (complete mixing defined as the point with 50?1% Donor Fluid and 50?1% Receiver Fluid). The inclusion of a membrane in the system dilates the time required for the looped system to reach equilibrium, at all follow rates simulated. D) CFD simulations were used to calculate the maximum shear stress at any surface within the modeled DCB; decreased porosity increases the resistance within the system and causes regions or points with higher shear stress as flow rate increases. These stress values allowed us to select adequate flow rates for subsequent experiments in the DCB that involve cells growing within the chambers??????????????????????... 219 Figure 8.3. Assessment of diffusion in the DCB. A) UV-crosslinked keratin membranes casted in custom 3D printed molds. B) DCB master parts ready for assembly, highlighting the inclusion of the keratin membrane and the silicone sealing gaskets. C) For the assessment of diffusion in the DCB inlets and outlets were clamped shut, and sampling ports (red arrows) were opened on each side and sealed with casted paraffin plugs as needed. This system was used to quantify diffusion of model molecules tartrazine and brilliant blue from a donor chamber to a receiver chamber either without an intermediate membrane (D), with a LCD membrane, or with a HCD membrane (E-F) D) Without a membrane, diffusion quickly equilibrates the common chamber, with no significant difference between the concentration of the donor and the receiver chambers by 10 min, for both molecules tracked (n=4). E- F) The inclusion of the membrane significantly delays the equilibrium of the system; equilibrium for both tartrazine (534 Da) and brilliant xvi blue (793 Da) takes at least 6105 min using a LCD membrane (n=4) and 4770 min using a HCD membrane (n=5). Tartrazine reaches equilibrium close to 0.5 (normalized concentration) between donor and receiver, but brilliant blue, which is 33% larger, is significantly and consistently stable around 0.8. G) The permeability of the system was -4 2 around 2x10 cm /s with no significant difference between the combinations of molecules and membrane CD. H-I) The keratin membranes recovered from the chambers showed differences between the low and high CD groups; LCD membranes were either not recovered or partially damaged, while the HCD samples were in considerable better state with some complete even after 8d. For all plots, statistical significance was determined as p<0.01 (**) or p<0.05 (*)?????????????????????????.. 222 Figure 8.4. Dynamic assessment of convection and diffusion in the DCB system. A) The DCB flow system is composed of a multi-channel rotary pump (4 or 8-channel pump head) in sequence with the parallel chamber inlets, outlets, 3-way connectors to a 1 ml syringe for sampling, and 50 ml reservoirs. B) Flow and mixing profiles in the DCB can be observed by flowing green dye through one of the lines (donor side) and water through the other (receiver side); i) at the beginning the flow dynamics of the system keep the flows separated, then as time progresses ii) donor fluid can be seen moving to the adjacent receiver line. To understand and characterize this movement, this dynamic bioreactor setup was used to quantify the role of convection and coupled convection-diffusion in the transport of dye molecules from the donor to the receiver. C) The change in volume of the loops after 24h running was used to assess bulk convection; at all flow rates tested, and independent of the use of membranes, the volume of the donor and receiver reservoirs did not significantly change (not significantly different from 0% change) indicating that convection in the flow system is symmetrical and stable throughout the pump runs (at least n=4). D) Having studied diffusion and convection separately, concentration of green dye in the donor and receiver were measured thoroughly over 24 h runs to assess both transport phenomena coupled together, at 1, 4, and 10 ml/min. The first TM assessment was of an ideally impermeable barrier (Parafilm ) setup, which indicated that the DCB flow lines remain completely independent from each other, as the concentration in the donor and receiver lines remain unchanged over time (donor not significantly different from 1.0 concentration, and receiver not significantly different from 0.0). For all plots, statistical significance was determined as p<0.01 (**) or p<0.05 (*)?????????..? 225 Figure 8.5. Assessment of coupled convection-diffusion in the DCB system. As a continuation of Figure 8.4, the concentration of green dye in the xvii donor and receiver were measured thoroughly over 24 h runs, studying the cases A) without a membrane, B) with LCD membranes, and C) with HCD membrane. From the concentration profiles it is possible to determine the points in time when the donor and receiver reach equilibrium (at least n=4). For the no-membrane cases, equilibrium was reached within 24h at 4 ml/min (at 12h) and 10 ml/min (8h), but no equilibrium was reached at 1 ml/min. The inclusion of the membrane significantly delays the equilibrium of the system; for the case with LCD membrane equilibrium within 24h only using 10 ml/min (8h), while no equilibrium was reached at 1 or 4 ml/min; last, with HCD membrane no state of equilibrium was reached for any rate within 24h. D) The state in which the membranes were recovered after these dynamic studies was determined by the CD; at all flow rates, the LCD membranes recovered were generally in worse state compared to the HCD samples. For all plots, statistical significance was determined as p<0.01 (**) or p<0.05 (*)??????????????.? 229 Figure 8.6. Cell cultures, growth, and differentiation within the DCB. A) The DCB system was simplified to reduce connections and ports that could lead to contamination; for cells studies that require incubation, the loops consist of the multi-channel rotary pump (4 or 8-channel pump head) in sequence with the parallel chamber inlets, outlets, and 50 ml reservoirs. B) Fibronectin- coated scaffolds were successfully seeded and proved to be manageable under sterile conditions for assembly into the DCB. C) L929 fibroblasts were cultured on both lines A and B for 7 and 28d; imaging of the scaffolds revealed that cells were healthy and growing, filling the concave curvatures of the pores, for at least 28d in the DCB (n=4). D) Fibronectin-coated scaffolds were seeded with hMSCs and grown in the bioreactor using hMSC growth media. After 7d, line A was changed to hMSC adipogenic media, while line B was kept with hMSC growth media for up to 28d; notice the different shades of media in the two lines (red arrows). E) Adipogenic differentiation was possible in the DCB for all cases studied, quantified by the normalized amount of lipids deposited as compared to a 2D non- differentiated population. Without a membrane, line B did differentiate but to a significantly lower degree compared to the direct differentiation of line A, the same trend presented in the LCD case. For the HCD membrane case, both lines were able to deposit similar amounts of intracellular lipids. Comparisons were done using ANOVA and Tukey?s multiple pairwise comparison; samples that do not share the same letter are significantly different (p<0.05)?????????????????????? 232 Figure 8.S1. Shear stress profiles in the DCB. CFD modeling results for the surface shear stress heat maps of the DCB. In general, shear stress is below 0.22 Pa throughout the DCB system, independent of membrane porosity (55 or 99%) or flow rates (1, 4, or 10 ml/min). Maximum (red points) and minimum (blue points) values of shear stress are presented. Maximum values are generally found at the inlets or outlets for the higher 4 and 10 ml/min flow rates; in the case of 1 ml/min, the xviii maximum and minimum values are inside the scaffold. The maximum values were found in in punctual cases and are not indicative of high stress regions in the scaffolds??????????????... 236 Figure 8.S2. Velocity profiles in the DCB. CFD modeling results for the flow velocity trajectories through the DCB. In general, velocity is higher at the inlets and outlets, were the resistance of the system is highest. The porosity of the modeled membrane barrier determines whether the interface of the flow lines touches (99% porosity, modeling no membrane) or remains mainly separated (55% porosity membrane). In all cases, an increase in flow rate (1, 4, or 10 ml/min) caused an increase in velocity in the system. All maximum velocities were found at the outlets, while minimum velocities (0 m/s) were found closer to the inlets or at the point where the inlet opens into the chambers and the resistance drops causing the velocity to decrease?????... 237 Figure 9.1. Curved profiles inside the DCB system. A) Comparison between the dual-chambered bioreactor (DCB, left) and the modification for imaging (imaging DCB, right). B) By reducing the thickness of the side wall, it serves as a clear window to image the common chamber of the DCB; i) this setup can be mounted on the microscope plate while connected to the running pump loop, and ii) the clear EShell material allows fluorescent excitation to track immunostained cells inside the DCB. C) Approach to the formation of curved interfaces in a DCB. The profile of flat porous scaffolds was modified to produce: i) 25% curved scaffolds, which have peaks that reach 25% into the depth of the adjacent chamber; and ii) 75% curved scaffolds, which have peaks that reach 75% into the depth of the adjacent chamber. iii) The inclusion of the keratin-based membrane has been proven to keep the flow lines separated producing stratified constructs, which can now be extended to the production of curved interfaces??????.. 247 Figure 9.2. Imaging and tracking flow inside the DCB for curved interfaces. The imaging DCB variation was used to track FITCd solution through the bioreactor in real-time for 1 h. The bottom line (line B) was perfused with the FITCd while simultaneously perfusing the top line (line A) with PBS. Four different types of scaffolds were assessed. The continuous scaffold, without the regulating membrane, shows mixing in the common chamber due to convective and diffusive transport as discussed in the previous chapter. Flat scaffolds with an intermediate HCD membrane show that the barrier keeps the convective flow of the lines separated for at least 1 h and maintains a flat interface between the two lines. Imaging of the 25% curved scaffold indicates that curved flow profiles can be created between the chambers and convective flow is just as efficiently separated while contouring to the curved surface as time progresses. The 75% curved scaffold similarly xix proves the previous point although a threshold to the curve contouring is probably indicated; here, even as time progresses for at least 1 h, convective flow does not fully contour to the shape of the membrane, leaving a pocket in the peaks of the curves (indicated with white arrows)?..?????????????????????? 250 Figure 9.3. Long-term observations of curved profiles in cell-laden DCBs. A) Cell-laden scaffolds, with the HCD keratin-based membranes (which auto-fluoresce close 500 nm wavelength), were collected after 9 d in running, incubated DCBs. i) Single continuous scaffolds without membranes, ii) flat scaffolds with intermediate membranes, iii) 25% curved scaffolds with intermediate membranes, and iv) 75% curved scaffolds with intermediate membranes were successfully recovered. No case showed signs of scaffold deterioration. Furthermore, the HCD membranes also remain intact after 9 d incubation subjected to cell activity, media, and 37 ?C temperature, an indication that the barrier effect and transport phenomena discussed previously is sustained in the DCB long-term. As in Figure 9.2, the 75% curved scaffolds seemingly exceed a curvature threshold, resulting in the deformation of the membranes into the scaffolds possibly due to the increased resistance and flow in the lines (indicated with white arrows). B) Different cell populations were sustained in the DCBs independent of scaffold curvature. In comparison to an empty EShell scaffold (no cells seeded) (i), proliferation and viability of NHDFs (ii-iv) was considerable, showing large populations and wide formation of intracellular actin structures. The populations of hMSCs induced into adipogenic differentiation (v), although present and viable, were smaller, showing lesser number of cells and intracellular actin structures. This populations are comparable to non-differentiated hMSC populations in terms of size and actin networks (vi-viii). Cell nuclei identification in the scaffolds was limited due to the auto- fluorescence of EShell at similar wavelengths???????.? 253 Figure 9.4. Expression of extracellular proteins in the DCB. The quantities of adipose-specific extracellular proteins adiponectin (lower detection threshold close to 52 pg/ml) and leptin (lower detection threshold close to 34 pg/ml), as well as bFGF (lower detection threshold close to 18 pg/ml) also secreted by adipocytes, were tracked in the media collected from both lines of the DCBs over time, specifically at seeding time (t0), after adipogenic induction at day 1 (D1), after 3 d (D3), and at end-point after 9 d (D9). Adiponectin was significantly expressed by hMSCs at the end-point, indicative of successful adipogenic differentiation in the DCB, and only occurred when separated from NHDF populations by an intermediate membrane, independent of interface curvature. Leptin was generally not expressed in the cases studied (no significant difference to non-seeded EShell in xx all cases and below the detection threshold in most cases); the only observable trend indicates that it is better expressed in NHDF-hMSC co-cultures, even higher without an intermediate membrane, although none of these observations are statistically significantly after 9 d. Last, bFGF is generally not expressed by day 9; quantifiable levels are only present at days 0 and 1 as it is a component supplemented into the NHDF growth media, but these levels decrease and disappear over time (no significant difference to non-seeded EShell in all cases and below the detection threshold in most cases). As with leptin, bFGF levels seemingly tend to increase with time but are still non-significant after 9 d. ANOVA and Tukey?s multiple pairwise comparisons were used to compare line A (NHDF-laden scaffolds) of all bioreactors at the end-point (black letters, groups that do not share the same letter are significantly different, p < 0.05). Comparison of end-point line B of all bioreactors was run independently (red letters, groups that do not share the same letter are significantly different, p < 0.05). Comparison of lines A and B of each bioreactor at each time point was done with two- sample t-test for the mean (significance using p < 0.01 (**) and p < 0.05 (*))???????????????????????. 257 xxi Chapter 1: Introduction Faces are singular. No two are exactly the same. Yet faces often resemble one another, and in this way, they indicate our lineage or whom we come from. - Heather Laine Talley, Saving Face (2014) 2 Every year 2 to 3 million people suffer severe burns in the United States, 3,4 with the world total reaching over 11 million patients . Facial burn statistics include 5 6,7 8 9 victims of violence (warfare, acid attacks, scalding ) and trauma (flame, electrical, chemical). Skin is the first barrier against external mechanical and biochemical agents, and is composed of the epidermis, dermis, and hypodermis layers, each with its distinct composition and function. When burned, skin cannot regulate temperature 4,10 or fluid transport, or stop bacterial infection. To avoid these issues, skin grafting, 4,11,12 the current gold standard, rapidly restores the skin barrier but does not restore skin function or features. Due to the complexity of facial topography and skin physiology, treatments usually result in hyperthropic scarring and lack of elasticity 4,10 and thus in loss of skin function and disfigurement. Facial burns have profound psychological and sociological effects; it is hard to quantify its most complex consequence: individuals trying to cope with disfigurements that alter their identity, 6,10,13 perception of self, and social status. 14?17 18? Layered skin scaffolds are produced for clinical use and dermal models, 27 but no approach restores both dermal function and facial features. To address these 1 limitations, the goal of this work was to develop new strategies to study stratified skin tissue equivalents that provide simultaneous physiological and topographical reconstruction. Throughout the course of this work we developed and studied two viable strategies for the reconstruction of complex facial burn wounds. First, we demonstrate the viability of a three-dimensional (3D) printing approach. We present a novel keratin-based photosensitive bioink that can be crosslinked using ultra violet (UV) light to form hydrogels. We can use this bioink in a lithography-based 3D printer to produce complex geometry face masks. The printed scaffolds were extensively characterized in vitro and in vivo to assess their viability as drug-delivery systems for the treatment of partial thermal burn wounds. Although the strategy can produce 3D structures and is viable for in vivo use, the lack of microstructural cues to the cells results in disorganized tissue that can be improved. Second, to address limitations in our first strategy, we have developed a novel dual-chambered bioreactor (DCB) to study stratified 3D cell populations. Our design allows the co-culture of two stratified layers of cell populations, aimed for the study of communication between dermis and hypodermis in the formation of layers or gradients. A key component of the DCB is the inclusion a regulatory membrane; we have developed and characterized a keratin membrane, engineered based on crosslinking parameters, that can alter the transport profiles of biomolecules between the cell populations in the bioreactor. Furthermore, the DCB-membrane system can be used to produce curved interfaces between the cell populations, allowing the formation of non-planar topographies, and to live-image the cells inside the DCB. 2 The strategies developed here elucidate on tissue stratification and aesthetic reconstruction. Furthermore, the keratin-based bioink, the membrane engineering, and the DCB technologies can be extended to study other stratified or gradient tissues and to fine-tune communication between cell populations in complex 3D constructs. Objectives The overall goal of this work is to develop engineering strategies to produce and study stratified skin tissue constructs that can simultaneously address physiological and topographical reconstruction of severe burn wounds, particularly those to the face. To this end, the main objectives of this work are: 1. To formulate a keratin-based bioink that can be used for 3D printing on a commercially available lithography-based 3D printer using UV light. 2. To use the photosensitive keratin-based bioink in the production of drug- laden face masks for the improvement of contracture, scarring, and aesthetics in severe skin wound healing. 3. To prove that dityrosine bonding chemistry and energy density during crosslinking can be used to fine tune the transport properties of keratin- based membranes for potential use in guided tissue regeneration of skin. 4. To assess the viability of our novel DCB for co-culturing adjacent cell populations with the inclusion of a regulatory keratin membrane. 5. To assess the viability of the DCB and membrane complex for the generation of curved interfaces between adjacent dermal and hypodermal cell cultures. 3 Chapter 2: Current and Future Perspectives on Skin Tissue Engineering: Key Features of Biomedical Research, 1 Translational Assessment, and Clinical Application 2.1. Introduction 2.1.1. The premise of tissue engineering The tissue engineering field was originally established 25 years ago by Langer and Vacanti in the aim of combining engineering design principles with our 28 understanding of biological mechanisms to replace or regenerate damaged tissue. Since then, tissue engineering has been leveraged for a variety of biomedical applications including disease modeling, resource sustainability, novel clinical therapies, and has also facilitated the development of powerful technologies such as 29 gene editing, bioreactor culture, and 3D bioprinting among many others. Central to this field is the principle that successful tissue formation involves the synergistic activity of many cell types, not just the isolated effects of any single population. Furthermore, these cells communicate with each other in a 3D system through both living and non-living components. This overarching theme of combining cells, 3D scaffolds, and environmental signals represents a promising strategy to create tissues for studying or treating diseases, but ? like many other biomedical technologies ? it often faces challenges in translation into clinically 1 Adapted from: Yu, J. R., Navarro, J., Coburn, J. C., Mahadik, B., Molnar, J., Holmes IV, J. H., Nam, J., & Fisher, J. P. (2019). Current and Future Perspectives on Skin Tissue Engineering: Key Features of Biomedical Research, Translational Assessment, and Clinical Application. Advanced healthcare materials, 1801471. (DOI: 10.1002/adhm.201801471). 4 effective therapies. Examples of such challenges include the accurate recapitulation 29 of tissue physiology, scalability to meet clinical needs, and financial cost. These requirements are particularly important for skin tissue engineering, where there has been high demand driven by clinicians as well as the cosmetics and pharmaceutical industries for personalized, functional, and cost-effective skin substitutes. Current strategies for fabricating such constructs will be discussed in detail, accompanied by recent noteworthy examples pertinent to clinical regenerative applications and in vitro skin models. We also include an analysis of current challenges and potential breakthroughs from the perspectives of major contributors to this field. 2.1.2. Normal skin structure and function As the outermost layer of the body, the skin is the first line of defence against external mechanical, biochemical, and environmental factors. Mammalian skin comprises multiple stratified layers, broadly the epidermis, dermis, and subcutaneous 30 fat tissue (often referred to as the hypodermis). The thinnest and most external layer of skin ? the epidermis ? is avascular and composed of multiple layers of keratinocytes. This layer also contains pigment-producing melanocytes as well as 4,30? antigen-presenting Langerhans cells that play a role in the host immune response. 32 The underlying dermis is rich in blood vessels, nerve endings, and various glands, and is comprised mainly of fibroblasts that synthesize type I collagen for the 4,30?34 extracellular matrix (ECM). Lastly, the hypodermis, which may be considered part of the endocrine system, consists mostly of adipose cells that function in energy 30,31 storage and thermoregulation. The hypodermis is often dismissed in skin models as simply a system of fat storage, but it also functions as a complex lipid barrier rich 5 32,33,35?40 in stem cells, hormones, and growth factors. As this tissue layer provides the nerves and blood vessels that permeate into the upper layers, the hypodermis plays a 35,41?43 key role in re-epithelization, wound healing, and angiogenesis. The structure and major components of normal human skin are summarized in Figure 2.1 and Table 2.1 below. Figure 2.1. Schematic representation of the major structures and layers of skin tissue that are necessary for normal skin function. Table 2.1. Major structures, cell types, and ECM components present in each layer of normal skin tissue. a) Major ECM Layer Major structures Major cell types components Stratified squamous Keratin, type IV/VII Keratinocytes, melanocytes, Epidermis keratinized collagen (basement Langerhans cells, Merkel cells epithelium membrane) Blood vessels, Type I collagen, nerves, Fibroblasts, endothelial cells, elastin, mechanoreceptors, Langerhans cells, proteoglycans, type Dermis hair follicles, mechanoreceptor cells, smooth IV/VII collagen sebaceous glands, muscle cells, hair follicle cells (basement sweat glands membrane) Adipocytes, fibroblasts, Blood vessels, Type I collagen, Hypodermis endothelial cells, smooth muscle nerves, hair follicles elastin cells, hair follicle cells a) Besides multicellular structures and ECM, all skin layers also produce growth factors critical for tissue function. These include, among others, epidermal growth factor (EGF), transforming growth factor beta (TGF-?), fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), and insulin-like growth factor 1 (IGF-1). 6 2.2. Relevant applications of tissue engineering in the development of skin substitutes The purpose of tissue engineered skin is to replace or model skin tissue with a construct that mimics native physiological form or function. Such a construct could be used in research and product development to examine the potential effects of various stimuli on skin without using animal models. Alternatively, tissue engineered skin constructs could have potential application as wound dressings or skin substitutes in cases of severe skin injury, where patient survival and clinical outcome are highly dependent on restoring the skin?s normal barrier function in a timely 44?46 manner. 2.2.1. Use and development of skin models for industry Generally speaking, the field of tissue engineering remains in its early stages of development. It relies heavily on academic research advancement and, while startup companies are sprouting up and developing worldwide, successful clinical outcomes have not been consistently achieved and large-scale industrial production is often unattainable. However, the case of skin is unique. Advances in skin tissue engineering and modeling have been chiefly led by large commercial entities in the last several decades, particularly the cosmetics and pharmaceutical industries. Therefore, it is important to highlight their role in the advancement of this field. Skin care products, cosmetics, and other topical agents have been traditionally tested in animal models; publications assessing skin corrosion and irritation (such as the 47,48 Draize rabbit skin irritation test) date back to as early as the 1940?s. These testing methods have since evolved, in large part thanks to investments into developing 7 alternative models to in vivo animal and ex vivo human skin approaches. A critical th turning point occurred in the early 2000s with the introduction of the EU?s 7 Amendment to the Cosmetics Directive. This amendment prohibited animal testing of finished products or cosmetic ingredients, introducing a marketing ban regardless of 47?49 the availability of alternative non-animal tests. Industry was thus forced to find alternatives or develop new methods. The economic muscle behind the cosmetics industry energized efforts in developing living skin equivalents that could recapitulate part or all of the skin?s natural structure. These skin equivalents typically consist of allogeneic skin cell populations that are grown in layers and seeded on scaffolds derived from ECM 50?53 proteins. Patents for this type of approach have been registered since the 1990s, but the technology was expedited into commercialization in large part due to the push from the cosmetics industry to develop suitable alternatives to animal testing. For instance, one of the most prominent manufacturers of tissue engineered skin is a ? wholly owned subsidiary of the cosmetics company L?Oreal. Their Episkin product line and other similar products are examples of reconstructed skin models that are widely available and extensively used as in vitro substitutes for human skin (referenced more than 565 times in scientific literature, as self-reported in L?Oreal?s literature database). These efforts led by industry have resulted in major scientific advances in skin tissue engineering, particularly in the development of skin models, living tissue equivalents, and protocols to assess skin properties. These include, among others, reproducible, in vitro assays using engineered human skin constructs to assess 8 54,55 chromosomal damage from topically applied agents, full-thickness skin 56 equivalents to serve as complex skin models, compromised skin assays to study 57 chemical penetration through wounded skin, and skin models to study the use of 58 LED light for acne therapy. Furthermore, since 2004, the Organization for International Cooperation and Development (OECD) has developed several in vitro methods for testing dermal corrosivity and irritation based on commercially available 59 products. Several US and EU agencies recognize these alternate test methods as a 60 way to reduce animal testing and increase global harmonization. Such efforts have also extended to pre-competitive cooperation between major companies. For example, Proctor & Gamble, L?Or?al, Johnson & Johnson, GlaxoSmithKline, Unilever, and Novartis, among others, have combined their efforts to develop alternative models to animal testing, which have resulted in several joint 47,61?63 publications. Overall, the innovation and R&D departments of major companies have produced substantial advancements in the use and development of skin models, some of which are further highlighted below (Table 2.2). 9 Table 2.2. Select listing of tissue engineered skin models developed or used in industry Industry Skin model used Research application investigator The Procter & In vitro nonanimal, skin-based genotoxicity assay for 54,55 MatTek EpiDerm Gamble Co. cosmetics In vitro nonanimal, skin-based genotoxicity assay for Episkin products cosmetics and other topically applied compounds AGE-modified collagen 3D in vitro model of advanced glycation end (AGE) hydrogels product accumulation in aging skin 64?67 L?Or?al S.A. Primary keratinocyte Effect of resveratrol on keratinocyte proliferation and cultures senescence Reconstructed human skin model to select cosmetics Episkin RHE ingredients on the basis of metabolism, efficacy and/or safety In vitro model to study the effects of topical acitretin SkinEthic RHE for treatment of severe psoriasis Stiefel? GlaxoSmithKline Primary human Human living skin equivalent for identifying proteomic 68,69 (GSK) keratinocyte-derived changes downstream of filaggrin deficiency relevant to living skin equivalent the pathogenesis of atopic eczema Cell-seeded silk/collagen In vitro trilayered skin equivalent (epidermis, dermis, skin equivalent and hypodermis) Franz-type cell chamber Johnson & In vitro model of compromised skin for the study of with skin compromised Johnson chemical penetration through skin 56? by tape-stripping Consumer Inc. 58 Effect of retinol on hyaluronic acid production and skin MatTek EpiDerm moisture for anti-aging skin product development Effect of low-level red light therapy as treatment for MatTek HEE acne Novartis a) SkinEthic RHE and Comparison of topical drug penetration into International b) 70 Organogenesis Apligraf reconstructed human skin equivalents AG a) Acquired by L?Oreal S.A. as an Epskin product; b) Previously known as Graftskin. RHE, reconstructed human epidermis; HEE, human epidermal equivalent. 2.2.2. Wound healing Perhaps an even more obvious application of skin tissue engineering is to augment or develop replacements for skin grafts used to treat patients with serious cutaneous injuries. In the clinical setting, skin grafts may be used to treat extensive 10 tissue defects by restoring normal barrier function while stimulating wound repair responses. However, if normal tissue healing is impaired, or if there are insufficient amounts of healthy donor tissue available, tissue engineered constructs may be 71 necessary. While some products have been shown to reduce morbidity and improve clinical outcomes after injury, no single skin substitute currently on the market has been demonstrated to fully restore normal skin structure and physiological 45,72 function. When the skin is extensively injured, it loses its ability to prevent bacterial 4,10 infection and regulate temperature or fluid transport. The natural response to severe skin injury in adults, involving tissue granulation and re-epithelialization, is characterized by a rapid proliferation of fibroblasts that deposit randomly-oriented collagen fibers to fill the tissue defect, followed by the migration of keratinocytes and 73 contraction of myofibroblasts that restore the barrier. This collection of disorganized tissue results in a fibrotic scar, and is often accompanied by lack a of sensation and elasticity as well as flawed features ? in effect, ?healing? does not restore native skin 4,10,31,74 function, histological structure, or aesthetics. Another point to consider is that other cell types normally present in the skin may be slower to regenerate, or do not grow back at all. For example, even if sebaceous glands are transplanted in skin grafts, normal secretory function typically 71,75 does not occur for months. Similarly, sensory and autonomic nerves present in neighboring areas of healthy skin may ingrow to eventually re-innervate the wound 71 area, but the process is slow and never fully complete. This leads to patches of skin that may experience abnormal sensation or sweat function. Finally ? and perhaps 11 even more importantly to patients ? the loss of melanocytes leads to changes in skin pigmentation, which can be disfiguring and difficult to treat with current cosmetic 76 techniques. 2.2.2.1. Current strategies for acute wounds Cutaneous wounds may be classified as acute or chronic, depending on the etiology. Some of the most common causes of acute skin injury include mechanical 45,76 trauma, burns, or the surgical excision of skin malignancies. The current gold standard for treating such wounds is autologous skin grafting, which ? while able to cover the tissue defect and restore barrier function using the patient?s own skin tissue ? suffers from the same limitations as described above in that the wound site 4,11,12,31 experiences significant contraction and haphazard tissue remodeling. Furthermore, the procedure is restricted by the availability of appropriate harvest sites from the patient, as well as the fact that the donor site becomes another wound requiring management. Studies have also indicated that hypertrophic healing and keloid formation may occur unpredictably, especially among those who already have 31,73 a genetic bias. The availability of autologous skin is also a limitation in cases where a patient?s wounds exceed more than 60% of their total body surface area; in these cases the injuries cannot be adequately covered by autografts due to the lack of 12,77 enough harvestable tissue. Treatment thus requires the use of alternative 45 startegies, most commonly cadaveric allografts. These function mainly as a temporary dressing to protect and stimulate healing in the wound bed before an 45,76 autograft can be placed. 12 2.2.2.2: Current strategies for chronic wounds In contrast to acute skin injuries, chronic wounds develop due to a deviation 78 from the normal wound healing process. Examples include diabetic ulcers, venous leg ulcers, and pressure sores. In each of these cases, an underlying comorbidity prolongs inflammation and delays the closure of an open wound, leading to an increased risk of infection. Difficulty in healing is often further compounded by tissue 78,79 ischemia or continual pressure on the site. Treatments for chronic wounds usually involve addressing the underlying condition, mechanically offloading the affected area, and debriding infected sites. In 79 extreme cases, amputation may be indicated. To try to prevent this, a wide range of clinical products to aid in the rate of wound closure and tissue granulation have been developed, although their use is generally limited due to unproven clinical efficacy, 45,72 high cost, and extensive time required for in vitro cell expansion. Examples of these products include biologic dressings, cultured epithelial autografts, and 45,75 composite skin substitutes. 2.3. Advanced tissue engineering approaches to regenerate skin At the most basic level, a successful tissue engineered skin construct ideally captures the complexities of the native 3D structure and fulfills the functions of natural skin tissue. Furthermore, it should support vascularization and provide supportive cues to cells present in the local environment. Lastly, if implanted in vivo, it must also be capable of integrating into the host with minimal scarring while generating a controlled inflammatory response. 13 In recent years, a diverse variety of strategies have been developed to try to achieve these goals. Many involve the delivery of cells or cues capable of stimulating or participating in tissue repair. The goal of these substitutes may be to directly replace cells previously lost at the defect site, deliver stem or progenitor cells that differentiate into the native tissue type, or stimulate pro-healing behavior by other 46,78 cells already present in situ. 2.3.1. Scaffolds to guide regeneration The reconstruction of skin in tissue engineering has for the most part been focused on the development of stratified constructs mimicking the bilayered structure 18?27 of the epidermis and dermis. Early approaches used synthetic components to minimize fluid loss and mechanical stress while maintaining structural stability at the wound site. Nylon and silicone composites proved popular and could be further coated with biomolecules and skin cells, leading to the emergence of products such as TM ? ? 15,80?82 Biobrane , Transcyte , and Integra . Scaffolds using only natural materials have also gained popularity because they contain protein motifs that facilitate cell adhesion, and demonstrate better compatibility and degradation in vivo, particularly when incorporating biomolecules 75 20,83 84 already naturally part of the skin ECM. Proteins such as collagen, gelatin, 23,85 86?88 89 90 plasma-based fibrin, keratin, chitosan, and dextran have been used both separately or in combination to culture fibroblasts and keratinocytes in efforts to mimic the dermis and epidermis, respectively. In general, these naturally-derived biomaterials are used to produce porous, soft substrates by a variety of methods 85 84 86,91 including self-assembly, , chemical crosslinking, freeze-drying, 14 22,89 22 electrospinning, and knitting. These constructs may also incorporate growth factors and cells of interest (generally fibroblasts, keratinocytes, or stem cells grown in vitro) in order to facilitate native cell ingrowth or the proliferation of seeded cells from autologous or allogeneic sources. Such growth factor- or cell-laden hydrogels 20 83 are widely used to study skin properties such as immunoreactivity, wound closure, 84,85,89 84,86 86,90 epithelialization, angiogenesis, or hair growth. The inclusion of specialized cells and growth factors in scaffolds, as well as their immunomodulatory roles, will be further described in subsequent Sections 2.3.2 and 2.3.4. ECM-based scaffolds are commonly used in vitro for modeling aspects of skin physiology and transport phenomena to take advantage of characteristic properties found in protein-based materials. While these models are helpful for addressing specific properties of native skin tissue, they are generally not comprehensive in that they take a narrow approach towards a singular goal while neglecting the complexity of skin physiology as a whole. For instance, Uchino et al. developed a cell-laden 3D human skin model containing vitrified collagen that supported the culture of dendritic 20 cells in a layered construct. In another recent publication, Sakamoto et al. used a pliable gelatin hydrogel sheet that sustained the release of basic fibroblast growth 84 factor and conformed to the shape of the wound. This construct was shown to accelerate epithelialization, granulation tissue formation, and angiogenesis in mice. In these and other similar publications, there is thorough characterization and careful study of a specific property ? in these cases, formation of either stratified or vascularized tissue ? yet to be successfully translated for clinical wound healing and 15 tissue regeneration, such models must be further developed to study both these factors and more simultaneously. As compartmentalized as these models may be, they have justified the use protein-based scaffolds in clinical trials, which have generally reflected the positive trends observed in vitro. In 2016, for example, Loan et al. published a clinical cohort study on the use of keratin-based scaffolds for superficial and partial thickness burn 88 injuries. When compared to the current clinical standard of care, keratin-based products provided faster re-epithelialization rates, reduced scarring, as well as improved clinical parameters such as reducing healing time, inpatient time, outpatient appointments, and antibiotic use. While many other commercialized clinical ECM constructs, including ? ? ? TM TM ? Dermagraft , Apligraf , Integra , AlloDerm , MatriStem , MatriDerm , ? TM PriMatrix , and PELNAC have been marketed as dermal equivalents or degradable dressings that aid in accelerating wound closure, the cosmetic results still typically 92 remain poor. This perhaps reflects the heavy focus of skin regeneration research on detailed cell behavior and molecular pathways. Translation from a series of cellular functions to the macroscopic processes of scarring and wound contraction is often difficult to achieve. However, as fabrication techniques and biomaterial options continue to expand, skin substitutes that are functional both at the micro- and macroscale may be expected to emerge in the near future. 2.3.1.1. Engineering multilayered tissue ?68,93?95 TM16,24,54 ?16 Products like EpiSkin , EpiDerm MatriDerm , and ?14 Apligraf are bilayered scaffolds designed to mimic the stratified structure of 16 human skin. A typical production process usually includes cultivating fibroblasts inside a hydrogel (generally type I collagen), upon which a layer of keratinocytes is seeded. The constructs are then submerged in growth media until the populations are mature, and the level of media eventually decreased to expose the keratinocytes to an air-liquid interface. This stimulates the cells to proliferate, stratify, and keratinize to 96 form the epidermal layers. Living skin equivalents are used throughout literature as in vitro skin barrier models and have also been used clinically with generally positive 16 14 results in re-epithelialization and wound closure, although, as before, scarred skin is the norm. These engineered scaffolds typically recapitulate only the epidermis and dermis, making them ideal for addressing injuries such as first- and second-degree burns. As such, third- (involving the epidermis, dermis, and hypodermis) and fourth- degree burns (affecting all the layers down to the muscle and bone) are less often considered. While not currently used for clinical applications, a number of trilayered constructs featuring a hypodermis-like layer have been developed for in vitro models 97,98 of human skin. Air-liquid interface culturing is typically utilized to develop these constructs, which are used to study skin tissue properties such as barrier function or 98,99 cell behavior. Some examples include, among others, the use of these trilayered models to investigate the role of adipocytokines in inhibiting fibroblast proliferation 82 97 and scarring, the regenerative potential of adipose-derived stem cells, and the role 56,98 of the hypodermis in drug absorption and metabolism. The fabrication and regeneration of the hypodermis layer has not yet been fully explored and thus carries great potential both in the lab and the clinic for skin 17 growth and regeneration. Past studies have reported that incorporation of other cell types such as mesenchymal stem cells (MSCs) within scaffolds may support these 100,101 endeavors. As will be discussed in the following section, MSCs are highly advantageous owing to their potential for multipotent differentiation, their immunomodulatory effects, and ease of patient isolation and expansion in vitro. 2.3.2. Cell therapies The use of autologous or allogeneic cell populations to aid in replacing or 29 regenerating tissue has long been a common strategy for skin tissue engineering. However, the current methods that strive deliver them while maintaining high tissue complexity often require extensive time and effort to generate the construct, and 45,72 therefore have limited use in point-of-care settings. To address this, one potential alternative to the current clinical products that deliver cells in flat sheets or scaffolds ? is an autologous skin cell suspension spray commercially marketed as RECELL . This product uses non-cultured autologous cells harvested from a patient which are subsequently suspended in solution and sprayed on the wound, allowing the cells to 102 adhere to the target tissue surface. The cell suspension predominantly contains keratinocytes (> 64%), but also includes significant populations of fibroblasts and 103 viable melanocytes. Holmes et al. completed a comparative clinical study of ? RECELL and autologous split-thickness skin grafting in the treatment of acute burns, where the clinical outcomes of the former were as effective as the latter while 104 requiring almost 40 times less donor tissue. Patients in this study reported significantly less pain at the donor sites, perceived greater improvement in wound appearance, and expressed overall higher satisfaction than patients receiving skin 18 ? grafts. It is important to note, however, that RECELL is not a standalone solution for regenerating functional skin and does not directly address the 3D positioning of the transplanted cell types or the multilayered nature of skin tissue. Additionally, ? RECELL can potentially exhibit inter-user variability, as it is a manual point-of-care process, and there is a steep learning curve for physicians. As will be discussed, patients typically have very favorable opinions of innovative skin tissue engineering technologies based on the desire to improve their outcomes. However, there are often differing opinions within the clinical community based on various factors including clinicians? familiarity with the new technologies or amount of experience in other techniques. Besides the active replacement of skin cells or structure using major cell types present in skin tissue, other cellular therapies under development instead aim to use stem cells or cytokines to induce natural tissue regeneration. One of the most popular choices for this strategy are mesenchymal stem cells. MSCs were originally isolated and characterized from mouse bone marrow by Friedenstein et al. in 2970, and 105,106 categorically defined by the International Society for Cell Therapy in 2006. They have been shown to readily differentiate into osteoblasts, chondroblasts, and 105,107 adipocytes when exposed to various stimulating factors. Studies also indicate that MSCs have the potential to differentiate into other cell types outside of the mesodermal germ layer including endothelial cells, keratinocytes, and skin appendage 108?110 cells. These cells are still most commonly derived from adult bone marrow 111?115 (BM), but they can also be isolated from many other tissues in the human body 116 117 118,119 such as adipose tissue, umbilical cord blood, or peripheral blood. While 19 they are perhaps most well-known for their use in cartilage and bone repair therapies, 106 MSCs have also been extensively researched for immunomodulatory applications. The diverse range of activities either induced directly by MSCs or indirectly stimulated by their pro-healing cytokines are summarized in Figure 2.2. Figure 2.2. Mesenchymal stem cells (MSCs) participate in many aspects of wound healing. They directly and indirectly promote several cellular functions by: (Clockwise from top) releasing pro-angiogenic cytokines; recruiting macrophages and producing immunomodulatory cytokines; releasing chemokines as well as factors for cell proliferation and remodeling; differentiating into fibroblasts and even skin appendage cells. In their undifferentiated state, MSCs exhibit immunoprivileged properties and have been previously leveraged for allogeneic implantation in multiple human clinical trials, as unprimed MSCs have a tendency for immune homeostasis and exhibit little 120 immunomodulatory activity unless triggered. On the other hand, MSCs can also be primed or licensed to become either pro- or anti-inflammatory based on their 121,122 microenvironment. These cells produce a wide variety of cytokines and growth factors, many with immunomodulatory functions. This has made them an interesting research topic for adjunct to tissue engineered constructs and skin regeneration cell 20 therapies; rather than replacing the host cells, MSCs used in this way can affect or 110,123?125 facilitate a therapy simply by their presence. Unsurprisingly, the cytokine production, immunomodulatory behavior, and differentiation potential of MSCs have long been investigated for beneficial effects on wound healing. In the mid-2000s, several groups showed that healing in various types of cutaneous wounds (e.g. excisional, burn, radiation damage) could be accelerated 126?128 and improved with application of autologous bone marrow MSCs. These pro- healing effects may even persist over significant periods of time ? studies have shown that, when introduced systemically, exogenous MSCs can localize in damaged areas 129,130 and maintain viability for up to 6 years after implantation in humans. However, the responsiveness of MSCs to immune signaling is mostly localized to their microenvironment, requiring either induction of endogenous MSCs migration, or 131 direct placement of exogenous cells at the site to maximize therapeutic benefits. Many hydrogel and polymer scaffolds have been thus been developed to induce activity and maintain MSC viability to promote the production of angiogenic, 132,133 immunomodulatory, matrix remodeling, or other regenerative cytokines. In skin tissue engineering in particular, MSCs can function to promote wound healing when immobilized in hydrogels placed over the wound site or when added as an 134,135 intermediary layer in split-thickness skin graft procedures. Consequently, the function and benefit of adding MSCs or MSC-conditioned media directly to a tissue engineered construct is an ongoing research topic. Their angiogenic, immunomodulatory, and paracrine signaling functions are also of immense interest, as well as their multipotent differentiation capabilities. However, 21 while promising, the general efficacy of MSC-based therapies is often difficult to determine due to the phenotypic variation of cells that occurs both between donors 136,137 and even within the same individual. This perhaps contributes to the fact that according to data reported by the US National Institutes of Health (http://www.clinicaltrial.gov), there are 624 ongoing or completed clinical trials using MSCs as of November 2018, yet none have so far been able to successfully bring a product to market. Nevertheless, as techniques for assessing MSC phenotype and understanding their capabilities become more advanced, specific and therapeutically active populations of cells may be isolated and used to develop clinically efficacious procedures. 2.3.3. 3D Printing and biofabrication of skin tissue constructs Conventional fabrication techniques such as manual dispensing, molding, freeze drying, and porogen leaching have been used extensively in skin tissue 138?141 engineering for the fabrication of cellular scaffolds. Novel approaches have 21 19 used free-form deposition or modeling on PDMS chips to miniaturize in vitro models. Although easy to implement, they lack the engineering control required to fabricate architecturally complex tissues. 3D printing is an additive manufacturing technique that enables precise layer-by-layer deposition of materials to fabricate complex designs in a highly repeatable manner. Bioprinting refers to the 3D printing 142 of biological materials and cells for the generation of living tissues. Owing to the highly stratified and complex structure of the skin, bioprinting offers unique advantages for developing clinically relevant skin constructs that capture native heterogeneity and architecture. Several reviews have extensively covered the main 22 aspects of 3D bioprinting and its relevance to tissue engineering and skin 143?146 regeneration. Here, we will recap the key aspects as they pertain to skin bioprinting as well as some of the latest advances in this field. 2.3.3.1. The 3D printing process In order to print a physiologically relevant and transplantable skin construct, many design criteria have to be met. For example, it needs to recreate the necessary dermal layers and components, maintain the flexibility and elasticity observed in native skin tissue, and spatially conform to irregularly-shaped wound surfaces with varying dermal layer requirements. Additionally, the graft has to mimic protective 147 barrier functions of the stratum corneum, integrate into the host, and exhibit low 144 immunogenicity. Consequently, regardless of whether the skin substitute is used as an in vitro model for pharmaceutical testing or as a graft for clinical use, choosing the right combination of geometry, compatible bioinks, printing technique, and post- printing maturation are critical. 2.3.3.2. Imaging and 3D model design As a preliminary step towards the printed skin construct, a 3D computer-aided design (CAD) model of the geometry with the desired internal architecture and spatial position of the bioink components is created. Conventional parametric mechanical and product CAD software or a variety of 3D printing-specific design software packages can create the desired geometries. However, parametric programs cannot easily create complex accurate, patient-specific models that are sometimes needed for clinical implantation. Non-invasive imaging techniques such as magnetic resonance 148 149 150 imaging (MRI), computed tomography (CT), ultrasound, and optical 23 146 coherence tomography (OCT) are used to scan patient features and map the 143 architecture to be printed. MRI and ultrasound imaging are more commonly used for soft tissue components due to their ability to distinguish between the various skin 143 layers. Patient imaging can then be integrated with software into a digital 3D 151 reconstruction of the skin tissue. New techniques such as active dynamic thermography (ADT) are constantly being developed and tested for accurate surface 152,153 measurements. These imaging techniques are particularly important when recreating challenging shapes and features such the contours of the face and 154,155 155,156 digits, mapping texture and depth, or accurately determining skin color and 157,158 pigmentation levels. The medical images are subsequently converted to the 3D printable stereolithography (STL) format where specific bioinks can be assigned for each printed layer or section. The efficacy of these techniques relies upon factors such as image resolution, depth penetration, as well as cost in order to be uniformly applicable for patients. However, the accuracy with which multilayered structures can be scanned and translated into printable grafts will undoubtedly improve with time and will play an important role in the development of personalized tissue engineered treatments in the future. 2.3.3.3. Bioinks and Material Selection A number of biomaterials, both natural and synthetic, have been examined 78,143,144,159 and reviewed in literature as potential skin substitutes. Naturally-derived polymers such as collagen, gelatin, alginate, fibrinogen, and chitosan have the 24 advantage of being biodegradable, decorated with functional peptides, and structurally similar to native ECM. Due to the abundance of collagen and proteoglycans in native skin tissue, they are a popular choice for skin grafts. However, their poor mechanical properties and rapid degradation rate limit the long- 144 term stability and applicability of the graft. Electrospinning is a 3D printing technique commonly used for fabricating composite scaffolds with biofunctionality augmented by synthetic polymers. These polymers, including polylactic acid 160 161,162 (PLA), poly(?-caprolactone) (PCL), and poly(lactic-co-glycolic acid) 163,164 (PLGA), exhibit superior mechanical properties and are commonly integrated with the softer, natural components described previously. However, the exact choice of biomaterials can also determine the printing technique being used, as they consequently affect the incorporation of cells within the scaffolds. For example, the harsh solvents used in electrospinning are not conducive to cells or some naturally 165 occurring polymers such as collagen. An alternative approach for bioink material selection explored by Kim et al., involves the use of decellularized porcine skin (S- 166 dECM). A significant advantage offered by dECM bioinks is the retention of matrix proteins critical to cell functionality and an improved cellular response to native cytokines and growth factors. Here, the researchers formulated a printable S- dECM bioink that was laden with fibroblasts for the dermal layer and inkjet-printed keratinocytes for the epidermal layer. Extensive in vitro analysis revealed favorable mechanical and rheological properties of the S-dECM bioink. Notably, minimal construct shrinkage was observed, a problem often associated with collagen. Functional evaluation of the cell-seeded constructs revealed improved cell 25 attachment, higher expression of proteins such as fibronectin, decorin, and type I collagen, as well as thicker dermal and epidermal layers compared to collagen controls. Similar results were observed in vivo where 3D-printed skin patches of S- dECM laden with adipose-derived stem cells and endothelial progenitor cells were grafted on mice for a cutaneous wound healing model. Overall, the cell-laden skin patch promoted neovascularization and re-epithelialization while accelerating wound closure. These results highlight the importance of proper bioink selection for optimal clinical translation. The choice of cells within the bioink also has a significant impact on the functionality of the 3D printed construct. Keratinocytes, which deposit keratin and are a major component of the epidermis, are the most abundant cell type being 144 investigated for most skin tissue constructs. However, a fully functional skin substitute uses additional cell components to develop aspects of native physiology such as ECM deposition, vascularization, pigmentation, and gland formation. Fibroblasts are routinely used for their high proliferative capacity and their broad- 167 spectrum matrix deposition (collagen, elastin, proteoglycans). Cubo et al. presented 3D printed plasma-derived fibrin scaffolds for the standardized production of skin equivalents. Using a custom-modified open source 3D printer (Printrbot), the researchers achieved systematic, layered deposition of human fibroblasts and keratinocytes, obtained from skin biopsies of healthy donors, human plasma, and calcium chloride in a single continuous print. The 3D printed skin construct consisted of a dermis (plasma-derived fibrin loaded with fibroblasts) and an epidermis (keratinocytes) layer that were matured in vitro. The ability to fabricate large skin 26 2 constructs (up to 100 cm ) rapidly using this platform is an important advantage from a clinical perspective. When grafted on to the backs of immunodeficient mice, these scaffolds showed promising results with respect to skin morphology, the various layers characteristic of healthy skin tissue (stratum basale, spinosum, granulosum, and corneum) as well as the presence of keratin and organized collagen fibrils binding the dermis and epidermis. Additionally, neovascularization of the 3D printed skin construct indicated full functional recovery and integration with native tissue. The ability to manufacture stratified skin tissue rapidly that has a large surface area and using human-derived cells has great potential for successful in vivo integration and 23 clinical translation. 157,168 Incorporation of melanocytes (for desired skin pigmentation and color), 169 adipose derived stem cells and adipocytes (for adipose tissue development), and 170 endothelial cells (for vascularization) are currently active areas of research and poised to have a significant impact on the final outcome of tissue engineered skin scaffolds. The promise of advanced skin tissue regeneration to facilitate wound repair or in vitro model development relies on the successful application of such multi- cellular, multi-material bioink. 2.3.3.4. 3D printing techniques A number of 3D printing techniques are currently available for 142 biofabrication. Latest trends include ink-jet deposition and laser-assisted bioprinting, which allow for tight control over microstructures and spatiotemporal deposition of biomaterials and cells ? especially stem cells ? for biomimicry and 27 159 miniaturization studies. Micro-extrusion based printing is more cell-friendly and allows for incorporation of biological molecules, but is limited to a printing resolution 171 of > 100 ?m for cell-based inks. Techniques such as electrospinning, and more 172,173 recently, scaffold-free spheroid-based fabrication have also been explored. The various 3D printing techniques are summarized below (Figure 2.3). The large number of available printing technologies has provided multiple options to optimize layers at the micro- and nano- scales. Choosing the appropriate printing technique for the structure and application can be a key part of conceptualizing a project or product. Xiong et al., for example, reported the extrusion-based 3D printing of gelatin micro- 174 porous scaffolds coated with silk fibroin (SF) and its sulfonated derivative. These composite scaffolds, designed to sequester and concentrate fibroblast growth factor (FGF-2), resulted in pore sizes of 100 to 200 ?m when coated with pure SF and pore sizes of 400 to 500 ?m using sulfonated SF. Scaffold performance was evaluated both in vitro and in vivo in a full-thickness rat skin defect model. FGF-2 incorporated scaffolds exhibited higher rates of cell proliferation, migration, and favorable morphology in vitro, particularly in sulfonated SF coated scaffolds. Similarly, in vivo analysis revealed improved would repair and vascularization after 28 days compared to control groups. High-magnification imaging indicated that surface roughness could also be varied using this method. A higher collagen content with organized collagen fibers as well as significantly thicker re-epithelialization was observed in these scaffolds. Overall, the scaffolds promoted full-thickness skin healing, and the incorporation of FGF-2 enhanced cell proliferation rate, tissue morphology, collagen fibril assembly, and vascularization. 28 Figure 2.3. Common techniques of 3D printing for skin tissue engineering. a) In electrospinning, the extruded polymer solution is subjected to voltage differences that generate filaments at the micro- or nano-scale, depending on the printing conditions. The fibers can be deposited into a planar surface or woven onto non-planar structures. b) In microextrusion printing, the polymer solution containing cells and other biologics is extruded through a needle and deposited layer-by-layer on the platform. Multiple layers can be assembled by controlling the needle movement. c) Ink-jet printing enables the dropwise deposition of the bioink. Typically, low viscosity solutions are used and the droplets can be generated via localized temperature or pressure variations. d) In Laser-Induced Forward Transfer (LIFT), also called Laser- Assisted Bioprinting (LAB), a focused laser beam is pulsed on top of a donor layer containing the desired bioink formulation. Energy is transferred from the laser and through the energy absorbing layer (typically metal-coated glass) to create localized vapor pockets that dislodge the donor layer bionk in the form of droplets. Extrusion printing was also used by Kim et al. to produce artificial skin phantoms mimicking human skin as catalogued by color and tone (Fitzpatrick skin types I?VI) in order to match the corresponding optical and mechanical properties in 175 laser tattoo removal applications. Epidermal-dermal phantoms were printed using gelatin and agar with different concentrations of coffee and TiO2 added to mimic the 29 melanin variations responsible for skin color and tone. Bilayered scaffolds were designed to match the thicknesses of epidermis and dermis, 150 ?m and 1 mm respectively. The resins were successfully extruded into 30 ?m thick layers, producing overall a 138 ?m thick epidermis and 810 ?m thick dermis with optical 175 properties emulating various tones of the human skin. This addresses an important aspect of skin tissue engineering regarding the complete recapitulation of native skin pigments and texture which is discussed further in the clinical section. Nevertheless, would closure and healing remains the priority. Other novel systems for micro-3D printing of skin include the Integrated Composite tissue/organ Building System (ICBS) presented by the Cho group and the 176,177 Laser-assisted BioPrinting (LaBP) system reported by the Chichkov group. These systems are reported to be capable of producing layered epidermal-dermal scaffolds with high spatial resolutions, and also organizing sequential layers of human primary dermal fibroblasts (HDFs) and epidermal keratinocytes (HEKs) with their respective ECM compositions. Using the LaBP technique, researchers fabricated stratified layers of fibroblasts and keratinocytes embedded in a collagen gel to 177 engineer a 10 mm x 10 mm scaffold with a full thickness of 2 mm. The layers were TM printed on a sheet of Matriderm as a proof of concept for post-print clinical translatability. Cell viability, construct structure, and cell junctions were maintained over a period of 10 days in vitro. More importantly, the presence of laminin suggested the potential for the formation of the basal lamina, an important facet of skin tissue. Although recapitulating the dermal and epidermal layers is critical for a successful skin graft, a full-thickness skin model consisting of the hypodermis more closely 30 mimics native skin physiology and functionality. Recent efforts by Kim et al. have attempted to recapitulate native skin architecture by 3D printing perfusable, 178 vascularized skin constructs consisting of all three layers. Building upon their 176 previous transwell platform, they successfully co-printed a PCL transwell chamber with sequential layering of the hypodermis (preadipocytes-embedded adipose-derived dECM-fibrinogen bioink), vascular channels using sacrificial bioinks (endothelial-cell embedded gelatin with thrombin), and the dermis (human dermal fibroblast (HDF)- encapsulated skin-dECM-fibrinogen bioink) to create a vascularized skin construct. The epidermal layer consisting of inkjet-printed keratinocytes was added after 7 days of construct maturation. Histological analysis revealed the presence of distinct layer- specific markers for the hypodermis, dermis, and the epidermis, as well as laminin 178 representative of a basal layer. The vascular channel was capable of supporting the underlying hypodermis while enabling an interface with the dermis. Improved epidermal stratification and higher expression of p63, a skin stemness marker, was also observed compared to a two-layered scaffold lacking a hypodermis or vascularization. Overall, such a full-thickness, vascularized skin construct that closely mimic native skin physiology holds great promise as an in vitro diagnostics platform or for investigation of skin pathologies. Unlike in vitro printing, in situ 3D printing would bring the printer into the operating suite to build a construct directly on the patient. The idea proposes a portable solution that is on-demand for clinicians, while circumventing the time required for tissue maturation in patients with immediate needs. This was first explored by Binder et al. where full-thickness skin defects were first created on the 31 179 dorsa of athymic mice. Human keratinocytes and fibroblasts embedded in collagen/fibrinogen hydrogel precursors were printed layer-by-layer directly on the skin defect using a modified 3D printer after the defect topography was mapped. The study showed promising results with complete closure of the wound by 3 weeks, as well as the formation of an organized dermal collagen layer with a fully formed epidermis. A follow-up of this technology investigated the use of either amniotic fluid? derived stem cells or MSCs as bioink components for immunomodulatory effects. Similar to the previous work, improved wound site recovery and 180 neovascularization was observed. The development of the Biopen handheld surgical device by O?Connell et al., is another example of rapid construct 181 translation. Here, the researchers devised a custom-made tool capable of bioprinting cell-laden gelatin methacrylamide/hyaluronic acid-methacrylate hydrogels and demonstrated that it could dispense adipose stem cells in a manual, direct-write fashion while maintaining high cell viability. Despite the exciting possibilities of in situ printing, creating a wound bed-following, precisely shaped contour may ultimately be less important than developing methods for reducing the maturation time between printing the scaffold and achieving a useable construct. 2.3.4. Delivery of immunomodulatory cues Another element of tissue engineering involves delivering biochemical cues to constructs to stimulate tissue regeneration. This may be promoted by drugs, cytokines, or growth factors, or mediated by the material properties of the scaffold 182,183 itself. In the physiological process of wound healing, a critical factor that dictates the outcome is the host?s immune response. When the skin is injured, the 32 body will attempt to heal the wound by engaging inflammatory and regenerative 73 processes in an ordered sequence. In the first 24-48 hours post-injury, neutrophils infiltrate the tissue and play a critical role in early host defense by clearing necrotic tissue and bacteria from the site. Circulating monocytes then enter the tissue and differentiate into macrophages. These macrophages may further polarize to different phenotypes ? M1 macrophages are pro-inflammatory but necessary for early host defense, while M2 macrophages are anti-inflammatory and stimulate tissue healing 184 via cytokines such as IL-10, TGF-?, and VEGF. In chronic wounds, however, the polarization is predominantly M1, resulting in the secretion of cytokines such as IL- 1? and TNF-? that maintain a state of chronic inflammation and prevent M2- 79,185,186 mediated tissue healing from occurring. Thus, the presence of underlying comorbidities may affect a wound?s ultimate outcome, for instance prolonging inflammation and inducing chronic ulceration instead of closing the wound. Major research efforts have focused on the use and release of signaling ligands or small molecule analogues to modulate the behavior of the immune system 187 locally and over extended periods of time. One such example is sphingosine-1- phosphate (S1P), whose receptors are highly expressed on monocytes and macrophages. This sphingolipid plays a major role in their proliferation, phenotype, 188,189 and migration in both the central nervous system and peripheral blood. When exposed to S1P in vitro, these cells preferentially adopt anti-inflammatory phenotypes 184,189 and display a reduced secretion of pro-inflammatory cytokines if stimulated. Lim, et. al. used S1P in combination with the antifungal agent ciclopirox olamine, 190 which also displays pro-angiogenic activity. The group found that injection of the 33 two agents into a polyvinyl alcohol sponge implanted in diabetic fatty rats supported endothelial migration and the formation of functional vessels. Similarly, fingolimod (also known as FTY720 and Gilenya) is a small molecule drug that acts as an agonist for several S1P receptors and has been previously shown to support endothelial cell 188,191 function and stabilize microvasculature. Past work in a muscle ischemia model using thin films of PLGA to control fingolimod release showed that the drug preferentially recruits anti-inflammatory monocytes and M2 macrophages via stromal cell derived factor-1 alpha- (SDF-1?) mediated chemotaxis and supports local 192 arteriogenesis. As mentioned previously, physiological wound healing is a multi-step process. While these and many other immunomodulatory strategies focus on influencing the cellular effectors of the host immune response and their downstream effects on tissue regeneration, other have aimed to directly combat sources of inflammation and the key biochemical signals that prolong this response. For instance, curcumin, a naturally-occurring polyphenol found in tumeric, has gained some interest as a potential agent for stimulating wound healing. Although its full mechanism of action has not yet been elucidated, it has been previously shown to 187,193 demonstrate some anti-inflammatory as well as anti-microbial potential. Tong, 194 et al developed a cellulose nanocrystal film to release curcumin. The group demonstrated that this system was able to inhibit bacterial growth when applied topically to streptozotocin (STZ)-induced diabetic rats with full-thickness skin defects. Furthermore, the treatment resulted in a significant increase in wound closure rate compared to controls, and regrowth of skin layers as well as glands and hair 34 follicles. Likewise, resveratrol, another natural polyphenol, has also been investigated for its anti-inflammatory and bacteriostatic properties. Berce, et al fabricated chitosan-sodium hyaluronate-resveratrol sponges that were shown to support the 195 formation of granulation tissue with reduced neutrophilic infiltration in mice. Furthermore, the construct displayed a lack of bacterial contamination compared to the control, and supported local angiogenesis and re-epithelialization. Beyond the delivery of factors to stimulate wound healing, another potential strategy may be to instead locally inhibit signals for inflammation and tissue damage at the wound site. Towards this aim, Kasiewicz, et al used lipidoid nanoparticles 196 loaded with siRNA targeting TNF-?. Transfection with these nanoparticles decreased TNF-? production by macrophages as well as MCP-1 by fibroblasts in co- culture. Although this strategy was only tested in an in vitro co-culture model, it represents a promising method for reducing inflammation locally, especially since systemic anti-TNF therapy carries a risk of global immunosuppression and opportunistic infection. Lastly, scaffold material properties may directly influence immune cell response and the resulting effects on tissue regeneration. For instance, Waters et al. investigated the in vitro response of macrophages cultured on oxidized keratin 197 isolated from human hair. The group found that this material induced polarization to an M2-like phenotype as characterized by both surface marker expression and cytokine production. Similarly, Sun described the use of a dextran- isocyanatoethylmethacrylate-ethylamine (DexIEME) hydrogel to stimulate skin 198 regeneration in both porcine and mouse models. This dextran-based, bioabsorbale 35 hydrogel was also shown to promote healing at pre-existing scar sites and promote the formation of hair follicles. The author examined macrophage polarization in response to DexIEME macromers and found that this predominantly led to M2 polarization, suggesting that the material is able to modulate the behavior of macrophages to affect wound outcome. Other scaffold factors such as topographical patterning and surface chemistry may also alter the microenvironment of immune cells to directly influence their 183,199 phenotype. Additionally, peptide motifs may be used to create immunomodulatory scaffolds, as evidenced by the self-assembling hydrogel composed of substance P and other bioactive peptides fabricated by Kim, et al. which was shown to recruit MSCs and facilitate wound closure in a diabetic mouse 200 model. Characterizing the immune profile of chronic wounds and determining how various biochemical and material factors may modulate it to promote wound healing represents a promising direction of research that has yet to be fully explored. 2.4. Current perspectives and future directions of skin tissue engineering As might be expected, the criteria for successfully developing a skin substitute and translating it to patient use varies widely depending on the perspective of the evaluator. Although most would agree that regenerative efforts must focus on achieving wound closure and proper tissue healing at a suitable rate, there are further considerations to be made in order for a tissue engineered construct to be deemed effective and equally accepted by the biomedical research community, regulatory 36 agencies, industry, clinicians, and patients. These will be discussed in detail below and are briefly summarized in Figure 2.4. Figure 2.4. Major considerations for a successful tissue engineered skin product from the perspectives of biomedical research organizations (e.g. academia and industry), regulatory agencies, and the clinic. 2.4.1. The biomedical research community To academic institutions and industrial R&D, the underlying goals of biomedical research are not entirely dissimilar and in fact quite frequently overlap. Both strive to harness fundamental biomolecular mechanisms in order to develop novel technologies that can be translated into clinically effective therapeutics. In the case of skin tissue, this pertains to expanding our current understanding of how 37 various micro-environmental factors affect cell phenotype, ultimately leading to epidermal-dermal stratification and tissue regeneration. As previously discussed, the stratified structure of the skin is critical for its function. Blood vessels and neural networks grow through the interface of the hypodermis and dermis forming the deep vascular plexus; capillaries spread into the layers providing metabolic and gaseous exchange, while oxygen and nutrients will 4,36 only reach the epidermis by diffusion. A lack of healthy, synergistic layer development would result in unviable nerve growth or vascularization which would 30,32 impair the sensing and thermoregulatory functions of skin. Irregular interfaces between the layers could also lead to improper adhesion, fluid collection (blisters or 30 bullae) or separation. Therefore, when developing strategies for regenerating skin in vitro, recapitulation of normal tissue architecture is critical. Techniques for monitoring tissue development and structure almost always involve histological assessment using stains such as hematoxylin/eosin and Masson?s trichrome. Epidermal stratification can also be tracked via expression of layer-specific markers 23,201 such as involcrin, filaggrin, locricrin, and cytokeratins 5, 10, and 14. In recent years, the role of the host immune response has also emerged as a vital topic for consideration. As previously discussed, major efforts have focused on modulating cell behavior in order to facilitate a normal sequence of inflammatory, tissue granulation, and remodeling processes. These cell types include circulating monocytes, macrophages, and T cells ? especially regulatory T cells ? that perform direct effector functions in addition to maintaining a complex milieu of regulatory 202,203 growth factors and cytokines. 38 One further consideration for the industry involves the logistics of scaling up 204 to mass-production. This includes ensuring that consistency can be maintained between lots ? for example, non-autologous components (e.g. ECM-based scaffolds) should exhibit similar mechanical properties between batches. Likewise, the procedures used in manufacturing, processing, and characterizing the products must show repeatable and reliable results. Finally, if autologous cells are to be included in the therapy, consistent methods of harvesting cells from patients at high yields will need to be developed. These cells then need to be expanded with minimal manipulation in vitro, as extensive passaging may cause them to change their phenotypes or senesce to become less effective. To further complicate this process, this entire procedure must be carried out in the minimal amount of time possible so as to expedite patient treatment, and at minimal production costs so that the constructs can be marketable. 2.4.2. Regulatory agencies The timeline for approval of innovative solutions can be quite long, as translation from bench to bedside often requires regulatory approval to determine the safety and efficacy of the therapy, as well as review of quality controls. In the US, therapeutics containing human cells, tissues, or tissue-based products (HCT/Ps) intended for implantation, transplantation, infusion, or transfer to a human recipient are generally regulated by the FDA?s Center for Biologics Evaluation and Research (CBER) (21 CFR 1271.3d). In 1998, the Center proposed regulations, implemented in 2005, that defined different types of HCT/Ps, product processing, and regulatory requirements for the manufacturers of those products (21 CFR 1271). Two parts of 39 the Public Health Service Act govern the regulation of HCT/Ps ? Section 361 aims to prevent the introduction, transmission or spread of infectious disease, whereas Section 351 provides FDA with the authority to regulate biological products. A tiered, risk-based approach governs which regulations apply to a given HCT/Ps. Very briefly, if a HCT/P is deemed to be minimally manipulated, be for homologous use, not combined with another article (with some limited exceptions), and depending on its systemic effect and dependence on metabolic activity of living cells, CBER only regulates the product so as to prevent the introduction, transmission, or spread of 205 infectious disease. Otherwise, the HCT/P is regulated as a drug, device, and/or biological product. These rules and regulations have been the subject of considerable 206 st discussion since their proposal. However, as part of its commitment to the 21 Century Cures Act and recognition of the expanding nature of the field, FDA finalized guidance on the scope of these regulations in 2017to give additional clarity 205,207 to where each HCT/P product falls in the regulatory framework. Many tissue-engineered products, including the skin products described here, must go through regulatory approval processes: e.g. either as a Premarket Approval (PMA), Investigational New Drug (IND), or Biologics License Application (BLA). To help product developers navigate the regulatory system, CBER provides an overview of what is required for a review of their products, including preclinical trial 208,209 design, assessment outcomes, and the progression through clinical trial phases. There are also programs and designations available to expedite the development and review of eligible biological products through Priority Review and Accelerated 210 Approval. Stakeholders are also able to contact CBER early in their product 40 development cycle through the INitial Targeted Engagement for Regulatory Advice 211 on CBER producTs (INTERACT) program before they are ready to submit a more formal meeting request. Early interaction in this way may help ensure that proposed testing yields the necessary information for a premarket submission. 2.4.3. Clinicians and patients The current gold standard treatment for closing severe skin wounds is skin grafting (full- or split- thickness), though for the past 40 years huge efforts have been invested into engineering an alternative solution. From a clinical standpoint, it is desirable to replace ?like with like.? In other words, the ideal scenario would be to treat skin defects with autologous skin or skin-like substitutes. These constructs would ideally include all the native components of skin: the epidermal and dermal layers, the tissue-specific combination of proteins and cells, as well as skin appendages such as hair follicles, sweat glands, and sebaceous glands. Most tissue engineered scaffolds and skin substitutes, however, are still early in research and product development stages. These include 3D printed constructs and scaffolds developed or matured using lab-on-a-chip or bioreactor technologies, none ? of which are readily available yet for clinical use. Products such as Dermagraft and ? Apligraf are cellularized, matrix-cultured products that are wound healing adjuncts but should not be considered true skin scaffolds. Rather, they should be catalogued as ?smart dressings? that are intended to rapidly degrade and induce primary healing by ? TM influencing the local cell population. Other products such as Integra , Alloderm , TM ? ? TM MatriStem , MatriDerm , PriMatrix , and PELNAC are bioengineered or decellularized matrices that incorporate into the wound and provide a template for 41 cellular infiltration. However, based on clinical experience, there is considerable patient-to-patient and product variation in the results and outcomes. These products can be considered a step in the right direction ? they provide a starting place and generally facilitate positive outcomes with proper clinical handling. But, as mentioned, they do not cover the requirements for functional, layered skin and still depend on additional grafting to fully reconstruct complex wounds. Clinical needs point towards a one-step technique, an off-the-shelf product that can recapitulate the original tissue structure (including its microenvironment). As mentioned previously, autografting has the additional complication of adding a second injury at the donor site, which is painful and may sometimes heal with additional deformity. This quandary has generated intense clinical interest in the capabilities of 3D printing of skin substitutes to avoid donor site morbidity: it introduces a potential solution that does not require significant harvest from the patient. Bioprinting techniques can produce constructs in a wide range of sizes and material combinations. In the near future, this method could feasibly be leveraged to 2 achieve perhaps as much as 500 cm of full thickness composite skin. The direct printing of skin on a patient has also been introduced as a potential method, but ? while an alluring idea ? is generally questioned by clinicians due to its steep costs. As might be expected in a burn or trauma center, patients may at times be incapable of making decisions for their treatment, so physicians must in these cases make a judgment on the techniques to be used. Many clinicians tend to avoid complex tissue engineered constructs because they are often quite expensive and not uniformly effective; in fact, the clinical burn community has not fully embraced new 42 technologies for decades. The surgical treatment of burns has changed little since the late 1970s and there have been arguably few major operative burn advancements ? since then. The first was Integra , which entered clinical use as a dermal template but not as the all-in-one skin substitute that clinicians hoped for. Next came Epicel, a cultured epidermal autograft intended for use only in deep, widespread burns that 76 produces widely variable results. The most recently approved burn treatment, ? RECELL , received approval in September 2018. It began clinical trials in 2010 and has since been used on approximately 250 patients in the US, either as part of the 104 clinical trial protocol or under compassionate use. Clinical trials for regenerative medicine therapies often require a long time enroll enough subjects to meet the study requirements or long-term endpoints to ensure safety. Several measures have been taken to expedite the review of these therapies including the FDA?s INTERACT program discussed earlier and the regenerative medicine advanced therapy (RMAT) st 212 designation defined by the 21 Century Cures Act. When able to participate in coordinating their own care, patients in general seem more accepting of tissue engineered solutions than clinicians. This is perhaps due to the straightforward desire of wanting to get better faster, or the perception that newer technology will yield better outcomes. This outlook also extends to participation in clinical trials for tissue engineered products ? the public appears to be generally optimistic of this field, and volunteers are frequently excited to be a part of the developmental process. From experience, many patients have also come forward asking for regenerative medicine solutions, often requesting that clinicians try these strategies after hearing about these strategies from the press or internet. However, 43 there remains a high expectation for success fueled by misinformation, and thus it is important to explain the current capabilities of tissue engineering ? what it is, what is isn?t, and how it works. None of the approaches discussed so far is a panacea for every burn or wound; they each are a tool in a fast-growing toolkit for clinicians. However, there are still a couple of tools missing. Aspects of skin tissue engineering research that could be further addressed include pigmentation, nerve regeneration, and mechanisms of adult human scarring. For example, skin pigmentation is a significant area that remains to be addressed in detail. Grafted skin ? whether or not regenerative technologies are used ? often looks different even if the texture of the skin appears normal without scars. Such pigmentation differences can be very apparent and may be more troublesome cosmetically to some patients than scars. Improving this outcome could make a substantial difference in the quality of life for some patients, especially when abnormal pigmentation might be considered disfiguring or not part of a patient?s self- identity. Broadly speaking, there are no current consistently effective solutions to treat skin hyper- or hypopigmentation. While there are techniques to either bleach or tattoo the affected areas, these are generally incomplete and often generate unsatisfactory cosmetic results. Case series suggest that cell suspension products ? containing viable melanocytes (e.g. RECELL ) are especially promising for 213 addressing dark skin tones, as well as for the treatment of vitiligo. However, there is still a disparity between case series using novel tissue engineered products versus robust, prospective, randomized clinical trials. Only by extensively studying and rigorously testing these products will their true efficacy be revealed. 44 2.5. Conclusions Being able to meet or exceed the quality of current gold standard autologous skin grafts with off-the-shelf, composite, full-thickness constructs represents the ?Holy Grail? of skin tissue engineering. For clinical applications, there is the added requirement of minimizing or altogether eliminating scar formation, as well as the need for broad effectiveness across a wide range of patient populations and wound types. Other qualities to the ideal skin substitute include the integration of functional appendages into these substitutes, as well as the ability match patient-specific pigmentation. The regenerated skin not only must look like native skin but also has to function appropriately; the clinical and physiological properties of the skin layers and structures have to be just right. So, what is the route to this ?Holy Grail?? The potential for accelerated and complete skin regeneration from the field of tissue engineering has greatly expanded over the last few decades, with many novel strategies and viable technologies reaching the product market. However, the current options available still remain limited in a number of ways ? for instance, too often the tradeoff between efficacy and cost is too high for a product to be regularly used. Further efforts to achieve an ideal skin substitute will require continued communication and collaboration among researchers, clinicians, and regulatory bodies to ensure that the final product optimally attains the wide range of objectives discussed here. In this way, skin substitutes will become more widely accepted as a viable solution for reducing the number of animals used for commercial testing, or for improving the quality of life in patients with serious skin injuries. 45 Chapter 3: Bioinks for Three-Dimensional Printing in 2 Regenerative Medicine 3.1. Introduction The advent of three-dimensional printing (3DP) is revolutionizing approaches to fabricating tissue mimics for therapeutic replacement, drug discovery, and fundamental biological understanding. The potential niche for 3DP in tissue engineering is seemingly infinite as we have at hand the ability to provide on-demand fabricated patient-specific designs rapidly and at low-cost. However, 3DP technology was first intended for industrial settings. The translation to tissue engineering applications is hindered by major hurdles that include technical printing issues or, more importantly, biocompatibility. Consequently, we are limited by the number of materials available that can satisfy both the 3DP and compatibility requirements. Here we provide a historical perspective of 3DP and summarize the different techniques; consider the important characteristic properties a bioink must have to fit bioprinting criteria; summarize recent bioink and biomaterial advancement used for 3D bioprinting; and discuss future directions to address current limitations for clinical impact. Generally, printable biomaterials, or bioinks, are materials that can be used in 3DP techniques that include or will include biological features. The term, bioink, may lead to some confusion as some may only consider the material a bioink if it is cell- 2 Adapted from: J Navarro*, G Calderon*, J Miller, JP Fisher: ?Bioinks for Three-Dimensional Printing in Regenerative Medicine? in Principles of Regenerative Medicine, 3rd Edition, Ch. 46, Ed. A Atala, R Lanza, R Nerem and A Mikos. Academic Press, pp. 805-830, 2018. (*equal contributors). 46 laden or contains some matrix or matrix-mimicking component. However, we would like to expand its definition to encompass any printable material that 1) will interface with biological components (e.g. tissues, cells, proteins, growth factors) during or after the actual print, or 2) is involved in the structural construction of scaffolds that will interface with biological components. Bioinks must comply with the 3DP technique as well as provide a biocompatible environment mimicking a desired tissue and ideally degrade controllably without any harmful byproducts. It is important to point out that harmful byproducts may not exclusively originate in degradation but may also come from temporary bioinks that play structural roles during the printing process. Unfortunately, toward satisfying these criteria, material properties work against each other requiring some compromise between desired printability and satisfactory biological features. However, bioengineers do have a selection of materials compatible with several different 3DP platforms. Therefore, in our discussion we will cover bioinks in the context of which printing technique is capable of printing with the described material. We will describe the diversity of available bioinks and biomaterials for 3DP under three main categories: 1) matrix or matrix-mimicking, 2) 214 sacrificial, and 3) support . 3.2. Fundamentals of 3D printing 3DP is an additive manufacturing technique originally applied to plastic and metal manufacturing but has progressed to adapt to biomedical engineering applications within the last few decades. Over three decades ago, Charles Hull 47 patented a new technology for rapid prototyping ? stereolithography. This described system to fabricate three-dimensional (3D) constructs utilizes liquid photopolymerization for building desired objects in a step-wise manner and is 215,216 considered the birth of 3D bioprinting . 3DP and bioprinting have revolutionized the ability to create objects with any shape or size on demand. The exciting promise toward medicine is that we can customize patient specific tissue scaffolds, fabricate on-demand medical devices, and reliably reproduce constructs for high-throughput screening. The key term that defines modern bioprinting is control. Casting approaches have surface resolution defined by the mold itself and no control over the internal structure of the casted sample. Similar salt leaching or electrospinning protocols for porous structures have little to no control over the internal pore size and 217 distribution . Modern technologies allow for control deposition and building; there is control over detailed characteristics such as location and content of material deposition. This includes microstructures to bear stress, cell encapsulation, and growth factor or other biochemical functionalization. Being able to fully design the bulk shape and the internal microstructure has allowed producing structures that more closely resemble nature?s complexity: controlled internal microstructures (pores, 218?223 gradients and layering) , micro channel and micro patterning for pre- 224,225 225? vascularization , and simultaneous deposition of different types of materials 227 219?221,225,228 or different cell lines . The technology allows for control and the mechanics are there to exploit; however, the current hurdle to overcome is defining the bioinks that can be adequately used with these systems. There exists an ever- 48 expanding list of 3D bioprinting technologies (and bioinks to match) that are differently suited for specific desired materials or applications. These techniques vary based not only on materials (e.g. affected by cross linking mechanisms to determine biocompatibility and/or mechanical properties), but also vary over resolution of architecture and speed of fabrication. Four 3DP methods have been adapted for biomedical application: 1) extrusion-based printing, 2) particle fusion-based printing, 3) inkjet printing, and 4) stereolithography or photopolymerization methods. Recent technological advances have resulted in novel methods or modifications of the four previously mentioned, including acoustic droplet ejection, direct-write assembly, 216 laser-guided direct writing, and 3D powder printing . Each of these approaches have advantages and limitations. However, given an identified desired application, often one or a combination of technologies proves to match the sought biomaterial and 216 architecture. A 2016 review by the Kaplan Lab summarized the evolution of bioprinting and additive manufacturing technologies. 3.2.1. Extrusion based printing Extrusion based printing is the most commonly employed method for 3DP, but it requires thermoplastic materials that are only cell compatible if printed at physiological temperatures. The technology forces a viscous ink through a nozzle that can solidify once deposited onto a build platform. The extruded material is deposited to form individual lines in a predefined path as dictated by a generated computer model to form a 3D object in a layer by layer fashion. Typically, each different ink is extruded out of the nozzle at a specific temperature and pressure so that the material can flow through the nozzle (e.g. polycaprolactone (PCL) will extrude at 80 ?C and a 49 220 pneumatic pressure of 400 kPa, whereas alginate can be deposited at 20 ?C ). Materials that are most commonly used for this technique possess sharp solid-to-melt transition such that the material can flow and rapidly solidify once passed through the nozzle. Progress has been made in developing thermoplastics that extrude at lower, more physiological temperatures and pressures by exploiting shear thinning 229 properties (i.e. PCL ). Other approaches include using materials that can be extruded at lower temperatures, do not rely on thermal setting but require additional 230 231 crosslinking mechanisms such as ionic bonding , pH alteration , or ultraviolet (UV) photopolymerization, among others. The inclusion of cells, however, can ultimately affect several aspects of the resulting print. Not only do the materials need to be able to flow through the nozzle at low temperature, but parameters such as high extrusion forces and narrow nozzle diameters, which usually help improve the print?s resolution for synthetics at high temperatures, affect cell viability. Decreasing pressure and/or increasing nozzle diameter may help improve cell viability by reducing cells? experienced shear stress but at the cost of potential nozzle clogging and print resolution. Additionally, this technique has some difficulty printing overhanging structures without supporting filler structures. The printed filaments can sag or even collapse without underlying support. While some groups utilize extrusion printing with PCL or poly(propylene fumarate) (PPF) for creating highly controlled 232?234 porous scaffolds , others generate structures as a sacrificial template with similar extrusion based methods to precisely pattern architectures which can be better suited 224,235 for some of the biocompatibility issues with this 3D printing technique . Overall, 50 extrusion-based printing has been applied to many tissue engineering applications and is most widely available. 3.2.2. Selective laser sintering Another printing technique, particle fusion printing methods such as selective laser sintering (SLS), uses a directed laser to raise the temperature of a powder material beyond their melting temperature to locally fuse or sinter the particles. The laser fuses each patterned layer in a repeated process in order to create 3D 236 structures . The construct?s resolution is only limited by the laser resolution and the powder size. This technique has had its most successful applications to bone tissue engineering by sintering composite materials. Examples include ceramics, PCL, or 237?240 hydroxyapatite . Even though the tissue engineering field has seen promise with this technique, the printed objects are difficult to incorporate with living cells and instead utilize a two-step process of first building the volume followed by the addition of cells. Additionally, there is some difficulty in controlling porosity which 216 can in turn affect the longevity of the biomedical application . Although, a major benefit is the technique?s capacity to be self-supportive of printed structures by the powder bed enabling the printing of complex overhanging geometries. And similar to extrusion based printing approaches, SLS also features groups who utilize the 240 241 technique additively and sacrificially . 3.2.3. Inkjet bioprinting Inkjet bioprinting is considered the cheapest bioprinting technique and was 242,243 patented as one of the first strategies for cell printing . Originally, inkjet 51 bioprinters were created by replacing traditional 2D ink-based printers? cartridges with biological solutions. Instead of paper used as the recipient of the ink, a moveable 244,245 stage added a third dimension to build a construct layer by layer . Inkjet bioprinters can deposit picoliter droplets with positional accuracy <30 ?m, allowing high precision control in positioning different materials and cells into specific 246 microenvironments . To deposit liquid onto a substrate, some inkjet printers electrically heat the print head to a range between 200 and 300?C. The high heat raises concerns for the viability and function of the cells after printing. Still, some studies using this technique show viability for mammalian cells attributing survival to 247,248 the short duration of the high temperature exposure . In inkjet printing a liquid droplet is solidified after deposition on a substrate, a process that must occur quickly to control spatial resolution of the printed volume. Important material properties of the ink are the viscosity and the surface tension to determine final shape, size, resolution, and accuracy of the print. Therefore, an important mechanism to consider for improving the final print is the crosslinkability of the bioink. Methods of effective 249,250 crosslinking for this application are via chemical, pH, or ultraviolet crosslinking . There is a delicate balance of crosslinking the deposited droplets for rapid structural organization and limiting the toxicity introduced to cells. The chemical modification to achieve the desired crosslinking capability can decrease cell viability and affect the chemical and mechanical properties of the material. These changes can alter the 251 biomimicry that was relevant prior to modification . Additionally, inkjet printing uses a nozzle head to deposit the bioinks and therefore is limited by the possibility of clogging. Bioinks for inkjet printing should have low viscosities (below 10 52 centipoise) because the necessary pressures or heat that would eject higher viscosities 252 would negatively affect cell viability . 3.2.4. Stereolithography The last 3D printing technique to be discussed, stereolithography, brings us back to the origins of 3D bioprinting. Stereolithography-based printing utilizes light as a tool to solidify liquid materials in a photochemical reaction. With a laser often used as the light source, a projection of light is shown onto the liquid material in a 253,254 specific pattern to solidify the exposed region . By means of using a light source, photosensitive material, and a controlled axis stage, one can print complex 3D structures. Since stereolithography depends on photosensitive material to print constructs, there are only a limited number of biomaterials that can be utilized. Common materials for this application are poly(ethylene glycol) diacrylate (PEGDA) based materials and gelatin methacrylate (gelMA). The acrylation modification to 255,256 PEG and gelatin render the material photosensitive for printing . Combinations of already self-assembling proteins, such as keratin or decellularized extracellular 257 matrix (ECM), with photo-initiators have allowed crosslinking of soft hydrogels . This method has the advantage of excellent structural integrity because there are no artificial interfaces that result from droplets (inkjet printing) or lines (extrusion printing). Also, the electromagnetic spectrum and its multiple energy wavelengths, allow for a broad range of chemistry alterations; various wavelength lasers, UV sources and the visible light spectrum are common energy sources used in this technique. The use of light, with all available intensities and wavelengths, results in very fast and precise builds. Resolution of this technique ranges from 25-200 ?m for 53 216,257 commercially available printers down to ~10 ?m for two-photon polymerization 258 setups . But because the material must be photosensitive, many biomaterials cannot 257,259 be used, or they must be chemically modified for photopolymerization . A photopolymerizable material like PEGDA is not in itself good for cell viability. Additional surface modifications are necessary to allow for cell attachment and 260 material degradation by incorporating peptide sequences . Additionally, resolution is determined mostly by the laser spot size and therefore has high 3D resolution, but 216 the prints often are warped as mechanical properties are typically weak . Overall, 3D bioprinting techniques vary in approach and can also result in a wide array of medical applications from tissue repair to modeling disease. In the last three decades, there has been tremendous progress in the development of not just the technique but also the biomaterials synthesized to expand the palette of available 3D printable bioinks. 3D bioprinting has exciting translational potential to produce implantable structures for regenerative medicine and high throughput, reproducible drug screening. However, to realize this medical impact, researchers must continue to explore the architecture, the biocompatible yet printable materials, and the inclusion of proliferating and differentiating cells for fabricated living tissues to reach a desirable function. 3.3. Bioinks 3D bioprinting may offer the potential to fabricate physiological tissue mimics; however, progress toward therapeutic application relies heavily on its integration with bioinks. Therefore, the development of biocompatible yet printable 54 bioinks requires tremendous consideration to closely match physical and functional aspects of the desired tissue. Since 3DP technologies originally were designed for nonbiological applications, some of the materials used as inks for printing such as thermoplastic polymers, ceramics, and metals cannot translate to supporting living cells. Hence, one of the greatest challenges of the field is to find materials that are both biocompatible and printable. As defined previously, printable biomaterials, or bioinks, encompass any printable material that 1) will interface with biological components during or after the actual print, or 2) is involved in the structural construction of scaffolds that will interface with biological components. These bioinks are distinguishable from other printable inks in that bioinks are to not produce any toxic element that could be detrimental to living cells or physiological function of the printed tissue. Generally, materials used in the field of regenerative medicine are divided between natural and synthetic materials. Natural materials have all the advantages of being physiological and inherently bioactive. However, natural materials lack in tunability, batch-to-batch consistency, and often physical properties necessary for printing. On the other hand, synthetic materials benefit from a high degree of tailoring to specific physical property needs with inherent consistency to meet the printing technique?s criteria. Still, synthetic materials often fail to match the biocompatibility of natural materials sometimes leading to toxic degradation products or lack of cell- binding sites. Some groups have compromised the divided material set by synthesizing a semi-synthetic class of materials such as gelMA or methacrylated 55 261,262 hyaluronic acid (MeHA) . Of our palette of biomaterials, only a subset is additionally suitable for bioprinting (Table 3.1.). Table 3.1. Bioink materials compatible with associated printing techniques Material Extrusion Stereolithography Inkjet Sintering Fibrin 235,263 264,265 Multiphoton Collagen 235,266,267 265,268 222 crosslinking 219,220,227,230,235,269? 273 Alginate 272 Hyaluronic Acid 274, Thiolated HA + 216 275 Acrylated HA (HA) thiolated gelatin Natural Gelatin 270,274,276,277 278 Keratin 257 Agarose 279 Hydroxyapatite 280 216,238 Carbohydrate 224 glass Gelatin Modified methacrylate 281,282 256,269,283,284 285 natural (gelMA) (Semi- synthetic) Methacrylated 262 285,286 HA (MeHA) 220,232,233,287,288 , hybrid with Polycaprolactone 227,289starch , co- 239,241 272 (PCL) printed with PU , hybrid with 290 hydroxyapatite Poly(glycolic 291, hybrid with 290 acid) (PGA) hydroxyapatite Poly(ethylene Synthetic 292, PEGX hybrid glycol) (PEG) or 216,279,294?296 297 with multiple PEG-diacrylate 293 proteins (PEGDA) Pluronic F127 225,298 Poly(propylene 234 223,299,300 fumarate) (PPF) Poly(vinyl 301,302 216 alcohol) (PVA) 56 For a bioink to be not only biocompatible but also printable, the material must have the capacity to be accurately and precisely deposited with spatial and temporal control. Each bioprinting technique may require a different subset of material properties. For example, inkjet printing requires bioinks to possess low viscosity to avoid nozzle clogging; extrusion printing benefits from shear thinning properties to fluidize through the nozzle and quickly solidify once deposited; selective laser sintering must be able to powderize finely and have an attainable melting temperature; and stereolithography requires photosensitive bioinks. These mentioned material properties to promote printability often come at the cost of compromising biocompatibility. As an example: a photosensitive material that crosslinks in the presence of a photo-initiator that can be highly cytotoxic. Therefore, the available bioinks are chosen to meet the demands of the particular printing process but also for its ability to shield encapsulated cells from a possibly harmful printing process. The requirements for printing depend on a variety of properties, including rheological behavior, the gelation process, or available biological interactions. From a rheology perspective, only specific ranges of viscosities match well with either inkjet or extrusion printing, but shear thinning is an example of a rheological property ideally suited for extrusion. The gelation process, or crosslinking, can greatly influence geometric integrity of the print. Gelation can occur through ionic, thermal, enzymatic, or photo-crosslinking mechanisms and these ultimately dictate which printing technique the bioink is compatible with. Biological interactions might need to be enhanced, especially for synthetic materials, by incorporating cell binding motifs or inclusion of an additional natural material. 57 We categorize the available bioinks into three categories: 1) matrix or matrix- 214 mimicking, 2) sacrificial, and 3) support . As seen in Table 3.1, the same bioink can fall under several bioprinting strategies even if the bioink is utilized in a different way. Each category requires specific workflows but are ultimately brought together in the printing process as illustrated in Figure 3.1. Figure 3.1. Bioink categories: From left to right, a desired final geometry volume can be fabricated with three different bioink approaches to result in the ultimate final print. The print setup shows how different bioinks would be incorporated during the print fabrication. 3.3.1. Matrix or matrix-mimicking bioinks Matrix or matrix-mimicking bioinks are printed and remain as part of the scaffold system. A scaffold can be printed and consist of the matrix bioink material only 58 (acellular scaffold), the matrix material with its surface chemically altered during or after the printing process (functionalized acellular scaffold), or the matrix printed with a loaded cell population (cell laden scaffold). In all cases, the matrix bioink is the material that provides the mechanical structure for cells to adhere to, and that will then be used to enhance cellular communication, proliferation, migration, and differentiation, and ultimately determine the function of the system. As a structural element in a biological environment, there is a delicate balance between achieving the rheological and mechanical properties needed to print a self- supporting structure and the eventual effects that the material may have on the biological development of the cellular component. This balance is specific to the properties and function ultimately desired for the printed sample, anything ranging 277,293,303 from soft and porous hydrogels for in vitro culture and assessment of cells to 223,288,304,305 very strong and durable scaffolds for in vivo bone regeneration . The desired mechanical and biological properties of the matrices are nevertheless restricted by the capacities of the 3DP technologies and the bioinks associated to each method. As it has been reiterated in literature over the years, synthetic materials can be engineered to provide strong scaffolds with tunable mechanical and chemical 288 properties . Nevertheless, these materials have been traditionally associated to low 303 biocompatibility , complex and demanding manufacturing processes (high 246,306 temperatures, high pressures, strong solvents, etc.) , very low degradation 288,289 306 rates , and, in some cases, cytotoxicity or harmful by-products . On the other hand, natural materials inherently provide the adequate biological cues that cells need for proper development. The perfect combinations of amino acid sequences, protein 59 ratios, growth factors, and cytokines are found in natural materials from fauna (ECM combinations, collagen, elastin, fibrin, keratin, hyaluronic acid, chitosan etc.) and 216,303,307,308 flora (alginate, agarose, agar, silk, etc.) . However, natural materials are also associated to weaker mechanical properties and high batch-to-batch 220,258 variability . Natural variability is unavoidable and perfectly defined manufacturing protocols for natural materials are virtually impossible, thus truly tunable properties are difficult to predict and the reason why results in studies using ECM usually differ on a case to case scenario. As it will be discussed in the next sections, synthetic and natural materials cannot be defined by positive and negative characteristics; researchers have been working on improving upon the weaknesses of both either by biochemically altering the individual materials or by implementing synthetic-natural hybrids or combined prints that use the strengths of each. 3.3.1.1. Synthetic materials As mentioned before, 3D bioprinting is an adapted technology; the original 215 patent filed by Hull in 1986 proposed to use the stereolithographic method to optimize prototype manufacturing of plastic parts for industrial applications. Since, the processes have evolved and revolutionized industry and bled into many other manufacturing applications including bioengineering, regenerative medicine, and tissue engineering. The original Hull patent was intended for synthetic materials, named ?UV curable materials?, which could be processed as a ?fluid medium capable 215 of solidification in response to prescribed stimulation? . This definition is technologically viable today and could be applicable to most of modern bioprinting methodologies, even if the solidification is not via UV crosslinking. The materials 60 have greatly evolved and new ones have arisen allowing researchers to incorporate synthetic materials and printing technologies into biomedical and tissue engineering. The greatest strength of synthetics is that the manufacturing processes are well known and can be engineered to specific mechanical and biochemical 303,306 properties . Polymer engineering allows control over molecular weights and distributions, as well as crosslinking densities, which can be tailored to control mechanical properties such as yield stress and strain, ultimate stress and strain, and 306 elastic modulus . This tailoring can occur as part of the polymer synthesis process but can be further modified in the printing or post-printing processes with curing or cross-linking steps. In the end, robust mechanical properties can be used to sustain high loads or to adequately respond to elastic deformation, ideal for structural scaffolding components of biological constructs. The synthesis of tunable mechanical properties also means that synthetic materials can be used in multiple 3DP techniques (Table 3.1) and result in constructs with consistent macro and microscopic definition. The pore distribution in a biological scaffold is an important parameter that will define the presence or absence of vascularization for oxygen, nutrient, and metabolic 305 waste transport in tissue regeneration . The superior print resolution and fidelity of synthetics has been widely explored to produce complex morphologies that may be applied as bio-mimicking scaffolds in regenerative medicine or as structural supports in co-printing applications, a concept that will be further detailed in following sections. In theory, just as the synthetic print resins can be modified to facilitate manufacturing processes, surface modifications can be implemented to allow better 61 interactions with biologic components. However, the manufacturing and modification processes and variables are usually demanding and work in very narrow ranges to achieve specific properties. Often, printing techniques will involve high temperatures, toxic organic solvents or crosslinking agents, rendering them incompatible with living cells and biological materials such as growth factors and proteins that aid cellular 306 function and survival . Synthetic materials in general do not support cell adhesion without additional surface functionalization for adhesion ligands such as arginine- glycine-aspartic acid (RGD), widely identified as binding motifs for proteins such as 257,303,306 fibronectin, osteopontin, and fibrinogen . Even with consistent morphology control and compatibility with surface modification strategies, synthetic polymers do not innately mimic ECM and that remains as its weakest characteristic for clinical 216 translation of synthetic bioprinted scaffolds . Still, synthetic materials that are commonly used in bioprinting applications include poly(ethylene glycol) (PEG) and poly(ethylene glycol) diacrylate (PEGDA), poly(?-caprolactone) (PCL), poly(D,L- lactic-coglycolic acid) (PLGA), poly(L-Lactic acid) (PLA), and poly(propylene fumarate) (PPF), among others. Poly(ethylene glycol) (PEG) has been long used as a coating on medical 306 devices to control host-immune responses or alter degradation rates in vivo . Furthermore, PEG is commercially available in many physical (linear, branched, molecular weight variation) and chemical variants (diacrylated variant PEGDA), and is FDA-approved as a biocompatible material, making it a versatile polymer for 293 bioengineering . Here, biocompatibility, a broad term, means the material does not kill cells or induce an aggressive immune reaction, but it does not necessarily mean 62 that it induces cell adhesion or proliferation. As other synthetic materials, PEG lacks attachment sites that cells need to adhere to a substrate. PEG requires chemical immobilization of binding motifs in order to support cell adherence and stem cell 216,306 differentiation . These characteristics usually result in PEG being used as a secondary plasticizer component in bioinks: even more so, PEG is often modified with acrylate groups to create photopolymerizable PEGDA, a variation that is commonly used with extrusion or stereolithography approaches, and can be easily 216,293 coupled with natural biomolecules for cell-laden bioinks . Poly(?-caprolactone) (PCL) is synthesized by ring-opening polymerization of 289 ?-caprolactone . PCL is a high molecular weight semi-crystalline polymer that has good solubility, a low melting point, thermoplastic behavior, and an extended hydrolysis-induced degradation profile in vivo. This polymer is stable in the body for 309 over 6 months , and then exhibits non-enzymatic, hydrolysis degradation of 2 to 4 years (depending on molecular weight) at physiological pH and temperature, without 288,289,308 leaving any cytotoxic byproducts . Extended degradation is ideal for providing long-term load bearing support during healing and regeneration 220 processes . PCL has a melting point close to 60?C, a relatively low temperature in 289,306,308 manufacturing industry, which allows its easy processing . Upon heating, PCL has viscoelastic properties ideal for extrusion-printing of constructs with elastic mechanical behavior, a characteristic that improves the brittle properties of polymers 309 like PLA and polyurethanes . The Hutmacher group printed cylindrical scaffolds of medical-grade PCL?tricalcium phosphate (mPCL-TCP) by fused deposition modeling that required added growth factors to increase osteogenic potential of seeded 63 288 mesenchymal stem cells (MSCs) . The growth factor coated PCL scaffolds successfully completed up to 12 months of unrestricted load bearing in vivo within 288 large tibial defects in sheep . The use of additional osteogenic growth factors is indicative of the biochemical limitations of PCL; other than hydrophobic non-specific binding of cells, PCL lacks binding motifs that provide specific binding sites for 306 cells . The combination with natural materials or other functionalized materials is the usual approach to address this limitation. The Atala group, for example, concurrently prints PCL with hydrogels based on gelatin, hyaluronic acid, or fibrin; 306 the hydrogels provide the biochemical cues for cellular adhesion and viability . Starch, a natural polysaccharide, has also been widely used to improve 289 biocompatibility of PCL . PCL-starch prints can enhance and stimulate osteoblast proliferation for bone regeneration, support hippocampal neurons and glial cells to treat spinal cord injury, or support bovine articular chondrocyte adhesion and 289 proliferation, and glycosaminoglycans for cartilage tissue engineering . Poly(lactic acid) (PLA) is a well-known established aliphatic polymer used in 308,309 temperature-based extrusion methods . It has a melting temperature close to 175 309 ?C, so it can be extruded in melt-based systems between 200 and 230 ?C . However, PLA glass transition occurs around 60?C and easily interacts with many plasticizers and solvents to change viscosity, a characteristic that allows printing at lower 308 temperatures . The resulting mechanical properties are usually high, with elastic 308,309 modulus around 1.5?2.7 GPa, but tends to be brittle . PLA is commonly used in orthopedic implants and drug delivery systems thanks to its biocompatibility and 308,309 biodegradability . Nevertheless, its degradation is via hydrolysis of ester bonds 64 which releases acidic byproducts; in vivo, this may cause the localized decrease of pH 309 through the release of lactic acid, inflammation and cell death . Poly(lactic-coglycolic acid) (PLGA) is the copolymer of lactide and glycolide, obtained via ring-opening polymerization, synthesized to address individual 308 limitations and uncontrolled degradation of PLA and poly(glycolic acid) (PGA) . Popular polymerization of d- and l-configurations of lactide yield the poly(D,L-lactic- coglycolic acid) variation, which is frequently used due to its improved toughness and 308 easy manipulation of hydrolysis-driven degradation rates . Shah group synthesized ?hyperelastic bone? (HB), a particle?laden 3D bioink that combines hydroxyapatite, a 290 highly bioactive ceramic, and either PCL or PLGA . The extrusion-printed structures exhibited mechanical and physical properties that allow further manipulation (sheets that can be rolled, folded, or cut). The hybrid with PCL showed highly elastic properties, capable of reaching a 61.2 ? 6.4% strain and a tensile elastic modulus of 10.3 ? 1.3 MPa, a behavior superior to the PLGA combination (36.1 ? 4.3% strain, 4.3 ? 0.4 MPa elastic modulus). In terms of cell interaction, the PLGA combination showed better results; both hydroxyapatite-PLGA and hydroxyapatite- PCL scaffolds supported human MSC adhesion and proliferation, and induce osteogenic differentiation in the absence of engineered growth factors after 28 290 days . Poly(propylene fumarate) (PPF) is a biodegradable polymer, broadly applied in tissue engineering due to its ability to form crosslinked networks through its 223,234 234 223,300 carbon?carbon double bond . The Mikos and Fisher groups have extensively studied PPF and its crosslinking capabilities for 3DP and tissue 65 engineering. Because it is biocompatible and can be photo-crosslinked, PPF is a 234,300 prime candidate for 3DP via stereolithography , but can also be as a viscous 234 bioink for extrusion and cured using an UV source . In the first case, the printing process is driven by the intensity of the light source and the proportions of photo- initiator and photo-inhibitor in the bioink, but the resulting mechanical properties of the constructs heavily depend on the amount of printing or post-printing exposure to 223,300 UV which determines polymer crosslinking density . In the case of extrusion, PPF resins exhibit shear-thinning behavior and concentration of PPF drives the viscosity level; other factors such as fiber spacing during deposition and pressure affect pore size and fiber diameter, respectively, but interplay among the factors can 234 also alter scaffold architecture . Melchiorri et al. report that human umbilical vein endothelial cells (HUVECs) and human umbilical vein smooth muscle cells were 300 seeded on stereolithography-printed PPF surfaces and proliferate in a 7-day study . Similarly, MSCs have been cultured of PPF scaffolds over 7 days and exhibit levels of metabolic activity that are not statistically different from cells cultured on standard 223 tissue culture polystyrene . In vivo, using 3DP PPF grafts to treat a coarctation of the aorta in mice for 6 months, printed PPF experienced a 40.76 ? 8.37% decrease in mass and full endothelialization of the inner lumen on the grafts was observed even 300 without preceding cell-seeding or surface modifications . 3.3.1.2. Natural materials Based on the definition earlier proposed, bioinks will either interact with biological components (e.g. tissues, cells, proteins, growth factors) during or after the actual print, or serve as structural components during the printing of scaffolds that 66 will interface with biological components. Without considering the specifics of in vivo or in vitro applications, there is an imminent interaction between cells and tissues with the bioink or its by-products. Natural materials are taken from animal or plant sources; these are materials that naturally developed to sustain cellular life cycles, nutrient and waste transport, and healing processes. They are composed of the perfect combinations of amino acid sequences, protein ratios, growth factors, and cytokines, thus intrinsically providing safe and nurturing interactions with cells. Composition provides the proper biochemical environment for cells to adhere or feel attracted to, subsequently allowing individual cells the healthy completion of the cell cycle and then induce cellular proliferation, migration, and differentiation. This is the basic definition of cytocompatibility, and natural materials provide a high intrinsic level of 216,303,306 it . Just as important, the composition and biochemistry of a natural material is designed to be degraded by physiologically-viable processes, through natural enzymatic and chemical processes, and to be discarded by natural metabolic activity, 303 leaving no significantly harmful by-products behind . Bioinks from these materials can be further biochemically enhanced by encapsulating tissue-specific growth factors, genes, and other controlled-release chemical-regulation factors. Similarly, the surfaces of printed hydrogels can be functionalized by adding the same biochemical factors with both approaches aiming to recreate environments more like those of in 290,303,306,307 vivo tissues . The balance between cytocompatibility and degradation means that these materials naturally go through the proper cycles and rates needed to 216 induce healthy integration with host tissue . 67 Natural materials can be used as isolated, purified proteins (e.g. collagen, fibrin, keratin, or elastin,) or as the natural protein combinations already present in the ECM, combinations that are specific to each type of tissue and determine the type of cells present, the bulk mechanical properties, and its function. The methods to obtain and alter natural materials in laboratory mainly consist of enzymatic cleaving, ionic 216 interactions, or temperature and pH variations . These methods are used to crosslink the bioinks via multiple 3DP techniques (Table 3.1) and produce hydrogels with, theoretically, fine-tuned biochemical and mechanical properties. As it will be further discussed (see Section 3.3.1.4. Cell-laden bioinks), encapsulated cells or those that later migrate into the printed scaffolds are greatly impacted by the mechanical cues imparted by the surrounding material. Cellular adhesion, morphology, migration, and specially differentiation, has been widely proven to be affected by the stiffness of the 282,306 substrate . In general, 3DP natural materials result in weak hydrogels difficult to manipulate into specific ranges of physical properties, limited by the inefficient or 306 low-energy crosslinks achieved by the traditional methods mentioned . Weak mechanical properties have been used as an advantage when attempting to model or regenerate soft tissues or substrate for cell culture, but it is a severe disadvantage when the applications relate to load-bearing hard (e.g. bone, cartilage) or elastic tissue (e.g. muscle, skin, vascular and gastrointestinal tissues, ligaments, tendons). This major weakness has led to combinations with strong and elastic synthetic materials, either by co-printing or as hybrids, as will be detailed later (see Section 3.3.1.3. Co- printing and hybrid bioinks). 68 The intrinsic biocompatibility of natural materials is the main reason that these have been used to formulate bioinks to use in a wide range of in vitro and in vivo bioengineering and regenerative medicine applications. The most popular materials are generally proteins from mammalian origin like collagen, gelatin and gelMA, fibrin, hyaluronic acid (HA), elastin, and keratin; similarly, popular polysaccharides from plant sources include alginate, starch, agarose, and silk, among others. Collagen, particularly type I collagen, is the most abundant ECM protein in 222,306,310 tissues . The most common types of collagen are the fibril-forming collagens (e.g. type I, II, III, V, and XI), and they are the main component in the ECM of 222,310 tissues such as bone, tendon, ligament, skin, muscle, or cornea . Because of its natural abundance in all types of tissues, it has variations that interact with most types of cells and perform a wide range of mechanical roles in either soft, elastic, or hard tissues. For this, collagen is arguably the material that most researchers have tried to adapt for bioengineering applications and has been reported as the most used for cell 306 235,267 and tissue culture . It has been used in 3DP techniques such as extrusion , 222 263 stereolithography , and inkjet , mostly using variations of pH-triggered or 310 303 temperature-triggered gelation that range from hours to minutes . Crosslinking of collagen by pH alteration using solubilized sodium bicarbonate (NaHCO3) solution has been used to construct multi-layered cell?hydrogel composites. This method provided a novel approach to printing both fibroblasts and keratinocytes in a single 228 experiment to model dermal/epidermal-like distinctive layers in a 3D hydrogel . Natural material hydrogels have typically been reported to have subpar printing 69 resolution when compared to synthetics. Nevertheless, Bell et al. report printing line widths of ?1 ?m using multiphoton crosslinking of type I collagen with a flavin mononucleotide photosensitizer, which confers structural control at a microscale 222 level . Collagen has also been widely used in combinations with other natural and synthetic materials, principally bringing the strong biocompatibility aspect to the 226,303,311 mixture . Gelatin is the denatured form of collagen that has undergone partial 222,306 hydrolysis . As collagen, gelatin is characterized by its wide availability, biocompatibility, predictable enzymatic degradation, non-toxic by-products, and 282 inherent cell binding motifs . It has been involved in engineering of soft and hard tissues, ranging from liver to bone, either by itself or as part as hybrids such as 307,312 gelatin-alginate, gelatin?fibrin, or gelatin?HA . Gelatin is widely regarded as the easiest protein to print, mainly because of thermally responsive behavior that allows 276,306 extrusion at temperatures below 20?C and hydrophobic crosslinking . However, the melting temperature of gelatin (30-35 ?C) is below physiological temperature, 307 severely limiting its clinical application in vivo . Even with high resolution obtained 282 by extrusion-based printing , gelatin hydrogels are usually soft and limited by temperature, thus requiring further crosslinking either by post-printing approaches 307 306 (e.g. using glutaraldehyde or thrombin ) or by the addition of functional groups . Like the acrylate modification on PEG that produces the versatile PEGDA, methacrylamide photo-initiator groups can be used on gelatin to obtain gelatin methacrylate (gelMA) to produce a photo-crosslinkable resin. This modification enables irreversible cross-linking, generally by UV irradiation, that preserves printed 70 282 architectures under physiological conditions . UV exposure time and gelMA concentration regulate printability, while the degree of methacrylation determine the mechanical properties and additional acetylation can be used to further influence 282,308 rheological properties of the bioink . Fibrin is a glycoprotein comprised of fibrinogen monomers, synthesized in the liver by hepatocytes. In the body, it has important roles in blood clotting and 306 wound healing . The clotting pathway has been replicated as a crosslinking method for 3DP; thrombin is used to rapidly polymerize fibrinogen into crosslinked 225,306 fibrin . As a glue-like gel, fibrin has been used clinically as surgical hemostatic 306 agents and sealants . The enzymatically-quick crosslinking rates have been 263,264,306 exploited with extrusion and inkjet-based printing , but the mechanical 306 properties of the constructs have been paradoxically described as both robust and 258 weak regardless of the concentration of the reagents . Fibrin-based hybrids materials with natural or synthetic components are usually reported to fine-tune mechanical properties depending on the application, including crosslinking with PEG, 258 the addition of PGA fibers, PLGA, hydroxyapatite, or demineralized bone matrix . Alginate is a natural polysaccharide derived from algae or seaweed. Sodium alginate is generally crosslinked in calcium chloride (CaCl2) aqueous solution, via an 216,313 ion exchange reaction between sodium and calcium . This chemically efficient reaction results in biocompatible, low polymer density and high water content 216 hydrogels . Traditionally, cell encapsulation in calcium alginate hydrogels has been the main application of alginate in tissue engineering and bioengineering models, despite the controversial effects of CaCl2, the crosslinking reagent, as well as sodium 71 citrate and ethylenediaminetetraacetic acid (EDTA), commonly used chelators, on 306 cell viability . This ionic crosslinking approach has been implemented in bioprinting, working particularly well in extrusion-based systems that extrude 219,220,227,230,272 alginate resin into CaCl2 reservoirs . Cells can be suspended in a solution of sodium alginate in cell-specific culture medium, after which crosslinking is induced by incubation in CaCl2 and results in a hydrogel construct laden with cells. This approach has been successful in bioprinting, for example, human cardiac-derived 230 cardiomyocyte progenitor cells (hCMPCs) for an in vitro committed cardiac tissue ; heterogeneous scaffolds with MSCs and chondrocytes (in alginate with osteogenic or chondrogenic differentiation medium respectively) for osteochondral tissue 219 engineering ; or encapsulated HepG2 liver cells printed directly on a polydimethylsiloxane (PDMS) chamber for a microfluidic pharmacokinetic liver 313 model . Nevertheless, Carrow et al. state major challenges for bioprinting alginate: (1) the difficulty to control the ionically-driven process which results in unpredictable microstructures, and (2) the high solubility of alginate as a disadvantage when 308 printing thick structures by extruding directly into CaCl2 aqueous solutions . Hyaluronic acid (HA), or hyaluronan, is a hydrophilic non-sulfated 306 glycosaminoglycan present in the ECM of tissues . The Atala group has regularly used HA in bioprinting processes by adding photo-crosslinkable methacrylate groups that can undergo free radical polymerization when irradiated with UV light. This modification allows printing soft hydrogels via stereolithography or extrusion with 274,275,306 additional post-printing UV curing . Although not mechanically robust on their own, HA hydrogels have served in cutaneous and corneal wound healing, 72 prototype vessel structure bioprinting, tumor modeling, and 3DP of cell-laden 306 structures . Other lesser-used natural materials for bioprinting include proteins and 310 257 289 216,303 polysaccharides like elastin , keratin , starch , or agarose . Despite the success of using isolated natural polymers, there has been growing interest in using the innate combination of proteins in the ECM. The ECM not only allows structural support and anchoring to cells, but provides a substrate for transport and communication, ultimately affecting the survival and differentiation of cells. The cell?ECM interactions are extremely complex and cannot be fully and precisely replicated in vitro or engineered from isolated proteins. Several groups have presented decellularized ECM (dECM) bioinks derived from adipose, cartilage, heart, bone, or skin tissues. The combination of proteins in the ECM can be understood as a hybrid of multiple natural materials; therefore, the available crosslinking methods and bioprinting approaches have been successful in producing 3D dECM scaffolds. Pati et al., developed dECM bioinks that can be extruded as filaments; printed scaffolds can then undergo gelation at physiological temperatures, remaining in the solution below 312 15 ?C and crosslinking by incubation at 37 ?C . 3.3.1.3. Co-printing and hybrid bioinks Approaches attempting to individually print synthetic and natural materials have produced scaffolds with mechanical and biochemical properties that very differently affect cells and tissues, and thus can be used in different types of in vitro and in vivo applications. A common generalization in the field is that synthetics are used for their strong, finely-tunable mechanical properties; nevertheless, it has been 73 proven that they can provide tunable degradation rates, functionalization capabilities, and various degrees of biocompatibility and print resolution. On the other hand, natural polymers have proven to be highly compatible with a wide variety of cells and biological components, mostly due to their inherent composition and function; as with synthetics, it is hard to generalize the negative characteristics of natural bioinks, but it is fair to say that the properties (mechanical or biochemical) are rarely fine-tunable and usually presented as ranges and wide error margins commonly associated with batch-to-batch variability. Over all, natural material properties, printing quality, and in vitro or in vivo behavior can be described as unpredictable and difficult to replicate. Combining both types of materials has been an increasingly popular hypothesis that relies on the positive properties of each. In theory, synthetic materials provide structural integrity and printing definition, while natural polymers can be 307,309,312 used to incorporate cells and other biological components . Two broad categories for combinations of synthetics and natural materials as bioinks for 3DP applications are (1) co-printing, the individual but parallel printing of natural and synthetic resins, and (2) hybrid bioinks, in which the resin is a uniform solution of both materials printed as a single construct. Co-printing approaches rely on printing synthetic scaffold structures with robust mechanical properties onto which natural hydrogels can be printed. This addresses the common limitation of natural materials, the inability to maintain uniform 3D structures in vivo (e.g. to allow tissue load bearing, or provide a specific porosity or microstructural pattern) or in vitro (e.g. to be robustly handled in 227 bioreactors, or as cell substrates), by integrating a synthetic scaffolding . 74 The main challenge is that co-printing relies on technologies that dispense more than one material during the printing process, sometimes with radically different deposition necessities and crosslinking mechanisms. In extrusion-based systems, for example, the rheology of the materials is the driving principle, and variables such as viscosity, flow rate, temperature, and pressure determine extruded line width, 220 fabrication time or print resolution . Shim et al. used a multi-head tissue/organ building system, possessing six dispensing heads to individually dispense thermoplastic PCL and alginate hydrogel in the same structure, to produce constructs 220 containing two different cell types for osteochondral tissue regeneration . As they report, the viscosity of alginate solution is about 10 Pa.s and needs low driving forces but high force control to achieve high resolution; on the other hand, viscosity for PCL ranges from 1020 to 2560 Pa.s at 80-120 ?C (a temperature high enough to damage 220 cells) which requires high driving forces to extrude . The same PCL-alginate approach was reported using a multihead deposition system (MHDS), they printed PCL and chondrocyte-laden alginate, with and without transforming growth factor 227 ? . Here, PCL was extruded at 80mm/min at 80 ?C using a 650 kPa pneumatic pressure. Sodium alginate was deposited at room temperature between lines of PCL, 227 at 400mm/min, and then crosslinked in sodium chloride . These cases illustrate the complexity of printing two different materials in the same structure; as stated before, the key relies in the independent extruding heads, in which the variables of the process (temperature, speed, pneumatic pressure, architectural patterns, etc.) can be controlled independently for each different material. 75 The capability of controlling each material separately, but still building a single construct, has high impact in resolution. In particular, the ability to place cells and materials with different properties in specific patterns confers high control over the resulting mechanical and biochemical behavior of the whole construct. Complex tissue constructs have been achieved via co-printing, such as the muscle-tendon unit 272 (MTU) approach recently reported by the Atala group . The MTU is the interface between muscle, elastic and fibrous in nature, and the tendon, very stiff and sparsely cellular. In this approach, natural hydrogels were composed of gelatin, HA and fibrinogen. Elastic polyurethane and hydrogel laden with C2C12 myoblasts were chosen for the muscle side, while stiffer PCL and NIH/3T3 fibroblasts gels were 272 selected for the tendon group . The co-printing approach allowed the controlled construction of the interface where two mechanically different tissues, with different cell populations, flawlessly meet. They report that the construct was not only able to mimic the complex mechanical behavior of the MTU, but successfully retain cell viability in both hydrogel portions (C2C12 cells with 92.7?2.5% viability and 272 NIH/3T3 cells with 89.1?3.3% after 7d ). The PCL-hydrogel co-printing approach has also been used to mimic mandible bone, ear cartilage, and skeletal muscle for 304 tissue engineering . Not only restricted to the production of tissue scaffolds for eventual in vivo applications, co-printing allows the construction of complex models for in vitro testing, particularly vascularization, microfluidic, and tissue-on-a-chip models. Having multiple heads depositing various materials and cells, under strict spatio- temporal control, has allowed researchers to produce highly complex models that 76 more closely resemble the behaviour of biological systems in vitro. These efforts usually require the use of cell laden hydrogels (see Section 3.3.1.4) and complex types of bioinks such as sacrificial (see Section 3.3.2) or supportive (see Section 3.3.3) bioinks. An interesting example of such structures can be observed in the Lewis group work, that use co-printing of natural and synthetic materials to develop 225 in vitro models of tissues and vascularization . They concomitantly print natural materials that include cell-laden castable and printable ECM composed of fibrinogen and gelatin, crosslinked via thrombin and transglutaminase (TG) enzymatic reactions, while the synthetic parts include silicone chip bases and pluronic sacrificial 225 materials . These materials are deposited and crosslinked independently and sequentially to produce a highly organized vascularized tissue analogue based on the strong characteristics of both natural and synthetic materials. Hybrid bioinks are the second approach trying to integrate synthetic and natural materials. In this case, there is a single bioink solution that includes both types of materials as solutes. Generally, hybrid bioinks are composed of a synthetic substrate solution with specific mechanical and rheological properties into which a natural component is mixed to alter biochemical and biocompatibility properties. For synthetic materials, adding natural groups to the bioink usually results in improved compatibility with cellular processes, by inclusion of binding sites and growth factors 227 or by reducing the high hydrophobicity of synthetics . For natural materials, the benefits are usually observed as structural or mechanical, but the inclusion of synthetic polymers to the protein chains also allows processing natural materials using the techniques and equipment designed for synthetics. The weak ionic 77 interactions or unpredictable enzymatic processes reserved to process alginate, fibrin or collagen, can be changed for optimized and finely tunable techniques like photo- 306 crosslinking or high resolution extrusion . The hybridization of the materials can be achieved either by mechanical entanglement of the materials in solution or by chemically joining the polymer and protein chains. The first is a common approach to improve the mechanical or rheological properties of natural materials. Narayanan et al. use human adipose tissue 314 stem cells loaded in alginate bioink with suspended PLA nanofibers . The cell laden alginate solution can be prepared separately from the nanofibers, but are vortexed together into a single solution, printed and crosslinked, trapping the PLA within the hydrogel without any crosslinking interaction between the two. This approach was successful in producing constructs that allow stem cell differentiation down the chondrogenic pathway, but more interestingly it proposes a method to use the distribution and alignment of the PLA nanofibers to stimulate orientations within the 314 ECM-mimicking hydrogels . The second approach relies on chemically altering and crosslinking the synthetic and natural chains, and is commonly used to improve the biocompatibility of the synthetic portion or to print the natural material using synthetic methodologies. The chemical modification allows personalization and optimization of the resulting bioink chain, which means higher specificity of the printed materials to cell or tissue functions. Shah group, for example, uses functionalized PEG to include a variety of proteins in extrudable, tunable, and cell 293 compatible bioinks . PEG with reactive ends (PEGX) is used to bridge protein and polymer chains in a variety of configurations, producing mixtures such as PEGX- 78 collagen, PEGX-gelatin, PEGX-fibrinogen, or PEGX-PEG, among many others, that 293 can be successfully loaded with cells and printed via extrusion . 3.3.1.4. Cell laden bioinks Current definitions of bioinks refer to resins that are loaded with cells and printed. As described before, we expanded the definition of bioink to include several categories of printable materials and not necessarily considering cells as the determinant bio factor. Nevertheless, the importance of cells for bioengineering and regenerative medicine is undisputable, and cell laden bioinks are crucial for the development of 3D bioprinting technologies and the goal of printing functional in vivo and in vitro tissues and organs. Synthetic and natural materials have been proven to have various degrees of success in cell compatibility, tissue integration, and tunable mechanical and biochemical properties, so why incorporate the complex additional factor of cells? It is commonly accepted that the acellular scaffold approaches have poor translation in vivo, mostly due to the limitation of cells only adhering to the surface of the constructs. The success of this approach is unpredictable, locations and concentrations of growth factors or chemo-attractants within the constructs cannot be 227 guaranteed and cell behavior cannot be controlled . We have mentioned before that the key term that defines modern bioprinting is control. Being able to control where cells, matrix, growth factors, and other biological components are placed, results in structures with higher orders of specificity and functionality. If materials and cells can be located and properly stimulated to construct gradients, strata, or clusters, there is a higher chance for success without relying on the unpredictable colonization of 79 native cells. Fedorovich et al. exploit this control feature to reproduce the specific 219 spatiotemporal distribution of cells and ECM in osteochondral tissue . The bioinks consisted of alginate solution in either osteogenic or chondrogenic differentiation medium, where MSC or chondrocytes were added respectively. After successful extrusion mimicking the adjacent bone and cartilage portions, ionic crosslinking, and subcutaneous implantation in mice, the dual, heterogeneous scaffolds show two different cell lineage commitments, with each type of cell remaining in its printed 219 position and depositing lineage-committed ECM . Another multiphase approach to osteochondral tissue engineering was presented by the Demirci group, aiming to 221 study the tissue interfaces in the anisotropic composition of fibrocartilage . Human MSCs were encapsulated in gels and printed by droplet deposition in an arrangement with zone-specific biochemical factors and ECM components (TGF-?1 for fibrocartilage and BMP-2 for bone regions). Again, cells show different lineage commitment by upregulation of osteogenesis- and chondrogenesis-related genes defined by the position and matrix they are printed in, yet constructing a single 221 heterogeneous scaffold . Cell laden bioinks are generally hybrid or natural bioinks that can be 3DP into hydrogels. The materials provide innate cell-binding motifs, hydrophilic surfaces, and 306 low cytotoxicity to promote cell adhesion . The hydrogel structures provide soft, degradable, and swelling networks that mimic ECM and allow cell migration, 282,309 metabolism, and differentiation with minimal restriction . The mechanical properties of 3DP hydrogels can be modified by regulating cross-linking density, the linking chemistry, or polymer concentrations to match properties close to those of 80 306 native ECM . Structural properties of the microenvironment, such as stiffness or composition, can deliver biochemical cues by mechanotransduction to regulate cell 293,303,306 shape, migration, and differentiation lineage selection . As an example, bioinks used for high-resolution prints generally produce stiffer gels ideal from a 3DP and structural standpoint, but the high elastic modulus will drive stem cell differentiation towards the stronger tissue lineages (bone, cartilage), making it an 282 obsolete approach to produce soft tissues . Hydrogels seem to have the ideal characteristics for cell adhesion and sustenance, but the material-cell tandem must also work with 3DP methods and account the impact of the processes on cell function after printing. First and foremost, no part of the bioink, the printer setup, additional crosslinking mechanisms, or by- products can be cytotoxic and have to be sterile-compatible; this seems straightforward, but it considerably reduces that available materials and processes that 216 can be used . Current 3D bioprinting methods mostly rely on physical forces or temperature to deliver materials. The most popular methods, extrusion-based and inkjet printing, rely on some mechanism of pressure that pushes the bioink through a nozzle. This setup translates pressure on the cells first as a compressive force within the cartridge and then as shear stress while they are moving through the nozzle. Varying the pneumatic pressure, extrusion speed, and nozzle diameter regulate the stress delivered to cells and have been proven to affect cell viability during and after 227,306 printing . A variety of printing protocols have studied these parameters and successfully printed multiple types of cells and materials with a very high viability rate: extrusion of MSCs and chondrocytes in alginate with 89% viability 5h after 81 219 246 print or 97% cell viability after thermal inkjet deposition . High shear rates and shear stress have been proven to harm the cells, but high viability is explained by: (1) stress causes protein denaturation by damaging the tertiary or quaternary structures of 246 the chains, a process that is reversible with time , or 2) natural or hybrid gels have shear thinning behavior, which decreases stress on the cells even at high shear rates or 219,312 pressures . Bioink viscosity and surface tension can also be modified to reduce 246,307 shear stress using solvents or surfactants , although these must comply with minimum cytotoxicity requirements. In the case of stereolithography, the UV intensities required to initiate and sustain photo-initiator and photo-crosslinking 216 reactions can negatively alter cell morphology and viability . The amount of energy radiated can also cause irreversible damage to the cells due to increasing temperatures and dehydration. Temperatures different from physiological 37 ?C, both high or low, and 306 drastic changes will alter cell metabolic activity and may cause cell death . This is a critical consideration for co-printing approaches, where the temperature required to deposit synthetic materials could easily be above 60-100 ?C and then solidify on cold 217,219,225,227 surfaces, both extremes having the potential for irreversible cell damage . Chemically or pH induced crosslinking by ionic and physical mechanisms can also harm cells. Additional crosslinking or post curing approaches, such as post print glutaraldehyde or EDTA chemical crosslinking or additional UV irradiation, which guarantee 3D geometries and mechanical properties, are commonly seen as additional 217 negative factors . 82 A critical element to maintain cells in the bioinks is the permanent need for proper oxygenation and metabolic transport. In vivo, any cell will be located within 100-200 ?m of a capillary, the maximum distance for adequate gas and metabolic 217,220,303 exchange . Thick casted or printed hydrogels without pores do not allow diffusion deep enough to supply oxygen or nutrients and result in necrotic cells encapsulated in the center of the structure. Generally, the design of 3D printed architectures is envisioned with pores or channels that provide open transport pathways and open vascularization channels. Interconnected pores, constructed by weaving strands or layered patterns, with dimensions ranging for tens to hundreds of micrometers usually allow transport through the scaffolds and report high cell 220,223,276,307 viability and development . Another common approach is to induce fast vascularization of the constructs, generally by adding endothelial progenitor cells or 217,225 growth factors and using bioreactors . Overall, accounting for the strengths and limitations of printing cell laden bioinks, numerous approaches regularly report very positive effects of this approach on cell behavior for a wide range of applications. As an in vitro example: Gaetani et 230 al. printed a model of undifferentiated, but committed cardiac tissue . Here, sodium alginate was dissolved in culture medium and mixed with hCMPCs. 3D constructs were obtained by printing strands into layers, stacking them to obtain different degrees of porosity, and crosslinking with CaCl2. Comparing to regular 2D cultures, bioprinting had no effect on cell viability and proliferation, but increased cardiac lineage commitment by upregulation of early and late cardiac transcription factors 230 and markers . On the other hand, an in vivo application aiming to produce an 83 implantable bioartificial liver: Wei et al. used a gelatin/fibrinogen matrix loaded with rat hepatic cells to produce 3D porous constructs via extrusion-printing and thrombin- 263 induced gelation . After extrusion, ~98% of the hepatic cells are reported as viable, steadily producing albumin and dissolving the surrounding gelatin matrix throughout 263 the culture time . 3.3.2. Sacrificial bioinks Sacrificial bioinks allow the fabrication of complicated structures and open geometries without dealing with many of the difficulties related to satisfying biological requirements. Using a sacrificial bioink for a print material, the print volume is created initially and will subsequently be washed away. The bioink provides space filling volume and support that will be evacuated. Some groups refer to their sacrificial bioinks as fugitive inks to further suggest its temporary role in 315 printing in the scope of the final structure . Therefore a sacrificial bioink needs only to be nontoxic which will also further along not induce any harmful byproducts; however no further biological features are necessary such as cell adhesiveness or 214 biodegradability . Here, by nontoxic, we idealize success cases where the byproducts are additionally non-cytotoxic. Still, sacrificial bioinks ideally match these specifications: high print fidelity, ease of removal, and lack of toxicity. To enable ease of removal, an important material property is its gelation process, in other words under which conditions will the printed material wash away. Some examples of bioinks such as Pluronic F127 or gelatin have a thermally reversible gelation process. Therefore, while printing can occur at one temperature, the printed material can evacuate when another temperature is attained. The Lewis lab 84 is one of the leaders utilizing Pluronic F127 as a fugitive ink to create perfusable networks in tissue mimics. Pluronic F127, a poloxamer, is solid at 37?C but can be liquefied when cooled to 4?C. They take advantage of this material property to flush out the fugitive bioink with cold cell media to leave behind perfusable channels. Resulting print structures resemble thick vascularized tissue mimics with 225 endothelialized lumens viable after 45 days of perfusion . The same group applies this strategy to other tubular tissues where most recently the technique was applied to 298 fabricating renal proximal tubules . 279 269 Other common sacrificial bioinks include agarose , alginate , and 278 gelatin . The Khademosseini lab applies sacrificial templating strategies with different materials. For agarose, the geometry is extruded to form a solid network at 4?C. Subsequently, the agarose fibers can be either manually removed or lightly vacuumed as the material does not adhere to surrounding photo-crosslinked 279 hydrogel . With a different material, the Khademosseini cohort uses sodium alginate to fabricate a sacrificial network. The gelation of alginate occurs ionically with 269 calcium chloride and can be removed with EDTA treatment . Others apply this strategy to 3DP other tissue mimics like aortic valves and bone; however, these 270,316 groups do not print with this bioink sacrificially . An additional naturally derived sacrificial bioink, gelatin, is thermally reversible proving useful for a sacrificial templating strategy. Lee and colleagues used inkjet printing to form vascularized tissue constructs dropwise and layer by layer with gelatin and collagen. Collagen layers were printed and polymerized at 4?C with a pH altering crosslinking agent (sodium bicarbonate) so that the thermal responsive nature does not take precedence 85 over the intended material evacuation. Within the collagen layers, gelatin was left uncrosslinked and therefore removed by liquefying the printed structure when the 278 final print structure was raised to room temperature . These natural sacrificial bioinks described are all able to be evacuated easily fitting one of the crucial characteristics of a sacrificial bioink. So far, the discussed sacrificial bioinks suffer from poor mechanical strength potentially leading to complications in maintaining print fidelity after the print. Some address this complication by using an alternative sugar material that has high 224,317 mechanical stiffness and is water soluble . Miller and colleagues developed an approach to use carbohydrate sugar glass to fabricate complex vascular designs. This strategy circumvented issues related to poor mechanical strength. However, a similar evacuation approach is utilized where the printed perfusable network is evacuated by running cell media through the channels in order to provide a fluidic network. Even though the carbohydrate glass lattices were printed at high temperatures (110?C, a temperature physiologically unviable), the finished print volume can be brought down to physiological temperature with complete cell media for 10 minutes to dissolve 224 away the carbohydrate glass . Post evacuation, these open channels can be perfused with cell suspensions for long term cell culture. However, the extrusion-based printing technique used to fabricate structures is not without its limitations. One example of a limitation is as the build volume relies on depositing layer by layer material, this technique is unable to create overhanging structures without support. Some in the bioengineering community utilize SLS as a different 3D bioprinting technique extending the list of bioinks while addressing the overhang 86 limitation of extrusion-based approaches. SLS-based prints have been commonly used with materials such as PCL and PLA to, for example, produce bone-mimicking scaffolds. However, there is emerging research for this printing strategy falls under 318,319 the scope of sacrificial bioinks . Limited in resolution only by the smallest powder size of the bioink and the laser used, SLS exhibits micron-sized fabricated 241 structures with complicated overhangs with a sacrificial templating workflow . Using PCL, a lattice structure can be selectively laser sintered and dissolved away to leave behind a fluidic network. However to provide perfusable channels, the printed construct requires dissolving in dichloromethane which may or may not induce 241 toxicity to future seeded cells . More gentle evacuation for future cell encapsulation may be necessary to translate to a cytocompatible workflow. While others utilize sacrificial templating strategies with plastics (i.e. acrylonitrile butadiene styrene) to fabricate intricate microfluidic systems and subsequently remove with acetone, these 320 removal approaches may not lend well to tissue incorporation . Although there is a short list of sacrificial bioinks, these materials share one thing in common ? the need to have multi-material integration for a finalized structure. Inevitably for the sacrificial ink to be evacuated, a requirement is an encapsulating material in which the flushed material can leave behind a space. The selection of the encapsulation material, the matrix or matrix-mimicking bioink (see Section 3.2.1), is crucial to incorporate cells or interface with biological tissue therefore might require cell adhesive sites or degradation properties. Additionally, the gelation process of the encapsulation should ideally be at a mismatch to the sacrificial 87 bioink so that the intended material is the one being sacrificed at the point of dissolving. 3.3.3. Supporting bioinks and supporting baths Supporting bioinks and supporting baths are used during 3DP to improve mechanical properties and expand geometric capacities of the bioprinted scaffold. High viscosity is commonly a desired material property for supporting baths to provide structural integrity as the print is fabricated. This material property functions to hold the print in place in a way that the printed material alone (without the support bath) would be unable to maintain (i.e. overhanging features or ultra-thin features). Consistent with requirements for other bioinks, the material should not possess toxicity for assurance of biocompatibility. The Feinberg lab takes an innovative approach to bioprinting by taking advantage of material properties to extrude bioprinted features within a hydrogel support bath. Their approach, coined freeform reversible embedding of suspended hydrogels (FRESH), uses a gelatin slurry hydrogel bath embodying thermo-reversible properties and Bingham viscosity to print soft, fragile constructs within the bath material. With the bath?s Bingham plasticity, the material acts as a solid unless a yield shear stress is attained, in which case the material will flow as a viscous fluid locally to the applied shear. This way, extruded crosslinked printed structures (i.e. fibrin, alginate, and collagen type I via enzymatic, ionic, and pH mechanisms respectively) within the hydrogel bath will immediately be surrounded by the supporting viscous material enabling improved complexity in fabrication without collapsing or deforming under the printed structure?s own weight. Additionally, the gelatin slurry 88 possesses thermo-reversible and biocompatible properties. The temporary support bath can be removed by raising the temperature from room temperature maintained during the printing process to a physiological range at 37?C (for gelatin specifically) to release the 3DP object supported within. Because the gelatin slurry is biocompatible, cells can be incorporated in the extruded bioink within the bath 235 without concern of its effect on cell viability . Similarly, others have achieved bioprinting of intricate structures that can otherwise not be achieved without support baths. Carbopol is another support bath material possessing desirable rheological properties in that the transition from locally fluidized to solidified afforded by its Bingham plastic nature allows for the support of printed structures. While the writing medium is more permanent (via photo- 301,321 crosslinking or other means), the support bath can be dissolved with water . The Angelini group has even expanded their carbopol support material to be cytocompatible for a range of cells (i.e. MSCs, human aortic endothelial cells) with enough elasticity to support cell division with the inclusion of cell media adjusted to a 322 pH of 7.4 . For those utilizing inkjet printing for their bioprinted constructs, a common support bath system is the use of a CaCl2 solution that can dually function as a crosslinking agent and a support material. In this way, the alginate ink-jetted deposition with or without cells are fabricated with the support bath instantly 273 polymerizing the material to allow for complex structures with overhangs . Yet another extrusion based bioprinting approach, the Burdick lab developed a hydrogel bioink format that is directly written into a self-healing support hydrogel 89 based on guest-host complexes. Their system is based on noncovalent and reversible bonds with the application of shear. By directly writing into this material, the bonds of the support bath are disrupted by the physical stimulus but are quickly reformed after the shear is removed. This enables the capacity of injectable hydrogels and extrusion 3DP. Their writing bioink based on HA was chemically modified for photo- crosslinking and, more importantly, the support bath possesses shear thinning 262 properties to provide structural support to print free-standing structures . Support baths must interact with the finalized structure in such a way that the material itself must still provide minimal baseline of compatibility. Certainly, some approaches have been utilizing the support bath as a reservoir of cells or perhaps crosslinking agents that will ultimately become part of the printed construct, therefore the level of cytocompatibility is correspondingly going to increase. Regarding toxicity, beyond the baseline of cell compatibility in cell-laden prints, studies in the future might want to consider long term effects on cytotoxicity. If residual components of support baths are cytotoxic, a few days after the print is not long enough to prove long term cell viability and functionality. Finally, in parallel to sacrificial inks, these materials must be easily removable to decouple from the finalized printed construct. 3.3.4. Current translation of 3D bioprinting Three-dimensional bioprinting has opened the door for opportunities in directly translating to the clinic. 3DP provides us with the ability to print tissue analogues by controlling the delivery of living cells and matching the appropriate material in a defined and organized manner. The control is promising because it is 90 beyond defining the exact location, but also the defining of sufficient number of cells in a multi-material environment. 3DP has applications to tissue-engineering scaffolds, constructing cell-based sensors, physiological screening for drug and toxicity, and 312 modeling tissue disease and tumors . Here we will broadly characterize translation applications to in vitro and in vivo examples. 3.3.4.1. In vitro applications In general, lab-on-a-chip style models can represent tissue analogues well by incorporating the many fluidic networks of our body (i.e. vasculature, lymphatics). An example of lab-on-a-chip, miniaturized functional tissue mimics was discussed in the sacrificial section previously from the Khademhosseini lab. In their work, they 279 incorporate embedded vascular networks in their 3DP constructs . Many others have also dedicated their research in incorporating vascular fluidic networks because 224,241,298,315 vascularization remains a critical challenge in tissue engineering . As discussed, beyond a few hundred microns of the diffusion limit, cells will not remain viable. However, even with many research groups studying vascularization strategies, many challenges remain such as what vascular geometries to print and reaching truly multiscale vasculature. By miniaturizing human tissue functional elements, studies can more accurately predict drug toxicity over animal models. Khademhosseini and colleagues delve into more complex in vitro models by developing liver-on-a-chip constructs. By developing a liver model, one can assess drug toxicity more adequately as the liver plays the most important role in drug metabolism. Their liver model was biofabricated as a perfusable bioreactor with a direct write printer to create hepatic 277 spheroid-laden hydrogel structures . 91 In addition to better drug screening, in vitro models fabricated with 3DP approaches can better represent tumor models to help researchers elucidate cancer mechanisms. Efforts are being made to three-dimensionally represent tumors in models as the progression of tumor metastasis is significantly different than a 2D 323 counterpart . The West lab models tumor angiogenesis with layer-by-layer tunable 324 PEG hydrogels . 3DP can provide cancer researchers with control over specific 3D microenvironments influencing nutrient transport and fluid shear stresses. With this kind of physiological mimicry, researchers can identify mechanisms directing tumor metastasis. A major challenge of 3DP is scale for tissue engineering translation ? printers often have difficulty achieving large, clinical-size organ analogues with the micro- detail of the cellular organization necessary to be functional. The Atala lab reports an integrated tissue-organ printer that prints cell-laden hydrogels of desired mechanical stiffness with sacrificial hydrogels. Fabricated constructs presented include mandible bone, cartilage, and skeletal muscle. The printed anatomical shapes can be composed of multiple biomaterials and cells resulting in structures that in the future can be 325 vascularized and included in more complex, solid organs . Until then, these constructs may be difficult to incorporate in the clinic as encapsulated cells will quickly necrose without proper vascularization. For in vitro 3DP, clinical translation is mostly limited to drug screening and disease modeling. These applications can provide important information on addressing drug toxicity and for drug development. 92 3.3.4.2. In vivo applications The promise of 3DP to fabricate tissue parts for implantation in the body is taking shape in in vivo applications. Tissue engineering has the potential to facilitate tissue regeneration by replacing injured parts and encouraging regrowth with the appropriate healing environment (i.e. growth factors, vascularization, stem cells). With 3DP, a diversity of critical sized defects can be addressed tailored to the needs of the individual. In this section, we discuss studies that utilize 3DP to fabricate scaffolds applied in vivo. As mentioned previously in the chapter, rapid vascularization is one of the limitations in applying tissue mimics in vivo as cells will necrose soon after implantation without access to nutrients and oxygen. Several studies achieve rapid vascularization in vivo enabled by pre-vascularization strategies fabricated with 3DP 326,327 techniques . Using sacrificial templating with a sacrificial carbohydrate bioink, an open microvascular network can anastomose in-line with the rat femoral artery 224,326 with a surgical technique . Similarly, others utilize 3D printed biodegradable scaffolds to directly and surgically anastomose built-in vasculature in an AngioChip 327 to the host vasculature of the rat hind-limb femoral vessels . Patency is maintained through the fabricated vessels within the scaffold. These studies demonstrate the potential of pre-vascularizing tissue analogues with 3DP for rapid vascularization in vivo. Other tissues have also been fabricated and implanted in vivo such as bone. Bone scaffolds 3DPed with MSCs, bone morphogenetic protein growth factor, and PCL as the bioink have been implanted in critical sized bone defects of sheep. After 93 three and twelve months, bone healing progresses with signs of vascularization 288 ingrowth, mineralization, and appropriate mechanical stiffness . By considering important aspects of biocompatibility and degradability, many of the bioink principles described in previous sections affects the success of 3DP tissue mimics for in vivo implantation. The opportunities to apply 3DP to in vivo applications are still being explored. Integration with the host tissue is of utmost importance for all in vivo studies. For researchers, great care is taken to select the most appropriate bioink during the fabrication process to allow the best possibility for integration success. Future work is necessary to ensure functionality of the intended tissue analogue taking consideration the role of the parenchyma. 3.4. Conclusion and future directions This chapter introduces current 3DP techniques and describes the materials used to enable bioengineering applications. For bioprinting, the materials, or bioinks, used for fabrication must be biocompatible in addition to printable. Bioinks are printing materials that will either interface with biological components or are part of the fabrication process of a construct that will come into biological contact. We describe three separate categories for bioinks: 1) matrix or matrix-mimicking, 2) sacrificial, and 3) supporting bioinks. Although there has been great progress in synthesizing biocompatible and printable materials, there is future work needed to address the greatest limitations in bioengineering. For one, vascularization must be rapidly integrated. Many of the 94 challenges for vascularization strategies involve scale to realize multiscale hierarchical vessels. Reaching capillary-sized vessels require microscale resolution in printing. However, many of the printing techniques that can even reach these small levels are not able to print rapidly. Another limitation in current studies is the use of physiologically relevant cell types. Fortunately, induced pluripotent stem cells hold much promise in alleviating this limitation. Materials available as bioinks are rather limited. Generally, the bioinks used are a compromise between structural strength and biocompatibility. Therefore, current and future work follow a multi-material approach to achieve the desired properties. The advent of 3DP applied to bioengineering and tissue engineering has pushed researchers to create complex biological structures with incorporated fluidic networks with multiple cells and materials for the most physiologically mimicking environment. This is an exciting time for the field to realize approaches that have clinical impact. From rapidly reiterating studies for drug screening to achieving high spatial resolution to fabricate tissue analogies, 3DP holds much promise for therapeutic translation. The choice of bioinks will continue to play a fundamental role in determining ultimate biocompatibility. 95 3 Chapter 4: Modeling Skin: Epithelial Barrier Models and 4 Bioreactors 3 4.1. Skin as an epithelial barrier model 4.1.1. Introduction Epithelial tissues are key regulators of physiologic homeostasis within humans, providing protection from foreign substances while also regulating transport of nutrients necessary for survival. Epithelial cells comprise the largest tissue in the human body, ranging from skin and vasculature to the placenta. Some instances of the physiological barrier are multilayer and multicell-type in nature and significantly influence the regulation of transport across the barrier. These added complexities are particularly important in various instances. In skin, for example, the multilayered barrier is critical in regulating temperature and water transport into and out of the body. Therefore, understanding the tissue-specific physiology and regulation will allow for improving the design process of therapeutics and biomedical treatments. This is particularly true for the use of tissue models to gain new knowledge about substance transport across these tissues, which is invaluable in drug development for optimizing delivery kinetics, testing efficacy, and potentially reducing necessary dosages, potential side effects, and cost. With regard to endogenous substances or 3 Adapted from: N Arumugasaamy, J Navarro, JK Leach, P CW Kim, JP Fisher: ?In Vitro Models for Studying Transport Across Epithelial Tissues Barrier?. Annals of Biomedical Engineering, Vol. 47, No. 1, pp. 1-21, 2019. 4 Adapted from: Yu, J. R., Navarro, J., Coburn, J. C., Mahadik, B., Molnar, J., Holmes IV, J. H., Nam, J., & Fisher, J. P. (2019). Current and Future Perspectives on Skin Tissue Engineering: Key Features of Biomedical Research, Translational Assessment, and Clinical Application. Advanced healthcare materials, 1801471. (DOI: 10.1002/adhm.201801471). 96 exogenous drugs, the use of tissue models provides a robust and rigorous method for investigating molecular interactions, and in some instances can enable studies that otherwise have key ethical limitations. Recent efforts to model epithelial tissues have utilized both two-dimensional 328?331 (2D) and three-dimensional (3D) approaches . Traditionally, 2D models are cells seeded on a transwell insert, creating apical and basolateral compartments as occurs 332 in vivo . Over the past few years, 2D-like organ-on-a-chip models have grown in prominence. These platforms facilitate the biomimicry of key aspects relevant to the tissue-specific interface, such as contractions and the resulting biomechanical 328,329 stimulation on lung cells within the lung-on-a-chip . Furthermore, these models can incorporate convective flow to mimic barriers interacting with fluid, such as occurs in the lungs with air or the epithelium with blood. Organ-on-a-chip models are 333 more extensively discussed in a 2016 review by Sakolish et al. . Also in recent years, 3D models have gained prominence due to the underlying premise that cells 334 behave differently in 3D and 2D environments . An engineered 3D environment can provide an improved spatial, biomechanical, and biochemical niche compared to 2D 334 environments , though it should be noted that engineered 2D approaches could also provide some of these cues. In particular, interest in tissue organoids has grown over 335 the past few years . Tissue organoids are 3D systems that mimic the spatial 330,331 arrangement of cells, which may be more realistic and useful for certain tissues , though these models are not without their concerns of reproducibility and limitations 335,336 of cellular complexity . Regardless of the approach taken, an emphasis on 97 biomimicry drives many of the approaches we discuss and will help to refine models used within each of these fields. Figure 4.1. Four types of interfacial barriers in human physiology. (A) The air-tissue interface shows molecular transport (purple diamonds) from air into the tissue space (i.e. skin or lungs). (B) The air-liquid interface shows molecular transport (yellow diamonds) from air into liquid, which is often the vasculature of the body, represented by blue/red vessels (i.e. skin). (C) The liquid-tissue interface shows molecular transport (green diamonds) from liquid into the tissue space (i.e. the brain). (D) The liquid-liquid interface shows molecular transport (orange diamonds) from liquid into liquid (i.e. the placenta). The epithelial nature of skin is comparable to other epithelial tissues such as the gastrointestinal (GI) tract, the lungs and blood-air barrier, the endothelium and blood-brain barrier, and the placenta. A key question arises from collective 98 comparison of these epithelial tissues: do the differences in each of the tissue-specific interfaces ultimately affect how tissue models are developed, and if so, how? These tissues represent multiple interfaces: air-to-tissue, air-to-liquid, liquid-to-tissue, and liquid-to-liquid (Figure 4.1). Though these interfaces are often considered unique, they may be more similar than many realize. Recent reviews have presented similar comparisons for epithelial tissues, though one focused on nanoparticles and the other 333,337 on organ-on-a-chip approaches . 4.1.2. Modeling skin epithelium As presented in previous chapters, skin is the first and largest barrier the body has to defend against external mechanical and biochemical agents. Mammalian skin is composed of the epidermis, dermis, and hypodermis layers. In skin anatomy the epithelium refers to the epidermis, the external layer of skin, which ranges in thickness from 0.04 to 2mm, is avascular, and composed of ~95% keratinized stratified squamous epithelial cells: keratinocytes. The epidermis is further stratified into the innermost stratum basale, stratum spinosum, stratum granulosum, and the outermost stratum corneum. As they move from the basale to corneum strata, keratinocytes increasingly organize, form desmosome junctions, and secrete keratin 4,338 and lipids reinforcing the mechanical and biochemical barrier . The epidermis and the dermis are attached via the intermediate basement membrane, a mesh of fibronectin, laminins, collagen IV, and proteoglycans that allows water retention and 4,33 anchors skin appendages such as hair follicles and sweat glands . The dermis, responsible for the mechanical properties of skin, is a 1 to 4 mm thick layer that consists mostly of fibroblasts synthesizing extracellular matrix composed of collagen 99 4,33,338 (~70%), elastin, and proteoglycans . Last, the innermost hypodermis is generally neglected in skin models as fat storage for thermal regulation. However, this complex lipid barrier has significant contributions to the overall function of skin. It is the layer from which nerves and larger blood vessels permeate the upper layers and is a rich source of stem cells, hormones, and growth factors, which are key players in re- 56,98,99,338 epithelization, wound healing, and angiogenesis . The combination of properties between these distinct layers of skin results in a highly efficient barrier. Any damage to it results in immediate compromised thermoregulation, massive fluid shifts and risk of bacterial sepsis. Understanding the transport through the skin barrier has been pursued to understand not only homeostasis and hydration, but also how medications can be optimized for topical or systemic delivery and how the body is shielded against, or responds to, pathogens and radiation. Ultraviolet (UV) radiation, for example, is an unavoidable agent that has 339 340 been studied with quantum dot penetration or percutaneous absorption models and proven to change the skin barrier and immune functions. Throughout the dermal layers, the permeability function is defined by different combinations of extracellular matrix (ECM) and lipids (ceramides, cholesterol, cholesterol esters, and free fatty acids). The lipids assemble into multi-lamellar sheets that fill the spaces between cells 33,341 and the ECM, the main route of permeation for most compounds . Models of skin have struggled to mimic the characteristic distribution of layers, ECM, lipids, and cells that form the dermal barrier. Because of the complex layering and its multiple functions, in vitro models have generally been too simple to adequately capture all aspects of the physiologically relevant transport phenomena 100 through skin. The earliest approaches to understand the barrier functions were based on ex vivo human and animal skin, particularly pig, mouse, rat, guinea pig, and snake 54,339,341?344 models . Pig skin is the traditional model of human skin, as it has a 341,345 comparable stratum compactum and broadly the same lipid content and ratios . Even with the advantage of being a natural barrier, explanted tissues are limited by donor availability, dermal differences between species, lack of control over specific cells or pathways, and the overall problem of maintaining a living tissue in optimal condition to render adequate results. Studies have further elucidated the limitations of animal models, particularly on the higher permeability of their skin due to differences such as thinner stratum corneum, different sweat gland or hair follicle density, and 54,346 lipid composition . Furthermore, the use of animals and animal skin for testing of cosmetic ingredients has been banned since 2009 by the European Union, motivating a critical need for alternative models to mimic the dermal barrier. Tissue engineering has been used to construct models based on specific questions, separating or combining the different dermal layers, cell populations, matrix compositions, and growth factors as needed. Such constructs have reached commercial grade as ?living skin equivalents? or ?dermal equivalents?. Products such 68,94,95 TM 24,54,347 as Episkin? , EpiDerm , MatriDerm?, and Graftskin are bi-layered scaffolds constructed by cultivating fibroblasts inside a hydrogel (generally collagen 24,54 type I), upon which a layer of keratinocytes is seeded . The constructs are cultivated submerged in growth media until the populations are mature, then the level of media is decreased to expose the keratinocytes to air (effectively an air-liquid 24,54 interface), stimulating them to proliferate, stratify, and keratinize . 101 Made from human cell lines and natural materials and grown together, commercial skin equivalents provide the biological processes and metabolism of 33 native human skin and are generally considered accurate representatives of skin . Nevertheless, the barrier function has been proven to be more permeable in skin equivalents, a difference commonly linked to the subtle variations in lipid content in 54,93 living equivalents, particularly a higher content of di- and triglycerides . Fleischli et al. reported a comparative study between Episkin? and human skin, assessing the transport of topically applied substances like caffeine, benzoic acid, salicylic acid, 94 and octyl methoxycinnamate . Differences between the tissues included the changes in the local concentration profiles in the stratum corneum, the water profile, the lipid 94 content, and the presence of natural moisturizing factor in living equivalents . Models including the hypodermis fat layer are rarely proposed and are not 56,97?99 commercial . In particular cases, collagen, fibrin, or silk gels have been used to encapsulate adipocytes or adipose-derived stem cells, which are cultured separately 98,99 and then stacked with bi-layered scaffolds in a process that can take up to 35 days . The addition of the adipose layer has been used to study potential complex drug 98 56 absorption models or the role of skin in adipose tissue metabolism . Such approaches using complex layers are compared against standardized controls that include synthetic membranes (PDMS, Silescol?, silicone, or dialysis membranes) or 340,348,349 naturally-derived gels (Matrigel, cellulose) . These systems are generally used as controls for transport variables or cell response, yet they cannot fully model the cell-lipid-matrix barrier complexity of skin. 102 These engineered living skin equivalents are used throughout literature as skin 24,54,68,93?95,347,350 barrier models, becoming a popular alternative to explanted tissues . As illustrated in Figure 4.2, traditional systems used for modeling transport across the skin can be broadly divided into three categories: i) a liquid-to-liquid interface system, usually Franz diffusion cells (Figure 4.2A), ii) topical application using an air-to-liquid interface system (Figure 4.2B), or iii) systemic application in an air-to- liquid interface system (Figure 4.2C). The Franz diffusion cells are standardized 93 systems for permeation studies and allow the use of living equivalents or the inclusion of split-thickness (epidermis only) or full-thickness (epidermis and dermis) 351 345,346,348,352,353 human or animal tissue . The air-to-liquid interface systems more accurately replicate transport through the dermal barrier, mimicking the exposure of the stratum corneum to air and environmental factors, the transport across the layers, and the blood or systemic components with the liquid/media chamber. Inclusion of living equivalents in these systems fully completes a skin model that can be grown and used in vitro. The topical and systemic exposure variations have been used to study the permeation of medications and active molecules across the dermal barrier to 350 further optimize drug delivery systems , or study phenomena such as cytotoxicity of 24,54,347 compounds generally applied to skin , dermal pathologies and wound 68,95 339,350 healing , or the effects and treatment of radiation . The Curren group used TM EpiDerm in air-to-liquid interface systems to develop the reconstructed skin micronucleus assay. Here, the induction of micronuclei in a dose-dependent fashion was considered the initial steps in developing an in vitro assay for chromosomal 103 damage in human skin, bypassing the need for in vivo genotoxicity tests of cosmetics 54,347 ingredients in animals . Figure 4.2. Models of transport through skin. Traditional models for skin transport include (A) liquid-to-liquid interface systems, usually Franz diffusion cells where the donor chamber acts as the apical side of the membrane and the receiver side acts as the basolateral side. Skin models still widely rely on variation of the transwell inserts (B-C), where the exposure of the barrier to air allows assessing the physiological behavior of keratinocytes and the external role of skin. (B) Topical application is modeled by applying the treatment on the air-exposed (apical) side of the barrier, while (C) systemic application is modeled by delivering the treatment via the liquid- exposed (basolateral) side. Common air-to-liquid interface systems generally provide the necessary setup to perform electrical resistance (ER) or trans-epithelial electrical resistance (TEER), tritiated water flux (TWF), and trans-epidermal water loss (TEWL) assays without 353,354 major modifications . Variations of these systems usually refer to modifications of the tissue layer, ranging from the use of synthetic membranes or laminated dialysis 104 348,349 membranes to novel 3D printed (3DP) constructs. 3DP technologies have been applied to skin modeling to exploit the spatiotemporal placement of cell lines and materials into stratified constructs such as specialized skin equivalents or stratified 355?357 epidermis . More robust 3DP approaches include the Integrated Composite 358 tissue/organ Building System (ICBS) presented by the Cho group or the Laser- 359 assisted BioPrinting (LaBP) system reported by the Chichkov group , which have been discussed in previous chapters. Most recently, the development of organ-on-a-chip technologies has allowed the introduction of skin-on-a-chip systems: miniaturized models of dermal cell 360?363 populations, structures, and functions . Mori et al. presented a method for fabricating perfusable, endothelialized vascular channels within a cultured collagen skin-equivalent chip. This system was intended for drug development and validated with the percutaneous absorption of caffeine and isosorbide dinitrate. Simulation of topical exposure indicated that both molecules first reached the vascular channel and then the bottom of the skin-equivalent at rates consistent with those of human skin and conventional models, successfully introducing the highly relevant parameter of 364 vascular transport into dermal models . Alternatively, Wurfuer et al. developed a multilayered skin-on-a-chip system with epidermal, dermal and endothelial 365 components . The microfluidic device was designed for co-culture of human skin cells, and each layer was separated by using porous membranes to allow interlayer communication. This novel approach allows modeling of skin inflammation and edema by disrupting the layers with tumor necrosis factor alpha (TNF-?) and subsequent therapeutic drug testing. Tissue engineering principles have successfully 105 modeled the multilayered structure of skin, including its complex combinations of cells, lipids, and ECM proteins. The progression of these constructs into liquid-to- liquid and air-to-liquid interface systems allows one to model the broad behavior of transport phenomena through skin for topical and systemic medications or the effects of radiation. Other specific parameters of skin transport have been studied, including 365 the viability of tight junctions and permeability of the endothelium . Organ-on-a- chip or 3DP approaches have enabled increased complexity of the models. For example, future efforts in modeling skin transport can include elements of the immune system, alterations in skin microstructure due to scarring and open-wound healing, and interactions of skin with underlying tissues. A more in-depth review by Abaci et al. discusses the use of induced pluripotent stem cell (iPSC) technology and multi-tissue organ-on-a-chip platforms for drug and chemical permeability, both 366 additional areas for future endeavors in skin transport modeling . 4 4.2. Bioreactor systems for skin 4.2.1. Introduction 367 Traditionally used in microalgal cultures , tissue engineering has adopted the bioreactor technology for the culture of cell populations. Nevertheless, the term ?bioreactor? has been loosely used to describe any cell-culture system more complex than a 2D Petri dish where the cells can be provided with a constant source of oxygen 368 and nutrients . Traditional 2D surfaces, such as tissue culture polystyrene (TCPS) plates, provide an adequate, non-cytotoxic adhesion surface for cells, but are far from 106 368,369 mimicking the complex microenvironment of real tissue . As such, functionality 369 and specificity of cells cultured in vitro was limited . Modern bioreactors are designed to provide cultured cells with 3D, dynamic, aseptic, complex biochemical and biomechanical cues that better mimic the physiology or pathophysiology of tissues. Now, bioreactor systems are not only for cell culture and expansion but can be used to induce lineage specificity for tissue engineering, tissue modeling, or cell 368,370,371 biology studies . Furthermore, bioreactors provide controlled and reproducible 368 environmental conditions specific to different tissue necessities . Increasingly complex bioreactors have been designed to sustain a variety of cell types such as 370,372 371 373 chondrocytes , NIH3T3 fibroblasts , human mesenchymal progenitor cells , 374 375 rat primary hepatocytes , stromal and erythroleukemia cells , human bone marrow 376 377 stem cells , or megakaryocytes , and even ex vivo tissue samples including 370 378 379 tendon , articular cartilage , and bone . The increasing complexity of these systems should be considered an attempt, or progression, to mimic the in vivo environment. As described before: there is a growing body of evidence to show that the in vivo environment presents the most optimal environment for engineered tissue 368 implants to grow . Simply stated, bioreactor engineering is striving to reproduce the microenvironment of the human body to grow tissues in vitro. 4.2.2. Complex and multi-chambered bioreactors Bioreactor engineering has produced devices that can dynamically stimulate tissue-specific constructs by introducing complex biomechanical cues such as cyclic 372 371 compression or tensile strain , or more joint-specific knee or intervertebral disc 370 368 bioreactors . Throughout bioreactor literature researchers highlight the need to 107 understand the dynamics of such culturing systems and their effects on the cell populations. Dynamic parameters can include flow, transient transport phenomena, cyclic mechanical cues, or synergistic paracrine signaling. This last point, the communication between to cell populations, has been particularly studied to determine how one population can determine or regulate the development and 373 374 375 373 functionality of a second population . For example, Nguyen et al. used a tubular perfusion bioreactor to study the effects of downstream paracrine signaling on the osteogenic differentiation of mesenchymal progenitor cells. The position of cell- laden alginate beads within the axial chamber of the bioreactor relative to each other and relative to a bone morphogenic protein 2 (BMP-2) source significantly determined the gene expression levels of osteogenic markers alkaline phosphatase 373 (ALP) and BMP-2 . As studied here, cells in the bioreactor are affected by an upstream signaling source. Other bioreactors have been designed to study two-way communication between to populations by relying in adjacent co-culture in multi- chambered bioreactors. Orienting or redirecting flow between chambers has also been used to induce gradients of cells, oxygen, or biomolecules within the bioreactor in attempt to model paracrine signaling and communication. These types of bioreactors have been widely reported; some examples include a bioreactor that combined cyclic compressive stimulation with a two compartment system for the formation of glucose gradients in 378 articular cartilage explants ; a two-chamber modular bioreactor with an intermediate membrane to study intestinal epithelium, particularly solute transport across the cell 380 barrier, and the effect of flow and stress on the permeability of the barrier ; or a 108 two-chambered microfluidic device used to produce platelets from stem cell?derived 377 megakaryocytes by tuning the shear stress profile on the cells . These designs aim to model some variation of transport phenomena between two, or more, cell populations. 4.2.3. Bioreactor systems for the study of skin. Compared to other tissues such as cartilage or bone, bioreactors have been used in lesser extent for the maturation or study of skin. As we have described in previous sections of this document, growing dermal layers and modeling the epithelial dermal barrier has been extensively studied and has resulted in multiple methods that have been successfully scaled to industrial manufacturing proportions. Nevertheless, there are still some limitation that have been addressed using bioreactor technology, mostly the issues of vascularization and elasticity of skin, and the high demand for large surface area constructs for the clinic. The study of skin vascularization in bioreactors has greatly advanced with the 381 emergence of 3D bioprinting. Mori et al. developed an epidermis/dermis equivalent with a vascular system that developed endothelial tight junctions with constant perfusion in a bioreactor. The skin constructs grown in the bioreactor had higher 381 barrier function, and maintained greater thickness than non-perfused samples . Bioreactors have become a popular tool for vascularization as can provides constant flow of nutrients and oxygen through channel systems and induce shear stress, a 382 known promoter of endothelialization and vascularization . Consequently, the presence of dynamic flow is key to developing a full-thickness skin tissue with the required degree of vasculature for rapid in vivo integration. However, one challenge in maintaining the required air interface characteristic to the skin case is the scaffold 109 contraction during the process of maturation. This leads to inconsistent interface levels that could potentially hamper the tissue maturation process. Achieving a static interface requires advanced capabilities and active fluid level monitoring, which have not yet been fully explored. This leads to the fact that one of the key differences of skin with respect to other organs is that it can be accessed from outside the body, the air interface, which allows approaches such as using a bioprinter to extrude skin 383 grafts directly over a wound site , rather than having to mature a vascular system within the skin scaffold. This, in essence, means replacing the in vitro bioreactor for the ideal in vivo bioreactor of the body. Nevertheless, bioreactors have provided interesting advances in developing the elasticity of skin equivalents and increasing the size of the scaffolds produced. Although skin is a highly elastic tissue, it is not under constant or cyclic loads in contrast to other tissues such as bone, muscle, or cartilage. Early work by Atala et al. used bioreactor systems for the expansion of living skin matrices to increase the surface area of skin available for reconstructive purposes, while simultaneously demonstrating that the mechanical properties of native skin were not adversely 384 affected using this method . Following trials showed similar success in patient- 385 derived skin samples, promising a path for clinical applications . Interestingly, upon a 5-day stimulation of cultured skin constructs (epidermal keratinocytes and dermal fibroblasts on porous silicone sheets) via stretching, Tokuyama et al. observed that the constructs exhibited a thicker epidermal layer and higher expression of ECM proteins compared to non-stimulated controls. The basement membrane structure was also more developed, underlining the impact of mechanical stimulation on skin 110 386 physiology . Although promising, the materials typically used for skin tissue engineering ? mainly collagen, fibrin, gelatin ? are not mechanically durable in in vitro conditions, and therefore there is limited work done with the dynamic culture of skin constructs. Lastly, bioreactors have been studied as an option to develop large-scale manufacturing product of semi-mature skin tissue. The high degree of control that the bioreactor provides to the researcher over the system parameters and the cues that the tissue constructs are receiving could be an ideal control system for manufacturing. As 387 discussed by Wendt et al. , ideally, a fully automated bioreactor system could provide an ?aseptic, closed, standardized, operator-independent system? that could 368,387 improve the manufacturing process . This approach was implemented for Dermagraft? (Advanced Tissue Sciences); within a closed bioreactor bag dermal cells were cultured on scaffolds, matured, cryopreserved, and transported to the clinic. The bioreactor setup allowed each bag to hold eight grafts and twelve bags to be cultured in parallel, potentially producing up to 96 skin grafts in a single 368,387?389 setup . This system failed due to variability between the batches and high costs, but nevertheless demonstrated the viability of bioreactors for high-volume production of skin grafts. Overall, bioreactors have been implemented to address vascularization, flexibility, and manufacturing limitations of skin tissue equivalents. Nevertheless, it is important to state that there is no standardized bioreactor for skin and the systems used have been for epidermis/dermis constructs; no system currently addresses the interaction of these layers with the hypodermis layer or any other subdermal tissues. 111 Chapter 5: Development and Characterization of a 3D Printed, 5 Keratin-Based Hydrogel 5.1. Introduction Naturally-derived biomaterials have proven to be physiologically 390?393 biodegradable and allow adequate cell attachment, proliferation, and migration . These characteristics make this group a viable option for the development of a biocompatible support matrix that can provide an in vivo-like environment for cell 394 growth in regenerative applications . However, the cost and availability of these materials have hindered their widespread adoption in regenerative medicine. Keratin, a naturally-derived polymer, can be derived from a renewable resource typically 395,396 considered waste: the cortex of human hair . Still, keratin-based biomaterials have had limited use for tissue engineering, contrasting with its increasing use in drug 397,398 or growth factor delivery applications . Given that keratin can be derived from a low-cost source and has the ability to self-assemble or be crosslinked to form hydrogels, it has the potential for broad 399,400 impact on the field of tissue engineering . By self-assembling, keratin proteins form porous gels which have been well characterized in terms of their microstructures 401?404 and macroscopic properties . These gels are formed by disulfide bonding between the cysteine residues of the alpha and gamma keratin proteins, which have 405 high and low molecular weights respectively . As such, research has usually 5 As published in: JK Placone, J Navarro, GW Laslo, MJ Lerman, AR Gabard, GJ Herendeen, EE Falco, S Tomblyn, L Burnett, JP Fisher: ?Development and Characterization of a 3D Printed, Keratin- Based Hydrogel?. Annals of Biomedical Engineering, Vol. 45, No. 1, pp. 237-248, 2017. 112 focused on varying the alpha-to-gamma keratin ration to alter crosslinking density 406?408 and the resultant mechanical properties . Keratin-based materials have not been limited to just self-assembled hydrogels; keratin fibers, macro-porous scaffolds, and sheets have been used in applications such as bone, muscle, skin, and nerve 409,410 regeneration . In addition to intrinsic properties that allow keratin to become an easily manufactured hydrogel, keratin contains cell binding motifs such as Leucine-Aspartic acid-Valine (LDV), Glutamic acid-Aspartic acid-Serine (EDS), and Arginine- 411,412 Glycine-Aspartic acid (RGD) . LDV and EDS binding residues have been shown 412 to support cellular attachment ; RGD and LVD have been widely identified as binding motifs for integrin-ligand binding for proteins such as fibronectin, 413 412 osteopontin, and fibrinogen , and can also be found in keratin-based materials . As such, keratin biomaterials intrinsically contain these motifs thereby enabling cellular attachment and proliferation. Removing the need for additional post-processing to incorporate binding motifs, potentially decreasing the complexity of the manufacturing process, increases the appeal of keratin for tissue engineering applications. Furthermore, even with current 3D printing technology, the ability to manufacture complex geometries with high resolution using naturally-derived gels is 414 relatively limited . Thus, there is a need for a naturally-derived, biocompatible scaffold that can be manufactured via 3D printing rather than by traditional casting. The success of current casting strategies is limited by resolution of the molds. Each application requires a custom mold, which can drive up the cost of fabricating patient- 113 specific hydrogels, and repeated use of a mold wears out its resolution, critically limiting the accuracy of sequential samples. 3D printing, however, allows for the generation of custom hydrogels with specified geometries using CAD software and provides an avenue to rapidly produce multiple samples in parallel. This study investigates a novel method that allows for the fabrication of keratin-based hydrogels using an off-the-shelf, commercially available 3D printer. The form of keratin used in this work is the oxidized form; thus, the cysteine groups have been chemically modified. Due to this modification, the alpha and gamma 397 keratins cannot form disulfide bonds to strengthen the resultant hydrogel . Only entanglement of keratin and any crosslinks formed during the 3D printing process are expected to contribute to the mechanical integrity of the hydrogels. We hypothesized that a photosensitive initiator-catalyst-inhibitor solution could be coupled with keratin to produce complex 3D constructs via UV crosslinking in a lithography-based 3D printer. The viability of this thesis was studied with a factorial experimental design that allowed us to determine the adequate keratin printing resin formulation, and further assessed by comparing the properties of the printed hydrogels with those of traditionally casted keratin samples. After, the structural and biocompatible characteristics of the printed keratin were tested to assess the viability of the method in a tissue engineering application. 114 5.2. Methods 5.2.1. Keratin preparation Oxidized keratin was prepared using a process developed and patented by KeraNetics, LLC (Winston-Salem, NC). Briefly, human hair was oxidized with peracetic acid, and proteins were extracted with Tris and ultrapure water. Protein extracts were purified with filtration and dialysis. Purified extracts were lyophilized and sterilized with gamma ray irradiation. The resulting intermediate bulk was stored at -20 ?C. Although this oxidized form is often called keratose, we will refer to this as keratin for consistency in the manuscript. 5.2.2. Resin formulation A combination of alpha and gamma keratin (95:5) was dissolved overnight in phosphate buffered saline (PBS) at a concentration of 4% wt/vol. The accompanying photosensitive initiator-catalyst-inhibitor solution was composed of riboflavin (Sigma-Aldrich Co., St. Louis, MO), sodium persulfate (SPS), and hydroquinone (Sigma-Aldrich), respectively. Factorial designs were utilized to determine the appropriate formulation; a first design varied catalyst (70-160 mM) and initiator (1-5 mM) without inhibitor and a second varied catalyst (70-200 mM) and inhibitor (0.0002-0.0014% wt/vol) using initiator at a concentration of 1 mM. Each formulation was prepared in a dish, half of each covered with aluminum foil, and crosslinked in an ultraviolet (UV) box for 20 min total in 5 min increments. The covered half, unable to photo-crosslink, allowed observing the extent of undesired crosslinking under the foil. Samples were evaluated by the depth of crosslinking, 115 relative strength of gelation, the extent of undesired crosslinking, and the clarity or resolution of crosslinking at the edge-resin interface. 5.2.3. Keratin printing The formulated resin, as detailed in the Results section, was used to fabricate 4%, 5%, and 6% wt/vol keratin hydrogels using a lithography-based EnvisionTEC Perfactory 4 DLP printer (EnvisionTEC Inc., Dearborn, MI). Approximately 100 mL of resin were used in an M-type resin tray, modified with PTFE strips to reduce the 2 resin volume in the build area. UV light intensity was set to 375 mW/dm , with an exposure time of 240 s per 100 ?m layer. Printed samples were collected and stored in PBS for 24h at 4 ?C. 5.2.4. Keratin casting A 3D printed mold was used to cast 4%, 5%, and 6% wt/vol keratin hydrogels using the formulated resin. The mold was a 5 mm thick rectangular block with 10 circular holes, 5 mm in diameter, passing completely through. A glass slide was clamped to the bottom of the mold, 98.2 ?L of keratin resin were added to each well, and then a second glass slide was clamped onto the top. The setup was allowed to crosslink in a UV Box for 210 min to replicate the amount of energy introduced to the samples during printing. After, the top of the mold was uncovered and the entire apparatus was soaked in PBS (pH 7.4) overnight. A negative complement to the original mold was printed such that cylinders matching the inner dimensions of the mold could fit inside the holes and push the casted samples out. Casted samples were collected and stored in PBS for 24h at 4 ?C. 116 5.2.5. Resolution assessment To evaluate the resolution of keratin prints, squared and circular geometries were designed in CAD and printed using 4% and 6% keratin resin. The following geometries were attempted (n=6 each): cylinders (diameter x height in mm): 1x1, 2x2, 4x2; cylinder pyramids which stack 4x0.5, 2x0.5, 1x0.5, and 0.5x0.5 steps; cubes (length x width x height in mm): 1x1x1, 2x2x2, 4x4x2; and cube pyramids which stack 4x4x0.5, 2x2x0.5, 1x1x0.5, and 0.5x0.5x0.5 steps. After 24 h in PBS, each sample was placed on a glass slide and submerged in PBS droplets, which allowed imaging using a fluorescence microscope (Axiovert 40CFL, Zeiss, Thornwood, NY) fitted with a digital camera (SPOT Insight 1120, Diagnostics Instruments, Sterling Heights, MI) at 2.5x. Sample dimensions were measured using ImageJ software and compared against the designed CAD values, reporting the percent difference. 5.2.6. Mechanical testing Mechanical tests were carried out using a Q800 DMA (TA Instruments, New Castle, DE) and an Instron 5565 with a 50N load cell (Instron, Norwood, MA) to compare the compressive moduli of printed and casted samples. Cylinders 5 mm in diameter and 5 mm in height were printed and casted using 4%, 5%, and 6% resin. All samples were preloaded with 0.001 N prior to mechanical testing. For tests with the Instron, the compression rate was fixed at 1 mm/min, and printed and casted samples (n=6 each) were compressed until yielding. The linear regime of the stress 2 strain curve was determined with a linear fit (R ? 0.95) to obtain the compressive modulus. Two conditions of dynamic mechanical analysis (DMA) testing were done 117 on printed samples (n=6): 1) strain sweep at a fixed 1 Hz frequency to determine the stress strain curve and the compressive modulus, and 2) dynamic loading with a frequency sweep to determine the storage (G?) and loss (G??) moduli, and the phase angle (tan(?)). 5.2.7. Swelling degree and uptake capacity To compare the saturated swelling capacity of printed and casted samples, 4%, 5%, and 6% keratin resin was used to obtain samples with each method (n=3). Samples were oven-dried at 37?C and the dried masses were recorded (Md); after, samples were fully rehydrated in PBS for 24 h at 4 ?C and the swollen masses were recorded (Ms). The swelling degree (QD) was calculated as the difference between Md and Ms, normalized over the initial Md. To study the time frame during which the hydrogels uptake and saturate with solution, printed 4% keratin samples were flash- frozen in liquid nitrogen for 20min and then lyophilized in a FreeZone Lyophilizer (LABCONCO, Kansas City, MO) at a 0.018 mBar vacuum and -48 ?C for 48h, after which the dried masses were recorded. Samples were then rehydrated in 100 times their dried mass of either Minimum Essential Medium (MEM) (Life Technologies, Frederick, MD) or PBS (n=5 each) for a period of 5 d. Mass of the samples was measured at 1, 3, 5, 10, 15, 20, 30, 45, and 60 min, completing the detailed uptake during the first hour, and then at 1.5, 2, 2.5, 3, 24, and 124 h. The swollen mass was normalized against the initial dried mass to obtain the relative uptake of PBS or MEM over time. 118 5.2.8. Crosslinking density approximation The crosslinking density (Ve) of printed 4%, 5%, and 6% keratin samples was estimated using a previously established relationship for keratin hydrogels 406 crosslinked via disulfide bonds . This equation uses the compressive modulus (K), the universal gas constant (R), temperature (T), and the swelling ratio (Q) based upon the density of keratin (?k) and the PBS (?s), and the swollen (Ms) and dry (Md) masses of the scaffolds: ? ?? ?? ?? = 1 ? = 1 + ( ) ( ? 1) ??? ?3 ?? ?? 5.2.9. Cytotoxicity analysis Based on ISO Standard 10993-5, the cytotoxic response of the printed keratin was evaluated by direct contact and conditioned media tests. Mouse fibroblasts (L929s) (ATCC, Manassas, VA) were seeded in 24-well plates per the manufacturer?s protocol using Minimum Essential Medium (MEM) (Life Technologies, Frederick, MD) supplemented with 10% horse serum (HS)(ATCC), and allowed to reach ?75% confluence. For the direct contact test, printed 5x5 cylinders of 4% keratin (n=9) were placed directly on top of the confluent cell layer. For the conditioned media test equal cylinders (n=9) were incubated with MEM+HS at a concentration of 0.1 g/mL at 37 ?C for 24 h under dynamic conditions, after which it was used to replace the media on the confluent cells. Control groups (n=9 each) included non-cytotoxic high density polyethylene (HDPE) (McMaster-Carr, Elmhurst, IL), Blank (MEM+HS only), Live (cells cultured under normal conditions), and Dead (cells incubated with 70% ethanol for 20 min prior to XTT assay). All groups were incubated at 37 ?C and 5% CO2 for 119 24 h after which cell metabolic activity was quantified with XTT assays (Cell Proliferation Kit II - XTT, Roche, Mainheim, Germany). Briefly, XTT labeling and electron coupling reagents were thawed and mixed (50:1 ?L ratio). Once samples were removed from the wells, XTT labeling mixture was added in with fresh MEM+HS (1:2) and incubated at 37 ?C and 5% CO2 for 4 h. After, absorbance was measured using a M5 SpectraMax plate reader (Molecular Devices, Sunnyvale, CA), using the 475 nm wavelength for the peak value and 690 nm as reference. Specific absorbance was calculated as (Peak475 ? Reference490 ? Blank475) and normalized to the mean of the Live group, obtaining the relative cell metabolic activity (%). 5.2.10. Live/Dead staining To visualize the viability of cells cultured on printed keratin, Live/Dead solution was prepared with 4 ?M of calcein AM (Invitrogen, Carlsbad, CA) and 2 ?M of ethidium homodimer (Invitrogen) in PBS, as detailed by the manufacturer. L929 cells were cultured: 1) as a 4% keratin conditioned media assay, and 2) directly on top of the printed 4% keratin scaffolds inside non-treated culture wells encouraging cells to remain on the hydrogels. After the culturing period, samples were rinsed with PBS and incubated with the Live/Dead solution in the dark for 20 min prior to imaging. Control groups included cells cultured under normal conditions on tissue culture polystyrene (TCPS), and an aggressive cytotoxic control incubating cells in 70% ethanol for 20 min prior to staining. Images were obtained with a fluorescence microscope (Axiovert 40CFL, filter set 23, Zeiss, Thornwood, NY) fitted with a digital camera (SPOT Insight 1120, or SPOT Idea 2920, Diagnostics Instruments, Sterling Heights, MI). 120 5.2.11. Microstructure observations SEM micrographs were obtained using an ESEM FEI Quanta 200F (FEI, Hillsboro, OR) to observe the porous structure of printed keratin. Prior to imaging, printed 4%, 5%, and 6% samples were hydrated overnight in PBS, flash-frozen in liquid nitrogen for 20 min, and lyophilized (0.018 mBar, -48 ?C, 48 h). After, samples were coated for 10 seconds with Au-Pd. Image acquisition dwell time was 10 ?s at 10kV or 300 ns at 5kV. 5.2.12. Statistical analysis Statistics for quantitative tests were performed using ANOVA and Tukey?s multiple pairwise comparison (p < 0.05 for significance) unless otherwise stated. Values reported are mean ? standard deviation, and significant differences are specified in figures and tables. 5.3. Results The goal of this work was to formulate a viable keratin resin and use it for 3D printing on an off-the-shelf, commercially available lithography-based 3D printer. As hypothesized, the oxidized form of keratin would not self-assemble into a hydrogel network and instead needed to be coupled with a photosensitive initiator-catalyst- inhibitor solution to gel via UV crosslinking. The proportions of a riboflavin-SPS- hydroquinone (initiator-catalyst-inhibitor) solution were determined with factorial combinations. As summarized in Table 5.1, the first combination of factors showed an increased crosslinking quality using 160 mM SPS with 1 mM riboflavin without 121 an inhibitor. Larger concentrations of riboflavin would result in subpar crosslinking, with low or no depth or strength of gelation. Then, 1 mM riboflavin was fixed in a second combination of factors that resulted in better crosslinking using 0.0008% hydroquinone and 200 mM SPS. In this second assessment, as expected, as crosslinking time increased so did the extent of undesired crosslinking. To account for this undesired crosslinking, the exposure time per layer was reduced to 4 minutes and the inhibitor concentration was increased to improve the X-Y resolution. The final formulation for the print resin was determined to be keratin (concentration ranging from 4% to 6% wt/vol) dissolved overnight in PBS (pH 7.4), 1 mM riboflavin, 200 mM SPS, and 0.001% wt/vol hydroquinone. Table 5.1. Formulation design of the keratin resin The first combination varied catalyst (SPS) and initiator (riboflavin) without an inhibitor. The second combination varied catalyst (SPS) and inhibitor (hydroquinone) using 1mM riboflavin initiator as determined from the first combination. Samples were tested for 20 min in 5 min increments. At each time point, qualitative observation of the depth of crosslinking, relative strength of gelation, the extent of undesired crosslinking, and the clarity or resolution of crosslinking at the edge-resin interface, contributed to the quality ranking of crosslinking. Time points are shaded red (poor) to yellow to green (best) to indicate quality of crosslinking. White boxes indicate that no crosslinking occurred. * denotes the best concentration pair due to the highest quality of crosslinking throughout all time points. Fine tuning of the inhibitor concentration and exposure time per layer was completed on the 3D printer to. 122 Using the formulation 4%, 5%, and 6% wt/vol keratin was successfully printed and casted, as seen in Figure 5.1. The UV exposure time during printing was 2 verified by testing resin crosslinking at 375 mW/dm for varying amounts of time. A 100 ?l drop of resin was dropped onto a glass slide and exposed to UV from the printer for 1-7 min; after each trial, the integrity of the sample was assessed and the crosslinked thickness measured, determining 4 min as the optimum exposure time. During printing, samples are lifted from the resin tray and exposed to air, causing the scaffolds to be dehydrated by the end of the process; the same phenomenon was seen in casted samples, where the exposed resin surface in the mold came out concave (Fig. 5.1). Both phenomena were reversed by hydrating the samples in at least 100 times their mass in PBS volume for 24h at 4 ?C, after which both printed and casted scaffolds became fully swollen and closely matched the design dimensions. The rehydration process was critical to not only match dimensions but also to rinse out any soluble resin not crosslinked. Cubical and cylindrical 3D geometries were designed in CAD and printed using 4% and 6% keratin to determine the smallest features that could be obtained with the formulated resin, as seen in Figure 5.2. Cubes (Fig. 5.2A-D), cylinders (Fig. 5.2E-J), and pyramids (Fig. 5.2K-L) with features as small as 1 mm were obtained by printing 4% resin, as well as cylinders using the 6% formulation. Although 1x1 cubes were also achieved with 6% resin, during imaging analysis it was impossible to discern straight edges or defined right angles that brand the larger cube or pyramid geometries. These observations, coupled with the significant difference (p < 0.05) between the measured dimension of 4% and 6% resins as summarized in Table 5.2, 123 indicates better results could be expected when using 4% resin for complex geometry prints. In most cases, the average printed dimensions of the cross-sections (XY measurements) were smaller but not significantly different (p < 0.05) from the design values. Nevertheless, the height component of the prints (Z) brought in significant variation (p < 0.05) in comparison to the design, suggesting the layer-layer crosslinking of the lithography-based print should be further studied. Figure 5.1. Printed and casted keratin hydrogels. Representative images of casted and printed keratin hydrogels (5 mm diameter and 5 mm height) using: A) 4, B) 5, and C) 6% wt/vol keratin resin. Printed samples show defined flat surfaces, while casted samples present concave upper surfaces due to the meniscus of the liquid resin in the mold during crosslinking. D) Comparison of height of printed versus casted hydrogels (n=6). Measurements were taken to the bulk of the hydrogel due to issues with the formation of a meniscus on the casted samples. Statistical significance (p<0.05) between fabrication methods is denoted with an *. 124 Figure 5.2. Resolution assessment of printed 4% and 6% keratin. Typical images of the different geometries printed with 4% and 6% keratin resin to assess resolution. (A-D) Cube sample dimension are noted as length x width x height (mm). (E-J) Cylinder sample dimensions are noted as diameter x height (mm). There was low success obtaining squared geometries with right angles using 6% keratin; there was no discernable difference between cubes and cylinders of the same dimensions, thus the lack of larger cube or pyramid samples using this resin. (K-L) Pyramid samples with cube and cylinder steps; the red boxes highlight the steps obtained, indicating proper printing of multiple size features within the same samples. All images show the cross-section (XY plane) of samples under 2.5x magnification. Table 5.2. Comparison between printed keratin samples and design dimensions Key 4% keratin 6% keratin design Geometry Printed Printed Printed Printed dimension dimension on dimension on dimension on dimension on [?m] XY [?m] XYZ [?m] XY [?m] XYZ [?m] Cylinder 1000 932.4 ? 51.5 NA 914.2 ? 204.9 NA 1x1 Cylinder # 2000 1903.4 ? 175.5 1764.4 ? 248.5* 1698.1 ? 160.9* 1594.3 ? 208.7* 2x2 Cylinder # 4000 3407.1 ? 272.4* NA 3797.8 ? 263.6 NA 4x2 Cube 1000 1003.1 ? 98.3 NA 1088.0 ? 7.3 928.7 ? 276.0 1x1x1 Cube 2000 1891.8 ? 186.9 1694.5 ? 327.3* NA NA 2x2x2 Cube 4000 3237.7 ? 240.8* NA NA NA 4x4x2 Dimensions of the samples were measured from 2.5x magnified images (as in Fig. 5.2) using ImageJ software and compared against the designed CAD values. The UV mask is projected on the XY plane, and the height dimension Z is obtained as the sample is pulled out from the resin tray. Even with accurate cross-sections (as measured on XY), samples may shear, losing layers and height, which results in larger XYZ deviations from the design or incomplete samples (NA). Statistical significance (p<0.05): # significant difference between 4% and 6% samples for a specific geometry; * significant difference between printed dimension and design value (t-test for the mean, p<0.05). 125 The printable keratin resin was further assessed by comparing the properties of the printed hydrogels to those of traditionally casted keratin samples. The compressive modulus was determined for 4%, 5%, and 6% printed and casted scaffolds in an Instron compressive test; the results are shown in Table 5.3. The compressive modulus increased with increasing concentration of keratin, but there was no significant difference (p < 0.05) between 5% and 6% samples using either printing or casting methods. In all cases, the compressive moduli were significantly (p < 0.05) higher for all printed samples in comparison to those casted at a given weight percent. The resultant swelling degree for the three resin formulations also showed the same significant difference (p < 0.05) between printing and casting, although the values were always higher for casting samples (Table 5.3). Table 5.3. Compressive modulus and swelling degree of printed and casted keratin 4% keratin 6% keratin Method 4% * 5% * 6% * 4% * 5% * 6% * 3D # + + 5.5 ? 0.7 14.8 ? 3.8 15.5 ? 2.8 16.8 ? 0.4 20.0 ? 1.7 18.6 ? 0.4 Printing # # Casting 3.1 ? 1.1 6.8 ? 2.4 6.9 ? 0.8 31.6 ? 3.8 32.7 ? 0.5 21.9 ? 0.3 Comparison of compressive modulus (from Instron data) and swelling degree between printing and casting methods for 4%, 5%, and 6% wt/vol keratin resins. Each property is presented with the mean value ? standard deviation (n=6 for compression moduli, n=3 for swelling degree). Statistical significance (p<0.05): * denotes significant difference between printing and casting methods; # denotes significant difference between the value for a specific formulation and the other two; + denotes significant difference between the two groups marked. Using the established equation for keratin hydrogels crosslinked via disulfide 406 bonds , the crosslinking density was estimated for each of the three printing resins based off of the measured swelling degree and mechanical properties. The apparent 126 crosslinking density increased with keratin concentration; for the 4% keratin gels, the 3 crosslink density was calculated as 1.17 ? 0.14 ?mol/cm , which was almost tripled 3 3 in the 5%, 3.02 ? 0.78 ?mol/cm , and 6% samples, 3.21 ? 0.58 ?mol/cm . The DMA frequency sweep was carried out on each of the samples with an applied stress within the linear regime as determined during the compression tests. For all three of the concentrations of keratin in the printing resin, the storage modulus (G?) was approximately linear across the frequencies observed (0.1-10 Hz), as shown in Table 5.4. Although there were some significant differences (p < 0.05) due to frequency (especially at 10 Hz) in the loss modulus (G??) and phase lag (tan(?)), none were observed for the properties between the different resin formulations when compared at 1Hz. Table 5.4. Dynamic mechanical analysis of printed keratin 4% keratin 5% keratin 6% keratin Dynamic Moduli 0.1 Hz 1 Hz 10 Hz 0.1 Hz 1 Hz 10 Hz 0.1 Hz 1 Hz 10 Hz 9.7 ? 9.8 ? 6.0 ? 12.0 ? 12.1 ? 8.6 ? 11.5 ? 12.3 ? 12.8 ? G? [kPa] 1.0 1.2 0.6 * 3.6 3.8 2.8 1.8 2.0 1.3 G?? 0.4 ? 0.6 ? 3.9 ? 0.6 ? 0.9 ? 4.0 ? 0.8 ? 1.3 ? 3.5 ? # [kPa] 0.1 0.1 0.8 * 0.1 0.0 1.2 * 0.1 0.1 0.8 * 0.04 ? 0.06 ? 0.66 ? 0.05 ? 0.08 ? 0.47 ? 0.07 ? 0.11 ? 0.27 ? tan(?) 0.01 0.01 0.07 * 0.02 0.03 0.07 * 0.02 0.03 0.0.4 * Dynamic loading results of 4%, 5%, and 6% printed keratin samples. Samples were loaded at a fixed strain within the linear regime determined from the compressive moduli in Table 3, with a preload of 0.001 N over a frequency sweep of 0.1 to 10 Hz. Values are presented with the mean value ? standard deviation (n=6). Statistical significance (p<0.05): # denotes significant difference between the value for a specific formulation and the other two; * denotes significant difference between the value at a specific frequency and the other two. 127 The PBS and MEM uptake behavior of printed keratin over time is shown in Figure 5.3A. It was observed that keratin hydrogels will uptake and saturate after approximately 20 min in both PBS and MEM. The relative uptake of PBS was higher than that of MEM, comparing the maximum uptake levels of 1773.8 ? 223.3% and 1581.9 ? 94.1% respectively. This behavior was detailed during the first hour, and the saturation plateau remained throughout the 5 d assay, as observed in the insert graph (Fig. 5.3A). Figure 5.3. Media uptake capacity and cytotoxicity of printed keratin. A) Uptake capacity of the 4% printed keratin hydrogels. Samples were lyophilized and then rehydrated in PBS or MEM (n=5 each) over a period of 5 d. During this period, the increasing mass of each sample was measured and normalized against the corresponding initial dried mass to obtain the relative uptake. The main graph shows the detailed uptake during the first hour, the insert graph shows the extended uptake over 5 d. B) Cytotoxic evaluation of the 3D printed keratin hydrogels by direct contact (DC) and conditioned media (CM) tests. L929 cells were cultured accordingly with keratin and a HDPE non-cytotoxic control and compared to those cultured under normal conditions in growth media (Live) and those treated with ethanol (Dead) (n=9, for all study and control groups). No statistical differences were found between Live, Keratin, and HDPE groups, but all three were different to the Dead group (Denoted *, p < 0.05). Due to the increased resolution capacity and superior mechanical properties, the microstructural and biocompatibility characteristics of the printed 4% keratin 128 scaffolds were tested to assess the viability of the method in cell culture applications. Cytotoxicity was assessed using direct contact (DC) and conditioned media (CM) tests on L929 cultures. The metabolic activity of L929s exposed to the keratin was found to be statistically not different to those in non-cytotoxic control (HDPE) and live groups as illustrated in Figure 5.3B. When exposed to keratin under DC, the relative cell metabolic activity was 96.3 ? 28.8%, and 90.9 ? 46.5% in response to exposure to keratin CM. These levels of metabolic activity were similar to those measured when the cells were exposed to HDPE, a material suggested for use per ISO Standard 10993-5 due to its proven non-cytotoxicity. The similarity between keratin and HDPE behavior is extended to the positive control live group, as no significant difference (p < 0.05) was recorded between the three groups. Significant differences (p < 0.05) were found in comparison to the negative control dead group. Live/Dead stains on the printed 4% keratin allowed the identification of cell proliferation throughout the scaffolds. As observed in Figure 5.4D, L929 cells can be seen alive on the hydrogel, a viable behavior compared to both live (Fig. 5.4A) and dead (Fig. 5.4B) controls in TCPS. The effect of leachable products, assessed with CM (Fig. 5.4C), also resulted in a majority of live cells with a minimum of dead cells comparable to the behavior of the live positive control. SEM imaging revealed honeycomb-like structures with open pores ?30-10 ?m in diameter (Fig. 5.4E), a size large enough for cells to fit through. This fact, along with the large amount of exposed surface area observed (Fig. 5.4F-G), implies the possibility of cell adhesion and migration throughout the keratin hydrogels. 129 Figure 5.4. Cytotoxicity and microstructural imaging of printed 4% keratin. (A-D) Live/Dead staining of L929 cells on: A) Positive control on TCPS. B) Negative control on TCPS. C) Cells with conditioned media. D) Cells seeded directly onto keratin scaffold. Dotted line represents edge of the scaffold. (E-G) SEM micrographs of the scaffolds. Hydrogels were dialyzed for 48 h with excess PBS, followed by lyophilization, sputter-coating with Au-Pd, and imaged to reveal a porous honeycomb-like structure. 5.4. Discussion To address the growing need for functional, naturally derived materials in tissue engineering and regenerative medicine, we have successfully 3D printed keratin-based hydrogels using additive manufacturing. To the best of our knowledge this is the first time keratin has been successfully 3D printed by formulating a photo- crosslinkable resin and using it in a UV lithography-based 3D printer. A recent 159 review on the advances of skin tissue engineering by Singh et al. adequately summarizes how current research leans towards bioprinting by extrusion for the 130 development of cell-laden (mainly fibroblasts and keratinocytes) scaffolds. These systems, with remarking resolutions as small as 1.57 ?m, and the more advanced laser assisted bioprinting systems, have managed to produce constructs with high geometrical complexity but limited to a handful of materials, mostly fibrin and 159 collagen I combinations . Keratin has no reported applications using 3D printing systems, but is used in a wide range of manufacturing protocols based on photo- 415?419 crosslinking and self-assembling casting . As presented before, casting strategies are limited by mold resolution, which in turn wears out due to sequential use. In the case of casting hydrogels, typically only the dimensions of the mold are reported and any micropatterning is limited to the mold parameters and the ability to remove the casted hydrogel from the mold without damage. Our approach has managed to produce scaffolds significantly close to the design dimensions (p < 0.05) with defined features of ?1 mm, using the preferred 4% keratin resin. Further efforts will be aimed at achieving greater resolution and refining the layer-layer crosslinking on the height (Z) dimension. These initial studies have shown that, using our printable keratin resin, we can obtain scaffolds with consistent reproducible dimensions, high resolution, and improved mechanical properties compared to those manufactured using traditional casting methods. In addition, we also achieved the ability to alter the crosslinking density within the hydrogels by modifying the keratin content in the printing resin. By controlling the crosslinking density, we are able to create scaffolds that are handleable, with compressive modulus in a range of 5.49 to 15.45 kPa. The modulus increased by almost a factor of three between the 4% and 6% keratin hydrogels, as did the calculated crosslinking density. However, it appears that the 5% and 6% 131 keratin hydrogels only have a small change in compressive modulus and in the calculated crosslinking density. This small difference may be due to steric hindrance of the reactive groups and their inability to come into close enough contact with other reactive species once a critical threshold was reached (Please see Sando et al. and 416,420 Fancy et al. for details on similar photo-crosslinking mechanisms ). Furthermore, as the hydrogels are printed, the keratin fibers can become entangled, thus, limiting their mobility and decreasing the probability of forming a crosslink with a neighboring keratin fiber. Previous studies have reported the mechanical properties of similar keratin casted hydrogels. Covalently crosslinked alpha-keratose scaffolds (via a photo-oxidative process catalyzed by blue light, a ruthenium complex, and sodium 416 persulfate) have tensile elastic modulus ranging from 9.5 to 53 kPa ; self-assembled keratin-collagen porous scaffolds report mechanical strength of 10.84 MPa, compared to the 50 kPa and 8.84 MPa strength of pure keratin and collagen scaffolds 417 respectively . Importantly, we observed a decrease in our casted samples mechanical strength when compared to literature values. We hypothesize this is due to the presence of the inhibitor in the resin solution used in the fabrication process thereby reducing the degree of crosslinking. Our scaffolds have mechanical properties similar to those reported for casted samples of self-assembled and crosslinked keratin, albeit using a new 3D printing approach. The frequency dependence at 10Hz shown in the DMA data also points to a successfully crosslinked network within the hydrogels. Also, the phase lag values between zero and one (0 ? tan(?) ? 1) match to the viscoelastic behavior of hydrogels 418 and indicate the formation of a stable hydrogel network . This behavior is 132 418 comparable to that observed for successfully self-assembled casted keratin . All three resins tested had storage modulus within 6.0 to 12.8 kPa; self-assembled casted keratin with suspended living cells has been reported to have storage moduli ranging 419 from 3 to 800 Pa, analog to soft tissues such as the brain . The mechanical properties obtained, in ample ranges that can be tuned with keratin concentration, support the viability of our printed scaffolds in tissue engineering applications. Successful crosslinking during printing results in consistent honeycomb-like structures, as observed by SEM imaging, that allow high uptake of solutions, a characteristic of great interest for drug or growth factor uptake and delivery, as well as in vitro or in vivo cell adhesion, proliferation, and migration. Printed keratin hydrogels uptake, both PBS and MEM, and saturate in 20 min. This saturation was sustained throughout a 5 d observation period without any qualitative changes to the scaffold?s dimensions or consistency, hence longer periods in this stable state can be expected. The uptake levels of PBS and MEM that are reached in 20 min, 1773.83 % and 1581.9 % respectively, highly exceed those of casted keratin. Alpha-keratose scaffolds reach equilibrium swelling at 26.4% after 24h in PBS, although these showed tendency to shrink down to -11.7 % (applicable for particular cases where 416 swelling highly undesirable, e.g. pressure-sensitive tissues such as nerves) . Similar hydrogels such as casted gelatin hydrogel in PBS swelled to 100% of its initial 415 volume in 4 h and to 240% in 24 h ; keratin?collagen scaffolds swell to 803% in 417 water, and the value drops to 7-8.5% at higher pH . The superior uptake and swelling properties of printed resin, particularly the 4% wt/vol keratin formulation, 133 could have a great impact in drug or growth factor uptake and delivery applications, as well as for cell culturing in tissue engineering and regenerative cases. Although the biological potential of keratin is known, the cytotoxicity assays allowed us to understand the impact of our hydrogel manufacturing process on cellular viability in vitro. The DC and CM variations of the cytotoxicity assay model the possible interactions of the printed samples with the cells. The direct contact test allows for the physical interaction of the material and the cells, and the conditioned media test evaluates the cytotoxicity of any leachable byproducts from the material by the simulation of clinical application. In all cases, the 3D printed hydrogels proved to interact with the cells in a manner comparable to previously identified non-cytotoxic material, HDPE. Specifically, cell reactions to the constructs did not present significant differences in metabolic activity when compared to the effects of HDPE or even to the unmodified cell monolayers of live controls. At the same time, keratin was shown to support a level of cell metabolic activity that was significantly different from an aggressive cytotoxic treatment, such as ethanol rinsing. These results match the widely reported ability of keratin-based materials to support different cell lines; these include myoblasts on casted keratin within the context of a flowable and 418 419 injectable system , L929s on cell-loaded casted keratin hydrogels , NIH/3T3 416 fibroblasts on alpha-keratose scaffolds , and RT4-D6P2T Schwann cells on keratin- 399 coated glass slides . These tendencies, coupled with the behavior observed in the Live/Dead stains, show that our proposed keratin-based printing resin and printing methodology have minimum cytotoxic interaction with different cell lines, and the 134 resultant hydrogels behave as well as other keratin-based constructs or commercial non-cytotoxic materials, such as HDPE. Future endeavors to optimize the printed keratin scaffolds can investigate a range of alpha and gamma keratin concentrations to help tune the mechanical properties. Additionally, kerateine, the unoxidized form of keratin, could be used to allow for additional crosslink formation. Although this will increase random crosslinking, the unoxidized cysteine residues could allow for the functionalization of the hydrogel and potentially increase the mechanical strength of the biomaterial. Drugs, growth factors, or other cytokines could be covalently linked to the keratin substrate, thereby enabling the hydrogels to not only have specific architecture and mechanical properties for specific cell cultures, but to play an active role through the release of these molecules. Future applications will need to take advantage of the platform developed here create cell line-specific 3D hydrogels with physiologically and clinically relevant structures and properties. The novel keratin-based printing resin and printing methodology presented have the potential to impact future research by providing an avenue to rapidly and reproducibly manufacture hydrogels with complex shapes without the need to generate molds each time. For instance, previous research on axon guidance utilizing keratin as a substrate could be impacted by the ability to generate scaffolds on a case- 409,410 by-case basis . Other research into the use of keratin hydrogels for bone, muscle, and skin regrowth can make use of this material to tune the mechanical properties while eliminating the need to generate trial-specific scaffolds. In each case, only a keratin concentration, with a set of UV photoinitiators and inhibitors, and an exposure 135 setting will need to be selected for each application to control the crosslinking density and the physical properties of the printed hydrogel. Thus, this methodology may aid in the translatability of this research to clinical applications where the aspect ratios and dimensions will depend on the individual patient. Not only is this methodology appropriate for the development of keratin hydrogels, but it can also be expanded to other extracellular matrix molecules that have appropriate chemistry for the crosslinking using a UV photoinitiator. For example, fibronectin and laminin could be used to manufacture 3D matrices using this same technology/methodology for cell growth. These 3D environments could then be used to research the effects of 3D microenvironments on cell culture. The development of a novel fabrication strategy for a low cost, naturally-derived hydrogel can be applied readily to current research endeavors to address the need for a wider variety of hydrogels in tissue engineering and regenerative medicine and can be expanded to other research avenues which wish to utilize other natural ECM molecules. 136 Chapter 6: In Vivo Evaluation of a 3D Printed, Keratin-Based 6 Hydrogels in a Porcine Thermal Burn Model 6.1. Introduction The development of scaffolds and matrices that simulate the structures and 412 functions of human tissues is a primary focus in the field of tissue engineering . Common materials used in the fabrication of these scaffolds include ceramics (e.g., hydroxyapatite, tri-calcium phosphate), synthetic polymers (e.g., polystyrene, 421 polyglycolic acid), and natural polymers (e.g., alginate, chitosan) . Protein-based materials have become attractive natural polymer options for tissue engineering and regenerative medicine (TERM) applications. These molecules can provide stable three-dimensional structures that facilitate cell attachment, proliferation, and 390?393 migration . Of note, keratin proteins and keratin-based materials are a viable option for tissue engineering scaffold development due to their biocompatibility, 412,422 structural characteristics, and abundance as a renewable resource . Keratins are a family of insoluble proteins found in epithelial tissues as intermediate filament proteins. These proteins can also be easily obtained from 397,412 epidermal structures such as feathers, hooves, wool, and human hair . Keratin proteins exhibit the ability to self-assemble or be crosslinked to form porous, fibrous 401?403,412,422 hydrogel scaffolds . Keratin-based biomaterials have been reported in 86,87,397,399,418,423?428 TERM applications for nerve, muscle, skin, and bone . Keratins 6 As prepared for: J Navarro, RM Clohessy, RC Holder, AR Gabard, GJ Herendeen, JP Fisher, RJ Christy, and LR Burnett: ?In Vivo Evaluation of a 3D Printed, Keratin-Based Hydrogels in a Porcine Thermal Burn Model?. 137 also contain motifs important for cellular attachment, such as leucine-aspartic acid- 413,429,430 valine (LDV) and glutamic acid-aspartic acid-serine (EDS) residues . Keratins 431 have been shown to promote cellular attachment of adipose-derived stem cells , 397,432 433 399 419 osteoblasts , hepatocytes , neural cells , and fibroblasts . Another desirable characteristic of keratin biomaterials is their functionality as tunable vehicles for small molecule or cell delivery to treat dermal wounds. Release of these entities into the wound space can improve healing outcomes via mechanisms such as signal cascade initiation or inhibition and elimination of potential microbial wound contamination. Research has demonstrated that keratin biomaterials have been successfully used for the delivery of growth factors (e.g., insulin-like growth factor-1 434,435 (IGF-1), basic fibroblast growth factor (bFGF)) , progenitor cells (e.g., skeletal 418 436 muscle myoblasts) , and antimicrobial agents (ciprofloxacin) , among other molecules and cells. For this work, we study the uptake and realease of Halofuginone from a keratin-based carrier. Halofuginone was selected for these studies based on its demonstrated effectiveness in preventing abnormal fibrillar collagen accumulation in 437?440 pathologies with fibrosis and wound contracture . Halofuginone is an FDA approved collagen I synthase inhibitor and has been shown to decrease collagen synthesis by inhibiting transforming growth factor beta-dependent Smad3 439 phosphorylation . Keratin can be extracted from the cortex of human hair using reductive or 405 oxidative chemistry . Reduced keratin can self-assemble due to the formation of disulfide bonds between cysteine amino acids in the protein chain; oxidation, on the other hand, causes capping of the cysteine groups and greatly decreases the self- 138 397 assembling capacity . We reported the formulation of a keratin-based resin for 3D printing by incorporating oxidized keratin into a photo-sensitive initiatior-catalyst- 422,441,442 inhibitor (riboflavin-sodium persulfate-hydroquinone) solution . Ultraviolet (UV) can then be used to induce dityrosine crosslinking of the keratin chains. 443 Dityrosine crosslinking has been reported for collagen and we proved it viable to 422,441,442 crosslink oxidized keratin . Our previous studies indicate that a keratin scaffold can be 3D printed and that this printed construct could be used to uptake and 422 release entrained drugs . The work here seeks to expand these findings. We aim to produce a keratin-based construct that is capable of the slow? release of a drug such as Halofuginone, providing an environment conducive to in vivo healing of burn wounds, and maintaining stability in customized packaging. Based on our previous work, we first assessed the viability of keratin-based photosensible bioink to print reproducible scaffolds for implantation, particularly the retention of geometrical features after the processes of lyophilization and sterilization. Second, we confirmed the viability of our 3D printed keratin-based hydrogel for prevention of burn contracture in vitro using a collagen gel contracture assay. Third, the efficacy of Halofuginone-loaded scaffolds in healing dermal burn wounds was assessed in vivo using a swine burn model. Healing parameters studied included collagen order, regrowth of dermal appendages, and lack of hyperkeratosis or hyperplasia. 139 6.2. Methods 6.2.1. Keratin extraction and bioink preparation Keratin extraction and purification was conducted following the protocol 404 described by de Guzman et al. . Briefly, human hair was immersed and oxidized in excess 2% peracetic acid for 10 h at 37 ?C on an orbital shaker. The treated hair was rinsed thoroughly with water, resuspended in excess 100 mM Tris base for 2 h, strained through a 500-?m sieve, and liquid recovered. The retained hair mass was then extracted with a 40-fold excess of ultrapure water for 2 hours at 37 ?C on a 150- rpm orbital shaker. The treated hair was again strained through a sieve and liquid recovered. Liquid from the Tris and ultrapure water extractions were combined, and pH adjusted to 7.4. Liquid was then centrifuged to remove insoluble particles, followed by filtration using 20-?m Whatman filter paper. The resultant filtrate was dialyzed against endotoxin-free water with a 5-kDa cellulose cartridge for 5 volume washes. The crude extract from the procedure was concentrated, pH-adjusted to 7.4, lyophilized, and ground into powder form. Crosslinkable keratin resin was prepared as detailed in our previous 422,441,442 publications . Briefly, extracted keratin was dissolved in phosphate buffered saline (PBS) at a concentration of 4% (wt/vol). The keratin solution was combined with a photosensitive initiatior-catalyst-inhibitor solution using a 4:1 ratio. This photo-sensitive solution is composed of 1 mM riboflavin (Sigma-Aldrich Co., St. Louis, MO), 200 mM sodium persulfate (SPS, Sigma-Aldrich), and 0.001% wt/vol 140 hydroquinone (Sigma-Aldrich). After thorough mixing on a magnetic stir plate, the resin is curable under UV light. 6.2.2. 3D printing ? continuous Digital Light Processing (cDLP) Keratin-based photo-sensitive resin was used to 3D print 4% wt/vol keratin hydrogels on a lithography-based EnvisionTEC Perfactory 4 DLP printer (EnvisionTEC Inc., Dearborn, MI). An M-type resin tray was modified with PTFE strips to reduce the resin volume (down to 80-100 ml) in the build area. UV exposure time was fixed to 240 s per 100 ?m layer, and intensity was calibrated for 350-375 2 mW/dm . Multiple sets were printed for samples with circular cross-section, 2 mm thick, with either 15 mm (characterization and contracture assays) or 30 mm (animal burn model) diameter (Figure 6.1A-B). Samples for rehydration and sol fraction studies had different thickness groups. Printed samples could either be rinsed in PBS and stored at 4?C, or loaded into custom 3D printed cases and frozen at -80?C overnight. The cases were designed to hold the scaffolds during freezing, lyophilization, sterilization, and transport, and printed on the Perfactory DLP printer using EShell 300 resin (EnvisionTEC) (Figure 6.1D). The samples could then be lyophilized in a FreeZone Lyophilizer (LABCONCO, Kansas City, MO) at a 0.018 mBar vacuum and -48?C for 48 h, after which they were sterilized under gamma radiation. 6.2.3. Rehydration and sol fraction of keratin hydrogels The effects of the manufacturing process, particularly lyophilization and rehydration, on the dimensions of 3D printed keratin hydrogels were assessed. 141 Samples with 15 mm diameter circular cross-section were printed with 2, 3, 4, or 5 mm thickness (designed values, n = 12). Mass and thickness of the samples were recorded (unrinsed values), after which they were moved into excess PBS for 48 h to rinse out uncrosslinked residues. Mass and thickness of the samples were measured (printed values) and samples were then frozen at -80?C overnight followed by 24h lyophilization. After, mass and thickness of the samples were recorded again (lyophilized values). Half of the samples were left unsterilized, while the other half were sterilized under gamma irradiation. Samples from all groups were then rehydrated in either PBS or methanol HH solution (25% methanol (Fisher Scientific, Hampton, NH), 25% milliQ water, 50% PBS) for 1h, after which mass and thickness of the samples were measured (rehydration values). All samples were then moved to excess PBS for 5d, after which the samples were lyophilized again, and mass and thickness were recorded for the last time (final rinsed values). The sol fraction [%] of the sample was calculated as the difference between the unrinsed mass and the final rinsed mass, over the total mass (n=12). 6.2.4. In vitro collagen gel contracture assay Impacts on fibroblast-mediated contracture were examined using a collagen gel contracture assay. A collagen gel working solution was prepared according to the manufacturer?s protocol (Cell BioLabs, San Diego, CA), and was mixed with 500,000 cells/mL suspension of adult normal human dermal fibroblasts (HDF, ATCC, Manassas, VA) in a ratio of 4:1 (collagen:cells). Samples of 500 ?L of the collagen/cell mixture were plated in 24-well plates and incubated (37 ?C). After 1 hour, 1.5mL of media was added to each well. The following day, collagen gels were 142 released from the walls and floor of the well using a 20 ?L pipet tip. Then, Falcon trans-well inserts with a pore size of 8 ?m were placed in each well over the collagen gels, carrying the keratin-based hydrogels to assess (3D printed keratin with or without Halofuginone). Overall, the groups studied were: unseeded collagen without hydrogel treatment (no cells, negative control); HDF seeded collagen without hydrogel treatment (HDF cells only, positive control); HDF seeded collagen with keratin hydrogel treatment (3D printed keratin); and HDF seeded collagen with Halofuginone-laden keratin hydrogel treatment (3D printed keratin + Halofuginone). Digital images were taken from a fixed position with fixed ruler in the field of view at time 0 (gel release), 12, and 24 h. 6.2.5. In vivo porcine burn model In vivo evaluation of the 3D printed keratin constructs was conducted using a 3 cm red Duroc porcine burn model. A total of 24 wound areas were marked on each pig (2 rows of 6 wounds on each side of the midline, at least 3 cm from the spine line). Each area was separated from other wound areas by approximately 3 cm to prevent site-to-site influence. The skin was tattooed over these markings using an electric tattoo machine (Spaulding and Rogers, Voorheesville, NY). Tattoo marks on the edge of the original wound allowed us to monitor the amount of skin contraction in the wound healing process. Clearly delineated wound bed edges were helpful in calculations of wound surface area and provide guidance as to where to take biopsies post-injury. Pigs were sedated with an injection of acepromazine (0.5? 1.5 mg/kg) and ketamine (10? 25 mg/kg,) into the clavotrapezius muscle just behind the ear. During the procedure, a nose cone attached to the anesthesia machine was used to 143 deliver isofluorane (3? 4%) with an oxygen flow rate of 3? 4 l/min. A 0.9% sodium chloride saline solution was delivered intravenously into an ear vein to maintain hydration. The animals were intubated, and isoflurane was maintained at 2% throughout the procedure. To perform the burn, brass cylinders (3 cm diameter) were heated in a dry bath incubator to 100?C, after which the external contacting surface of the cylinder was placed on the skin of the pig, as performed previously. The brass cylinder was held in contact with the skin for 17 seconds to produce a deep partial- thickness burn. Prior to bandaging, the burn wounds were allowed to cool and TM pictures were taken. Afterwards, Tegaderm dressing (3M, Maplewood, MN) was placed over the wound, followed by non-adherent TELFA gauze and taped into place with Elastikon surgical tape. A layer of antibiotic Ioban was then used to cover the entire dorsum/flank of the animal. On post-burn day 3, the animal was again anesthetized, the dressing was removed, and the following treatments applied: Group 1 (3D Halo), a 3D printed lyophilized keratin-based hydrogel loaded with 225 ug/mL Halofuginone; Group 2 (Keratin cream), a 5% non-lyophilized keratin topic solution; Group 3 (No treatment), negative control; and Group 4 (3D Keratin unloaded), a 3D printed lyophilized unloaded keratin-based hydrogel. Treatments were again covered with the described bandages. Bandages were removed, wounds observed, and select wounds biopsied on days 30 and 70 (n=6). 6.2.6. Wound evaluation On days 30 and 70 animals were anesthetized as detailed above and burns were assessed using an histomorphologic scale to quantify cutaneous scars after 444 burns, blinded to treatment received, as described by Singer et al. . Microscopic 144 observations were semi-quantitatively scored in a blind manner using a scale for the following categories: presence or absence of 1) epidermal hyperkeratosis, 2) epidermal hyperplasia, 3) inflammation, 4) vascular proliferation, 5) collagen orientation, 6) hair follicles, 7) apocrine glands, and 8) smooth muscle. Each parameter examined was assigned a score of (+1) for a normal finding otherwise, a score of (0) was given. The presence of normal collagen was assigned a maximum score of (+3), followed by a decreasing score of (+2) to (0) given to progressively deeper scar: (+2) involving the papillary dermis, (+1) for upper half of reticular dermis and (0) for upper and lower half of the reticular dermis. The total histology score was determined by adding all individual scores. The best possible outcome was a score of (10), which was equivalent to normal tissue with no residual scarring. A score of (0) was equivalent to a deep scar involving all levels of the dermis. Points are given for the findings as described in Table 6.1. Table 6.1. Scoring parameters for wound healing in vivo. Parameter Parameter Scoring Scoring examined examined Absent (1) Smooth Absent (0) Hyperkeratosis Present (0) muscle Present (1) Absent (1) Normal (3) Epidermal Collagen Abnormal Pap. Dermis (2) hyperplasia orientation Abnormal upper Reticular Dermis (1) Present (0) Abnormal upper/lower Reticular Dermis (0) Absent (0) Active Absent (1) Hair follicles chronic Present (1) inflammation Present (0) Apocrine Absent (0) Vascular Absent (1) glands Present (1) proliferation Present (0) 145 6.2.7. Histology At the appropriate explant timepoints, biopsy samples were immediately placed in 10% formalin for at least 48 h before processing, embedding in paraffin, and cut into 6 ?m thick sections, and mounted onto glass slides. Slides were stained with picrosirius red (PSR) or Masson?s Trichrome (MTC) to monitor structural changes in the tissue throughout the treatment time course. Histological examination of sectioned samples was performed by a veterinary pathologist. 6.2.8. Statistics Statistics for quantitative tests were performed using ANOVA and Tukey?s multiple pairwise comparison (significance using p < 0.05) for multi-group comparisons. Differences between individual groups and references were assessed with two-sample t-test for the mean. For the wound healing data, statistical analysis was conducted using GraphPad Prism 7.0 for Windows (La Jolla, California) with data expressed as the mean ? standard of error (SEM). Histological scores were considered nonparametric data, and studied with non-parametric, two-way ANOVAs and t-test using Holm-Sidak Method with p<0.05 considered statistically significant. 6.3. Results and Discussion The goal of this work was to assess the viability of a 3D printed keratin-based construct that is capable of the in vivo slow? release of Halofuginone for the treatment of dermal burn wounds. As presented in our previous work, oxidized 422,441,442 keratin has been successfully implemented into a UV photosensitive resin . 146 This keratin-based bioink was used in a lithography-based continuous digital light processing (cDLP) 3D printer, which crosslinks the polymer by projecting UV light through an array of micro-mirrors that independently control the light intensity to 445,446 create the designed microstructure pattern . This approach yielded biocompatible hydrogels with honeycomb-like structures that allowed high uptake of solutions (PBS or minimum essential media (MEM)) as well as cellular adhesion, proliferation, and 422 migration . Uptake and swelling capacities are characteristics of great interest for drug or growth factor delivery, as well as for cell culturing in tissue engineering and regenerative medicine. The printed scaffolds could uptake and saturated with PBS and MEM in 20 min, reaching maximum swelling values of approximately 1770 and 1580% respectively, exceeding those of casted keratin or gelatin reported in 415?417 literature on period without any qualitative changes to the scaffold?s dimensions or consistency. Furthermore, we showed that the keratin-based printing resin and printing methodology have minimum cytotoxic interaction with L929 mouse fibroblasts. Cell populations were viable when seeded on the scaffolds with sustained 422 metabolic activity, morphology, and proliferation rates . These results encouraged us to believe that the printed hydrogels could be used as drug-delivery systems for the treatment of dermal burn wounds. As shown in Figure 6.1A-B, our 3D printing protocol yields consistent batches of hydrogels with reproducible dimensions, as tracked in sol fraction and rehydration tests. Furthermore, we have a manufacturing protocol that yields large quantities of samples, ideal for standardized testing in vitro and in vivo (Figure 6.1C). 147 Figure 6.1. Keratin-based photosensible bioink was used to produce standardized scaffolds for in vitro and in vivo assessment of their viability as drug-delivery systems for the treatment of dermal burn wounds. A) 4% wt/vol keratin hydrogels were 3D printed on a lithography-based EnvisionTEC Perfactory 4 DLP printer. B) Crosslinked samples are stable and retain their printed dimensions while stored and rinsed in PBS. C) The 3D printing protocol yields consistent, large-quantity batches of hydrogels with reproducible dimensions for standardized in vitro and in vivo testing. D) Custom 3D printed cases were produced to hold and transport the hydrogels throughout the manufacturing process from laboratory to surgery. The cases successfully supported i) the printed samples that underwent rinsing; ii) freezing at -80 ?C and lyophilization; and iii) sterilization with gamma irradiation. E) Sol fraction at the end of the manufacturing process reveals significant effects of gamma irradiation on mass loss of thinner constructs. The 2- and 3-mm thick samples have less sol fraction before sterilization than 4- and 5-mm thick constructs, but gamma irradiation alters their crosslinked networks and results in higher soluble components at the end-point; statistical significance: * difference between non- sterilized and gamma irradiated sample (mean comparison t-Test, p?0.05). F) Sterilization and rehydration media also have significant effects on the thickness of the end-product hydrogels. All printed scaffolds differ from the designed constructs and lyophilization does not significantly further reduce the thickness, but gamma 148 irradiation does. Rehydration in HH or PBS cannot fully restore the dimensions, resulting in thickness loss of over 2 mm in the most extreme cases. Statistical significance: * mean thickness differs from the Designed value (t-Test, p?0.05); ? thickness differs from the Printed value (ANOVA per thickness group, p?0.05); # thickness differs from the ReHyd. PBS (No Ster.) value (ANOVA per thickness group, p?0.05). 6.3.1. Effects of sterilization and rehydration on keratin hydrogels The geometry of the scaffolds is a key feature in dermal wound healing; first, the dressing should fit the burn wound site, accommodating to the thickness of irregular wounds, to restore the barrier function of skin against irregular water loss or 4 bacterial infection . Even if the geometry of the wound is adequately scanned to print a complex dressing, the dimensions must hold throughout the manufacturing process, particularly the sterilization and rehydration steps before implantation. The custom 3D printed cases (Figure 6.1D) designed for the samples allowed us to track the samples along the manufacturing process from production in the laboratory to the surgery room. This process includes steps that can potentially alter the mass and dimensions of the 3D printed constructs, particularly the steps of rinsing (Figure 6.1Di), freezing and lyophilization (Figure 6.1Dii), and sterilization (Figure 6.1Diii). Tracking mass and dimensions of the samples throughout manufacturing allowed us to assess these alterations quantitatively. As seen in Figure 6.1E, the soluble fraction at the end of the process indicates significant effects of gamma irradiation on mass loss of thinner constructs. The 2- and 3-mm thick samples have lower sol fraction without sterilization (3.1 ? 6.4 and 5.0 ? 4.2 % respectively) than 4- and 5-mm thick constructs (13.5 ? 2.5 and 14.6 ? 1.7 % respectively), Gamma irradiation alters the crosslinked networks, particularly for the thinner samples, and results in higher 149 content of soluble components at the sterilized end-point. After irradiation, the 2 mm thick samples had a 478 % increase in sol fraction, while the 3 mm samples had a 391 % increase. On the other hand, the thick samples only had 17.5 and 3.9 % sol fraction increases (4 and 5 mm respectively). The soluble fraction (sol fraction) is measurement of an un-crosslinked mass in a hydrogel network. In this case the intermediate lyophilization and sterilization steps could be interfering with the crosslinked network and may produce changes in the shape and mass loss. As samples were crosslinked under the same conditions, particularly exposure time to UV per crosslinked layer, thinner samples initially had a higher crosslinking density, demonstrated with lower sol fraction. Nevertheless, thinner samples are also more vulnerable to gamma penetration which, as indicated in the sol fraction increase, results in breaking the crosslinked network. Higher sol fraction is indicative of higher mass loss and probably an increased degradation rate in vivo. Sterilization and rehydration also have significant effects on the dimensions of the hydrogels. Thickness was particularly tracked due to the layered nature of the lithography-based 3D printing process. As summarized in Figure 6.1F, all printed scaffolds are significantly thinner from the original designs. Lyophilization does not significantly further reduce the thickness, but gamma irradiation generally does. Lyophilized samples facilitate transportation and sterilization but require a step of rehydration to uptake the Halofuginone prior to implantation. Rehydration in methanol HH solution (to solubilize Halofuginone) or PBS was proven to not fully restore the dimensions of the printed samples in all cases, resulting in thickness loss of 2.19 mm in the most extreme case (5mm, gamma sterilized, rehydrated in HH). 150 The thinner samples, particularly the 2 mm group, were able to rehydrate back to the printed conditions in both HH and PBS. Sterilization with gamma irradiation had significant effects on the samples. In all cases, gamma resulted in significant reduction of thickness, and those samples then resulted in even further thinning after rehydration. The effects of lyophilization and sterilization observed in these studies are in line with our own studies on keratin-based hydrogels and with the reports in literature. In our previous studies with the crosslinkable keratin bioink, we detailed how rinsing in PBS purposely removes uncrosslinked components from the hydrogel, 441 reducing sample mass up to 57% . Similarly we observed that freezing and lyophilization cycles induce additional irreversible frustration in the already crosslinked network; dimensions of the scaffolds can be generally restored with rehydration in PBS, but the total mass uptake can be irreversibly decreased; this is indicative that swelling assessments, which require a lyophilization step, can be offset 441 due to additional entanglement . Sterilization methods have also been widely studied to determine their effects on hydrogels networks. Gamma radiation has been used before to sterilize keratin- 418,425,433,435,436 based constructs ; other methods include UV irradiation, ethanol, or 422,430,441 sterile water rinsing . Gamma is generally accepted as a safe method that 447,448 allows cell culture and reduces bacterial proliferation . On UV crosslinked multi- 448 arm poly(ethylene glycol) (PEG) acrylate hydrogels, Escudero-Castellanos et al. assessed the cytotoxic effects of gamma irradiation, disinfection with 70% ethanol, and steam sterilization in an autoclave. Here, gamma did not affect fibroblast viability 151 and showed significantly lower hemolysis levels (direct human red blood cells 448 incubation) than disinfected samples . On the other hand, similar studies on natural- 447 derived hydrogels were performed on collagen surfaces by Tyan et al. where they showed that gamma irradiation cleaves peptide linkages and alters collagen-bonded surfaces, suggesting that gamma irradiation of about 10 KGy may significantly degrade the bioactivity of crosslinked collagen. This sterilization method has also been proven to induce dimensional changes due to possible alterations on the 449 crosslinking chemistry. Kanjickal et al. studied the effects of sterilization methods on the properties of PEG hydrogels, including methods of ethylene oxide (EtO), hydrogen peroxide, and gamma irradiation. Samples sterilized with gamma had a higher concentration of radical species when compared to unsterilized or EtO constructs; gamma irradiated samples additionally presented significant decreases in 449 swelling ratio for all PEG molecular weights and formulations assessed . Hodder et 450 al. studied 3D printed alginate and methyl cellulose, finding that gamma irradiation strongly reduced the bioink viscosity and stability after extrusion, significantly reducing printing fidelity. Furthermore, gamma caused significant reduction of methyl cellulose mass over time, indicating that the crosslinked chemistry was altered 450 and eventually released all the component . Based on the data obtained and literature, it is important to highlight that the manufacturing process of keratin-based hydrogels can result in changes to the 3D printed scaffolds ready for implantation. Particularly the steps of lyophilization and rehydration can result in changes in the crosslinked network, loss of mass, and changes in the dimensions of the samples. These observations are important for the 152 eventual treatment of complex 3D wounds. Even if the 3D scanning and printing are precise, and the bioink is adequate, the manufacturing and preparation of implantable constructs can modify important parameters that could eventually impact healing characteristics such as wound coverage or contraction. 6.3.2. In vitro contracture of keratin-based hydrogels The viability of 3D printed keratin-based hydrogels for prevention of burn contracture was assessed in vitro using a collagen gel contracture assay. As illustrated in Figure 6.2A, the use of 3D printed keratin constructs was always positive to reduce relative contracture compared to a no treatment, HDF cells only group. Using unloaded keratin hydrogels saw a contracture reduction close to 53 % compared to the HDF only group. Inclusion of the Halofuginone further reduces contracture, by 83 %, compared to the cells? behaviour, coming very close to the no contracture (no cells) case although still significantly different. Furthermore, the inclusion of the Halofuginone significantly reduces contracture compared to the unloaded 3D printed keratin hydrogels. Contraction is a delicate process that heavily relies on timing. Contraction of the open wound is necessary to restore the dermal barrier, but uncontrolled contracture due to excessive migration of fibroblasts and disorganized 4,31,74 collagen deposition results in higher degrees of scarring . Increasing wound size and depth reduce wound closure (re-approximation); deep wounds that destroy progenitor fibroblast stem cells then rely on re-epithelialization from the wound 4 margins to close, which delays healing and results in higher scarring and contracture . Halofuginone has been proven to prevent abnormal fibrillar collagen accumulation in 153 437?440 wound contracture , by inhibiting collagen I synthase to decrease collagen 439 synthesis . Figure 6.2. A) In vitro contracture assay using a cell-laden collagen gel treated with loaded or unloaded keratin hydrogels. The use of 3D printed keratin constructs was generally positive to reduce relative contracture compared to no treatment group; the inclusion of Halofuginone further reduces contracture. Statistical significance: * difference when compared to HDF cells only group, p?0.01; # difference when compared to 3D printed keratin group, p?0.01. B) Combined histomorphologic scores at 30 and 70d post-implantation, expressed as average mean ? SEM, best possible outcome = 10, worst = 0. Statistical significance: + mean histomorphologic score is different at days 30 and 70, p?0.001. C) Representative macroscopic images showing treatment groups at day 0 (burn induction), day 3 (initial treatment applications), day 30 and day 70 (biopsy collection, end points). 154 Scaffolds or dressings, both in research and industry, have attempted to reduce 451,452 skin wound contracture, particularly collagen-based constructs . Commercial products such as MatriDerm or Hyalograft? C have been proven to reduce contraction while additionally improving elasticity, epithelialization, and basement 439 membrane formation in vitro . Collagen-glycosaminoglycan sponges were used to reduce wound contracture on mice; here, the lower concentrations of the 1-ethyl-3-3- dimethylaminopropylcarbodiimide hydrochloride (EDC) crosslinker resulted in better contraction processes, indicating that single components within a regenerative 453,454 scaffold can greatly impact contracture . Similarly, collagen matrices loaded with basic fibroblast growth factor (bFGF) were used on a rabbit model and in vitro on collagen gels; collagen matrix and bFGF independently promoted reepithelization and reduced contracture, results that were further improved when the two parts were 455 86 combined as a delivery system . On a case closer to our study, Xu et al. used freeze-dried keratin scaffolds in a Wistar rat animal model and report that this approach resulted in less contraction compared to non-scaffold treatment, additional 86 to earlier vascularization, thicker epidermis, and hair follicle formation . Compared to literature, our data here confirms reported results for hydrogels, loaded or not, reducing contracture in dermal wounds. We further confirm studies and effects of Halofuginone and keratin use, although our studies are the first to report them coupled together. As collagen constructs reported previously, the 3D printed keratin does not inhibit the effect of their loaded component, in this case Halofuginone, and it provides an adequate transport system with the additional characteristic of controlled geometry for delivery to complex dermal wounds. 155 6.3.3. In vivo assessments of keratin-based hydrogels The efficacy of Halofuginone-loaded scaffolds in healing dermal burn wounds was assessed in vivo using a swine burn model. Healing parameters studied included order of collagen microstructures, regrowth of dermal appendages, and lack of hyperkeratosis or hyperplasia, among others. Additionally, the progress of in vivo wound contracture was also imaged and studied. Animal models to assess excessive 456 dermal scarring include rabbit ear or porcine dermis . Red Duroc pigs in particular are a suitable model for hypertrophic scarring and excessive contraction following 457,458 cutaneous injury . The objective of this study was to evaluate the ability of 3D- printed keratin constructs to reduce burn-induced hypertrophic scarring and contracture in a porcine partial thickness burn model. Partial thickness burns were created on the dorsum of a Red Duroc pig using a brass burn block, and wounds were treated with 3D-printed 4% keratin constructs unloaded or supplemented with Halofuginone, a collagen synthesis inhibitor, at day 3 post-injury. In the current 444 study, we utilized a histomorphologic scale methodology to determine the efficacy of the three different treatment groups (keratin-based formulations) in reducing burn- induced hypertrophic scarring and contracture vs. no treatment (negative control). Injury to epithelial and mesenchymal dermal components was determined by examining the presence and/or absence of epithelial hyperplasia, epidermal hyperkeratosis, inflammation, vascular proliferation and appendage structures including hair follicles, apocrine glands, and smooth muscle, and scored as described in Table 6.1. Statistical histomorphological findings are summarized in Figure 6.2B; individual time-point results are shown as incidence formatted data (30 d in Table 156 6.2, and 70 d in Table 6.3) and as histologic images (MTC in Figure 6.3, and PSC in Figure 6.4). Table 6.2. Incidence of histological observation after 30d. Group 1: Group 2: Group 3: Group 4: 3D Parameter Ideal 3D printed Keratin No printed keratin examined keratin cream treatment + Halofuginone Hyperkeratosis Absent 1/6 4/6 1/6 1/6 Epidermal Absent 0/6 0/6 0/6 0/6 hyperplasia Hair follicle Present 6/6 5/6 6/6 4/6 Apocrine glands Present 6/6 6/6 6/6 6/6 Smooth muscle Present 0/6 2/6 3/6 2/6 Normal 0/6 0/6 0/6 0/6 (+3) Collagen Abnormal 0/6 0/6 0/6 0/6 orientation (+2) Abnormal 5/6 6/6 6/6 5/6 (+1) Inflammation Absent 1/6 2/6 3/6 2/6 Vascular Absent 0/6 1/6 0/6 1/6 proliferation Representative biopsy sections from each of the burn sites (n=24 per time point) at the 30 d time point, showed no statistically significant differences amongst treatment groups. As summarized in Figure 6.2B and Table 6.2, the mean histomorphologic scarring score of about (+4) was given to Groups 2 (4.33 ? 0.42), 3 (4.17 ? 0.48) and 4 (3.83 ? 0.60) (Keratin cream, No treatment, and 3D keratin unloaded, respectively), and slightly lower (3.17 ? 0.31) for Group 1 (3D Halo). 157 Table 6.3. Incidence of histological observation after 70d. Group 1: Group 2: Group 3: Group 4: 3D Parameter Ideal 3D printed Keratin No printed keratin examined keratin cream treatment + Halofuginone Hyperkeratosis Absent 6/6 5/6 4/6 4/6 Epidermal Absent 0/6 0/6 0/6 0/6 hyperplasia Hair follicle Present 3/6 5/6 6/6 4/6 Apocrine glands Present 5/6 5/6 6/6 6/6 Smooth muscle Present 1/6 2/6 1/6 2/6 Normal 0/6 0/6 0/6 0/6 (+3) Collagen Abnormal 2/6 4/6 2/6 4/6 orientation (+2) Abnormal 4/6 2/6 4/6 2/6 (+1) Inflammation Absent 4/6 4/6 4/6 3/6 Vascular Absent 4/6 4/6 2/6 3/6 proliferation Using macroscopic (Figure 6.2C) and microscopic histologic (Figures 6.3 and 6.4) data, epidermal hyperplasia was observed in all groups and the absence of hyperkeratosis occurred only once in Groups 1, 3 and 4 but with greater incidence (4 of 6 wounds) in Group 2. These findings may suggest that the keratin cream treatment (Group 2) has some beneficial effects in improving the short-term outcome of burn- induced healing. The presence of mesenchymal structures (dermal appendages) such as hair follicles, apocrine glands, smooth muscle (arrector pili) and blood vessels occurred variably but with similar incidence in all groups and thus, no significant treatment effect between treatments, including the untreated control. Specifically, hair 158 follicles and apocrine glands were observed with equal incidence (6 of 6 wounds) in all treatment groups. Smooth muscle (arrector pili) had variable incidence: (3 of 6 wounds) for Group 3, (2 of 6 wounds) for Groups 2 and 4, but not observed (0 of 6 wounds) in Group 1. Figure 6.3 at 30 d reveals an immature formation of collagen in the upper layers of the dermis, characterized by the purple stain of the collagen rather than with bright blue. Additional evaluation of collagen alterations using Picro Sirius Red stained sections under polarized light revealed equivalent collagen fiber disorganization in all treatment groups, particularly in the lower reticular dermis, and thus all wounds received a low score of (+1) (Figure 6.4). Less frequently observed changes such as the absence of inflammation and vascular proliferation occurred with similar incidence in all groups regarding of the treatment. The incidence of active chronic inflammation was similar in three groups: absent in (3 of 6 wounds) for Group 3, in (2 of 6 wounds) for Groups 2 and 4 but absent in (1 of 6 wounds) for Group 1. Vascular proliferation was absent in (1 of 6 wounds) for Groups 2 and 4 but present in (6 of 6 wounds) for Groups 1 and 3. 159 Figure 6.3. Representative histological sections (Masson?s trichrome stain, 20x magnification) of all treatment groups, showing prominent epidermal and dermal changes. At 30 days post-procedure there is marked epidermal hyperplasia and marked collagen degeneration (staining red and/or pale blue by Masson?s) predominantly involving papillary dermis and upper reticular dermis. After 70 days, there is considerably less apparent dermal changes characterized by intensity (blue staining) and uniformity of newly deposited collagen within the dermis but without significant improvement in the epidermis. Scale shown: 2 mm. 160 Figure 6.4. Representative (low-power, x20) histological sections (Picro Sirius Red stained) of all treatment groups. Collagen content and order is characterized by the intensity and pattern of polarized light induced birefringence of collagen fibers. Intense red-orange-yellowish birefringence indicates mature and organized collagen fibers (control, normal tissue, far left). Green birefringence has generally been correlated to immature, fine and less organized collagen, although color change can 1 be correlated to rotations of the samples and light incidence . Here, color was not considered to assess quality of the collagen but merely to identify it and its general orientation trends. At 30-days post-procedure, similar dermal changes are demonstrated by degree of collagen alterations in all samples; collagen is only present but disorganized in the lower reticular dermis. At 70 days, images show similar dermal changes in all groups, with collagen present throughout the dermal layers although disorganized. Scale shown: 1 mm. 161 At the 70-day timepoint, no major improvements on wound healing were recorded, mostly continuing the trends observed at 30 d, as summarized in Figure 6.2B and Table 6.3, and evaluation again showed no major statistically significant differences amongst treatment groups. A mean score of about (+5) was given to for Groups 1 (5.17 ? 0.54) and 3 (5.17 ? 0.79), and a score close to (+6) for Groups 2 (5.83 ? 0.70) and 4 (5.67 ? 0.67). Using macroscopic (Figure 6.2C) and microscopic (Figures 6.3 and 6.4) information, we observed the absence of hyperkeratosis occurred with slightly variable incidence amongst all groups but with a positive trend favorable to Groups 1 and 2, although the scores for Groups 3 and 4 were not far behind. Furthermore, the presence of epidermal hyperplasia was a frequent finding as it occurred with equal incidence in all wounds regarding of treatment. However, the presence of normal appendages structures varied between groups with no clear trend. Hair follicles were frequently observed with a variable incidence between treatment groups; in (3 of 6 wounds) for Group 1, (4 of 6 wounds) for Group 4, (5 of 6 wounds) for Group 2 and (6 of 6 wounds) for Group 3. The presence of apocrine glands occurred with equal incidence (5 of 6 wounds) for Groups 1 and 2; and (6 of 6 wounds) for Groups 3 and 4, respectively. Similar findings were observed for the presence of smooth muscle), with equal incidence occurring in (1 of 6 wounds) for Groups 1 and 3; and (2 of 6 wounds) for Groups 2 and 4. Collagen fiber alterations appeared to have diminished in all groups but with best overall improvements observed in Groups 2 and 4, when compared to lower scores in Groups 1 and 3 (Figures 6.3 and 6.4). There was an improvement in the incidence of collagen, both observed in the Massom?s trichrome and Pricosirius Red stains, score of (+2) 162 assigned to Groups 2 and 4 in (4 of 6 wounds) vs. (2 of 6 wounds) for Groups 1 and 3. A score of (+2) is characterized by abnormal weave pattern of collagen only limited to the papillary dermis, having collagen throughout the cross-section of the samples. Again, while these changes (improvement in collagen alterations) were not statistically significant between treatment groups at 70-days post-procedure, they may be considered favorable trends towards improving the outcome of burn-injury scarring. Other parameters examined such as inflammation and vascular proliferation showed similar histological findings between treatment groups and thus no statistical significance was found. The incidence of active chronic inflammation occurred equally in three groups; absent in (4 of 6 wounds) for Groups 1, 2 and 3; and absent in (3 of 6 wounds) for Group 4. Vascular proliferation occurred equally in two groups; absent in (4 of 6 wounds) for Groups 1 and 2 but at lower incidence (2 of 6 wounds) for Groups 3 and 4. Overall, except for the presence of epidermal hyperplasia, which was observed in all wounds regardless of the treatment and time points, most of the parameters examined appeared to have been improved over time when comparing the 30-day vs. 70-day time points. However, the referenced overall improvements were not considered statistically significant. While the overall improvement in collagen organization observed between treatment groups by the 70-day end-point was not considered statistically significant, there was general trend favorable to treatment with keratin cream and 3D keratin unloaded that was demonstrated by the improvement in the level of collagen fiber organization. The most important improvement was the statistically significant difference in the total outcome scores 163 for Group 1 (3D Halo) at 30 days (3.17 ? 0.31) vs. at 70 days (5.17 ? 0.54) post- procedure (p-value 0.0094). It is worth highlighting again that this group, the 3D printed keratin scaffolds loaded with Halofuginone, were the only samples to present significant improvement between days 30 and 70. This is indicative that the use of 3D printed keratin is not inhibiting the healing processes, and the inclusion of Halofuginone further indicates faster and more organized dermal healing after a burn. This is supported by our previous in vitro observations on contracture and reports in literature; 3D printed keratin is non-inferior to other reported natural, regenerative 86,451?455 scaffolds for skin wounds in vivo , and it includes the additional complex parameters of controlled geometry via 3D printing and drug delivery. 6.4. Conclusions Our previous studies on 3D-printed keratin had given us indications that the scaffolds could be used as a drug delivery mechanism for the treatment of dermal burn wounds. Here, we have further elucidated on the manufacturing process and the effects of critical steps such as lyophilization and sterilization. Keratin-based photosensible bioink can be used to produce large amounts of scaffolds for implantation, and we now have a better understanding of how 3D printed geometrical features, crosslinking properties, or mass are altered when the constructs undergo freeze-thawing cycles and gamma irradiation. We now have built a complete protocol to 3D print keratin scaffolds that can be used in vivo, including the presentation of a viable form of customized packing for production and transportation. Furthermore, keratin hydrogels are viable for the uptake and release of a contracture-inhibiting drug 164 such as Halofuginone, which provides a mechanism to reduce scarring of severe burn wounds. In vivo data shows that the Halofuginone-laden printed keratin is non- inferior to other similar approaches reported in literature. Although not entirely solving the issue of hyperplasia, the treatment does improve healing parameters such as hyperkeratosis, growth of apocrine glands and hair follicles, lesser effect on the growth of smooth muscles, absence of inflammation and vascular proliferation, and general improvement in collagen order over 70 d. The combined use of keratin and Halofuginone showed significant improvement in healing from days 30 to 70, compared to any other group studied. These studies are indicative of the potential of keratin bioink in dermal wound healing. Having proven its viability in vitro and in 422,441,442 vivo, here and in previous studies , we aim to keep increasing the complexity of the 3D printed constructs, particularly aiming to treat complex facial burn wounds and elucidate in complex topographical reconstruction of dermal tissues. 165 Chapter 7: Development of Keratin-Based Membranes for 7 Potential Use in Skin Repair * 7.1. Introduction Severe skin burns are reported worldwide for over 11 million patients every 3,4 5 6,7 8 9 year , including victims of violence (warfare , acid atacks , scalding ) and trauma (flame, electrical, chemical). Skin is the first barrier against external mechanical and biochemical agents, and is composed of the epidermis, dermis, and hypodermis layers, each with its distinct composition and function. When burned, skin cannot 4,10 regulate temperature or fluid transport, or stop bacterial infection . Wound contraction after severe burns in adults is characterized by fast proliferation of fibroblasts that deposit random collagen to rapidly restore the skin barrier against 31,74 external pathogens . Depending on the size of the wound, as well as time and genetic factors, unsystematic deposition of extracellular matrix (ECM) and clotting 4 can lead to scarred tissue . This disorganized, fibrous tissue is characterized by lack of sensation and elasticity, hypertrophy, and flawed features; in effect, ?natural healing? closes the wounds but does not restore skin function, layered microstructure, 4,10,31,74 4,11,12,31 or aesthetic features . Autologous grafting, the current gold standard , aims to restore the barrier using the patient?s skin but has the same limitations as 7 As published in: J Navarro, J Swayambunathan, MJ Lerman, M Santoro, JP Fisher: ?Development of Keratin-Based Membranes for Potential Use in Skin Repair?. Acta Biomaterialia, Vol. 83, pp. 177- 188, 2019. * Further complimentary reading: J Navarro, J Swayambunathan, M Santoro & JP Fisher: ?Assessment of the Effects of Energy Density in Crosslinking of Keratin-Based Photo-Sensitive Resin?. 2018 IX International Seminar of Biomedical Engineering (SIB), IEEE, pp. 1-6, 2018. doi:10.1109/SIB.2018.8467744 166 wound contraction. It is imperative to develop strategies that restore the layered structure of native skin. Toward this end, guided tissue regeneration (GTR) is a surgical technique developed for bone-related procedures in which a physical barrier is used to regulate tissue growth. Membranes are used to stop fast-proliferating fibroblasts and connective tissues from filling bone wounds, while allowing specialized cells (such as periodontal fibroblasts or bone cells) to migrate and correctly regenerate the osseous 459?463 tissues . Current GTR uses degradable membranes to restrict volumetric tissue 461,462,464 growth and prevent cellular migration and unwanted deposition of ECM . Here, we propose the use of crosslinked keratin membranes for GTR of skin. Keratins are a family of structural proteins that can be found either as a major cytoskeletal component of keratinocyte cells in the epidermis layers of skin (?soft? keratin) or as a fibrous extracellular protein in hair, wool, quills and horns (?hard? 32,405,465 keratin) . Keratin, particularly ?hard? keratin, has been used in scaffolds and 88 418 drug delivery carriers for skin , muscle , and nerve tissue 399,418,424,436,466?468 engineering . Furthermore, extracellular ?hard? keratin has been successfully used in the treatment of dermal burn wounds on animal models (split- 86 87 thickness burns in mice, rats , or pigs ) and on clinical patients (split-thickness 88 burns less than 10% of total body surface area ). We previously reported a method for three-dimensional (3D) printing hair-derived keratin hydrogels using a riboflavin- 422 sodium persulfate-hydroquinone (initiator-catalyst-inhibitor) photosensitive resin . Ultra-violet (UV) light was used to form dityrosine bonds and photocrosslink keratin 422 on a lithography-based 3D printer . 167 In this study we assessed the viability of UV-crosslinked keratin as a permeable membrane, particularly for the regulation of molecular transport. Under UV, the amount of energy delivered to the resin volume (energy density, ED) defines the amount of dityrosine bonds formed and, correspondingly, the crosslinking degree (CD). As such, we hypothesize that ED could be used as a design parameter to control microstructural properties of the hydrogels, including swelling, degradation, mechanical properties, and transport across the network. First, we assessed if dityrosine bonding could be exploited to regulate the CD of keratin samples by controlling ED. We then quantified how CD affects properties dependent on the hydrogel microstructure, such as mechanical and swelling behaviors. As transport among dermal layers is fundamental for skin physiology, we were particularly interested in how the membranes can alter the diffusion rates of molecules towards a target cell population. As such, we studied the effect of CD on the partition coefficient and permeability of model molecules with varying molecular weights. Subsequently, we evaluated transport of growth factors and nutrients across the membranes and how engineered membranes can be used to regulate cellular functions, specifically adipogenic differentiation of mesenchymal stem cells. 7.2. Materials and methods 7.2.1. Keratin preparation Keratin was prepared using a proprietary method by KeraNetics, LLC 418,422 (Winston-Salem, NC). As described before , peracetic acid was used to oxidize 168 human hair and keratin proteins were extracted using serial rinses with ultrapure water and Tris buffer, and purified by filtration and dialysis. Purified keratin was lyophilized, sterilized using gamma radiation, and stored at -20 ?C. 7.2.2. Keratin resin and curing The keratin-based photocrosslinkable resin was prepared as previously 422 reported . Briefly, keratin was dissolved overnight in phosphate buffered saline (PBS, pH 7.4) at a concentration of 4% wt/vol, and then mixed with a photosensitive solution at a 4:1 ratio. The photosensitive initiator-catalyst-inhibitor solution consisted of 1mM riboflavin (Sigma-Aldrich Co., St. Louis, MO), 200 mM sodium persulfate (SPS, Sigma-Aldrich), and 0.001% wt/vol hydroquinone (Sigma-Aldrich), respectively. After thorough mixing the resin is curable under UV light by formation of dityrosine bonds as summarized in Figure 7.1A. The crosslinking of keratin can be 2 controlled with the UV light intensity (I) [mW/mm ], exposure time (t) [s], resin 3 2 volume (V) [mm ], exposed area (A) [mm ], or sample thickness (h) [mm]. These are the parameters we routinely control during our casting or printing processes. Here, we 3 used them to define a unifying parameter, energy density (ED) [mJ/mm ], to produce, evaluate, and compare the hydrogels produced. The simplified function ED was defined as: ED = (I x A x t) / V = (I x t) / h (7.1) The resin was used to cast hydrogels using the ED parameter. For this, 3D printed molds were used, either with square (10 x 10 mm) or circular (8 or 13 mm diameter) holes passing through. A glass slide was clamped to the bottom of the 169 mold, and the sealed molds were then filled with resin as needed and exposed to UV 422 2 at constant intensity of 350 mW/dm , defining the thickness of the sample using a known resin volume (Fig. 7.1B). Samples were rinsed and stored in PBS at 4 ?C. Nomenclature was defined as exposure time-thickness (e.g. Sample 12-1.5 is a sample with thickness of 1.5 mm, UV exposed for 12 min) (Fig. 7.1C-D). Figure 7.1. Photocrosslinking of keratin membranes. A) Chemical mechanism for dityrosine bonding in oxidized keratin: i) keratose contains tyrosine within their peptide chains; ii) tyrosine has a susceptible hydroxyl group that be deprotonated upon interaction with free radicals formed between SPS and riboflavin upon UV exposure; iii) unbalanced terminals bond and recombine with unbalanced terminals in adjacent chains; iv) final chemical equilibrium is reached via keto-enol tautomerism when dityrosine bonds between keratin chains are stable; the crosslinking reaction is terminated once removed from a UV source due to the reducing effect of the hydroquinone photoinhibitor. B) Dityrosine chemistry was used to cast keratin membranes: i) a translucent sheet was covered with a thin layer of resin; ii) the casting mold was then laid over the flat sheet and the two were clamped together. The clamped setup was then crosslinked under UV to seal them together; iii) the sealed molds were filled with keratin resin as needed and exposed to UV; iv) the cured mold can be cut from the base sheet using fine thread. C) Examples of crosslinked circular 170 samples 13mm in diameter with different thicknesses (1, 1.5, or 2 mm). D) Final crosslinked samples 12-1.5 (exposure time 12 min, thickness 1.5 mm) stored in PBS. 7.2.3. Initiator consumption Cylindrical samples (13 mm diameter, 1.5 mm thick) were exposed to UV for 6, 12, 24, 48, and 96 min as described in the previous section, producing samples groups 6-1.5, 12-1.5, 24-1.5, 48-1.5, and 96-1.5, respectively. These groups had ED 3 values of 8.4, 16.8, 33.6, 67.2, and 134.4 mJ/mm , as presented in Figure 7.2A (further combinations of thickness and exposure time, and the resulting ED values, can be found in this graph). Samples were collected after casting and immediately stored in 20 ml excess PBS at 4 ?C in the dark for 4 days. Afterward, the conditioned PBS containing the uncrosslinked residues of the keratin resin was collected. 100 ?l samples from each solution were set in 96 well-plates for absorbance reading of riboflavin at 450 nm using a SpectraMax M5 plate reader (Molecular Devices, Sunnyvale, CA). Control samples included unreacted resin, riboflavin, SPS, hydroquinone, and keratin, in PBS (n=9). Absorbance of the samples was normalized against the unreacted resin control for quantitative comparison. 7.2.4. Sol fraction Thick cylinder samples (8 mm diameter) 8-3, 24-3, 48-3, and 96-3 were prepared as described above. Samples were collected after casting and lyophilized without rinsing. Next, samples were weighed (unrinsed total mass) and stored in excess PBS at 4 ?C. The samples were then rinsed following our rinsing protocol (See Supplementary data Methods and Figure 7.S1). Briefly, sequential rinses using fresh 171 PBS for 15 min, at room temperature, for a total of 9 cycles; a final tenth rinse was a single, extended overnight rinse (14h). All samples were then lyophilized again and weighed (rinsed mass), using a microbalance (Sartorius ME-5, Sartorius, Goettingen, Germany). The sol fraction [%] of the sample was calculated as the difference between the unrinsed total mass and the rinsed mass, over the total mass (n=12). 7.2.5. Fourier-transform infrared spectroscopy (FTIR) Thin membrane samples (10 x 10 mm) 12-1.5, 24-1.5, 48-1.5, and 96-1.5 were prepared as described above. After completing the rinsing protocol in PBS and lyophilization, samples were mounted onto MirrIR Low-e Microscope Slides (Kevley Technologies, Chesterland, OH). FTIR measurements were performed on a Smiths IlluminatIR II, imaging under attenuated total reflection (ATR) mode. 100 x 100 ?m sampling areas were scanned 128 times per sample. Background was removed and the diamond contact probe cleaned with 100% ethanol before each sample. Control samples included uncrosslinked keratin, PBS, and photosensitive solution only (n=9). Quantification of the peaks was done by taking the area under the curve within the -1 1400 to 1650 cm range and normalizing against the value for unreacted keratin. Variations in these results were used to assess alterations in the aromatic rings of tyrosine due to the formation of dityrosine bonds. Post-processing of FTIR spectra 469 was done using Spectragryph ? optical spectroscopy software . 7.2.6. Thermogravimetric analysis and differential scanning calorimetry Thin membrane samples (10 x 10 mm, 1.5 mm thick) 12-1.5, 24-1.5, 48-1.5, and 96-1.5 were prepared as described above. Additional samples for 12-2 and 96-1 172 groups were prepared, representative of low and high ED, respectively. All samples were subjected to our standard rinsing protocol in PBS and then lyophilized. The 12-2 and 96-1 samples, and uncrosslinked lyophilized keratin, were subjected to thermogravimetric analysis (TGA, TA Instruments, New Castle, DE - Energy Research Center, University of Maryland) heating the samples up to 350 ?C at 10 ?C /min under 100% N2. TGA was performed to identify variations between crosslinked and uncrosslinked samples, and to determine the temperature ranges where these occurred. The 1.5 mm thick samples, and uncrosslinked keratin controls, were massed using a micro-balance (Sartorius ME-5, Gottingen, Germany) and sealed into differential scanning calorimetry (DSC) pans (10-15 mg per pan). DSC was performed on a Q100 differential scanning calorimeter (TA Instruments, New Castle, DE) by heating the samples from 30 to 300 ?C at 10 ?C/min under 100% N2. Resulting thermograms were compared to those of uncrosslinked controls to 470 determine the crosslinking degree (CD) as previously reported in literature (n=9 to 15). Briefly, in DSC thermograms the area under the peak represents the maximum enthalpy change possible in the system. Changes in these areas are indicative of changes in the internal energy of the network and are used to quantify bond formation 470 and CD. As such, enthalpy change was used to calculate CD , as: ??(?0) ? ??(??) ??(??)[%] = ? 100 (7.2) ??(?0) 173 Where, ??(?0) is the reaction enthalpy of uncrosslinked reference (area under reference peak) and ??(??) is the reaction enthalpy of the crosslinked sample (area under sample peak). 7.2.7. Swelling Thin membrane samples (10 x 10 mm) were prepared for the combinations 12-2, 12-1.5, 12-1, 24-2, 24-1.5, 24-1, 48-2, 48-1.5, 48-1, 96-2, 96-1.5, and 96-1. All samples were rinsed in PBS over 24h following our standard protocol and then lyophilized. Initial dried mass was recorded. The samples were then rehydrated in 2 ml excess minimum essential medium (MEM, Life Technologies, Frederick, MD) at 37?C. Rehydrated mass of the samples was recorded at 1, 3, 5, 10, 15, 20, 30, 45, and 60 min, 3h, 6h, 24h, 2d, 3d, 5d, 7d, 14d, 21d, and 28d. At each time point the sample was taken out of solution, gently blotted to remove excess MEM, weighed (swollen mass), and then returned to the solution. The volume of MEM was kept at 2 ml by refilling every 3 days. Swelling was calculated as the difference between swollen mass and dried mass, over dried mass (n=9). 7.2.8. Mechanical properties Thick cylinder samples (8mm diameter) 8-4, 24-4, 48-4, and 96-4 were analyzed using an Instron 5565 (Instron, Norwood, MA). Mechanical compressive testing was done on the samples using a 50 N load cell at a compression rate of 1 mm/min until yielding. The ultimate stress and strain values were taken from the stress-strain curves, and the linear regime of the curves was determined with a linear 2 fit (R ? 0.95) to obtain the compressive modulus (n=7). 174 7.2.9. Partition coefficient Thin membrane samples (13 mm diameter), 12-1.5, 24-1.5, 48-1.5, and 96-1.5 were rinsed following our standard protocol and then equilibrated in PBS at room temperature for an additional 24h. Separately, fluorescein isothiocyanate-dextran (FITCd, 10, 150, or 2000 kDa molecular weight (MW)) was dissolved at 10 mg/ml in PBS, and equilibrated under the same conditions. Samples from the initial FITCd solutions were taken and measured for fluorescence in the SpectraMax M5 plate reader. The mass of equilibrated hydrogels was also recorded. The hydrogels were then removed from the PBS and added to 0.5 ml of the equilibrated FITCd solutions. After 24h, 10 ?l samples were collected from the solutions and diluted with 90 ?l of PBS. Samples were read for fluorescence alongside a FITCd concentration ladder, to calculate the concentration of the sampled solutions (n=9). The partition coefficient (K) is the steady-state relation between the concentration of a solute in a gel (??) and the concentration of the solute in the solution (??) in which the gel is equilibrated in. As such, K between crosslinked keratin and the FITCd solutions was obtained using 471 the mass balance of the FITCd between the hydrogel and the solution , which results in: ?? ??(?? ? ??) ? = = (7.3) ?? ???? Where ?? is initial concentration of FITCd in solution; ?? is stable final concentration of FITCd in solution; ?? is volume of solution; and ?? is volume of the 175 hydrogel, which is assumed equal to the equilibrated mass of the hydrogel as it is over 95% PBS (density ~1 g/ml). 7.2.10. Permeability Thin membrane samples (13 mm diameter), 12-1.5, 24-1.5, and 96-1.5 were prepared as described above, rinsed following the protocol, and equilibrated in PBS at room temperature for over 24h. Thickness and diameter of the equilibrated R membranes was recorded before setting them in Transwell inserts (Falcon Permeable Support for 12 Well Plate with 8 ?m transparent polyethylene terephthalate (PET) membrane, Thermo Fisher Scientific, Waltham, MA). The PET membrane of the Transwell was punctured five times using a 22G 1? needle to guarantee that any barrier effect was due to the keratin membrane. 3D printed stoppers are used to weigh down the membrane and assure sealing. The well (receiver side) was filled with 1.3 ml of PBS and allowed to re-equilibrate with the membrane for 12h. After, 1.3 ml of 10 mg/ml FITCd solution (10, 150, or 2000 kDa MW) in PBS was be added to the Transwell (donor side) at room temperature. At 0, 5, and 30 min, and 1, 2, 3, 6, and 24 h, a 5 ?m sample was collected from the donor and receiver sides and diluted with 95 ?l of PBS. Three samples were tested per FITCd solution, alongside a FITCd concentration ladder (n=9). Volume loss in the wells was 5.4 to 8.1%, thus we assumed constant volume and concentrations throughout the test. We further assumed that the donor had known initial concentration while the receiver concentration was zero, and that flux across the membrane quickly reaches steady state and does not change over time. As such, we used a pseudo-steady state model, in which the 176 concentration profile across the membrane is assumed invariant, to calculate the 472 permeability (P) of the hydrogels to each type of FITCd by plotting : ?? ? ?? ??? 1 1 ??? ( ) = ( ) ( + ) (7.4) ??0 ? ??0 ? ?? ?? Where ?? is donor concentration at time t; ?? is receiver concentration at time t; ??0 is initial donor concentration; ??0 is initial receiver concentration; ? is time; ? is membrane area; ? is membrane thickness; ?? is donor volume; and ?? is receiver volume. 7.2.11. Transport of adipogenic molecules across keratin membranes Thin membrane samples (13 mm diameter), 12-1.5, 24-1.5, and 96-1.5 were cleaned following our protocol for sequential PBS rinses and equilibrated in PBS at room temperature for over 24h. The membranes were used in Transwell inserts with stoppers using the same methodology described above. Human mesenchymal stem cells (hMSCs, Lonza, Walkersville, MD) were seeded in 12-well cultures plates at a 2 density of 5000 cells/cm and expanded to total confluence, as suggested by the provider for successful adipogenesis, by feeding them every 3 d with hMSC growth media (Dulbecco?s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS), 1% non-essential amino acids (NEAA), and 1% penicillin/streptomycin 100 U/100 mg (P/S)). Once at total confluence, the Transwell-membrane setups were set on the culture wells. The Transwell donor side was filled with 0.5 ml of hMSC adipogenic media (DMEM with 10% FBS, 100 U/100 mg P/S, 1 mM sodium 177 pyruvate, 1 ?M dexamethasone, 10 mg/ml insulin, 0.5 mM 3-isobutyl-1- 473,474 methylxanthine (IBMX), and 200 mM indomethacin) , while the receiver side with the cells was filled with 0.5 ml hMSC growth media. Media on both sides was replenished every 3d for 21d. As positive controls, wells with Transwell only or no Transwell were fed with 1.5 ml of adipogenic media, while negative controls without Transwell or membrane, were fed hMSC growth media to avoid differentiation. Adipogenic differentiation of hMSCs was assessed after 21d using AdipoRed? Assay Reagent (Lonza) to quantify intracellular lipid accumulation (n=9). 7.2.12. Statistics Statistics for quantitative tests were performed using ANOVA and Tukey?s multiple pairwise comparison (significance using p<0.05) for multi-group comparisons. Differences between individual groups and references were assessed with two-sample t-test for the mean (significance using p<0.01 (+) and p<0.05 (*)). 7.3. Results The goal of this work was to prove that dityrosine bonding chemistry could be used to engineer keratin membranes and to regulate molecular transport. As hypothesized, the energy density during the curing process was used as a design parameter to control the hydrogel?s physicochemical properties, including permeability across the membrane. As illustrated in Figure 7.1, the keratin photosensitive resin and the tyrosine available in oxidized keratin were successfully exploited to produce hydrogels. The ED parameter proposed (Equation 7.1) was used 178 to produce a variety of crosslinked samples by combining different thicknesses and 3 exposure times. As shown in Figure 7.2A, samples ranged from low ED (6.3 mJ/mm 3 minimum) to high ED (201.6 mJ/mm maximum). Samples within this range were used to assess how the crosslinking density (CD) of keratin could be regulated using the ED parameter. The extent of crosslinking of the resin was assessed by quantifying the uncrosslinked tyrosine residues within the hydrogels, and by measuring the CD of the samples. Unreacted riboflavin content was found to decrease as the ED increased (Figure 7.2B), indicating that the chemical initiation does occur. For the lowest ED 3 sample (8.4 mJ/mm ) the initiator residue was 72% the initial amount, while the 3 highest ED sample tested (134.4 mJ/mm ) had residues as low as 1.4%. The 3 riboflavin initiator is mostly consumed when crosslinked with a 67.2 mJ/mm ED (12% of the initial riboflavin is left unreacted) and close to absolute consumption 3 with 134.4 mJ/mm . The sol fraction of the samples was also decreased in a non- linear trend as the ED increased (Figure 7.2C); here, the decrease in sol fraction fell within a narrow range (73.2-68.6%), yet the differences between groups were significant (p<0.05). Both sets of data were acquired on samples that had thoroughly completed the rinsing protocol, after which no leachable products should be left in the hydrogels (See Supplementary data and Figure 7.S1). All subsequent sets of data were rinsed following the same protocol to avoid soluble leaching confounding in the results. FTIR spectra (Figure 7.2D) showed four characteristic peaks at 1400, 1450, -1 1525, and 1650 cm which are representative of C=C stretch patterns in the rings of aromatic compounds (Figure 7.1A); the comparison between groups and the 179 uncrosslinked control showed significant changes dependent on the ED (Figure 7.2E). TGA (Figure 7.2F) performed on representative samples of low and high ED showed that profiles of crosslinked samples differed from the unreacted reference within a temperature range of 150 to 275 ?C. DSC showed a characteristic peak (around 200 ?C) in the uncrosslinked keratin profile; for crosslinked samples the peak shifts to 220-230 ?C and changes in magnitude depending on the ED (Figure 7.2G). All samples present a defined peak, but as the ED increases the peak progressively disappears. As described in the Methods and Equation 7.2, the changes in potential energy, enthalpy, were used to calculate a DSC CD for the samples. As illustrated in Figure 7.2H, there is non-linear relation between ED and CD; the CD increases as ED increases at lower values until it reaches saturation point between 33.6 and 67.2 3 mJ/mm , after these values all samples have a maximum CD close 60 %. From this, it 3 was determined that samples with ED of 16.8, 33.6, 67.2, and 134.4 mJ/mm had a CD of 19.6, 46.9, 54.6, and 59.5 %, respectively. 180 Figure 7.2. The effects of energy density (ED) on the crosslinking density (CD) of keratin membranes. A) The ED parameter was used to produce a wide range of samples by combining thicknesses and exposure times at a fixed UV intensity of 350 2 mW/dm . Bars that share the same color were produced with different parameter combinations but have the same ED (listed on top on each bar). The consumption of riboflavin (B) and the sol fraction of the hydrogels (C) confirmed that the crosslinking reaction (Figure 7.1A) occurs and is ED-dependent. D) FTIR spectra -1 shows four peaks between 1400 and 1650 cm , indicative of changes in aromatic structures, that increasing for the higher ED samples. E) The area under the FITR peaks for the samples was normalized against the area under the peaks for unreacted keratin; this quantification shows the change in aromatic C=C stretch between samples and its relation to ED. F) TGA profiles for comparison of crosslinked and uncrosslinked keratin, indicating the formation of bonds and changes in the profiles between 150 and 275 ?C. G) DSC thermograms for keratin presents a characteristic peak at 210?C which changes in magnitude in crosslinked samples (blue region). Organized curves can be further compared to show how the peak shifts to 220-230?C and varies depending on the ED of the sample; low energy samples present a defined peak, which progressively decreases as the energy density increases. H) Quantification of these peaks can be related to changes in the enthalpy of the system, indicative of the formation of bonds and the crosslinking degree of the network dependent on the ED. For all plots, samples that do not share the same letter are significantly different (p<0.05) 181 The swelling of crosslinked samples in MEM was studied over the period of 1 month. As presented in Figure 7.3A, keratin hydrogels had a swelling capacity that varies around 1500% and was heavily determined by the ED of the sample. Swelling profiles showed a rapid increase during the first day and then a period of relative stability during the first 5 days. Stable swelling was determined at day 3, the point where most groups present either stability or a maximum peak (Figures 7.3A-B). For the early time points, the low ED samples had a higher swelling degree, quickly reaching peak values 2117?148, 1969?43, and 1735?99 % (25.2, 16.8, and 12.6 3 3 mJ/mm groups, respectively). The high ED groups (above 67.2 mJ/mm ) took longer to hydrate but remained consistently stable at their peak values throughout the 28 d. 3 Here, we considered that the system reached the maximum CD with 67.2 mJ/mm of ED, the level at which the swelling capability no longer increased (p<0.05). After 5d rehydrating, Figure 7.3B shows how low ED groups had a significantly relevant drop in the swelling capacity (p<0.01 and p<0.05), registering drops as steep as -81?10 % 3 3 (16.8 mJ/mm ) and -40?13 % (33.6 mJ/mm ). On the other hand, high ED groups 3 with at least 67.2 mJ/mm had no reduction in their swelling capacity (changes are not significant or increasing, p?0.05), indicating that these groups remained stable after 28d. The effects of lyophilization on the rehydration process of keratin-based hydrogels were studied to assess changes in diameter, thickness, and mass uptake. These studies showed some degree of irreversible frustration due to the drying process, which is the reason all subsequent studies were performed on samples that were always kept hydrated in PBS (See Supplementary data and Figure 7.S2). 182 Figure 7.3. Characterization of membrane properties dependent on the network microstructure. A) Swelling profiles of keratin hydrogels in MEM as a function of ED, showing how low ED samples reach their swelling saturation capacity faster than high ED groups yet they also start a process of degradation at earlier timepoints. B) The drop in swelling capacity between days 3 and 28 are indicative of degradation for low ED groups; high ED samples were proven to be highly stable when crosslinked 3 with at least 67.2 mJ/mm . C) Different combinations of thickness and exposure time were used to produce duplicated ED values to assess the viability of the ED parameter as proposed in Equation 7.1. Even if the duplicates follow close trends, statistical differences indicate that higher complexity in the equation is required (statistical significance p<0.01 (+) or p<0.05 (*)). D-E) Mechanical characterization of the hydrogels shows the relation between ED and elastic modulus, ultimate stress, and ultimate strain., Higher ED samples have higher mechanical properties, following non-linear trends that show a saturation profile that further elucidates on the crosslinking limitations and maximum capacities. For all plots, samples that do not share the same letter are significantly different (p?0.05). 183 The stress-strain curves for the crosslinked keratin samples (Figure 7.3D) showed different behaviors dependent on the ED of the groups. As exposure time increases and sample thickness decreases, the materials become stiffer. This observation was further confirmed by quantifying the compressive modulus as a function of ED (Figure 7.3D, inset). Additionally, as ED increases the ultimate stress increases and the ultimate strain decreases until both trends reach plateau levels (Figure 7.3E). The trends here confirm that the hydrogel networks keep changing 3 with ED up to 67.2 mJ/mm ; after that threshold, the mechanical properties measured reach their maximum values. Using the plateau of the plots, the ultimate stress registered for these groups was 18?2 kPa, while the ultimate strain was 59?3 %. Transport properties were studied to assess the viability of crosslinked keratin as membranes. The relation between ED and the partition coefficient (K) and permeability (P) of the hydrogels was studied using FITCd as a model molecule (Figure 7.4A). For high MW of FITCd (150 and 2000 kDa) there was minimal or no variation of K as ED increased, while the low MW 10 kDa solution presented a sharp increase and subsequent increments of K as ED increased (Figure 7.4B, p?0.05). On the other hand, the relation between ED and P had different trends. For the low MW (10 kDa) solutions, changes in the ED cause no significant variations in the P of the system (p?0.05); differently, for the high 150 and 2000 kDa MW samples there is continuous increase of K as ED increases (Figure 7.4C). The overall increase in P 3 3 between the low (16.8 mJ/mm ) and high (134.4 mJ/mm ) EDs was 28% (not significant), 289%, and 2576% for the 10, 150, and 2000 kDa solutions, respectively. 184 Figure 7.4. The effects of crosslink degree (CD) on transport phenomena across keratin membranes. A) Transport phenomena across the keratin membranes were studied using simplified models of the partition coefficient, permeability, and diffusion transport of adipogenic molecules. B) The partition coefficient of keratin hydrogels in equilibrium state with 10, 150, and 2000 kDa FITCd solutions; ED had greater significant effect on the coefficient for smaller molecules. C) Permeability, on the other hand, is the dynamic interaction of the gel with the solute and was affected by ED for the high MW molecules. D-E) Adipogenic differentiation of hMSCs was significantly affected by the variation of the CD of the membranes used. Imaging and quantification of intracellular lipids determined that membranes with higher CD allow better transport of nutrients and adipogenic molecules after 28 d. For all plots, samples that do not share the same letter are significantly different (p?0.05). 185 The transport of physiologically relevant molecules was studied with the diffusion of adipogenic media across the membranes. As observed in Figure 7.4D, feeding the cells through membranes with 20 and 47% CD resulted in sparse production of intracellular lipid vesicles in comparison with the more homogeneous production of populations fed through membranes with 55 and 59% CD. Quantification of these production levels (Figure 7.4E) shows that all membrane groups allowed adipogenic differentiation of hMSCs. Within the study groups, membranes with 55 and 59% CD were linked to increased lipid deposition, 5 to 8 times greater than the non- differentiated controls, in comparison to the 20 and 47% CD (p?0.05). 7.4. Discussion Membranes have been used in guided tissue regeneration (GTR) to regulate tissue growth. This technique is based on the idea that a physical barrier is needed to stop fast-proliferating cells (such as fibroblasts) from colonizing open wounds, while giving time to the slower specialized cells to migrate and regenerate particular 459?463 microstructures of native tissues . This is a common strategy in the treatment of 459 periodontal diseases and enossal dental implants , as well as the treatment of 461 460 critical-sized bone defects and the preservation of empty sockets . Current membranes provide biocompatibility, prevention of cell migration, integration to the surrounding tissues, clinical manageability, and structural strength in the case of 461,464,475?482 stronger tissues (e.g. bone) . As in other GTR approaches, the inclusion of a membrane could be conducive to the organized regeneration of layers in skin. The membrane should act as a physical barrier for cells, maintaining two populations 186 separated by inhibiting migration from one layer to the other. Simultaneously, the membrane should still be permeable to allow transport of biomolecules such as growth factors, hormones, or nutrients between the two sides. To this end, we proposed the use of keratin-based crosslinked membranes (Figure 7.1) and we aimed to prove that dityrosine bonding could be used to engineer their microstructural and transport properties. 7.4.1. The effects of energy density (ED) on the crosslinking density (CD) of keratin membranes As illustrated in Figure 7.1A, riboflavin activation (and formation of reactive species) occurs in the presence of UV light. The tyrosine amino acids in keratin have a susceptible hydroxyl group that can lose its hydrogen to free radical environments, especially in a persulfate radical catalyst like SPS. The tyrosil radicals formed can bind in pairs forming dityrosine bonds, bridging two keratin chains. When the resin combination is exposed to UV light, riboflavin initiates free radical reactions 416,422,483 mediated by the SPS that result in dityrosine crosslinking . Hydroquinone is an inhibitor of radical-initiated polymerizations, as it is a reducing agent. By using it, the crosslinking reaction is controlled and occurs only when riboflavin initiates under UV. This chemical sequence was confirmed with the consumption of riboflavin as ED increased (Figure 7.2B); sol fraction changes indicated that the structures were being crosslinked to different extents (Figure 7.2C). The riboflavin consumption data indicates that there is enough initiator to keep the reaction occurring with at least 3 134.4 mJ/mm , yet the sol fraction plot shows that the structure seems to reach 187 3 crosslinking saturation between 33.6 and 67.2 mJ/mm . These independent sets of data suggest that the limiting reagent in our crosslinking scheme is not the photoinitiator riboflavin but rather the amount of tyrosine content present in the keratin chains. Even with a wide range of ED, close to 70% of the mass of the hydrogel remained uncrosslinked; this limitation is indicative of the low tyrosine concentration and the amount of bonds available. Knowing that crosslinking was occurring, FTIR analyses (Figure 7.2D) were performed to confirm changes in chemical bonds according to the reaction shown in Figure 7.1A. The spectra showed peaks representative of C=C stretch patterns in aromatic rings, such as those found in tyrosine and dityrosine bonds (Figure 7.1A). The quantification of the peak magnitudes, which correlates to the amount of dityrosine bonds present, increased with ED until it reached a plateau value at 67.2 3 mJ/mm (Figure 7.2E). These results corroborate the findings of Figure 7.2B-C and allow us to determine that i) the dityrosine crosslinking reaction is occurring as designed; ii) the reaction is dependent on the ED, and iii) there is an ED threshold at 3 67.2 mJ/mm after which crosslinking is maximized because of the limited quantity of tyrosine bonds available. Furthermore, the data indicates that the hydroquinone inhibitor works; without it the free radical reaction would continue in the absence of UV light and all samples would display the same CD. We were further interested in obtaining a range of CDs that could be tested for molecular transport and a method that can standardize the production of keratin membranes with designed CDs. To this end, we used the thermal method used by 470 Hirschl et al to measure CD using DSC. In a preliminary calorimetric assessment, 188 TGA profiles of crosslinked samples differed from the unreacted reference; the higher temperatures reached without mass changes are indicative of the formation of bonds that require higher amounts of thermal energy to break (Figure 7.2F). The temperature range where these differences occurred was used for DSC testing. The DSC profiles showed a peak characteristic to keratin samples (crosslinked and uncrosslinked). The reduction of this peak as ED increased is indicative of the reduction of available bonding sites as ED increases; available bonding sites are less stable and have higher potential energy than bonded sites, which are portrayed in higher and wider peaks in the thermogram (Figure 7.2G-H). The relation between ED and CD is non-linear and mirrors the saturation behavior shown by FTIR data in Figure 7.2E. Furthermore, Figure 7.2H not only illustrates the ranges and variations of CD for keratin-based hydrogels, but further confirms the threshold ED between 3 33.6 and 672. mJ/mm . Comparison of CD data to other biologically-derived hydrogels is limited as indirect quantifications are generally reported or calculated as 416,484,485 theoretical crosslink density (in terms of molarity) , an approach we had 422 previously reported for the photosensitive resin . Other approaches in literature generally use indirect assessments of the CD degree (using swelling or degradation 300,483?488 data, sol fraction, mechanical testing, chemical spectra, etc.) or crosslinking 489,490 kinetic models to evaluate the CD of biologically-derived polymers. The non- linear relation between CD and ED can be further exploited as a design criterion; using Figure 7.2H we can now design the CD of a keratin hydrogel by knowing the required ED. This latter can be then achieved by various combinations of the 189 manufacturing parameters UV intensity (I), exposure time (t), resin volume (V), exposed area (A), and sample thickness (h). 7.4.2. Characterization of membrane properties dependent on the network microstructure Having determined the relation between ED and CD, we proceeded to engineer keratin membranes with different physicochemical properties, such as swelling, degradation, and mechanical behaviors. As presented in Figure 7.3A, keratin hydrogels exhibited an initial swelling inversely proportional to the ED of the sample. After 5 days the different groups started diverging and losing their swelling capacities at different rates (Figures 7.3A-B). These changes are indicative of ongoing degradation within the keratin network. The high ED groups (over 67.2 3 mJ/mm ) showed slower swelling kinetics than groups with low ED, likely due to the higher CD. Swelling was no longer ED-dependent in high ED groups, consistent with 3 the other data sets and the constant CD above 67.2 mJ/mm of ED (Figure 7.2H). As the assay is mass-based, mass loss due to degradation would cause the swelling values to drop. Based on these variations and macroscopic observations, degradation started to occur much faster for low ED groups. This is an important set of data for the engineered membranes; eventually, when subjected to physiological aqueous conditions the hydrogels will undergo ?four stages of degradation, namely hydration, 461 strength loss, loss of mass integrity, and solubilization via phagocytosis? . Excluding the last cell-mediated stage, the swelling profiles compiled here provide a clear picture of the behavior that different ED samples would eventually undergo in vivo. Low ED groups will likely undergo higher hydration but also fall into strength 190 loss and loss of mass integrity stages much faster than higher ED groups; furthermore, this can be an indication that low ED groups will be solubilized by phagocytosis faster that their higher ED counterparts. This set of data further proves our bonding threshold identified by sol fraction, riboflavin consumption, FTIR, and DSC assays. Here, we considered that the system reached the maximum CD with 67.2 3 mJ/mm of ED, the level at which the swelling capability no longer increased and remained stable at the 1500-2000 % range. In comparison, we previously reported the 422 swelling capacity for printed keratin scaffolds at 1581.9 % , and in both studies it improves the behavior of other similar protein-based hydrogels formed by 415 417 conventional casting including casted gelatin , keratin-collagen scaffolds , keratin 416 485 484 scaffolds , silk fibroin-chitosan scaffolds , or DTBP-modified chitosan . Swelling profiles were also used to assess the accuracy of the ED equation proposed (Equation 7.1), by comparing the behavior of duplicated ED samples produced with different combinations of thickness and UV exposure time (Figure 3 7.3C). For example, samples with ED of 25.2 mJ/mm were produced with 3 combinations 12-1 and 24-2; same for groups 50.4 mJ/mm (24-1 and 48-2) and 100.8 3 mJ/mm (48-1 and 96-2), as highlighted in Figure 7.2A with the paired columns. Figure 7.3C shows that duplicated groups behave similarly, following very close trends but presenting statistically significant differences. They are close, yet not equal. The ED approach to the production of crosslinked membranes is limited to five variables. These variables were chosen because they are the parameters that we routinely fine-tune during a simple casting process or in stereolithography-based 3D printing. The simplified ED equation proposed adequately incorporates the broader 191 parameters involved in the crosslinking process, but variations in the duplicated groups could be tracked back to more complex properties of the resin. The equation could be further optimized by including parameters such as the optical properties, homogeneity of the material, the penetration and attenuation of UV light through the keratin resin, and the UV absorption by the constituent molecules of the resin. Other publications have detailed on the effect of various complex parameters in light-based crosslinking, particularly for 3D printing purposes, including the critical energy 491,492 required to initiate polymerization, the penetration depth of the curing light , and 493 UV absorption . These reports focused on one or two variables and involved synthetic materials that have been extensively characterized. Other parameters such as initiator concentration, keratin concentration, the shapes and resolution, or 422 cytotoxicity, had been previously studied during the formulation of the resin . The ED parameter proposed does not depend on properties specific to the material but on variables involved in the curing process, and has been proven to be consistent using a highly complex protein-based material such as keratin. Based on the consistency of our results and the cross-examination with a variety of chemical and mechanical assays, the energy density parameter ED can be reliably used to characterize the physicochemical properties of crosslinked membranes (<3 mm thick). It is worth highlighting that any attempts to produce thicker samples (anything thicker than 4-5 mm) should consider expanding the ED equation to include exponential decay parameters that account for the decay of UV light intensity with increasing penetration depth during crosslinking, according to the Beer-Lambert?s law. For 192 instance, intensity parameter I could be replaced by , given by averaged integration of the Beer-Lambert?s law over the film thickness. Furthermore, the swelling profiles provide preliminary information on the degradation of the hydrogels and highlight a potential release limitation of the system. 436 Saul et al. previously discussed the considerable effects of diffusion- and degradation-mediated release from self-assembled keratin hydrogels. As observed in Figure 7.3A, decreases in swelling potential could be indicative of degradation and of degradation-mediated release. In this study, the primary goal was to understand the degree of diffusion-mediated release of uncrosslinked components (both uncrosslinked keratin and unreacted chemical resin components), as studied with data such as Figures 7.2B-C and 7.3A-C. Diffusion-mediated release could potentially impact diffusion transport across the membrane; understanding this phenomenon provided a testing time frame when the membranes had no uncrosslinked components leaching or major degradation. This was addressed by implementing a consistent rinsing protocol (Supplementary data and Figure 7.S1) to remove the bulk of the uncrosslinked components and by using the swelling profiles (Figure 7.3A) to determine a time frame (3-5 days) before any ED group presented a significant drop in mass (quantified as swelling degree). With time, which depends on biochemical cues and environmental factors, there will be degradation-mediated release. This scenario will release (1) crosslinked components that are breaking down or are released by bond cleavage, and (2) additional uncrosslinked components trapped within the crosslinked mesh. 193 Mechanical assessment of the membranes revealed the effect of ED on the compressive moduli and ultimate stress and strain of the keratin materials. The data correlates to the crosslinking dynamics of the system. Even if the energy delivered can keep increasing, the crosslink bonds are dependent on the available tyrosine, and thus it reaches a maximum CD and maximum structural properties. On one hand, swelling data indicates that the system is reaching the maximum capacity with 67.2 3 mJ/mm of energy, a level at which swelling is significantly more stable (Figures 7.3A-B). On the other, the elastic moduli and ultimate stress and strain registered for 3 samples with ED below 67.2 mJ/mm show a significantly increasing trend, indicating they are within a range of variable CD (Figures 7.3D-E). Overall, the swelling and mechanical profiles and the data compiled from them confirm our expectations of the UV-photosensitive chemistry and its saturation, and again confirm the ED threshold. We discussed the limitations of the ED parameter proposed, but the behavior in swelling and mechanical properties indicate that the keratin membranes can be fine-tuned over wide ranges using our simplified ED approach. It is worth highlighting again that the goal of this section was to prove that the ED parameter, which defines CD, can be used to produce keratin membranes with a variety of properties, including a wide range of swelling profiles, degradation behaviors, mechanical properties, and diffusion capabilities. Here, we present a range of options to our fellow researchers with the interest of showing the versatility of our ED approach and the membranes produced. Hydrogels produced with low ED might be 436 suitable for applications interested in fast release, such as the studies of Saul et al. for the sustained release of bioactive ciprofloxacin from keratin hydrogels, while we 194 expect to use the high ED hydrogels for sustained use as membranes that should have lower effects on diffusion applications. 7.4.3. The effects of crosslink degree (CD) on transport phenomena across keratin membranes The use of the keratin-based crosslinked hydrogels as membranes was proposed based on the criteria that it should 1) act as a physical barrier for cells so that two populations can grow adjacently without mixing; 2) be permeable to allow transport of biomolecules to allow cross-talk between the two layers; and 3) it should be able to degrade as the layers form, without leaching cytotoxic by-products or having major impact on skin properties such as water transport, elasticity, or scarring. Here, we focused on studying the second criterion using membranes with a range of CD that would impact molecular transport of model molecules, growth factors, and nutrients across the hydrogel (Figure 7.4A). Hydrogels were successfully implemented as membranes using Transwell inserts with the addition of 3D printed stoppers, which allowed sealing and ensured that all diffusion occurred through the membrane. K was determined by the size of the FITCd molecules in solution, but it was further affected by the ED of the membranes. As defined in Equation 7.3, K is the ratio between the concentrations of solute in the hydrogel and in solution. In a balanced gel-solution system, half of the solute would theoretically localize in the gel and half in the solution. The smaller 10 kDa groups resulted in an unbalanced system where most of the solute was uptaken by the hydrogel (K ? 0.5); the larger 2000 kDa groups were also unbalanced, although the majority of the FITCd molecules remained in solution (K ? 0.5). The 195 150 kDa molecules seemingly reached a balanced state of equilibrium, were the solute was closely equally distributed between gel and solution (K ? 0.5). This indicates that smaller molecules are uptaken and retained within the membranes even against a concentration gradient. The partition coefficient and permeability studies using FITCd elucidated on the effect of membrane design properties, mainly ED and CD, on the transport of different-sized molecules (Figures 7.4B-C). In general, taking the increasing trends between K and ED for all MWs, the higher ED allows higher uptake of solutes into the hydrogel. At any particular time the low MW samples are more uptaken by the hydrogels (K of up to 0.8) than the higher MW samples (maximum 0.3 and 0.5) no matter the ED. On the other hand, P increases as ED increases for the high MW. K is the property in equilibrium, while P is the product of the equilibrium (how much solute is uptaken and remains in the gel) and the dynamic transport (how much solute moves across the hydrogel). When pairing both sets of data we can conclude that for the low MW higher amounts of solute are accumulated within the keratin gel, a trend that does not change with increasing ED. This indicates the low MW samples remain within the membranes. On the other hand, larger molecules present higher P and lower K values through the gels which are indicative of transport across the barrier. For these, the effect of the ED of the membranes is clearer. As the ED of the membrane increases the P and K values also increase, indicating that increasing ED does represent better diffusion rates through the membrane. In terms of molecule size and shape, the Stoke?s radii for FITCd 10, 150, and 2000 kDa is reported as 494,495 approximately 23, 85, and 270 Angstroms, respectively . Low MW FITCd 196 (below 2 kDa) is rod-like, while chains with 2 to 10 kDa behave like flexible coils. FITCd larger than 10 kDa behaves as highly branched polymers and become 494,495 increasingly symmetrical . The size and shape indicate that the smaller molecules might be interacting chemically or electrostatically with the networks or are trapped within the smaller pores of the hydrogel. The larger molecules, due to their size and flexibility, are probably crossing the hydrogel via the larger pores, slower but without interaction or entrapment. Low CD membranes, which present lower values of P, may have unreacted portions within their chains that can be responsible of reducing transport; as the CD increases, interaction between the chains and the molecules is reduced, allowing higher transport rates. The transport of molecules across the membrane was further studied with the diffusion of adipogenic molecules for the differentiation of a target hMSC population. The healthy development of the adipose layer (hypodermis) in severe burn wounds is of particular interest. This is a rich layer of adipocytes and stem cells that synthesize 35 adipose-specific ECM (mostly collagen types I, III, IV, V, and VI ) and growth 32,36 37,38 factors (highlighting leptin and adiponectin hormones, and basic fibroblast 40 35 growth factor ) key players in re-epithelization, wound healing, and angiogenesis . If the hypodermis is producing all these factors in vivo, diffusion through the tissues is transporting them to the adjacent dermis layer. Ensuring the transport of biomolecules across keratin membranes will further help us to eventually replicate the transport between stratified hypodermis and dermis to prove its benefits in regeneration of skin. As presented in Figures 7.4D-E, the membranes with higher CD (55, 59%) allowed transport of molecules involved in adipogenic differentiation that 197 resulted in higher lipid deposition, in comparison to the 20 and 47% CD groups. Formation of intracellular lipid droplets are evidence that adipogenic factors were transported; some combination of the growth factors, hormones, amino acids, vitamins, sugars, and steroids found in the DMEM, FBS, P/S, sodium pyruvate, dexamethasone, insulin, IBMX, and indomethacin that make up the adipogenic media, were able to cross the membrane and reach the cells to induce adipogenic differentiation. Similar to behavior observed in the P and K experiments, some fraction of the adipogenic molecules is always transported as no membrane proved to fully inhibit molecular transport, a fact backed by all membranes allowing differentiation of the cells (as compared to a no-membrane scenario). Considering that the differentiation of hMSCs is a delicate process that requires a very specific combination of the components mentioned before, we can assume that the differentiation observed is the result of all, or most, molecules being able to cross the membranes. Based on our characterization (FTIR, DSC crosslinking density, and the quantification of structural and diffusion properties), there is a marked difference with 3 membranes produced with ED below the 67.2 mJ/mm threshold. At that ED level, all results indicate that the membranes have reached a maximum level of crosslinking (CD of 55 - 59%); over this ED threshold, the membranes allow better transport of growth factors that result in improved adipogenic differentiation of hMSCs. 7.5. Conclusions The goal of this work was to develop a GTR-based membrane with engineered microstructural and transport properties for future use in the development 198 of stratified tissues. We introduced an energy density (ED) parameter that incorporates simplified casting and printing parameters for the sequential production of thin keratin-based membranes. Our keratin-based resin has been optimized for the reproducible manufacture of membranes with defined CD and physicochemical properties, which correlate to the ED level selected. Keratin membranes allow diffusion of molecules such as media nutrients and growth factors, and the ED of the crosslinked network can be used to regulate the transport profiles. Further studies are warranted to quantify the cross-talk between cell populations across the membrane, evaluate the barrier effect of the membranes on cell migration, and the in vivo development of dermal layers. Overall, we can now engineer and manufacture our membranes and conclude that fine tuning the CD is valuable to control several microstructural properties of the hydrogels, including swelling and degradation profiles, mechanical properties, and transport across the network. 7.6. Supplementary Materials and Methods 7.6.1. Rinsing of the leachable components over time Cylindrical samples were exposed to UV for 6 min (13 mm diameter, 1.5 mm thick) and 96 min (5.3 mm diameter, 1.5 mm thick) as described in the previous section, producing samples groups 6-1.5 and 96-1.5, representatives of low and high energy density (ED) groups respectively. Six samples of each group were collected after casting, lyophilized without rinsing, and then weighed (unrinsed total mass). All other samples were subjected to sequential rinses in phosphate buffered saline (PBS, pH 7.4). Six samples of each group were collected after 1, 3, 5, 7, 10, 12, 15, and 20 199 rinses. Each rinse consisted of fresh PBS for 15 min, at room temperature. The tenth rinse was a single, extended overnight rinse (14h), part of our usual rinsing protocol. After collection, samples were lyophilized and weighed (rinsed mass) (n=4 to 6). The masses of all lyophilized samples were recorded using a microbalance (Sartorius ME- 5, Sartorius, Goettingen, Germany). 7.6.2. Effect of lyophilization on swelling properties Cylindrical samples (13 mm diameter, 1.5 mm thick) were exposed to UV for 12 and 96 min as described in the previous section, producing samples groups 12-1.5 and 96-1.5, representatives of low and high ED groups respectively. All samples were subjected to sequential rinses in PBS over 24h and then weighed and measured (Original mass, diameter, and thickness). Half of the samples from each group were returned to PBS and stored at 4?C; the other halves were lyophilized. All samples were measured again (Initial mass, diameter, and thickness) and then moved to 2 ml excess minimum essential medium (MEM, Life Technologies, Frederick, MD) at 37?C. The mass and dimensions of the samples were recorded after 30min, 2h, 20h, 46h, 70h, and 116h. At each time point the sample was taken out of solution, gently blotted to remove excess MEM, measured, and then returned to the solution (n=7). The volume of MEM was kept at 2 ml by refilling at the third day. 200 7.7. Supplementary Figures Figure 7.S1. Removal of all soluble mass from the crosslinked membranes. Low CD samples present a high loss of mass after the initial, but stabilize after the third or fifth rinse; on the other hand, high CD samples, have a lower mass loss and are stable after the first rinse. Measurement of these mass changes on the microbalance, and the low standard deviation errors, allow us to conclude that no mass changes are occurring once the rinsing protocol is complete, indicative of the fact that all leachable products left after the UV crosslinking reaction have been removed. For all plots, samples that do not share the same letter are significantly different (ANOVA with Tukey?s comparison, p<0.05). 201 Figure 7.S2. The effects of lyophilization on swelling of the crosslinked membranes. Lyophilized and non-lyophilized keratin membranes were rehydrated to assess any additional entanglement caused in the freezing and drying processes. A) During rehydration, there are significant differences due to lyophilization at all time points 3 3 for both low (12-1.5, ED 16.8 mJ/mm ) and high (96-1.5, ED 134.4 mJ/mm ) CD membranes. On the other hand, diameter (B) and thickness (C) are recovered over time and are comparable to non-lyophilized samples. Overall, dimensions can be broadly restored with rehydration, but the mass uptake can be irreversibly decreased; this is indicative that swelling assessments, which require a lyophilization step, can be offset due to additional entanglement. ANOVA with Tukey?s comparison was performed on each group independently, samples that do not share the same letter are significantly different within that group as denoted by color (p<0.05). Comparison between lyophilized and non-lyophilized samples was assessed at each time point using two-sample t-test for the mean (p<0.01 (+) or p<0.05 (*)). 202 Chapter 8: Dual-Chambered Membrane Bioreactor for Co- 8 Culture of Stratified Cell Populations 8.1. Introduction Skin is composed of multiple stratified layers that act as a protective barrier 33,496 against external mechanical and biochemical factors . Mammalian skin layers include the epidermis, dermis, and hypodermis, i.e. the subcutaneous fat tissue. The 4,31 outermost epidermis layer consists almost entirely of keratinocytes (~95%) . The dermis, which provides mechanical resistance, is mostly fibroblasts synthesizing 4,30?32 collagen (mostly type I, and minor types IV and VII), elastin, and proteoglycans . 32 Lastly, the hypodermis, considered part of the endocrine system , consists of mature 30,31 adipocytes that provide insulation for energy storage and thermoregulation . This 33,35 32,36 layer is an important source of stem cells , hormones such as leptin and 37?39 40 adiponectin , and growth factors , all players in the re-epithelization and 41,43,497 angiogenesis processes of skin . The combination of properties between the layers allows for skin to be an efficient barrier and any damage to it results in immediate compromised 4,10,496 thermoregulation, massive fluid shifts, and risk of bacterial sepsis . Furthermore, blood vessels and neural networks grow through the interface of the hypodermis and 4,36 dermis forming the deep vascular plexus . A lack of healthy, synergistic development of all the layers results in unviable nerve growth or deficient 8 As prepared for: J Navarro, J Swayambunathan, M Santoro, JP Fisher: ?Dual-Chambered Membrane Bioreactor for Co-Culture of Stratified Cell Populations?. [For Biotechnology and Bioengineering]. 203 30,32 vascularization which impair the sensing and thermoregulation functions of skin . Irregular interphases between the layers can also lead to improper adhesion, fluid 30 collection (blisters), or separation . Just as important as the layer-to-layer relation is for function in vivo, stratifying the hypodermis and dermis layers could play a major role in healing and regeneration processes in vitro. By allowing layer-specific factors to diffuse into an adjacent layer it would be possible to study how communication between layers can drive development or regeneration, but is imperative to have both tissue layers growing together and communicating. Nevertheless, hypodermis adipocytes and dermis fibroblasts are not cultured in vitro using the same media, nor are their progenitor stem cells. To produce a dermis-hypodermis complex that benefits from the synergistic development of both cell populations it is necessary to develop new technologies for stratified co-culturing. Here we propose the development of a novel dual-chambered bioreactor (DCB) to culture and study stratified 3D cell populations, building towards the formation of a dermis-hypodermis complex, as illustrated in Figures 8.1A-B. Perfusion bioreactors can be used to dynamically culture a population of cells 373,498,499 with specific growth media . If different scaffolds are grown in separate bioreactors, they each develop a homogeneous cell population with independent characteristics. The scaffolds can be bound together in vitro or in vivo using a variety 221,500?506 507? of layering techniques, usually chemical crosslinking , melting/sintering 510 511 157,512?516 , self-assembling , or by inducing extracellular matrix (ECM) deposition , but it is difficult to keep cell populations physically separated while allowing paracrine signaling among them (Figure 8.1Bi). The rationale behind the proposed 204 DCB consists on being able to flow two different types of media into the same chamber so that a single construct is developed. In the DCB, two parallel flow inlets connect to a common chamber, followed by two flow outlets (Figure 8.1Bii). As it is, the two flow lines gradually mix with each other due to diffusion between the chambers, and because the bioreactors run on loops. This mixing results in continuous gradients rather than a defined interface. The introduction of a permeable barrier within the DCB reduces flow mixing without altering the diffusion of small molecules from one side to the other (Figure 8.1Biii). The membrane reduces the formation of gradients by reducing cellular migration and unwanted deposition of ECM, but still allows communication between both sides by the diffusion of biomolecules such as growth factors or hormones originating from the cells and the media. Furthermore, by employing membranes with varying permeability to regulate diffusion between the lines, we aim to control molecular cross-talk between cell populations. To meet these requirements, we used crosslinked keratin-based membranes. We previously reported the development of keratin-based hydrogel 441 membranes that can be fine-tuned to alter diffusive transport across them . Here we assessed the viability of our DCB for growing adjacent cell populations with the inclusion of a keratin membrane. We hypothesized, as illustrated in Figure 8.1Biii, that using the keratin membranes in the DCB chamber regulates the flow separation between the cell populations but does not impede molecular diffusion, a feature that can be exploited to produce stratified or gradient constructs. To this end, the objectives of this work are to assess the viability of the DCB to: 1) perfuse two 205 adjacent layers with different media; 2) allow communication between layers via diffusion through a permeable barrier, both under static and dynamic conditions; 3) co-culture and sustain two adjacent cell populations; and 4) allow studying the effects that one cell population can induce in an adjacent layer. 8.2. Methods 8.2.1. Bioreactor design and 3D printing A computer-aided design (CAD) model file for the DCB was created in SolidWorks? (Dassault Syst?mes SE, V?lizy-Villacoublay, France). As illustrated in Figures 8.1Ci-ii, the DCB is assembled using a master part twice. This part consists of an inlet and an outlet that feed into a chamber with an open wall (internal dimensions: 10 x 10 x 8 mm), and channels for silicone rubber sealing gaskets (custom-made, 50A durometer, 2 mm thick) (Figure 8.1Cii). The master parts were 3D printed out of EShell 300 resin (EnvisionTEC Inc., Dearborn, MI) on a stereolithography-based 3D printer (Perfactory 4 DLP printer, EnvisionTEC). Stainless steel hardware (#10-24 0.5? machine screws and nuts) and the gaskets were used to close the DCB. When assembled, the cross-section reveals a common chamber (10 x 10 x 17.5 mm) with parallel inlet-outlet lines (Figures 8.1Ciii-iv). The inlet/outlet connections match 16-gauge pump tubing, which were then connected to 1/8? ID platinum-cured silicone tubing using 1/8?-1/8? Kynar connectors as needed (all Cole-Parmer, Vernon Hills, IL). Porous scaffolds (9.8 x 9.8 x 8 mm, with 600 3 ?m pores) were designed to fit into the chambers and 3D printed with EShell 300. For experiments involving cell culturing, all EShell 300 parts and scaffolds were 206 sterilized using ethanol:phosphate buffered saline (PBS) serial washes (100:0, 75:25: 50:50, 25:75 and 0:100 for 20 minutes at each step under ultraviolet (UV) light) and stored in fresh sterile PBS at 4 ?C until use. All tubing, connectors, stainless steel hardware, and the silicone gaskets were autoclaved. Figure 8.1. DCB design and 3D printing. A) Layered microstructure of human skin and our approach to the development of a simplified layered tissue engineering scaffold. B) Approach to the formation of strata and gradients in a DCB. i) When cultured independently, scaffolds have a single, homogeneous cell population; layering the scaffolds in vivo would result in poor integration and uncontrolled gradient interface. ii) The DCB allows two flow lines of media into a common chamber; without a barrier, the two flows would mix producing a gradient. iii) Inclusion of a membrane in the DCB would keep the flow lines separated producing a stratified construct; permeability of the membrane would regulate transport between the two sides. C) i) The DCB is composed of two equal master pieces that can be 207 sealed together; ii) the master piece is composed of an inlet, an outlet, and an open centerpiece, as well as including guide channels for sealing gaskets; iii-iv) when combined, the cross-section reveals a common chamber that provides an interface for communication between the two flow lines. D-E) The master pieces and custom porous scaffolds for cell culturing were 3D printed on a Perfactory 4 DLP printer (EnvisionTEC) using EShell 300 resin. F) The final DCB system assembled has been optimized to have maximum external dimensions of 64 x 40 x 23 mm (length, width, height), an effective internal volume of 3.8 ml (total volume as seen in Fig. Civ in green, shades distinguish the two halves that compose it), and inlet and outlet connections to standard 1/8? (3.175 mm) inner diameter tubing. 8.2.2. Computational modeling The CAD files created were used in SolidWorks? Flow Simulation add-in package to conduct computational fluid dynamics (CFD) simulations of the transport dynamics in the DCB. Lids were created out of the two inlets and outlets to define the fluid volume modeled. Parameters defined included gravity anti-parallel to flow (- 2 517 9.81 m/s ), type of flow (laminar and turbulent), and the fluids involved . The water profile was copied and renamed to create identical Donor Fluid and Receiver Fluid. Boundary conditions defined outlets at environmental pressure (101325 Pa), and inlets to equal flow rates (at 1, 4, or 10 ml/min). At inlet one, the initial volume fraction of Donor Fluid was set to 100%, while Receiver Fluid was set to 0%. Likewise, at inlet two the initial volume fraction of Donor Fluid was set to 0%, and Receiver Fluid to 100%. After running the mesh and calculation, the results were used to map the shear stress, velocity, and the volume fraction distributions in the DCB. The maximum shear stress in each chamber and the final volume fraction of both fluids at each outlet were recorded. These values were used to set up iterations modeling the DCB loops; the output volume fractions were used as the inlets volume fractions and ran again as the next iteration. This iteration process continued until the 208 volume fraction of each fluid in each chamber was 50 ? 1%, after which the system was assumed equilibrated. These iterative runs were repeated for every flow rate with and without a solid porous barrier. The barrier, to model the membrane, was a solid plate with holes to match 55% (100 holes, 420 ?m diameter) or 99% (4 ?m thick wire mesh) porosity. The iteration count was normalized to flow rate to get time- comparative differences between groups, assuming all chamber volumes were equal and constant. The final data set described the volume fraction of the donor and receiver at every normalized iteration point proportional to time. 8.2.3. Preparation of keratin membranes Keratin was extracted from oxidized human hair using a proprietary method 418,518 by KeraNetics LLC (Winston-Salem, NC) . The keratin-based 441,518 photocrosslinkable resin was prepared as previously reported . Briefly, keratin was dissolved in PBS (pH 7.4) at a 4% wt/vol concentration, and then mixed with a photosensitive solution at a 4:1 ratio. The photosensitive solution consisted of 1 mM riboflavin (Sigma-Aldrich Co., St. Louis, MO), 200 mM sodium persulfate (SPS, Sigma-Aldrich), and 0.001% wt/vol hydroquinone (Sigma-Aldrich). After thorough mixing the resin is curable under UV light by formation of dityrosine bonds. The resin was used in 3D printed molds to cast membranes (12 x 12 x 1.5 mm), exposing 2 to UV at constant intensity of 350 mW/dm . Membranes were produced with 20% low and 59% high crosslinking density (LCD and HCD) by regulating the total amount of energy delivered to the sample during crosslinking, as further detailed in 441 our previous publication . Samples were thoroughly rinsed to remove uncrosslinked components and stored in PBS at 4 ?C (Figure 8.3A). 209 8.2.4. Membrane degradation assessment LCD and HCD membranes were used under static and dynamic conditions in various tests described below. After each test, membranes are recovered from the DCBs and qualitatively assessed according to the criteria outlined in Table 8.1. Table 8.1. Qualitative assessment of keratin-based membranes recovered from DCB runs Qualification Criteria Complete Full membrane recovered in a single piece Membrane recovered in a single piece, parts are missing or Partial thinned out Membrane recovered in pieces, parts are missing or inside Broken the scaffolds None left No membrane left between or inside the scaffolds 8.2.5. Diffusion in the DCB The DCB chamber was assembled using two master parts, each fitted with a porous scaffold, and two silicone sealing gaskets (Figure 8.3B). Sampling ports were opened on each side of the DCB and kept sealed with casted paraffin plugs, as illustrated in Figure 8.3C. The DCB was filled with PBS, each side holding 1.6 ml, after which all inlets and outlets were sealed, and the system was left to stabilize for 1h. This setup was evaluated without a membrane, with LCD, or with HCD membranes (at least n=4). Next, 50?l of green food dye (McCormick, Baltimore, MD) which is composed of tartrazine (FD&C Yellow 5, 534.4 Da) and brilliant blue FCF (FD&C Blue 1, 792.8 Da), was added to the donor chamber. A 5 ?l sample was collected from the donor and receiver sides using the sampling ports and diluted with 95 ?l of PBS. Sampling was done every 3 min for 1 h for the cases without 210 membrane; the cases with membranes were sampled every 5 min for the first hour, then every 30 min for 8h, and followed with sampling every 2h during 10h segments for up to 8 days. Samples were tested in triplicate for absorbance values of tartrazine and brilliant blue molecules at 425 and 630 nm using a SpectraMax M5 plate reader (Molecular Devices, Sunnyvale, CA) alongside a concentration ladder. Concentration values were normalized to initial donor concentrations. Additionally, at the end- points, membranes were collected from the DCBs and assessed using Table 8.1. From the sampled data, the permeability of the system was calculated assuming constant volume (total volume loss in each chamber was 5.0 to 11.2%), known initial donor and receiver concentrations, and that flux across the membrane quickly reaches steady state. As such, we used a pseudo-steady state model, in which the concentration profile across the membrane is assumed invariant, to calculate the 472 permeability (P) of the molecules crossing the hydrogel membrane by plotting : Cd ? Cr PtA 1 1 ?ln ( ) = ( ) ( + ) (8.1) Cd0 ? Cr0 L Vd Vr Where Cd is donor concentration at time t; Cr is receiver concentration at time t; Cd0 is initial donor concentration; Cr0 is initial receiver concentration; t is time; A is membrane area; L is membrane thickness; Vd is donor volume; and Vr is receiver volume. 211 8.2.6. Bulk convection in the DCB A multi-channel rotary pump (Masterflex L/S with 4-channel or 8-channel pump head) was used to run parallel DCBs. As illustrated in Figure 8.4A the DCB requires two pump lines to function, each line connects in sequence the pump, the chamber inlet, chamber outlet, a 3-way connector to a 1 ml syringe for sampling, a 50ml reservoir, and back to the pump. The bioreactor loops are filled with PBS, removing all air from the lines and scaffolds, and equilibrated for 30 min at constant flow rate (1, 4, or 10 ml/min). The pumps were then stopped, and the initial volume of the reservoirs was recorded. Flow was restarted and ran undisturbed for 24h, the point at which the final volume of the reservoirs was recorded. This setup was evaluated without a membrane, with LCD, or with HCD membranes for each flow rate (at least n=4). The change in volume in each line (named donor and receiver) over 24h was calculated as the difference in the reservoir volume between 0 and 24h, over the initial reservoir volume, and reported as a percentage. 8.2.7. Dynamic assessment of transport in the DCB (convection and diffusion) Closed loop DCB setups using the rotary pump (same connection described above, Figure 8.4A) were used to quantify diffusion in the system under dynamic conditions. The bioreactor loops were filled with PBS, removing all air from the lines and scaffolds, and equilibrated for 30 min at constant flow rate (1, 4, or 10 ml/min) to have stable 40 ml volumes in the reservoirs. This setup was evaluated without a membrane, with a Parafilm? M (Bemis Co., Oshkosh, WI) impermeable barrier, with LCD, or with HCD membranes for each flow rate (at least n=4). After, 200 ?l of green dye (McCormick) was added to 40 ml PBS in the donor reservoir. A 300 ?l 212 sample was collected from the donor and receiver sides using the sampling syringe. Sampling was done every 30 min for two hours, then every hour for 8h, and followed with sampling every 2h to complete 24h. Samples were tested in triplicate for absorbance of tartrazine molecules at 425 nm using a SpectraMax M5 plate reader (Molecular Devices, Sunnyvale, CA) alongside a concentration ladder. Concentration values were normalized to initial donor concentrations. As before, membranes were collected from the DCBs at the end and assessed using Table 8.1. 8.2.8. Cell seeding and viability in the DCB 2 Printed porous EShell 300 scaffolds were coated with 3 ?g/cm fibronectin (from bovine plasma, Sigma-Aldrich) in PBS at 37?C for 12h to facilitate cell 519 attachment . Complete coating of the 3D structure was assured by pipetting the fibronectin back and forth through the scaffold and rotating every 30 min. The coated scaffolds were seeded with mouse fibroblasts (L929s, ATCC, Manassas, VA) at a 2 density of 10000 cells/cm for 4h. The scaffolds were then inserted into DCB chambers and into simplified closed loop setups using the rotary pump (Figure 8.6A- B). This setup was evaluated without a membrane (n=4). Additional control cultures included a 2D static culture, and a 3D static culture of a seeded scaffold in 12-well plate. Both lines A and B of the bioreactor (Figure 8.6A) were filled with L929 growth media consisting of minimum essential medium (MEM, Life Technologies, Frederick, MD) supplemented with 10% horse serum (ATCC), and ran for 7 and 28d at 4 ml/min. Fresh growth media was replaced in reservoirs A and B every 2d. At the end-point, culture viability of the DCB was assessed by imaging scaffolds using a 213 bright-field microscope (Axiovert 40CFL, Zeiss, Thornwood, NY) fitted with a digital camera (SPOT Insight 1120, Diagnostics Instr., Sterling Heights, MI) at 2x. 8.2.9. Adipogenic differentiation in the DCB across the keratin membrane As in the previous test, 3D printed porous EShell 300 scaffolds were coated 2 with a 3 ?g/cm fibronectin solution at 37?C for 12h. The scaffolds were then seeded with human mesenchymal stem cells (hMSCs, Lonza, Walkersville, MD) at a density 2 2 of 5000 cells/cm (total surface area of scaffold: 49.2 cm ). Control hMSCs were cultured statically on 2D 12-well cultures plates and statically on 3D coated scaffolds, 2 at the same 5000 cells/cm . The seeded scaffolds were loaded into DCB chambers and into simplified closed loop setups using the rotary pump as before (Figure 8.6A- B). This setup was evaluated without a membrane, with LCD, or with HCD membranes at 4 ml/min flow rate (at least n=3). Both lines A and B of the bioreactor (Figure 8.6A) were filled with hMSC growth media (high glucose Dulbecco?s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS), 1% non- essential amino acids (NEAA), and 1% penicillin/streptomycin 100 U/100 mg (P/S)). Fresh growth media was replaced in the reservoirs of both lines every 2d; same media and periodicity was used on the static controls. After 7d, once the controls reached ?90% confluence, the scaffolds in the DCB were assumed ready for differentiation. The media in line A was replaced with hMSC adipogenic media (DMEM with 10% FBS, 100 U/100 mg P/S, 1 mM sodium pyruvate, 1 mM dexamethasone, 10 mg/ml insulin, 0.5 mM 3-isobutyl-1-methylxanthine (IBMX), and 200 mM 473,474 indomethacin) , while line B was filled hMSC growth media. Media on both DCB lines and the static controls was refreshed every 2d for 21d. Cells were lifted by 214 continuously pipetting trypsin through the scaffolds, and re-seeded on white 96-well plates for 12h; controls were equally lifted and re-seeded. Adipogenic differentiation of hMSCs in DCBs and controls was assessed using fluorescent AdipoRedTM Assay Reagent (Lonza) to quantify intracellular lipid accumulation (n = 5). 8.2.10. Statistics ANOVA and Tukey?s multiple pairwise comparison (significance using p < 0.05) were used for multi-group comparisons in Minitab 18 (Minitab, Inc). Differences between individual groups and references were assessed with two-sample t-test for the mean (significance using p < 0.01 (**) and p < 0.05 (*)). Data is presented as mean ? standard deviation. 8.3. Results and Discussion The goal of this work was to assess the viability of our DCB for co-culturing adjacent cell populations with the inclusion of a regulatory keratin membrane. As illustrated in Figure 8.1B, we hypothesized that including a membrane in the DCB?s common chamber can regulate flow profiles and the separation of cell populations without impeding molecular diffusion for intercellular communication. Simply, we aim to keep cell populations separated but able to communicate with each other. The designed system (Figure 8.1C) provides two parallel flow lines into and out of a common chamber. The inclusion of a membrane in the chamber limits mixing and convective flow between lines, and regulates diffusive transport depending on its 441 crosslinking density (CD) . 215 Bioreactors can be used to dynamically stimulate tissue-specific constructs by 372 introducing complex biomechanical cues such as cyclic compression , tensile 371 368,370,377 strain , or shear stress . Orienting or redirecting flow has also been used to 376,378,380,520 induce gradients of cells, oxygen, or biomolecules within bioreactors . Other groups have introduced bioreactors that rely on transport between two compartments to cause either oriented flow or specific shear stress profiles on cells, 377,521 not necessarily having populations in each chamber . For our bioreactor, the two adjacent chambers were purposefully designed to harbor cell populations, and the setup was engineered to hold a membrane to control the flow profile in the system. As we aim to keep cell populations separated but communicating, we have characterized the transport phenomena in the DCB to prove that the balance between convective and diffusive transport can be used to regulate the exchange of biochemical molecules. 8.3.1. CFD modeling of the DCB membrane system The CFD modeling sought to prove that the transport patterns of convection and diffusion hypothesized were possible in a DCB-membrane complex (Figure 8.2A). The simulations produced heat maps of the volume fraction of donor fluid (red) and receiver fluid (blue) for different combinations of membrane porosity (55 or 99%) and inlet flow rate (1, 4, or 10 ml/min). Using the volume fraction, a quantification of the amount of fluid in a mixture volume, it is possible to observe and track changes in the mixing of the donor and receiver fluids. The heat maps (Figure 8.2B) provide an idea of the mixing state when the simulation reaches steady state, having the outlets open to environmental pressure. These qualitatively show that, 216 independent to flow rate, the inclusion of a 55% porosity membrane reduces the mixing of the fluids when compared to a 99% porosity membrane (our model of a non-membrane case). To recapitulate the effects of the looped flow in the bioreactor, the outlet data was sequentially input back into the inlets until an equilibrated mixing state was reached (1:1 concentration donor and receiver fluids in both chambers). As seen in Figure 8.2C, the comparative mixing profiles using the loop approach indicate the iteration points, proportional to time, when the DCB systems reach concentration equilibrium, and show that as flow rate increases the system takes less time take to reach this state. Furthermore, the inclusion of a membrane in the DCB dilates the time required for the looped system to reach equilibrium, independent to the flow rate studied. A 55% porosity membrane will increase the time required to reach equilibrium by 62% at 1 ml/min, 88% at 4 ml/min, and 84% at 10 ml/min. From the CFD data, we can expect to control the length mixing period within the DCB using the membrane porosity and flow rate. The idea of using a membrane is 380 comparable to other systems reported before. Giusti et al. proposed a similar two- chamber modular bioreactor with an intermediate membrane to study intestinal epithelium and solute transport across the cell barrier. The authors demonstrated how application of flow increased the permeability of intestinal epithelial Caco-2 cells in the cell-laden membrane, and higher permeability achieved with longer flow stimuli across the membrane. Here, time and flow are used to determine the level of permeability of the growing cell barrier; for the DCB, time and flow determine the mixing degree of the chambers. Furthermore, in our case the cells are not in the membrane but in the chambers, our membrane is used as a controlled barrier. 217 Multiple bioreactors have been designed to use flow to induce gradients and include 380 377 membranes in some way; the Giusti and Nakagawa bioreactors use the 521 membrane as the cell substrate to study, while the Dimitrijevich system can either use the membrane as the substrate or as a barrier to keep a cell-laden scaffold in place. As mentioned before, in the DCB the membrane is used to regulate the flow profile in the system, particularly the balance between convective and diffusive transport, to determine the biochemical cues stimulating the adjacent cell populations and eventually produce gradient or stratified models as needed. The CFD simulation results were also used to calculate the maximum shear stress at any surface within the modeled DCB (Figure 8.2D). In general, these maximum points were found at the inlets and outlets for the higher 4 and 10 ml/min flow rates, and only found inside the scaffolds for the low 1 ml/min cases (Supplementary Figure 8.S1). The outlets were also the points where the highest velocities of the system were reported for all flow rates (Supplementary Figure 8.S2). Decreasing the porosity of the membrane model (increasing the barrier effect) results in an increase in maximum shear stress values. We can assume that decrease in porosity increases the resistance within the system and causes increments in the shear stress as flow rate increases. These stress values allowed us to select adequate flow rates for subsequent DCB experiments that involve cells. The simulations indicated that the 10 ml/min run with an 55% porosity membrane induced the highest shear forces within the system at 1.1 Pa. Research on the effects of shear stress on hMSCs has reported that the threshold shear stress value for osteogenic 2 differentiation of hMSCs is close to 0.9 Pa (9 dynes/cm ) and the adhesion strength 218 522?524 for MSCs is between 0.3 and 0.6 Pa . The general shear stress maps of the DCB are approximately below 0.2 Pa (Supplementary Figure 8.S1), but the simulations indicate that using a 10 ml/min flow rate in cell culture studies may induce undesirable osteogenic differentiation or produce lifting of the cell populations in particular points in the system. The 1 and 4 ml/min flow rates, which have adequate mixing profiles with and without the membrane in the DCB, are expected to provide a shear environment that sustains cell adhesion and development. Figure 8.2. CFD modeling of the DCB membrane system. A) CAD model for the proposed DCB system with porous scaffolds inserted in the line chambers, detailing the position of the inlets (1 and 2) and the outlets (1 and) and the orientation with 219 respect to gravity; the membrane cavity houses the barrier between the two lines. B) Resulting heat maps of the volume fraction of Donor Fluid (red) and Receiver Fluid (blue), both with the identical properties of water, within the DCB for combinations of membrane porosity (55 or 99%) and inlet flow rate (1, 4, or 10 ml/min). Once stable, the simulations indicate different profiles of mixing at the interface of the two flow lines and the membrane. C) The volume fractions for Donor and Receiver Fluids at every normalized iteration point (proportional to time) were plotted to create mixing profiles as functions of porosity and flow rate; these profiles indicate the point in time, relative to the other cases, were the DCB systems reach concentration equilibrium (complete mixing defined as the point with 50?1% Donor Fluid and 50?1% Receiver Fluid). The inclusion of a membrane in the system dilates the time required for the looped system to reach equilibrium, at all follow rates simulated. D) CFD simulations were used to calculate the maximum shear stress at any surface within the modeled DCB; decreased porosity increases the resistance within the system and causes regions or points with higher shear stress as flow rate increases. These stress values allowed us to select adequate flow rates for subsequent experiments in the DCB that involve cells growing within the chambers. 8.3.2. Assessment of diffusive and convective transport within the DCB To understand and characterize the transport phenomena occurring in the DCB, static and dynamic variation of the 3D printed bioreactor were set up to quantify the role of diffusion, convection, and coupled convection-diffusion in the transport of tartrazine molecules from the donor to the receiver side of the common chamber. Diffusion was assessed without flow using a modified setup with sampling ports (Figure 8.3B-C) by tracking the concentration changes of model molecules tartrazine and brilliant blue in the donor and receiver chambers. For all the transport studies we defined reaching equilibrium as the point when there is no significant difference between the concentrations of the donor and the receiver (p<0.01 (**) or p<0.05 (*)). Without a membrane, diffusion equilibrates the concentration of both chambers, with no significant difference between the donor and receiver chambers by 10 min (Figure 8.3D). Nevertheless, using LCD or HCD membranes significantly 220 delays the equilibrium of the system (Figure 8.3E-F). Several pieces of information can be gathered from this data. First, both tartrazine, with molecular weight of 534 Da, and brilliant blue, 793 Da, are diffused for at least 6100 min across LCD membranes before reaching equilibrium. The HCD membrane allows a higher rate of diffusion, equilibrating both molecules by 4800 min. Second, the smaller molecule tartrazine reaches equilibrium at a concentration ratio close to 1:1 donor:receiver (0.5/0.5), but the brilliant blue, which is approximately 33% larger, is significantly and consistently stable at a ratio around 1:4 donor:receiver (0.2/0.8). This imbalance implies that there could be an additional level of interaction between the membrane and the tartrazine molecule. Last, all combinations of molecules and membranes -6 2 resulted in permeability of the system close to 3.8x10 cm /s, and no significant difference between the groups (Figure 8.3G). The membranes recovered at the end of the diffusion assay showed LCD samples were either partially damaged or completely degraded after 8d in the static DCB. On the other hand, HCD samples appeared complete after initial examination; at worst, parts were broken and remained in the scaffold when the membranes were lifted (Figure 8.3H-I). We had observed similar diffusion trends in previous characterization of the R keratin-based membranes in a Transwell model (12-well plate Transwell Falcon 441 Permeable Supports) . We had observed that for the low MW molecules (10 kDa , fluorescein isothiocyanate-dextran (FITCd)), higher amounts of solute are accumulated within the keratin gel, a trend that does not change with increasing energy density delivered to the keratin during crosslinking, indicating that smaller molecules could be chemically or electrostatically interacting with the crosslinked 221 441 networks or are trapped within pores of the hydrogel . In comparison to the permeability of tartrazine and brilliant blue, the permeability calculated for low MW -5 2 -5 2 FITCd was 2.6x10 cm /s across LCD membranes and 3.3x10 cm /s across HCD 441 without significant difference between groups (p?0.05) . The data here follows the same trend, although at a slower rate consistent with smaller molecules interacting with the membrane, of no significant difference in the permeability of low MW molecules across LCD or HCD membranes when tested in the DCB. Figure 8.3. Assessment of diffusion in the DCB. A) UV-crosslinked keratin membranes casted in custom 3D printed molds. B) DCB master parts ready for assembly, highlighting the inclusion of the keratin membrane and the silicone sealing 222 gaskets. C) For the assessment of diffusion in the DCB inlets and outlets were clamped shut, and sampling ports (red arrows) were opened on each side and sealed with casted paraffin plugs as needed. This system was used to quantify diffusion of model molecules tartrazine and brilliant blue from a donor chamber to a receiver chamber either without an intermediate membrane (D), with a LCD membrane, or with a HCD membrane (E-F) D) Without a membrane, diffusion quickly equilibrates the common chamber, with no significant difference between the concentration of the donor and the receiver chambers by 10 min, for both molecules tracked (n=4). E-F) The inclusion of the membrane significantly delays the equilibrium of the system; equilibrium for both tartrazine (534 Da) and brilliant blue (793 Da) takes at least 6105 min using a LCD membrane (n=4) and 4770 min using a HCD membrane (n=5). Tartrazine reaches equilibrium close to 0.5 (normalized concentration) between donor and receiver, but brilliant blue, which is 33% larger, is significantly and consistently -4 2 stable around 0.8. G) The permeability of the system was around 2x10 cm /s with no significant difference between the combinations of molecules and membrane CD. H-I) The keratin membranes recovered from the chambers showed differences between the low and high CD groups; LCD membranes were either not recovered or partially damaged, while the HCD samples were in considerable better state with some complete even after 8d. For all plots, statistical significance was determined as p<0.01 (**) or p<0.05 (*). Convection was assessed by characterizing the bulk flow downstream using the dynamic setup of the DCB (Figure 8.4A-B). The downstream reservoir volumes were tracked over 24 h and at all flow rates tested, independent of using no membrane, LCD, or HCD membranes, the volume of the donor and receiver reservoirs did not significantly change (not significantly different from 0% change). This consistency is indicative of either the convection transport in the two flow lines being independent and in steady state, or the system being symmetrical. The first option would state that the fluid in each line remains in its original line and does not change the volume of the loop over time. The symmetry option would indicate that even if one line is pulling flow from the other, the system pressures correct this and pull flow in the other direction so that the overall volume of the looped lines remains constant. From the CFD velocity trajectories in the DCB (Supplementary Figure 223 8.S2) we can get a qualitative idea of how the flows remain fully developed across the common chamber and the porous scaffolds without crossing or interfering with the adjacent line when the intermediate membrane is included. But without a membrane (99% porosity equivalent) the flow lines mix together and are hard to distinguish from each other. It is the membrane that restricts convection between the chambers. Still, as a symmetrical system the DCB that can be viable option to produce gradients, which require mixing, as initially hypothesized. Overall, our data indicates that convection in the system is not solely responsible for transport between the chambers. To further elucidate on this set of data, we proceeded to assess the coupled diffusive- convective transport of the main chamber under dynamic conditions. 224 Figure 8.4. Dynamic assessment of convection and diffusion in the DCB system. A) The DCB flow system is composed of a multi-channel rotary pump (4 or 8- channel pump head) in sequence with the parallel chamber inlets, outlets, 3-way connectors to a 1 ml syringe for sampling, and 50 ml reservoirs. B) Flow and mixing profiles in the DCB can be observed by flowing green dye through one of the lines (donor side) and water through the other (receiver side); i) at the beginning the flow dynamics of the system keep the flows separated, then as time progresses ii) donor fluid can be seen moving to the adjacent receiver line. To understand and characterize this movement, this dynamic bioreactor setup was used to quantify the role of convection and coupled convection-diffusion in the transport of dye molecules from the donor to the receiver. C) The change in volume of the loops after 24h running was used to assess bulk convection; at all flow rates tested, and independent of the use of membranes, the volume of the donor and receiver reservoirs did not significantly change (not significantly different from 0% change) indicating that convection in the flow system is symmetrical and stable throughout the pump runs (at least n=4). D) Having studied diffusion and convection separately, concentration of green dye in the donor and receiver were measured thoroughly over 24 h runs to assess both transport phenomena coupled together, at 1, 4, and 10 ml/min. The first assessment was of an TM ideally impermeable barrier (Parafilm ) setup, which indicated that the DCB flow lines remain completely independent from each other, as the concentration in the donor and receiver lines remain unchanged over time (donor not significantly different from 1.0 concentration, and receiver not significantly different from 0.0). For all plots, statistical significance was determined as p<0.01 (**) or p<0.05 (*). 225 The simultaneous effect of coupled convection and diffusion was assessed by thoroughly tracking the changes in concentration of the donor and receiver chambers TM over 24 h in the dynamic setup. With an impermeable Parafilm barrier between the chambers, the concentration in the donor and receiver lines remains constant; throughout the 24h, the donor is not significantly different from the 1.0 initial normalized concentration, and the receiver not significantly different from 0.0 (Figure 8.4D). This indicates, beyond statistical significance, that independent, parallel flow lines can be produced in the DCB if needed. The opposite case involved running the DCB without a membrane, where equilibrium was reached within 24h at 4 ml/min (at 12h) and 10 ml/min (8h), but no state of equilibrium was reached at 1 ml/min (Figure 8.5A). This is a significant change from considering diffusion only (Figure 8.3D), when equilibrium occurred within 10 min for all cases. When membranes are used, there are significant changes in the equilibrium time points. Using LCD membranes, equilibrium was reached within 24h only under 10 ml/min flow (at 8h) while no equilibrium was reached using 1 or 4 ml/min (Figure 8.5B). Using HCD membranes, the state of equilibrium was not reached at any flow rate within 24h (Figure 8.5C). In general, the inclusion of a membrane significantly delays the equilibrium of the system, which indicates that membranes can change mixing and transport dynamics between the chambers. All groups presented equilibrium of tartrazine at a concentration ratio close to 1:1 donor:receiver (0.5/0.5) or had trends that indicated eventual equilibrium at that ratio. We observed this behavior in the static diffusion assay as well, indicating that the diffusion equilibrium is conserved under dynamic conditions and dilated due to the convection component. 226 As compartmentalized in the diffusion-only and convection-only assessment, our coupled data indicates that convection in the system is not solely responsible for transport between the chambers, and the membrane and flow rates are the main regulators of this behavior. This observation is comparable to others reported in 378 literature. Spitters et al. developed a bioreactor that combined cyclic compressive stimulation with a two compartment system divided with the tested sample (no membrane); the system forms gradients across a sample by regulating the concentration of each compartment, a fact also proven using CFD. Using articular cartilage explants and CFD, the authors demonstrated that the bioreactor creates glucose gradients due to diffusion across the samples and not only by convective flow separating from the main chambers and moving through the sample. Our observations 378 are similar to Spitters? conclusion about their bioreactor ; the transport phenomena between compartments does not involve convective flow forced from one chamber to another, but rather perpendicular diffusion stemming from the bulk convective transport. Other complex bioreactors that rely on transport between chambers have included membranes, but their use does not mean that transport across them occurs by 377 diffusion only. Nakagawa et al. reported a two-chambered microfluidic device used to produce platelets from stem cell?derived megakaryocytes, inducing shear stress on a cell-laden membrane by having one side flow parallel to the membrane plane (main flow) and the other providing 60? inflow across the membrane (pressure flow). Here, the direction of the pressure flow was used to induce convection across the barrier, a behavior we avoided by maintaining the main directions of the DCB flows parallel. In 521 a similar patented system by Dimitrijevich et al. , the main direction of inlet flow is 227 perpendicular to the membrane plane, but relies on the fluid following the membrane surface and flowing back parallel to the inlet. In this case the main flow is first perpendicular and then parallel to the membrane, which the inventors use to induce shear stress and perfusion by convection and diffusion across a membrane barrier to cells on the other side. Overall, the transport phenomena in our DCB is determined by the CD of the membrane, flow rates, and the orientation of flow but, as other bioreactors in literature prove, these factors can be changed and combined to induce different transport profiles as needed. An important difference between the dynamic cases assessed used was the state of the hydrogel membranes recovered. At all flow rates, the LCD membranes recovered were generally in worse state compared to the HCD samples (Figure 8.5D). This trend confirms (1) the observations in the diffusion-only cases, where HCD samples appeared complete after initial examination and only broke when the membranes were lifted from the scaffolds (Figure 8.3H-I), and (2) our previous observations on the faster degradation and loss of swelling capacity of LCD 441 membranes . Under coupled convection-diffusion transport of the DCB, HCD membranes will provide a more reliable barrier effect over prolonged periods of time. Overall, we have confirmed that the DCB membrane system allows transport of molecules from one chamber to another, and thus potential intercellular communication between adjacent cell populations, via diffusion through a permeable barrier, a phenomenon that can be further regulated over time by changing the crosslinking density of the membrane used. 228 Figure 8.5. Assessment of coupled convection-diffusion in the DCB system. As a continuation of Figure 8.4, the concentration of green dye in the donor and receiver were measured thoroughly over 24 h runs, studying the cases A) without a membrane, B) with LCD membranes, and C) with HCD membrane. From the concentration profiles it is possible to determine the points in time when the donor and receiver reach equilibrium (at least n=4). For the no-membrane cases, equilibrium was reached within 24h at 4 ml/min (at 12h) and 10 ml/min (8h), but no equilibrium was reached at 1 ml/min. The inclusion of the membrane significantly delays the equilibrium of the system; for the case with LCD membrane equilibrium within 24h only using 10 ml/min (8h), while no equilibrium was reached at 1 or 4 ml/min; last, with HCD membrane no state of equilibrium was reached for any rate within 24h. D) The state in which the membranes were recovered after these dynamic studies was determined by the CD; at all flow rates, the LCD membranes recovered were generally in worse state compared to the HCD samples. For all plots, statistical significance was determined as p<0.01 (**) or p<0.05 (*). 229 8.3.3. Cell cultures, growth, and differentiation within the DCB The DCB system was optimized for assembly under sterile conditions and for use inside an incubator for at least 28 d, using a 4 ml/min flow rate in all cases (Figure 8.6A-B). Seeding the fibronectin-coated scaffolds with cells, we were able to study, first, if the DCB sustains cell growth, and, second, if we could induce changes on one population using transport of biomolecules from the adjacent line. Using the same type of cells, L929 fibroblasts, and same type of growth media in both lines for 7 and 28d, microscopy imaging of the scaffolds revealed healthy colonies of cells on both chambers of the DCB (Figure 8.6C). As observed previously in literature, the preferential localization of cells in concave curvatures or negative curvature 519 substrates is indicative of healthy populations within the bioreactor for at least 28d. As previously discussed, we had concerns about shear stress lifting adhered cells. Here we confirm cell attachment and further development under a 4 ml/min flow rate, which causes maximum shear stress of 0.23 Pa when a model barrier with 55% porosity is used (Figure 8.2D). Next, by using to two different media on the same type of cells we were able to confirm that the bioreactor system allows biomolecular diffusion between its chambers. Perfusing line A with adipogenic media and line B with regular growth media, we assessed diffusion of adipogenic factors by quantifying intracellular lipids deposited by hMSCs in both lines as compared to a 2D non-differentiated population (normalized lipid deposition of 1.0). Due to sustained diffusion between the chambers, the line without adipogenic media still received adipogenic factors and differentiated in all combinations studied. Without a membrane, hMSCs in line B did 230 differentiate but to a significantly lower degree compared to the direct differentiation of cells in line A (different normalized lipid deposition of 13.8 ? 0.1 in line A vs. 8.15 ? 2.3 in line B, p<0.05). The same trend was observed when using LCD membranes (different normalized lipid deposition of 7.9 ? 1.0 in line A vs. 4.9 ? 0.6 in line B, p<0.05). Differently, for the HCD membrane cases, there was no significant difference in the amount of intracellular lipids deposit by both lines (normalized lipid deposition of 7.7 ? 1.2 in line A and 5.8 ? 0.8 in line B, p<0.05) (Figure 8.6E). We have confirmed that a keratin-based membrane in the DCB allows diffusion of, at least, adipogenic factors. Even though the DCB had less lipid deposition than control scaffolds, with the static 3D cultures having significantly higher lipid deposition than the studied groups (p<0.05), it is important to highlight that the 2D and 3D controls can only be sustained with a single type of media. The DCB was successful differentiating hMSC populations indirectly and under dynamic conditions. This data is encouraging, it indicates that the DCB approach allows co-culturing cells lines simultaneously without abandoning their specific growth or differentiating media, and we aim to extend this to other co-culture setups. 231 Figure 8.6. Cell cultures, growth, and differentiation within the DCB. A) The DCB system was simplified to reduce connections and ports that could lead to contamination; for cells studies that require incubation, the loops consist of the multi- channel rotary pump (4 or 8-channel pump head) in sequence with the parallel chamber inlets, outlets, and 50 ml reservoirs. B) Fibronectin-coated scaffolds were successfully seeded and proved to be manageable under sterile conditions for assembly into the DCB. C) L929 fibroblasts were cultured on both lines A and B for 7 and 28d; imaging of the scaffolds revealed that cells were healthy and growing, filling the concave curvatures of the pores, for at least 28d in the DCB (n=4). D) Fibronectin-coated scaffolds were seeded with hMSCs and grown in the bioreactor using hMSC growth media. After 7d, line A was changed to hMSC adipogenic media, while line B was kept with hMSC growth media for up to 28d; notice the different shades of media in the two lines (red arrows). E) Adipogenic differentiation was possible in the DCB for all cases studied, quantified by the normalized amount of lipids deposited as compared to a 2D non-differentiated population. Without a membrane, line B did differentiate but to a significantly lower degree compared to the direct differentiation of line A, the same trend presented in the LCD case. For the 232 HCD membrane case, both lines were able to deposit similar amounts of intracellular lipids. Comparisons were done using ANOVA and Tukey?s multiple pairwise comparison; samples that do not share the same letter are significantly different (p<0.05). Our stratification approach using the DCB and membranes to separate populations while allowing transport communication will provide an interesting platform to study the development and relations of dermis-hypodermis layers. We are interested in studying the role of the hypodermis layer and its adipose-specific ECM 35 (mostly collagen types I, III, IV, V, and VI ) and adipose-specific growth factors and 32,36?39 hormones in skin healing processes. If the hypodermis is producing such factors in vivo, diffusion through the tissues and convection through the common deep vascular plexus is transporting them to the adjacent dermis layer. The DCB system will allow us to replicate this transport between hypodermis and dermis to prove the benefits of stratification and co-culturing in skin tissue engineering. The study of gradients and stratification has been well documented for other tissues 272,502,505,525 like the muscle-ligament-bone complex , osteochondral tissue 501,504,507,508,510,526 157,513,514,527 , or thin epithelial barriers such as the epidermis/dermis , 506,516 or gastrointestinal tissues . Most approaches are based on when and where cells should be deposited on scaffolds to improve their chances of forming layers or gradients. Several gradient formation techniques rely on the fact that cells will migrate from one layer to the other, either following or leaving behind a gradient trail 218,272,501 of ECM or growth factors, to construct continuous interphases . Multichambered bioreactors have studied intercellular communication between 233 different types of cells with various results. Fisher and coworkers used a tubular 498,499 perfusion system to grow stem cell-laden alginate beads . This approach increased surface area exposure to nutrients and oxygen, and each bead could be 373 loaded with different cells with the potential of spatial control . Nevertheless, the 379 system was limited by using a single type of media. Chang et al. presented a double-chamber bioreactor consisting of two tubular-shape glass chambers separated by a silicone-rubber septum that held a single scaffold. This setup could supply different types of culture media to support different kinds of cells on each half of the scaffold but relied on engineering the two sides of the scaffold before the bioreactor. All transport was due to limited diffusion of the media across the scaffold, which created a gradient for the migration of cells and resulted in the integration of both 528 sides of the scaffold. A similar setup was studied and patented by Berry et al. . 376,520 Similarly, the Tuan group developed a double chambered bioreactor that allowed constant communication across a bilayered scaffold. In this case, each side of the bioreactor held a different scaffold, with substrate and cells specific to either bone or cartilage for osteochondral tissue engineering. The scaffolds remained homogeneously different, but diffusion and the gradient deposition of ECM bound them together developing a tidemark after 6 weeks in the bioreactor. Different from the bioreactors presented earlier on, these last do not use membranes and rely on the formation of gradients. Our DCB approach aims to regulate cell migration and communication due to the inclusion of the membrane, allowing it when forming gradients or purposefully preventing it to create stratified layers. 234 8.4. Conclusions Here we presented our dual-chambered bioreactor (DCB) design with the inclusion of degradable membranes for the stratification of cellular cultures. We have characterized the transport phenomena within the bioreactor to understand the effect of the design and the membranes in the potential formation of stratified layers or gradients. As detailed before, the DCB can provide adjacent flow lines within a common chamber; the inclusion of the membrane can regulate flow layering or mixing, which could be translated to stratification or gradients, respectively. The crosslinking degree of the membranes not only regulates the permeability properties of the network but can be used to change convection and diffusion in the bioreactor. Our data suggest that the DCB can perfuse two adjacent layers with different media, allowing communication between the layers via diffusion through the membrane. Furthermore, the bioreactor is viable for co-culture and sustaining of adjacent cell populations, indicating that cross-talk between the populations can be used to induce changes on one another. Our overall approach to study the dermis-hypodermis complex in a bioreactor system aims to address current limitations in skin tissue engineering, but it can be used for other multilayered tissues such as blood vessels or gastrointestinal tissues. As a bioreactor it can be implemented to study 3D co-cultures and communication pathways between cell populations. We consider the system has applications in cellular biology, tissue engineering, or fluid dynamics research, as well as broader application in industry for cellular or bacterial co-cultures. 235 8.5. Supplementary Figures Figure 8.S1. Shear stress profiles in the DCB. CFD modeling results for the surface shear stress heat maps of the DCB. In general, shear stress is below 0.22 Pa throughout the DCB system, independent of membrane porosity (55 or 99%) or flow rates (1, 4, or 10 ml/min). Maximum (red points) and minimum (blue points) values of shear stress are presented. Maximum values are generally found at the inlets or outlets for the higher 4 and 10 ml/min flow rates; in the case of 1 ml/min, the maximum and minimum values are inside the scaffold. The maximum values were found in in punctual cases and are not indicative of high stress regions in the scaffolds. 236 Figure 8.S2. Velocity profiles in the DCB. CFD modeling results for the flow velocity trajectories through the DCB. In general, velocity is higher at the inlets and outlets, were the resistance of the system is highest. The porosity of the modeled membrane barrier determines whether the interface of the flow lines touches (99% porosity, modeling no membrane) or remains mainly separated (55% porosity membrane). In all cases, an increase in flow rate (1, 4, or 10 ml/min) caused an increase in velocity in the system. All maximum velocities were found at the outlets, while minimum velocities (0 m/s) were found closer to the inlets or at the point where the inlet opens into the chambers and the resistance drops causing the velocity to decrease. 237 Chapter 9: Studies on Complex Topographical and Physiological Reconstruction of Skin in a Dual-Chambered 9 Bioreactor 9.1. Introduction 3,4 Every year over 11 million people report severe burns worldwide . Facial 5 burns, also classified as a particular type of violence, involves victims of warfare , 6,7 8 acid atacks , scalding , and general trauma. Burns alone constitute 5-20% of all warfare military casualties, with ~70% of these being to the vulnerable, least covered 5 7 areas: face and hands . In acid attacks and scalding the target is usually the face . 7,9 Even with survival rates close to 96-98% , facial burns have severe psychological and sociological effects; it is hard to quantify its most complex consequence: individuals trying to cope with disfigurements that alter their identity and perception 6,10,13 of self, and social stigma and exclusion . When burned, skin cannot regulate temperature or fluid transport, or stop 4,10 bacterial infection . The natural healing response to severe burns in adults, wound contraction, is characterized by fast proliferation of fibroblasts that deposit random collagen to rapidly restore the skin barrier. The disorganized, fibrous collagen tissue (scar), is characterized by lack of sensation, hypertrophy, lack of elasticity, and flawed features; in effect, ?healing? does not restore skin function, stratification, or 4,10,31,74 aesthetic features . Autologous skin grafting, the current gold standard clinical 9 As prepared for: J Navarro, M Janes, N Arumugasaamy, M Kimicata, J Allbritton-King, M Santoro, JP Fisher: ?Studies on Complex Topographical and Physiological Reconstruction of Skin in a Dual- Chambered Bioreactor?. [In preparation] 238 4,11,12,31 treatment , also aims to restore the barrier using the patient?s skin but has the same limitation as wound contraction; as described by Bottcher-Haberzeth et al., particularly in children, there is hypertrophic scarring or keloid formation that is 31 frequently disabling and disfiguring . Tissue engineering has provided approaches to separately restore either 1) the cellular niches and physiology of skin, or 2) the complex 3D facial topography. Mammalian skin is composed of multiple stratified layers, broadly the epidermis and dermis, and the subcutaneous fat tissue, or hypodermis. We are particularly interested in the interactions between the last two. The dermis is a rich layer, 1 to 4 mm thick, that consists mostly of fibroblasts synthesizing extracellular matrix composed of collagen (~70%, mostly type I and small amounts of types IV 4,30?34 and VII), elastin, and proteoglycans . The hypodermis, considered part of the 32 endocrine system , consists of adipose tissue used in energy storage and 30,31 thermoregulation . The hypodermis is generally undervalued or neglected in skin models as fat storage, but it is also a complex lipid barrier and the layer where nerves and larger blood vessels permeate the upper layers from, as well as a rich source of 33,35 32,36 37?39 stem cells , hormones such as leptin and adiponectin and basic fibroblast 40 growth factor (bFGF) , key players in re-epithelization, wound healing, and 35,41?43 angiogenesis . We consider that including the hypodermis in 3D dermal models can: 1) improve skin stratification by providing other layers with adipose-specific growth factors and extracellular matrix (ECM), and 2) preserve reconstructed topographies in vivo. 239 Skin scaffolds that recapitulate this layered anatomy are produced for clinical 14?17 18?27 use and dermal models , but no approach restores both dermal function and facial features. To address these limitations, it is necessary to develop new strategies to study stratified skin tissue equivalents that provide simultaneous physiological and topographical cues for tissue reconstruction. We have previously reported our studies on a strategy that can combine the fundamental logic of guided tissue regeneration 441 (GTR) techniques and multi-chambered bioreactor technologies. This approach led to the development of a dual-chambered bioreactor (DCB) that incorporates a membrane to study stratified 3D cell populations for skin tissue engineering. The DCB provides two adjacent flow lines within a common chamber that can harbor cell-laden scaffolds; the inclusion of an intermediate membrane separating the common chamber and the scaffolds can regulate flow layering or mixing, which can be exploited to produce stratification or gradients of cell populations in the scaffolds. Here, we aim to further elucidate on the use of this DCB-membrane complex for the study of a dermis/hypodermis system that can recapitulate complex topographical and physiological parameters of skin. The objective of this work was to assess the viability of our DCB for growing non-planar adjacent cell populations with the inclusion of a regulatory keratin membrane. First, having studied the DCB with planar, flat interfaces in our previous publication, we now assessed its viability for perfusing non-planar, curved interfaces. These studies were performed in a variation of the original DCB which allows imaging of the inside of the bioreactor in real time. Second, we integrated both curvature and cell populations, to assess the synergistic development of adjacent dermis fibroblasts and hypodermis stem-cell-derived 240 adipocytes and evaluate whether including topography parameters would alter cell viability in the DCB or not. 9.2. Methods 9.2.1. Preparation of keratin membranes High crosslinking density (HCD) membranes were prepared as previously 441 reported . Briefly, Keratin extracted from oxidized human hair (KeraNetics LLC, 418,441,518 Winston-Salem, NC) was dissolved in phosphate buffered saline (PBS) (4% wt/vol) and then mixed with a photosensitive solution at a 4:1 ratio. The photosensitive solution consisted of 1 mM riboflavin (Sigma-Aldrich Co., St. Louis, MO), 200 mM sodium persulfate (SPS, Sigma-Aldrich), and 0.001% wt/vol hydroquinone (Sigma-Aldrich). The resin was used in 3D printed molds to cast 1.5 2 mm-thick membranes, exposing to ultraviolet (UV) light at 350 mW/dm to induce 59% HCD. Samples were thoroughly rinsed in PBS to remove uncrosslinked components. For experiments involving cell culturing, membranes were sterilized using PBS:sterile PBS serial washes (100:0, 75:25: 50:50, 25:75 and 0:100 for 20 minutes at each step under UV light) and stored in fresh sterile PBS at 4 ?C until use. 9.2.2. Dual-Chambered Bioreactor (DCB) setup and its variation for imaging As reported in our previous chapter, the DCB was assembled using a master part twice. This part consists of an inlet and an outlet that feed into a chamber with an open wall (internal dimensions: 10 x 10 x 8 mm), and channels for silicone rubber sealing gaskets (custom-made, 50A durometer, 2 mm thick). The master parts were 241 3D printed out of EShell 300 resin (EnvisionTEC Inc., Dearborn, MI) on a stereolithography-based 3D printer (Perfactory 4 DLP printer, EnvisionTEC). An alternative model of the master part was designed to allow real-time imaging of the inside of the bioreactor. As compared to the regular DCB in Figure 9.1A, the imaging DCB has one clear side (2 mm thick) which allows imaging with hand-held camera or mounted on a microscope (Figure 9.1Bi-ii). Stainless steel hardware (#10-24 0.5? machine screws and nuts) and the gaskets were used to close the DCBs. The inlet/outlet connections match 16-gauge pump tubing, which were then connected to 1/8? ID platinum-cured silicone tubing using 1/8?-1/8? Kynar connectors as needed (all Cole-Parmer, Vernon Hills, IL). A multi-channel rotary pump (Masterflex L/S with 4-channel or 8-channel pump head) was used to run parallel DCBs. Each bioreactor requires two pump lines to function (generally lines A and B); each line connects in sequence the pump, the chamber inlet, chamber outlet, a 50ml reservoir, and back to the pump. For experiments involving cell culturing, all EShell 300 parts were sterilized using ethanol:sterile PBS serial washes (100:0, 75:25: 50:50, 25:75 and 0:100 for 20 minutes at each step under ultraviolet (UV) light) and stored in fresh sterile PBS at 4 ?C until use. All tubing, connectors, stainless steel hardware, and the silicone gaskets were autoclaved. 9.2.3. Porous scaffolds for the DCB 3 Porous scaffolds with 600 ?m pores were designed to fit into the chambers and 3D printed with EShell 300 (EnvisionTEC). Four groups were printed: single continuous scaffolds, which fill the whole DCB common chamber (9.8 x 9.8 x 17.5 mm); flat scaffolds, which fill the line chamber only and require an intermediate 242 membrane (9.8 x 9.8 x 8 mm); 25% curved scaffolds, which have a peak the reaches 25% into the depth of the adjacent chamber; and 75% curved scaffolds, which have a peak the reaches 75% into the depth of the adjacent chamber. Scaffolds were sterilized as other EShell 300 parts using ethanol:sterile PBS washes detailed above. 9.2.4. Flat and curved interfaces in the imaging DCB Closed loop setups using the rotary pump (as described above) were used to image the profile of flat and curved interfaces in the imaging DCB. The bioreactor loops were filled with PBS, removing all air from the lines and scaffolds, and equilibrated for 30 min at constant flow rate of 10 ml/min. The setup was organized to have the window in the imaging DCB exposed to the microscope lens (Figure 9.1B) (Nikon Ti2 inverted microscope mounted with Nikon DS Qi2 camera, Nikon, Tokyo, Japan). This setup was evaluated with the four scaffold groups and HCD keratin-based membranes (n=3). Once stable, 5% fluorescein isothiocyanate-dextran (FITCd, 150 kDa MW, Sigma-Aldrich) in PBS was added to 20 ml PBS in the Line B reservoir. Continuous flow of the FITCd through the bioreactor was imaged at 510 nm wavelength. 9.2.5. Dermal-hypodermal constructs with curved interfaces in the DCB 2 Printed porous scaffolds from all groups were coated with 3 ?g/cm fibronectin (from bovine plasma, Sigma-Aldrich) in PBS at 37?C for 12h to facilitate 519 cell attachment . Complete coating of the 3D structure was assured by pipetting the fibronectin back and forth through the scaffolds and rotating every 30 min. Coated scaffolds were seeded with either neonatal normal human dermal fibroblasts (NHFSs, 243 2 Lonza, Walkersville, MD) at a density of 6000 cells/cm , human mesenchymal stem 2 cells (hMSCs, Lonza) at a density of 3500 cells/cm , or left without cells (empty) (n=7, at least). Additional control cell cultures were seeded scaffold in 12-well plates at the same densities. For the DCB setups, which include the HCD membranes when needed, line A chamber was always loaded with NHDF scaffolds, while line B chamber was loaded with either hMSC or empty scaffolds. Lines with NHDF scaffolds or empty scaffolds were filled with fibroblast growth media (FGM, composed of fibroblast basal media 2 supplemented with 0.1% insulin, 0.1% human basic fibroblast growth factor (bFGF), 0.1% gentamicin-1000, and 2% fetal bovine serum, Lonza). Lines with hMSC scaffolds were filled with hMSC growth media (high glucose Dulbecco?s Modified Eagle Medium (DMEM) with 10% fetal bovine serum (FBS), 1% non-essential amino acids (NEAA), and 1% penicillin/streptomycin 100 U/100 mg (P/S)). Lines were run for 24 h to allow cell attachment. After, the lines with hMSC scaffolds were replaced with hMSC adipogenic media (DMEM with 10% FBS, 100 U/100 mg P/S, 1 mM sodium pyruvate, 1 mM dexamethasone, 10 mg/ml insulin, 0.5 mM 3-isobutyl-1-methylxanthine (IBMX), and 200 mM 473,474 indomethacin) . The DCBs were then left running for 9d, with fresh media replaced in reservoirs A and B every 3 d. Samples of the media of both lines was collected at 1, 6, and 9 d and kept frozen at -20 ?C until use. At the end-point, the scaffolds were recovered from the DCBs for imaging. 9.2.6. Staining and Imaging The scaffolds and membranes recovered were moved to 12-well plates for cell fixing, permeabilization, and staining. Media was removed, and the scaffolds were 244 rinsed with Hank?s balanced salt solution (HBSS, Lonza) at 37 ?C for 15 min. The cell-laden scaffolds were fixed in excess 4% paraformaldehyde (Sigma-Aldrich) plus 1% sucrose (Sigma-Aldrich) for 30 min at room temperature. The samples were then rinsed with warm PBS three times, followed by permeabilization using permeabilization buffer (sucrose, sodium chloride, magnesium chloride hexahydrate, HEPES buffer, and 0.5% Triton X-100) for 20 min at room temperature. Samples were again rinsed with warm PBS three times, for 15 min each. In the dark, the scaffolds were then immunostained with 2.5% Alexa Fluor 594 Phalloidin (Invitrogen, Carlsbad, CA) and cell nucleus stain DAPI (Vector Laboratories, Burlingame, CA), for 1h incubation in each stain, followed by final rinsing in PBS. 9.2.7. ELISA quantifications To determine concentrations of intercellular adiponectin, leptin, and bFGF, ELISAs were performed. DuoSet kits (R&D Systems, Minneapolis, MN) for each target were purchased and used per manufacturer?s protocols. The media samples collected from the bioreactors at each time point were thawed at room temperature to run each assay. Absorbance was measured using a Tecan Spark Multimode Microplate Reader (Tecan Life Sciences, Morrisville, NC). 9.2.8. Statistics ANOVA and Tukey?s multiple pairwise comparison (significance using p < 0.05) were used for multi-group comparisons in Minitab 18 (Minitab, Inc). Differences between two specific samples were assessed with two-sample t-test for 245 the mean (significance using p < 0.01 (**) and p < 0.05 (*)). Data is presented as mean ? standard deviation. 9.3. Results and Discussion The objective of this work was to assess the viability of the DCB system for growing curved, adjacent cell populations with the inclusion of a regulatory keratin- based membrane. Having studied the DCB with flat interfaces in our previous publication, here we assessed its viability for perfusing non-planar, curved interfaces. As illustrated in Figure 9.1C, the DCB system as reported in the previous chapter was studied for the formation of curved interfaces. Here, we study simplified curvatures in the DCB. The flat porous scaffolds previously used were modified to produce 25 and 75% curved scaffolds. The 25% curved scaffolds have peaks that reach 25% into the depth of the adjacent chamber, while 75% curved scaffolds reach 75% into the depth of the chambers, two surfaces representative of a smooth (<45? tortuosity) and a coarse (>45? tortuosity) topography, respectively (Figures 9.1Ci-ii). Even if the initial scaffolds introduced in the DCB have curved topographies and a non-planar interface between them, these features will be lost due to the parallel convective flow. With the inclusion of the membrane in the common chamber, the flow contours around the membrane, following any curvature introduced (Figure 9.1Ciii). To the best of our knowledge, there are no reported studies in literature that can produce curved stratified layers. Our approach will allow us to address facial skin tissue regeneration without neglecting the facial features. The curvature of the 246 scaffolds can be 3D printed to match individual topographies of facial burn victims and still be developed into viable, stratified skin in the DCB due to the membrane. Figure 9.1. Curved profiles inside the DCB system. A) Comparison between the dual-chambered bioreactor (DCB, left) and the modification for imaging (imaging DCB, right). B) By reducing the thickness of the side wall, it serves as a clear window to image the common chamber of the DCB; i) this setup can be mounted on the microscope plate while connected to the running pump loop, and ii) the clear EShell material allows fluorescent excitation to track immunostained cells inside the DCB. C) Approach to the formation of curved interfaces in a DCB. The profile of flat porous scaffolds was modified to produce: i) 25% curved scaffolds, which have peaks that reach 25% into the depth of the adjacent chamber; and ii) 75% curved scaffolds, which have peaks that reach 75% into the depth of the adjacent chamber. iii) The inclusion of the keratin-based membrane has been proven to keep the flow lines separated producing stratified constructs, which can now be extended to the production of curved interfaces. 9.3.1. Imaging and tracking flow inside the DCB for curved interfaces The imaging DCB variation (Figures 9.1A-B) was used to track FITCd solution through the bioreactor in real-time for 1 h. The four types of scaffolds were assessed, in combination with HCD keratin membranes, to assess mixing and contours as a function of interface curvature. As shown in Figure 9.2, the continuous scaffold shows mixing in the common chamber due to convective and diffusive 247 transport coupled with the lack of a separating membranes, phenomena previously discussed (Chapter 8). Flat scaffolds with an intermediate HCD membrane show that the barrier keeps the convective flow of the lines separated for at least 1 h and maintains a flat interface between the two lines. Similarly, the 25% curved scaffold indicates that curved flow profiles can be created between the chambers and convective flow is just as efficiently separated while contouring to the curved surface as time progresses. On the other hand, for the 75% curved scaffold convective flow does not fully contour to the shape of the membrane, leaving a pocket in the peaks of the curves (last row, indicated with white arrows). As such, smoother topographies can be properly perfused in the DCB system, while the coarser surfaces might cause pockets of high recirculation that could complicate diffusion of nutrients and growth factors to the region. It would be of interest to study the threshold surface angle/inclination where such behavior starts to occur, elucidating on the limiting parameter to study complex topographies in the DCB. It should be highlighted that topographic reconstruction of facial features using tissue engineering is scarcely reported in literature, less so in bioreactors. Although it is not a physiologic priority during trauma, clinical intervention, and healing, the loss of facial features has severe emotional impact on the victims. Nearly all publications on burns or facial reconstruction include a statement regarding the immeasurable effects that scars and disfiguration have on the patient?s psyche and 13 social role; as described by Talley , without [a] face social interactions fall apart. Using multiple layers of dermis-epidermis constructs or live skin equivalents to attempt to reconstruct facial features has proven to be a poor approach. Patients are 248 subjected to multiple reconstructive surgeries over the years, yet these generally result 13 in various degrees of scarring and contracture, never regaining their features . Others, specially the vulnerable populations, can never afford the expensive 6 procedures . Methods to restore facial topography include hard implants that augment bone and cartilage volume aiming to replace the volume lost in muscle, fat and 529?531 530 skin . For example, Kim et al. report the use of 3D printed PCL coated with chondrocytes suspended in fibrin for rhinoplasty; the scaffolds were used for cartilagous augmentation in mice with partly successful results in maintaining the 530 expected shape and biocompatibility for the post-operative period . This approach changes facial volumes but still relies on covering the augmentations with compromised skin or dermal grafts. 249 Figure 9.2. Imaging and tracking flow inside the DCB for curved interfaces. The imaging DCB variation was used to track FITCd solution through the bioreactor in real-time for 1 h. The bottom line (line B) was perfused with the FITCd while simultaneously perfusing the top line (line A) with PBS. Four different types of scaffolds were assessed. The continuous scaffold, without the regulating membrane, shows mixing in the common chamber due to convective and diffusive transport as discussed in the previous chapter. Flat scaffolds with an intermediate HCD membrane show that the barrier keeps the convective flow of the lines separated for at least 1 h and maintains a flat interface between the two lines. Imaging of the 25% curved scaffold indicates that curved flow profiles can be created between the chambers and convective flow is just as efficiently separated while contouring to the curved surface as time progresses. The 75% curved scaffold similarly proves the previous point although a threshold to the curve contouring is probably indicated; here, even as time progresses for at least 1 h, convective flow does not fully contour to the shape of the membrane, leaving a pocket in the peaks of the curves (indicated with white arrows). 250 9.3.2. Combination of topography and physiology in the DCB The second aim of this work was to integrate curvature and cell viability in a single study, to assess the synergistic development of adjacent dermis fibroblasts and hypodermis stem-cell-derived adipocytes while simultaneously evaluating the possible effects of including topography parameters in the DCB on the already- proven cell culture viability of the system. As compiled in Figure 9.3, we tracked the state of the membrane curvature after incubation. Single continuous scaffolds, that run without membranes, showed no evident wear of the EShell structures (Figure 9.3Ai). Flat scaffolds and 25% curved scaffolds, both with intermediate membranes, showed no signs of scaffold deterioration and the HCD membranes remained intact after 9 d incubation (Figures 9.3Aii-iii). The bulk preservation of the membrane is an indication that smooth topography curvatures and the barrier effect can be retained and contoured in the DCB even under cell activity, media exposure, and sustained physiological temperature. Differently, reminiscent of the trend observed in Figure 9.2, the 75% curved scaffolds seemingly exceed a curvature threshold, resulting in the deformation of the membranes into the scaffolds possibly due to the increased resistance and flow in the lines (indicated with white arrows) (Figures 9.3Aiv). The trends observed without cells before were confirmed here after 9 d cultures: there is an indication of a threshold surface angle/inclination where flow cannot follow the curvature of the interface. As before, smoother topographies can be properly perfused and preserved under incubation conditions for at least 9 d in the DCB system, while the coarser surfaces induce deformation of the membranes possibly due to regions of 251 increased resistance overcoming the mechanical properties of the hydrogel membrane (these observations were confirmed in at least duplicate bioreactors). Nevertheless, viability of cell cultures in the DCBs is maintained, independent of scaffold curvature. In comparison to an empty EShell scaffold (no cells seeded) (Figure 9.3Bi), NHDF scaffolds sustained large populations and extended, thick formation of intracellular actin structures indicative of healthy proliferation under the dynamic conditions on flat, 25%, or 75% curved scaffolds (Figures 9.3Bii-iv). The hMSC scaffolds induced into adipogenic differentiation (adipo), had smaller populations, showing lesser numbers of cells and thinner intracellular actin structures on the surfaces with no observable differences due to curvature (Figure 9.3Bv). Similar populations were imaged on non-differentiated hMSC scaffolds, in terms of size and actin networks thickness, at all scales (Figures 9.3Bvi-viii). This indicates that fibroblasts greatly proliferate in the DCB while hMSC have smaller colonies and remain mostly as a mono-layer. The NHDF populations cultured alone or co-cultured with hMSC showed larger populations when compared to the other fibroblast L929s cultured previously in the DCB (previous chapter). In all cases, the viability of the bioreactor to sustain viable cell cultures was preserved even with the addition of the curvature variable. The effects of the curvature threshold, and the overall difference between smooth or coarse curves, should be evaluated in greater detail. Studies of curvature can be interpreted as studies of scaffolds that have profiles with continuously variable thickness; in the bioreactor, changes in thickness are comparable to changes in perfusable diameter which can induce regions of high or low resistance, shear stress, and flow velocity, as well as changes in the relative 252 distance between a source population and a receiver population in exchange of growth factors or nutrients. These variables have been proven to affect cell viability, 373,498,519 differentiation, and functionality . The behaviors and trends observed here are indicative that such topography assessments can be studied in the DCB for cellular co-cultures. Figure 9.3. Long-term observations of curved profiles in cell-laden DCBs. A) Cell-laden scaffolds, with the HCD keratin-based membranes (which auto-fluoresce close 500 nm wavelength), were collected after 9 d in running, incubated DCBs. i) Single continuous scaffolds without membranes, ii) flat scaffolds with intermediate membranes, iii) 25% curved scaffolds with intermediate membranes, and iv) 75% curved scaffolds with intermediate membranes were successfully recovered. No case showed signs of scaffold deterioration. Furthermore, the HCD membranes also remain intact after 9 d incubation subjected to cell activity, media, and 37 ?C 253 temperature, an indication that the barrier effect and transport phenomena discussed previously is sustained in the DCB long-term. As in Figure 9.2, the 75% curved scaffolds seemingly exceed a curvature threshold, resulting in the deformation of the membranes into the scaffolds possibly due to the increased resistance and flow in the lines (indicated with white arrows). B) Different cell populations were sustained in the DCBs independent of scaffold curvature. In comparison to an empty EShell scaffold (no cells seeded) (i), proliferation and viability of NHDFs (ii-iv) was considerable, showing large populations and wide formation of intracellular actin structures. The populations of hMSCs induced into adipogenic differentiation (v), although present and viable, were smaller, showing lesser number of cells and intracellular actin structures. This populations are comparable to non-differentiated hMSC populations in terms of size and actin networks (vi-viii). Cell nuclei identification in the scaffolds was limited due to the auto-fluorescence of EShell at similar wavelengths. A variety of approaches have been reported to change or restore volume of the facial soft tissues (skin, fat, and muscle). Fat itself has been traditionally used in such approaches. Subcutaneous volume restoration in the face using transplanted fat is 529 popular a procedure in cosmetic surgery, reportedly used as early as 1893 . In fat fillings, adipose tissue collected from excess deposits is injected into target areas to 532,533 alter thickness, and thus volume . Nevertheless, fat injected without a 533,534 predetermined structure results in partial loss of the injected volume . Because of its autologous nature, reabsorption or migration of the fat is not a concern and is corrected with follow-up injections. Donofrio et al. state that fat injected 529 subcutaneously restores large facial volume loss with relatively good persistence , while reported clinical experiences claim that fine-needle fat injections ensure long- 533 lasting, and in many cases permanent, results . The combination of the filling procedure, the reabsorption rates, and the synthesis of new ECM by cells surrounding and invading the area, means that fillings cannot guarantee a specific final topography. Still, the viability of fat has been studied and additional structuring cues 254 have been added to address the limitations discussed. Tissue-engineered adipose constructs are generally based on encapsulating adipogenic stem cells in hydrogel 35,534 534 networks . Alhadlaq et al. pre-conditioned hMSCs in adipogenic media and then re-suspended in a photo-crosslinkable PEGDA solution for casting under UV in adipogenic medium. Shape and dimensions of the scaffolds were maintained after 4- week implantation in mice. Furthermore, adipose tissue constructs have proved to be 35,534 a viable option to maintain predefined physical shape[s] and dimensions in vivo . The DCB is a viable tool to study geometry retention in hypodermis tissue equivalents, or to question the development of adipose tissues that can improve this parameter with or without the inclusion of a secondary cell population. As previously discussed, topography is not the only factor to consider; complex tissue geometries should still act as functional tissues. As compiled in Figure 9.4, the quantities of adipose-specific extracellular proteins adiponectin and leptin, as well as bFGF secreted by adipocytes and other types of cells, were tracked in the media collected from both lines of the DCBs over time. Here, we assessed the effect of co-culturing NHDF adjacent to hMSCs induced into adipogenic differentiation in the DCB system. Adiponectin was significantly expressed by hMSCs at the 9d end-point, indicative of successful adipogenic differentiation in the DCB. Adiponectin is an early marker of adipocyte function, and our expression 535 values are comparable to those previously reported. Martella et al. completed a study on the expression of this marker in the adipogenic differentiation of hMSCs; adiponectin is fairly undetectable before 7d in static, 2D culture, with significant 255 expression of ng/ml concentrations only measured at 10 d. Other studies report their 536 earliest expression timepoints (with concentrations close to 4 ng/ml) at 8 d . Leptin was generally not expressed in the cases studied (no significant difference to non-seeded EShell in all cases and below the detection threshold in most cases); the only observable trend indicates that it is better expressed in NHDF-hMSC co- cultures, even higher without an intermediate membrane, although none of these observations are statistically significantly after 9 d. This is consistent with other observation in literature, where variations in leptin concentration were only reported significant after 8 d of culture and treatment with chemokines on human 537 adipocytes . Such report indicates quantities of ng/ml for mature adipocytes, three degrees of magnitude higher than the concentrations quantified in the DCB. This is indicative that either dynamic culturing in the DCB or the co-culture with NHDF is decreasing the expression of leptin or slowing the maturation of the cells. Last, bFGF is generally not expressed by day 9; quantifiable levels are only present at days 0 and 1 as it is a component supplemented into the NHDF growth media, but these levels decrease and disappear over time (no significant difference to non-seeded EShell in all cases and below the detection threshold in most cases). As with leptin, bFGF levels seemingly tend to increase with time but are non-significant by day 9. Other specific molecular markers with higher specificity that should be considered to assess the development of the adipose tissue in the DCB could include peroxisome proliferator-activated receptor gamma (PPAR?), the CCATT/enhancer binding 535 protein (C/EBP), or fatty acid-binding protein 4 (FABP4) . 256 Figure 9.4. Expression of extracellular proteins in the DCB. The quantities of adipose-specific extracellular proteins adiponectin (lower detection threshold close to 52 pg/ml) and leptin (lower detection threshold close to 34 pg/ml), as well as bFGF (lower detection threshold close to 18 pg/ml) also secreted by adipocytes, were tracked in the media collected from both lines of the DCBs over time, specifically at seeding time (t0), after adipogenic induction at day 1 (D1), after 3 d (D3), and at end- point after 9 d (D9). Adiponectin was significantly expressed by hMSCs at the end- point, indicative of successful adipogenic differentiation in the DCB, and only occurred when separated from NHDF populations by an intermediate membrane, independent of interface curvature. Leptin was generally not expressed in the cases studied (no significant difference to non-seeded EShell in all cases and below the detection threshold in most cases); the only observable trend indicates that it is better expressed in NHDF-hMSC co-cultures, even higher without an intermediate membrane, although none of these observations are statistically significantly after 9 d. Last, bFGF is generally not expressed by day 9; quantifiable levels are only present at days 0 and 1 as it is a component supplemented into the NHDF growth media, but these levels decrease and disappear over time (no significant difference to non-seeded EShell in all cases and below the detection threshold in most cases). As with leptin, 257 bFGF levels seemingly tend to increase with time but are still non-significant after 9 d. ANOVA and Tukey?s multiple pairwise comparisons were used to compare line A (NHDF-laden scaffolds) of all bioreactors at the end-point (black letters, groups that do not share the same letter are significantly different, p < 0.05). Comparison of end- point line B of all bioreactors was run independently (red letters, groups that do not share the same letter are significantly different, p < 0.05). Comparison of lines A and B of each bioreactor at each time point was done with two-sample t-test for the mean (significance using p < 0.01 (**) and p < 0.05 (*)). The DCB system allows cell culture and development even with the inclusion of the curvature parameter. In almost all cases, the inclusion of smooth or coarse curves in the system did not significantly alter the expression of adipogenic markers compared to controls and flat scaffolds. There are indications that co-culturing with an adjacent fibroblast population, not curvature, does influence the development of the hMSCs. Particularly in the case of adiponectin, its highest expression levels in hMSC lines occurred when separated from NHDF populations by an intermediate membrane, independent of interface curvature (p < 0.05). Significant difference was observed with the removal of the membrane (single continuous scaffolds). Without the regulating membrane, mixing of the media seems to have negative effects on the development of the cells, an additional parameter supporting the inclusion of the hydrogel membranes in the DCB. As stated before, the adipocytes in the native hypodermis synthesize adipose-specific ECM (mostly collagen types I, III, IV, V, and 35 32,36?39 VI ) and adipose-specific growth factors and hormones . If the hypodermis is producing all these factors in vivo, diffusion through the tissues and possibly convection through the common deep vascular plexus is transporting them to the adjacent dermis layer, but in no case there is direct mixing. The DCB allows us to replicate this diffusive transport between stratified hypodermis and dermis to prove 258 its benefits in skin tissue engineering. At 9 d these effects are limited, although data trends toward positive, increasing concentrations of molecular markers. Sustained diffusion of between the cell populations over longer periods of time, preemptively at 473,474 least 21 d that mark mature adipogenic differentiation of hMSCs . Long-term presence of growth factors and hormones such as leptin, adiponectin, or bFGF in the dermis could have the potential to improve the formation of a common vascular network between layers and subsequent wound-healing processes, therefore increasing the viability of the stratified construct in vivo. 9.4. Conclusions Current skin constructs for facial wounds cannot restore dermal physiological functions and complex topographical features lost in burns. Our overall approach to study the dermis-hypodermis complex in a bioreactor system is a novel method to address both current limitations in skin tissue engineering and one that can be extended to other layered or gradient tissues. Stratifying the hypodermis and dermis layers would be beneficial for the overall development of a skin construct. Clinicians can treat and graft skin burn wounds, but, to the best of our knowledge, no viable approach has been able to successfully address both the dermal physiology and the facial topography reconstruction simultaneously. Throughout this work we have confirmed the viability of our DCB for growing non-planar adjacent cell populations with the inclusion of a regulatory keratin-based membrane. We can successfully introduce curved interfaces in the system, and we can perfuse sustain such profiles under cell culture conditions for at 259 least 9 d. An added value developed to assess this problem was the introduction of the real-time imaging DCB tool. Furthermore, we successfully assessed the integration of both curvature and cell viability questions in a single system. We can assess the synergistic development of co-cultured adjacent dermis fibroblasts and hypodermis stem-cell-derived adipocytes with the inclusion of curvature parameters that do not reduce cellular viability in the DCB. The complexity of including parameters of dynamic flow, three-dimensionality, co-culturing, bioreactor conditions, and now curvature, has allowed us to set up a complete system for the interrogation of dermis- hypodermis constructs for the overall goal of topographic and physiological reconstruction of facial tissues. 260 Chapter 10: Summary The overall goal of this work was to develop engineering strategies to produce and study skin tissue constructs that can simultaneously address physiological and topographical reconstruction of severe burn wounds, particularly those to the face. The motivation for this work was to enhance facial reconstruction, both of features and tissue functionality, for patients with severe burn wounds that compromise their identity and perception of self. Throughout the course of this work we developed and studied two viable strategies for the reconstruction of complex facial burn wounds. The first, a 3D printing keratin-based hydrogel approach to accurately capture complex shapes while serving as a drug-delivery mechanism to treat burn wounds. The second, aimed to address limitations of the first, aimed to develop a complex bioreactor system that could be used to study stratified tissues such as skin. The first part of this work was aimed to formulate a keratin-based bioink that could be used for 3D printing on a commercially available lithography-based 3D printer. The hydrogels obtained have 3D printing resolution close to 1mm and mechanical, biochemical, and microstructural properties comparable or superior to those of casted keratin scaffolds previously reported. This was the first time a keratin- based bioink for 3D printing was reported in literature. The next project aimed to use the keratin-based bioink in the production of drug-laden face masks for the improvement of contracture, scarring, and aesthetics in severe burn wound healing. We 3D printed scaffolds that were successfully loaded with Halofuginone, a collagen I synthase inhibitor to decrease collagen synthesis in 261 fibrosis cases. Manufacturing steps such as lyophilization and gamma irradiation sterilization were proven to cause significant changes the dimensions of the scaffolds, an important consideration to match wounds for in vivo testing. A red duroc pig model was used to prove that the benefits of Halofuginone-laden printed keratin are non-inferior to other similar products in literature and the clinic. We saw improvement of healing parameters such as hyperkeratosis, growth of apocrine glands and hair follicles, reduced inflammation and vascular proliferation over 70 d. Collagen order and stratification of the tissues were generally lacking, although these in vivo studies are indicative of the potential of the bioink in dermal wound healing. Next, based on our in vivo experience, we aimed to understand and reconstruct the layered structure of skin using a guided tissue regeneration (GTR) approach. In GTR, a membrane is used to regulate tissue growth by stopping fast-proliferating cells and allowing slower cells to migrate and reconstruct specialized microstructures. Here, we proposed the use of UV crosslinked keratin membranes based on our keratin bioink, Variables from the crosslinking process were grouped within an energy density (ED) parameter to manufacture and evaluate the membranes. Sol fraction, spectrographs, and thermograms were used to quantify the non-linear relation between ED and the resulting crosslinking degree (CD). Transport assays showed that the membranes allow molecular diffusion, dependent of CD. This was confirmed with lower transport of adipogenic growth factors to hMSCs when using low ED membranes; high ED samples resulted in increased production of intracellular lipids. Overall, we can engineer keratin membranes with specific CD, a valuable tool to tune microstructural and transport properties. 262 Next, we presented a dual-chambered bioreactor (DCB) for co-culturing adjacent cell populations with the inclusion of a regulatory keratin membrane. The DCB can provide adjacent flow lines within a common chamber; the inclusion of the membrane can regulate flow layering or mixing, which can be exploited to produce stratification or gradients of cell populations in the scaffolds. Molecular transport across the membrane was defined by a balance of convection and diffusion. This balance allowed the perfusion of two different fluids, with the membrane defining the mixing degree between the two. Furthermore, the bioreactor sustained two adjacent, healthy cell populations for at least 28 d, and was used to induce indirect adipogenic differentiation of hMSCs due to molecular cross-talk between the populations. Last, we assessed the viability of the DCB and membrane complex for the generation of curved interfaces between adjacent dermal and hypodermal cell cultures. We proved the DCB can incorporate curved interfaces and perfuse such profiles under cell culture conditions for at least 9 d. An added value developed to assess this problem was the introduction of the imaging DCB tool, a variation that allows real-time imaging of the inside of the bioreactor. Furthermore, we successfully assessed the integration of both curvature and cell viability questions in a single system with co-cultured adjacent dermis fibroblasts and hypodermis stem-cell- derived adipocytes. The strategies developed here elucidate on tissue stratification and aesthetic reconstruction. Furthermore, the keratin-based bioink, the membrane engineering, and the DCB technologies can be extended to study other stratified or gradient tissues and to fine-tune communication between cell populations in complex 3D constructs. 263 Chapter 11: Future Directions This work has highlighted the need for skin tissue engineering strategies that address simultaneous topographical and physiological reconstruction of complex facial burn wounds. Current treatments of facial burn wounds result in hyperthropic scarring and lack of elasticity and thus in loss of skin function and disfigurement. Clinicians can treat and graft skin burn wounds, but, to the best of our knowledge, no viable approach has been able to successfully address both the dermal physiology and the facial topography reconstruction simultaneously. The 3D printing approach developed in Chapters 5, 6, and 7 resulted in the formulation of a novel keratin-based bioink. We have extensively characterized its properties and optimized the lithography-based 3D printing process with the material. Based on our in vitro and in vivo assessments, keratin demonstrated great potential for tissue engineering and drug-delivery applications. As it proved to be non-inferior or superior to other reported hydrogels used in vivo, the keratin scaffolds have potential for regeneration and drug delivery in other soft tissue applications. Furthermore, the dityrosine crosslinking chemistry employed could be used to create hydrogels using other proteins. The bioreactor approach developed in Chapters 7, 8, and 9 resulted in the development of the DCB-membranes system. After extensive characterization and optimization, by the end of this work, the DCB system can include complex parameters such as dynamic flow, three-dimensionality, co-culturing, bioreactor conditions, and curvature, which sets up a complete system for the interrogation of 264 dermis-hypodermis constructs for the overall goal of topographic and physiological reconstruction of facial tissues. Our overall approach to study the dermis-hypodermis complex in a bioreactor system is a novel method to address current limitations in skin tissue engineering. The effects of curvature, particularly the effect of increasing resistance within the bioreactor chambers is an interesting parameter to study. Variable resistance in the bioreactor has the potential to form regions of high or low velocity or shear stress, which could be exploited to induce hMSC differentiation into hard tissues such as bone or cartilage, or the more complex osteochondral interface. Another area to explore is vascularization of the gradient and layered structures. The bioreactor interface can be used to induce the formation of a common vascular network between layers, as the common vascular nexus between dermis and hypodermis. The inclusion of endothelial cells or endothelial growth factors in the system could be benefitted with the oriented flow into the formation of such vascular networks. Both the variable resistance and the vascularization projects have the potential to compliment previous work our research lab has completed on bone tissue engineering and bioreactor development. Additionally, future studies should question the effect of the scaffold material. So far we have used synthetic EShell 300 3D printed porous scaffolds in the DCBs; it would be interesting to move forward into testing nature-derived materials as supports in the bioreactors. We are particularly curious of including layer-specific extracellular matrix (ECM). 3D printed scaffolds made from dermis ECM and hypodermis/adipose tissue ECM would provide additional layer-specific cues that can further elucidate on the development of the two adjacent tissues. 265 In general, the DCB system can be used to study molecular transport between adjacent cell populations, and to engineer stratified tissue constructs. Our current application is for the stratified co-culturing of dermis and hypodermis tissue equivalents for the physiologic and topographic reconstruction of facial soft tissues, but it has the potential to extend to other layered issues, gradients, multi-tissue interfaces, or simply tissues with complex topographies that complicate their development in vitro. As a modeling system it has immediate application to (1) study layered tissues, (2) the co-culture of two adjacent different cell lines, and (3) the study of crosstalk communication between adjacent populations. We expect this technology will have broad impact on future biomedical research, tissue engineering, and tissue modeling applications. 266 References 1. Lattouf, R. et al. Picrosirius red staining: a useful tool to appraise collagen networks in normal and pathological tissues. J. Histochem. Cytochem. 62, 751? 8 (2014). 2. Arno, A., Barret, J. P., Harrison, R. A. & Jeschke, M. G. Face Allotransplantation and Burns. J. Burn Care Res. 33, 561?576 (2012). 3. World Health Organization (WHO). Burns: Fact Sheet (Reviewed September 2016). (2016). 4. Markeson, D. et al. Scarring, stem cells, scaffolds and skin repair. J. Tissue Eng. Regen. Med. 9, 649?668 (2015). 5. Atiyeh, B. S., Gunn, S. W. A. & Hayek, S. N. Military and civilian burn injuries during armed conflicts. Ann. Burns Fire Disasters 20, 203?15 (2007). 6. Weiss, J. Women?s rights in Colombia: Acid attacks on the rise. World Aff. 177, 50?57 (2014). 7. Mannan, A., Ghani, S., Sen, S. L., Clarke, A. & Butler, P. E. M. The problem of acid violence in Bangladesh. J. Surg. 2, 39?43 (2004). 8. Seifert, D. et al. Violence against children: Further evidence suggesting a relationship between burns, scalds, and the additional injuries. Int. J. Legal Med. 124, 49?54 (2010). 9. American Burn Association. National Burn Awareness Week 2017 Fact Sheet. (2017). at 10. Peck, M., Molnar, J. & Swart, D. A global plan for burn prevention and care. Bull. World Health Organ. 87, 802?803 (2009). 11. World Health Organization (WHO). Burn Prevention: Success Stories, Lessons Learned. (2011). 12. Orgill, D. P. & Ogawa, R. Current methods of burn reconstruction. Plast. Reconstr. Surg. 131, 827e?36e (2013). 13. Talley, H. L. Saving Face?: Disfigurement and the Politics of Appearance. (New York University Press, 2014). 14. Drosou, A. et al. Use of a bioengineered skin equivalent for the management of difficult skin defects after pediatric multivisceral transplantation. J. Am. Acad. Dermatol. 52, 854?858 (2005). 15. Jones, I., Currie, L. & Martin, R. A guide to biological skin substitutes. Br. J. Plast. Surg. 55, 185?193 (2002). 16. Cervelli, V. et al. The use of MatriDerm and skin grafting in post-traumatic wounds. Int. Wound J. 8, 400?405 (2011). 17. Wood, F. M. Skin regeneration: The complexities of translation into clinical practise. Int. J. Biochem. Cell Biol. 56, 133?140 (2014). 267 18. Schneider, R. K. M. et al. Three-dimensional epidermis-like growth of human mesenchymal stem cells on dermal equivalents: Contribution to tissue organization by adaptation of myofibroblastic phenotype and function. Differentiation 76, 156?167 (2008). 19. Lee, W. et al. Multi-layered culture of human skin fibroblasts and keratinocytes through three-dimensional freeform fabrication. Biomaterials 30, 1587?1595 (2009). 20. Uchino, T., Takezawa, T. & Ikarashi, Y. Reconstruction of three-dimensional human skin model composed of dendritic cells, keratinocytes and fibroblasts utilizing a handy scaffold of collagen vitrigel membrane. Toxicol. Vitr. 23, 333?337 (2009). 21. Lee, V. et al. Design and fabrication of human skin by three-dimensional bioprinting. Tissue Eng. Part C. Methods 20, 473?84 (2014). 22. Hartmann-Fritsch, F. et al. Collagen hydrogels strengthened by biodegradable meshes are a basis for dermo-epidermal skin grafts intended to reconstitute human skin in a one-step surgical intervention. J. Tissue Eng. Regen. Med. 10, 81?91 (2016). 23. Cubo, N. et al. 3D bioprinting of functional human skin: production and in vivo analysis. Biofabrication 9, 015006 (2016). 24. Wills, J. W. et al. Genetic toxicity assessment of engineered nanoparticles TM using a 3D in vitro skin model (EpiDerm ). Part. Fibre Toxicol. 13, 50 (2016). 25. Brohem, C. A. et al. Artificial skin in perspective: Concepts and applications. Pigment Cell Melanoma Res. 24, 35?50 (2011). 26. T?ot, L., Otman, S., Trial, C. & Brancati, A. The use of noncellularized artificial dermis in the prevention of scar contracture and hypertrophy. Wound Repair Regen. 19, 49?58 (2011). 27. Ali, N. et al. Skin equivalents: Skin from reconstructions as models to study skin development and diseases. Br. J. Dermatol. 173, 391?403 (2015). 28. Langer, R. & Vacanti, J. Tissue engineering. Science (80-. ). 260, 920?926 (1993). 29. Khademhosseini, A. & Langer, R. A decade of progress in tissue engineering. Nat. Protoc. 11, 1775?1781 (2016). 30. McLafferty, E., Hendry, C. & Farley, A. The integumentary system: anatomy, physiology and function of skin. Nurs. Stand. 27, 35?42 (2012). 31. B?ttcher-Haberzeth, S., Biedermann, T. & Reichmann, E. Tissue engineering of skin. Burns 36, 450?460 (2010). 32. Kolarsick, P. A. J., Kolarsick, M. A. & Goodwin, C. Anatomy and Physiology of the Skin. J. Dermatol. Nurses. Assoc. 3, 203?213 (2011). 33. Zhang, Z. & Michniak-Kohn, B. B. Tissue engineered human skin equivalents. 268 Pharmaceutics 4, 26?41 (2012). 34. Konig, A. & Bruckner-Tuderman, L. Epithelial-mesenchymal interactions enhance expression of collagen VII in vitro. J Invest Dermatol 96, 803?808 (1991). 35. Kang, J. H., Gimble, J. M. & Kaplan, D. L. In Vitro 3D Model for Human Vascularized Adipose Tissue. Tissue Eng. Part A 15, 2227?2236 (2009). 36. Wong, R., Geyer, S., Weninger, W., Guimberteau, J. C. & Wong, J. K. The dynamic anatomy and patterning of skin. Exp. Dermatol. 25, 92?98 (2016). 37. Shibata, S. et al. Adiponectin Regulates Cutaneous Wound Healing by Promoting Keratinocyte Proliferation and Migration via the ERK Signaling Pathway. J. Immunol. 189, 3231?3241 (2012). 38. Salathia, N. S., Shi, J., Zhang, J. & Glynne, R. J. An in vivo screen of secreted proteins identifies adiponectin as a regulator of murine cutaneous wound healing. J. Invest. Dermatol. 133, 812?21 (2013). 39. Jin, C. E., Xiao, L., Ge, Z. H., Zhan, X. B. & Zhou, H. X. Role of adiponectin in adipose tissue wound healing. Genet. Mol. Res. 14, 8883?8891 (2015). 40. Hebert, T. L. et al. Culture effects of epidermal growth factor (EGF) and basic fibroblast growth factor (bFGF) on cryopreserved human adipose-derived stromal/stem cell proliferation and adipogenesis. J. Tissue Eng. Regen. Med. 3, 553?561 (2009). 41. Makino, T. et al. Basic fibroblast growth factor stimulates the proliferation of human dermal fibroblasts via the ERK1/2 and JNK pathways. Br. J. Dermatol. 162, 717?723 (2010). 42. Rivera-Gonzalez, G., Shook, B. & Horsley, V. Adipocytes in skin health and disease. Cold Spring Harb. Perspect. Med. 4, 1?18 (2014). 43. Hamuy, R. et al. One-stage, simultaneous skin grafting with artificial dermis and basic fibroblast growth factor successfully improves elasticity with maturation of scar formation. Wound Repair Regen. 21, 141?154 (2013). 44. Darwish, A. in Skin Grafts - Indications, Applications and Current Research 35?42 (2011). doi:10.5772/21916 45. Chua, A. W. C. et al. Skin tissue engineering advances in severe burns: review and therapeutic applications. Burn. Trauma 4, 3 (2016). 46. Madaghiele, M., Sannino, A., Ambrosio, L. & Demitri, C. Polymeric hydrogels for burn wound care: Advanced skin wound dressings and regenerative templates. Burn. Trauma 2, 153 (2014). 47. Robinson, M. K. et al. Non-animal testing strategies for assessment of the skin corrosion and skin irritation potential of ingredients and finished products. Food Chem. Toxicol. 40, 573?592 (2002). 48. Holmes, A. M. et al. Rising to the challenge: applying biofabrication approaches for better drug and chemical product development. Biofabrication 269 9, (2017). 49. Almeida, A., Sarmento, B. & Rodrigues, F. Insights on in vitro models for safety and toxicity assessment of cosmetic ingredients. Int. J. Pharm. 519, 178?185 (2017). 50. Hansbrough, J. F. & Naughton, G. K. TEMPORARY LIVING SKIN REPLACEMENT. (1995). doi:US006040467A 51. Sorrell, J. M. & Caplan, A. I. MULTILAYERSKIN OR DERMAL EQUIVALENT HAVING A LAYER CONTAINING MESENCHYMAL STEM CELLS. (2002). doi:10.1038/incomms1464 52. Eisenberg, M. COMPOSITE LIVING SKIN EQUIVALENT. (2000). doi:10.1016/j.(73) 53. Kemp, P., Bell, E., Kagan, D. T., Mason, V. & Cavallaro, J. PREPARATION OF TISSUE EQUIVALENTS BY CONTRACTION OF A COLLAGEN GEL LAYERED ON A COLLAGEN GEL. (1996). doi:10.1227/NEU.0000000000000466 54. Curren, R. D., Mun, G. C., Gibson, D. P. & Aardema, M. J. Development of a method for assessing micronucleus induction in a 3D human skin model (EpiDerm TM ). Mutat. Res. 607, 192?204 (2006). TM 55. Mun, G. C. et al. Further development of the EpiDerm 3D reconstructed human skin micronucleus (RSMN) assay. Mutat. Res. - Genet. Toxicol. Environ. Mutagen. 673, 92?99 (2009). 56. Bellas, E., Seiberg, M., Garlick, J. & Kaplan, D. L. In vitro 3D Full-Thickness Skin-Equivalent Tissue Model Using Silk and Collagen Biomaterials. Macromol. Biosci. 12, 1627?1636 (2012). 57. Davies, D. J., Heylings, J. R., Gayes, H., McCarthy, T. J. & Mack, M. C. Further development of an in vitro model for studying the penetration of chemicals through compromised skin. Toxicol. Vitr. 38, 101?107 (2017). 58. Li, W. H., Fassih, A., Binner, C., Parsa, R. & Southall, M. D. Low-level red LED light inhibits hyperkeratinization and inflammation induced by unsaturated fatty acid in an in vitro model mimicking acne. Lasers Surg. Med. 50, 158?165 (2018). 59. National Toxicology Program - U.S. Department of Health and Human Services. Alternative Methods Accepted by US Agencies. at 60. WSSD Global Partnership for Capacity Building to Implement the Globally Harmonized System of Classification and Labelling of Chemicals (GHS). The GHS and the Global Partnership: a success story from Rio to Rio. Achievements, lessons learned and future directions. in (2012). at 61. Hewitt, N. J. et al. Use of Human In Vitro Skin Models for Accurate and 270 Ethical Risk Assessment: Metabolic Considerations. Toxicol. Sci. 133, 209? 217 (2013). 62. Pfuhler, S. et al. In vitro genotoxicity test approaches with better predictivity: Summary of an IWGT workshop. Mutat. Res. - Genet. Toxicol. Environ. Mutagen. 723, 101?107 (2011). 63. Matts, P. J. et al. The COLIPA in vitro UVA method: A standard and reproducible measure of sunscreen UVA protection. Int. J. Cosmet. Sci. 32, 35?46 (2010). 64. Flamand, N. et al. Development of genotoxicity test procedures with Episkin?, a reconstructed human skin model: Towards new tools for in vitro risk assessment of dermally applied compounds? Mutat. Res. - Genet. Toxicol. Environ. Mutagen. 606, 39?51 (2006). 65. Pageon, H. et al. Biological Effects Induced by Specific Advanced Glycation End Products in the Reconstructed Skin Model of Aging. Biores. Open Access 4, 54?64 (2015). 66. Ido, Y. et al. Resveratrol prevents oxidative stress-induced senescence and proliferative dysfunction by activating the AMPK-FOXO3 cascade in cultured primary human keratinocytes. PLoS One 10, 1?18 (2015). 67. Eilstein, J. et al. Xenobiotic metabolizing enzymes in human skin and SkinEthic reconstructed human skin models. Exp. Dermatol. 24, 547?549 (2015). 68. Hsia, E., Johnston, M. J., Houlden, R. J., Chern, W. H. & Hofland, H. E. J. Effects of topically applied acitretin in reconstructed human epidermis and the rhino mouse. J. Invest. Dermatol. 128, 125?30 (2008). 69. Elias, M. S. et al. Proteomic analysis of filaggrin deficiency identifies molecular signatures characteristic of atopic eczema. J. Allergy Clin. Immunol. 140, 1299?1309 (2017). 70. Schmook, F. P., Meingassner, J. G. & Billich, A. Comparison of human skin or epidermis models with human and animal skin in in-vitro percutaneous absorption. Int. J. Pharm. 215, 51?56 (2001). 71. Seyhan, T. in Skin Grafts - Indications, Applications and Current Research 3? 17 (2011). at 72. Debels, H., Hamdi, M., Abberton, K. & Morrison, W. Dermal matrices and bioengineered skin substitutes: A critical review of current options. Plastic and Reconstructive Surgery - Global Open 3, (2015). 73. Gurtner, G., Werner, S., Barrandon, Y. & Longaker, M. Wound repair and regeneration. Nature 453, 314?321 (2008). 74. Hu, M. S. et al. Tissue engineering and regenerative repair in wound healing. Ann. Biomed. Eng. 42, 1494?507 (2014). 75. Turner, N. J. & Badylak, S. F. The Use of Biologic Scaffolds in the Treatment 271 of Chronic Nonhealing Wounds. Adv. wound care 4, 490?500 (2015). 76. Thornton, J. & Gosman, A. Skin Grafts and Skin Substitutes and Principles of Flaps. Sel. Readings Plast. Surg. 10, (2004). 77. Loss, M., Wedler, V., K??nzi, W., Meuli-Simmen, C. & Meyer, V. E. Artificial skin, split-thickness autograft and cultured autologous keratinocytes combined to treat a severe burn injury of 93% of TBSA. Burns 26, 644?652 (2000). 78. Han, G. & Ceilley, R. Chronic Wound Healing: A Review of Current Management and Treatments. Advances in Therapy 34, 599?610 (2017). 79. Nunan, R., Harding, K. G. & Martin, P. Clinical challenges of chronic wounds: searching for an optimal animal model to recapitulate their complexity. Dis. Model. Mech. 7, 1205?1213 (2014). 80. Freyman, T. M., Yannas, I. V. & Gibson, L. J. Cellular materials as porous scaffolds for tissue engineering. Prog. Mater. Sci. 46, 273?282 (2001). 81. Metcalfe, A. D. & Ferguson, M. W. J. Tissue engineering of replacement skin: the crossroads of biomaterials, wound healing, embryonic development, stem cells and regeneration. J. R. Soc. Interface 4, 413?437 (2007). 82. Choi, J., Lee, E. H., Park, S. W. & Chang, H. Regulation of transforming growth factor B1, Platelet-derived growth factor, And basic fibroblast growth factor by silicone gel sheeting in early-stage scarring. Arch. Plast. Surg. 42, 20?27 (2015). 83. Wang, H. M. et al. Novel Biodegradable Porous Scaffold Applied to Skin Regeneration. PLoS One 8, 2?12 (2013). 84. Sakamoto, M. et al. Efficacy of gelatin gel sheets in sustaining the release of basic fibroblast growth factor for murine skin defects. J. Surg. Res. 201, 378? 387 (2016). 85. Llames, S. G. et al. Human plasma as a dermal scaffold for the generation of a completely autologous bioengineered skin. Transplantation 77, 350?355 (2004). 86. Xu, S., Sang, L., Zhang, Y., Wang, X. & Li, X. Biological evaluation of human hair keratin scaffolds for skin wound repair and regeneration. Mater. Sci. Eng. C 33, 648?655 (2013). 87. Poranki, D. et al. Evaluation of skin regeneration after burns in vivo and rescue of cells after thermal stress in vitro following treatment with a keratin biomaterial. J. Biomater. Appl. 29, 26?35 (2014). 88. Loan, F., Cassidy, S., Marsh, C. & Simcock, J. Keratin-based products for effective wound care management in superficial and partial thickness burns injuries. Burns 42, 541?547 (2016). 89. Sarkar, S. D., Farrugia, B. L., Dargaville, T. R. & Dhara, S. Chitosan-collagen scaffolds with nano/microfibrous architecture for skin tissue engineering. J. Biomed. Mater. Res. - Part A 101, 3482?3492 (2013). 272 90. Sun, G. et al. Dextran hydrogel scaffolds enhance angiogenic responses and promote complete skin regeneration during burn wound healing. Proc. Natl. Acad. Sci. 108, 20976?20981 (2011). 91. Coburn, J. C., Brody, S., Billiar, K. L. & Pandit, A. Biaxial mechanical evaluation of cholecyst-derived extracellular matrix: A weakly anisotropic potential tissue engineered biomaterial. (2007). doi:10.1002/jbm.a.30943 92. Webber, M. J., Khan, O. F., Sydlik, S. A., Tang, B. C. & Langer, R. A Perspective on the Clinical Translation of Scaffolds for Tissue Engineering. Ann. Biomed. Eng. 43, 641?656 (2015). 93. Rozman, B., Gasperlin, M., Tinois-Tessoneaud, E., Pirot, F. & Falson, F. Simultaneous absorption of vitamins C and E from topical microemulsions using reconstructed human epidermis as a skin model. Eur. J. Pharm. Biopharm. 72, 69?75 (2009). 94. Fleischli, F. D., Morf, F. & Adlhart, C. Skin Concentrations of Topically Applied Substances in Reconstructed Human Epidermis (RHE) Compared with Human Skin Using in vivo Confocal Raman Microscopy. Chim. Int. J. Chem. 69, 147?151 (2015). 95. Rodrigues, F. et al. Are coffee silverskin extracts safe for topical use? An in vitro and in vivo approach. Ind. Crops Prod. 63, 167?174 (2015). 96. Bernstam, L. I., Vaughan, F. L. & Bernstein, I. A. Keratinocytes grown at the air-liquid interface. In Vitro Cell. Dev. Biol. 22, 695?705 (1986). 97. Trottier, V., Marceau-Fortier, G., Germain, L., Vincent, C. & Fradette, J. IFATS Collection: Using Human Adipose-Derived Stem/Stromal Cells for the Production of New Skin Substitutes. Stem Cells 26, 2713?2723 (2008). 98. Monfort, A., Soriano-Navarro, M., Garc?a-Verdugo, J. M. & Izeta, A. Production of human tissue-engineered skin trilayer on a plasma-based hypodermis. J. Tissue Eng. Regen. Med. 7, 479?490 (2013). 99. Sugihara, H., Toda, S., Yonemitsu, N. & Watanabe, K. Effects of fat cells on keratinocytes and fibroblasts in a reconstructed rat skin model using collagen gel matrix culture. Br. J. Dermatol. 244?253 (2001). 100. Alapure, B. V et al. Accelerate Healing of Severe Burn Wounds by Mesenchymal Stem Cell Seeded Biodegradable Hydrogel Scaffold Synthesized from Arginine-based Poly(ester amide) and Chitosan. Stem Cells Dev. scd.2018.0106 (2018). doi:10.1089/scd.2018.0106 101. Mirzaei-parsa, M. J. et al. Nanofiber-acellular dermal matrix as a bilayer scaffold containing mesenchymal stem cell for healing of full-thickness skin wounds. Cell Tissue Res. 1?13 (2018). doi:10.1007/s00441-018-2927-6 102. Marshall, J. M., Grant, I. & Cederholm-Williams, S. A. SPRAY DELIVERY OF CELLS. (2002). doi:10.1016/j.(73) 103. Wood, F. M., Giles, N., Stevenson, A., Rea, S. & Fear, M. Characterisation of 273 the cell suspension harvested from the dermal epidermal junction using a ReCell? kit. Burns 38, 44?51 (2012). 104. Holmes IV, J. H. et al. A Comparative Study of the ReCell? Device and Autologous Spit-Thickness Meshed Skin Graft in the Treatment of Acute Burn Injuries. J. Burn Care Res. 694?702 (2018). doi:10.1093/jbcr/iry029 105. Friedenstein, A. J., Chailakhjan, R. K. & Lalykina, K. S. The development of fibroblast colonies in monolayer cultures of guinea-pig bone marrow and spleen cells. Cell Tissue Kinet. 3, 393?403 (1970). 106. Dominici, M. et al. Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy 8, 315?317 (2006). 107. Muraglia, A., Cancedda, R. & Quarto, R. Clonal mesenchymal progenitors from human bone marrow differentiate in vitro according to a hierarchical model. J. Cell Sci. 113, 1161?1166 (2000). 108. Kim, E. Y. et al. Neuronal cell differentiation of mesenchymal stem cells originating from canine amniotic fluid. Hum. Cell 27, 51?58 (2014). 109. Singaravelu, K. & Padanilam, B. J. In vitro differentiation of MSC into cells with a renal tubular epithelial-like phenotype. Ren. Fail. 31, 492?502 (2009). 110. Sasaki, M. et al. Mesenchymal Stem Cells Are Recruited into Wounded Skin and Contribute to Wound Repair by Transdifferentiation into Multiple Skin Cell Type. J. Immunol. 180, 2581?2587 (2008). 111. Pers, Y. M., Ruiz, M., No?l, D. & Jorgensen, C. Mesenchymal stem cells for the management of inflammation in osteoarthritis: State of the art and perspectives. Osteoarthr. Cartil. 23, 2027?2035 (2015). 112. Le Blanc, K. & Davies, L. C. Mesenchymal stromal cells and the innate immune response. Immunol Lett 168, 140?146 (2015). 113. Amann, E. M. et al. Systemic recovery and therapeutic effects of transplanted allogenic and xenogenic mesenchymal stromal cells in a rat blunt chest trauma model. Cytotherapy 20, 218?231 (2018). 114. Rodriguez-Menocal, L., Shareef, S., Salgado, M., Shabbir, A. & Van Badiavas, E. Role of whole bone marrow, whole bone marrow cultured cells, and mesenchymal stem cells in chronic wound healing. Stem Cell Res. Ther. 6, 24 (2015). 115. Sargent, A. & Miller, R. H. MSC Therapeutics in Chronic Inflammation. Curr. Stem Cell Reports 2, 168?173 (2016). 116. Gaur, M. et al. Mesenchymal Stem Cells from Adipose Tissue in Clinical Applications for Dermatological Indications and Skin Aging. Int. J. Mol. Sci. 18, 208 (2017). 117. Madrigal, M., Rao, K. S. & Riordan, N. H. A review of therapeutic effects of mesenchymal stem cell secretions and induction of secretory modification by 274 different culture methods. J. Transl. Med. 12, 260 (2014). 118. Mansilla, E. et al. Bloodstream Cells Phenotypically Identical to Human Mesenchymal Bone Marrow Stem Cells Circulate in Large Amounts Under the Influence of Acute Large Skin Damage: New Evidence for Their Use in Regenerative Medicine. Transplant. Proc. 38, 967?969 (2006). 119. Fu, W. L., Zhou, C. Y. & Yu, J. K. A new source of mesenchymal stem cells for articular cartilage repair: MSCs derived from mobilized peripheral blood share similar biological characteristics in vitro and chondrogenesis in vivo as MSCs from bone marrow in a rabbit model. Am. J. Sports Med. 42, 592?601 (2014). 120. Squillaro, T., Peluso, G. & Galderisi, U. Clinical Trials With Mesenchymal Stem Cells: An Update. Cell Transplant. 25, 829?48 (2016). 121. Waterman, R. S., Tomchuck, S. L., Henkle, S. L. & Betancourt, A. M. A new mesenchymal stem cell (MSC) paradigm: Polarization into a pro-inflammatory MSC1 or an immunosuppressive MSC2 phenotype. PLoS One 5, (2010). 122. Bernardo, M. E. & Fibbe, W. E. Mesenchymal stromal cells: Sensors and switchers of inflammation. Cell Stem Cell 13, 392?402 (2013). 123. Conget, P. et al. Replenishment of type VII collagen and re-epithelialization of chronically ulcerated skin after intradermal administration of allogeneic mesenchymal stromal cells in two patients with recessive dystrophic epidermolysis bullosa. Cytotherapy 12, 429?431 (2010). 124. Cerqueira, M. T., Marques, A. P. & Reis, R. L. Using Stem Cells in Skin Regeneration: Possibilities and Reality. Stem Cells Dev. 21, 1201?1214 (2012). 125. Dong, H. K. et al. Gene expression profile of cytokine and growth factor during differentiation of bone marrow-derived mesenchymal stem cell. Cytokine 31, 119?126 (2005). 126. Li, H. et al. Adult bone-marrow-derived mesenchymal stem cells contribute to wound healing of skin appendages. Cell Tissue Res. 326, 725?736 (2006). 127. Chen, L., Tredget, E. E., Wu, P. Y. G., Wu, Y. & Wu, Y. Paracrine factors of mesenchymal stem cells recruit macrophages and endothelial lineage cells and enhance wound healing. PLoS One 3, (2008). 128. Fu, X. & Li, H. Mesenchymal stem cells and skin wound repair and regeneration: Possibilities and questions. Cell Tissue Res. 335, 317?321 (2009). 129. Eggenhofer, E., Luk, F., Dahlke, M. H. & Hoogduijn, M. J. The life and fate of mesenchymal stem cells. Frontiers in Immunology 5, 1?6 (2014). 130. Hoogduijn, M. J. et al. Donor-derived mesenchymal stem cells remain present and functional in the transplanted human heart. Am. J. Transplant. 9, 222?230 (2009). 131. Lee, D. E., Ayoub, N. & Agrawal, D. K. Mesenchymal stem cells and 275 cutaneous wound healing: novel methods to increase cell delivery and therapeutic efficacy. Stem Cell Res. Ther. 7, 37 (2016). 132. Rustad, K. C. et al. Enhancement of mesenchymal stem cell angiogenic capacity and stemness by a biomimetic hydrogel scaffold. Biomaterials 33, 80? 90 (2012). 133. Lee, D. E., Ayoub, N. & Agrawal, D. K. Mesenchymal stem cells and cutaneous wound healing: Novel methods to increase cell delivery and therapeutic efficacy. Stem Cell Research and Therapy 7, (2016). 134. Eke, G., Mangir, N., Hasirci, N., MacNeil, S. & Hasirci, V. Development of a UV crosslinked biodegradable hydrogel containing adipose derived stem cells to promote vascularization for skin wounds and tissue engineering. Biomaterials 129, 188?198 (2017). 135. Chen, L. et al. Pre-vascularization enhances therapeutic effects of human mesenchymal stem cell sheets in full thickness skin wound repair. Theranostics 7, 117?131 (2017). 136. Samsonraj, R. M. et al. Establishing criteria for human mesenchymal stem cell potency. Stem Cells 33, 1878?1891 (2015). 137. Trounson, A. & McDonald, C. Stem Cell Therapies in Clinical Trials: Progress and Challenges. Cell Stem Cell 17, 11?22 (2015). 138. De France, K. J., Xu, F. & Hoare, T. Structured Macroporous Hydrogels: Progress, Challenges, and Opportunities. Advanced Healthcare Materials (2018). doi:10.1002/adhm.201700927 139. Li, L., Fukunaga-Kalabis, M. & Herlyn, M. The three-dimensional human skin reconstruct model: a tool to study normal skin and melanoma progression. J. Vis. Exp. (2011). doi:10.3791/2937 140. Souto, L. R. M. et al. Model for human skin reconstructed in vitro composed of associated dermis and epidermis. Sao Paulo Med. J. 124, 71?6 (2006). 141. Yildirimer, L., Thanh, N. T. K. & Seifalian, A. M. Skin regeneration scaffolds: A multimodal bottom-up approach. Trends in Biotechnology (2012). doi:10.1016/j.tibtech.2012.08.004 142. Murphy, S. V. & Atala, A. 3D bioprinting of tissues and organs. Nature Biotechnology (2014). doi:10.1038/nbt.2958 143. Yan, W. C. et al. 3D bioprinting of skin tissue: From pre-processing to final product evaluation. Advanced Drug Delivery Reviews (2018). doi:10.1016/j.addr.2018.07.016 144. Vijayavenkataraman, S., Lu, W. F. & Fuh, J. Y. H. 3D bioprinting of skin: a state-of-the-art review on modelling, materials, and processes. Biofabrication 8, 032001 (2016). 145. Tarassoli, S. P. et al. Skin tissue engineering using 3D bioprinting: An evolving research field. Journal of Plastic, Reconstructive and Aesthetic 276 Surgery (2018). doi:10.1016/j.bjps.2017.12.006 146. El-Serafi, A. T., El-Serafi, I. T., Elmasry, M., Steinvall, I. & Sj?berg, F. Skin regeneration in three dimensions, current status, challenges and opportunities. Differentiation 96, 26?29 (2017). 147. Lee, S. H., Jeong, S. K. & Ahn, S. K. An update of the defensive barrier function of skin. Yonsei Medical Journal (2006). doi:10.3349/ymj.2006.47.3.293 148. Lee, C.-Y., Chang, T.-F., Chang, N.-Y. & Chang, Y.-C. An automated skin segmentation of Breasts in Dynamic Contrast-Enhanced Magnetic Resonance Imaging. Sci. Rep. 8, 6159 (2018). 149. Kim, H., Chang, A. R., Cho, S. & Ye, S. J. A patient-specific three- dimensional couplant pad for ultrasound image-guided radiation therapy: A feasibility study. Radiat. Oncol. (2018). doi:10.1186/s13014-018-1098-7 150. Kleinerman, R., Whang, T. B., Bard, R. L. & Marmur, E. S. Ultrasound in dermatology: Principles and applications. J. Am. Acad. Dermatol. (2012). doi:10.1016/j.jaad.2011.12.016 151. Liu, P., Zhu, J. Y., Tang, B. & Hu, Z. C. Three-dimensional digital reconstruction of skin epidermis and dermis. J. Microsc. (2018). doi:10.1111/jmi.12671 152. Lee, S. L. & Lu, Y. H. Modeling of bioheat equation for skin and a preliminary study on a noninvasive diagnostic method for skin burn wounds. Burns (2014). doi:10.1016/j.burns.2013.10.013 153. Prindeze, N. J. et al. Examination of the early diagnostic applicability of active dynamic thermography for burn wound depth assessment and concept analysis. J. Burn Care Res. (2015). doi:10.1097/BCR.0000000000000187 154. Mourits, D. L., Remmers, J. S., Tan, S. H., Moll, A. C. & Hartong, D. T. An Individualized 3-Dimensional Designed and Printed Conformer After Dermis Fat Grafting for Complex Sockets. Ophthal. Plast. Reconstr. Surg. (2018). doi:10.1097/IOP.0000000000001113 155. Lee, K., Kim, M. & Kim, K. 3D skin surface reconstruction from a single image by merging global curvature and local texture using the guided filtering for 3D haptic palpation. Ski. Res. Technol. 24, 672?685 (2018). 156. Hani, A. F. M., Eltegani, N. M., Hussein, S. H., Jamil, A. & Gill, P. in Visual Informatics: Bridging Research and Practice (eds. Zaman, H. et al.) 5857, 243?253 (Springer, 2009). 157. Ng, W. L., Qi, J. T. Z., Yeong, W. Y. & Naing, M. W. Proof-of-concept: 3D bioprinting of pigmented human skin constructs. Biofabrication (2018). doi:10.1088/1758-5090/aa9e1e 158. Choi, W. et al. The fibroblast-derived paracrine factor neuregulin-1 has a novel role in regulating the constitutive color and melanocyte function in human 277 skin. J. Cell Sci. (2010). doi:10.1242/jcs.064774 159. Singh, D., Singh, D. & Han, S. S. 3D printing of scaffold for cells delivery: Advances in skin tissue engineering. Polymers (Basel). 8, 1?17 (2016). 160. Blackstone, B. N. et al. Inflammatory response and biomechanical properties of coaxial scaffolds for engineered skin in vitro and post-grafting. Acta Biomater. 80, 247?257 (2018). 161. Tra Thanh, N. et al. Optimization and characterization of electrospun polycaprolactone coated with gelatin-silver nanoparticles for wound healing application. Mater. Sci. Eng. C 91, 318?329 (2018). 162. Pal, P. et al. Bilayered nanofibrous 3D hierarchy as skin rudiment by emulsion electrospinning for burn wound management. Biomater. Sci. 5, 1786?1799 (2017). 163. Blackwood, K. A. et al. Development of biodegradable electrospun scaffolds for dermal replacement. Biomaterials 29, 3091?3104 (2008). 164. Mo, Y. et al. Controlled Dual Delivery of Angiogenin and Curcumin by Electrospun Nanofibers for Skin Regeneration. Tissue Eng. Part A 23, 597? 608 (2017). 165. Zeugolis, D. I. et al. Electro-spinning of pure collagen nano-fibres ? Just an expensive way to make gelatin? Biomaterials 29, 2293?2305 (2008). 166. Kim, B. S. et al. 3D cell printing of in vitro stabilized skin model and in vivo pre-vascularized skin patch using tissue-specific extracellular matrix bioink: A step towards advanced skin tissue engineering. Biomaterials 168, 38?53 (2018). 167. Sriram, G., Bigliardi, P. L. & Bigliardi-Qi, M. Fibroblast heterogeneity and its implications for engineering organotypic skin models in vitro. European Journal of Cell Biology (2015). doi:10.1016/j.ejcb.2015.08.001 168. Dai, N.-T. et al. Restoration of skin pigmentation after deep partial or full- thickness burn injury. Adv. Drug Deliv. Rev. 123, 155?164 (2018). 169. Klar, A. S., Zimoch, J. & Biedermann, T. Skin Tissue Engineering: Application of Adipose-Derived Stem Cells. Biomed Res. Int. 2017, 1?12 (2017). 170. Dai, N.-T. et al. Development of a Novel Pre-Vascularized Three-Dimensional Skin Substitute Using Blood Plasma Gel. Cell Transplant. 27, 1535?1547 (2018). 171. Panwar, A., Tan, L., Panwar, A. & Tan, L. P. Current Status of Bioinks for Micro-Extrusion-Based 3D Bioprinting. Molecules 21, 685 (2016). 172. Patra, S. & Young, V. A Review of 3D Printing Techniques and the Future in Biofabrication of Bioprinted Tissue. Cell Biochem. Biophys. 74, 93?98 (2016). 173. Moldovan, N. I., Hibino, N. & Nakayama, K. Principles of the Kenzan Method for Robotic Cell Spheroid-Based Three-Dimensional Bioprinting . Tissue Eng. 278 Part B Rev. 23, 237?244 (2017). 174. Xiong, S. et al. A Gelatin-sulfonated Silk Composite Scaffold based on 3D Printing Technology Enhances Skin Regeneration by Stimulating Epidermal Growth and Dermal Neovascularization. Sci. Rep. 7, 1?12 (2017). 175. Kim, H., Hau, N. T., Chae, Y. G., Lee, B. Il & Kang, H. W. 3D printing- assisted fabrication of double-layered optical tissue phantoms for laser tattoo treatments. Lasers Surg. Med. 48, 392?399 (2016). 176. Kim, B. S., Lee, J.-S., Gao, G. & Cho, D.-W. Direct 3D cell-printing of human skin with functional transwell system. Biofabrication 9, 025034 (2017). 177. Koch, L. et al. Skin tissue generation by laser cell printing. Biotechnol. Bioeng. 109, 1855?1863 (2012). 178. Kim, B. S., Gao, G., Kim, J. Y. & Cho, D. W. 3D Cell Printing of Perfusable Vascularized Human Skin Equivalent Composed of Epidermis, Dermis, and Hypodermis for Better Structural Recapitulation of Native Skin. Advanced Healthcare Materials 1801019 (2018). doi:10.1002/adhm.201801019 179. Binder, K. W. et al. In situ bioprinting of the skin for burns. J. Am. Coll. Surg. 211, S76 (2010). 180. Skardal, A. et al. Bioprinted Amniotic Fluid-Derived Stem Cells Accelerate Healing of Large Skin Wounds. Stem Cells Transl. Med. 1, 792?802 (2012). 181. O?Connell, C. D. et al. Development of the Biopen: a handheld device for surgical printing of adipose stem cells at a chondral wound site. Biofabrication 8, 015019 (2016). 182. Ogle, M. E., Segar, C. E., Sridhar, S. & Botchwey, E. A. Monocytes and macrophages in tissue repair: Implications for immunoregenerative biomaterial design. Exp. Biol. Med. 241, 1084?1097 (2016). 183. Sridharan, R., Cameron, A. R., Kelly, D. J., Kearney, C. J. & O?Brien, F. J. Biomaterial based modulation of macrophage polarization: A review and suggested design principles. Materials Today 18, 313?325 (2015). 184. Olingy, C. E. et al. Non-classical monocytes are biased progenitors of wound healing macrophages during soft tissue injury. Sci. Rep. 7, 447 (2017). 185. Rodero, M. P. & Khosrotehrani, K. Skin wound healing modulation by macrophages. Int. J. Clin. Exp. Pathol. 3, 643?53 (2010). 186. Hesketh, M., Sahin, K. B., West, Z. E. & Murray, R. Z. Macrophage phenotypes regulate scar formation and chronic wound healing. International Journal of Molecular Sciences 18, (2017). 187. Rosique, R. G., Rosique, M. J. & Farina Junior, J. A. Curbing Inflammation in Skin Wound Healing: A Review. Int. J. Inflam. 2015, 1?9 (2015). 188. Chun, J. & Hartung, H. P. Mechanism of action of oral fingolimod (FTY720) in multiple sclerosis. Clinical Neuropharmacology 33, 91?101 (2010). 189. Di Dario, M. et al. Myeloid cells as target of fingolimod action in multiple 279 sclerosis. Neurology 2, e157 (2015). 190. Lim, N. S. J., Sham, A., Chee, S. M. L., Chan, C. & Raghunath, M. Combination of ciclopirox olamine and sphingosine-1-phosphate as granulation enhancer in diabetic wounds. Wound Repair Regen. 24, 795?809 (2016). 191. Rezaie Shirmard, L. et al. Nanoparticulate fingolimod delivery system based on biodegradable poly (3-hydroxybutyrate-co-3-hydroxyvalerate) (PHBV): design, optimization, characterization and in-vitro evaluation. Pharm. Dev. Technol. 22, 860?870 (2017). 192. Awojoodu, A. O. et al. Sphingosine 1-phosphate receptor 3 regulates recruitment of anti-inflammatory monocytes to microvessels during implant arteriogenesis. Proc. Natl. Acad. Sci. 110, 13785?13790 (2013). 193. Basnet, P. & Skalko-Basnet, N. Curcumin: An anti-inflammatory molecule from a curry spice on the path to cancer treatment. Molecules 16, 4567?4598 (2011). 194. Tong, W. Y. et al. Antimicrobial wound dressing film utilizing cellulose nanocrystal as drug delivery system for curcumin. Cellulose 25, 631?638 (2018). 195. Berce, C. et al. Cutaneous wound healing using polymeric surgical dressings based on chitosan, sodium hyaluronate and resveratrol. A preclinical experimental study. Colloids Surfaces B Biointerfaces 163, 155?166 (2018). 196. Kasiewicz, L. N. & Whitehead, K. A. Silencing TNF? with lipidoid nanoparticles downregulates both TNF? and MCP-1 in an in vitro co-culture model of diabetic foot ulcers. Acta Biomater. 32, 120?128 (2016). 197. Waters, M., VandeVord, P. & Van Dyke, M. Keratin biomaterials augment anti-inflammatory macrophage phenotype in vitro. Acta Biomater. 66, 213?223 (2018). 198. Sun, G. Pro-Regenerative Hydrogel Restores Scarless Skin during Cutaneous Wound Healing. Advanced Healthcare Materials 6, 1700659 (2017). 199. Hotchkiss, K. M. et al. Titanium surface characteristics, including topography and wettability, alter macrophage activation. Acta Biomater. 31, 425?434 (2016). 200. Kim, J. E., Lee, J. H., Kim, S. H. & Jung, Y. Skin Regeneration with Self- Assembled Peptide Hydrogels Conjugated with Substance P in a Diabetic Rat Model. Tissue Eng. Part A 24, 21?33 (2018). 201. Pourchet, L. J. et al. Human Skin 3D Bioprinting Using Scaffold-Free Approach. Adv. Healthc. Mater. 6, 1?8 (2017). 202. Mosser, D. M. & Edwards, J. P. Exploring the full spectrum of macrophage activation. Nature Reviews Immunology 8, 958?969 (2008). 203. Lech, M. & Anders, H. J. Macrophages and fibrosis: How resident and 280 infiltrating mononuclear phagocytes orchestrate all phases of tissue injury and repair. Biochimica et Biophysica Acta - Molecular Basis of Disease 1832, 989? 997 (2013). 204. McAllister, T. N., Dusserre, N., Maruszewski, M. & L?Heureux, N. Cell-based therapeutics from an economic perspective: primed for a commercial success or a research sinkhole? Regen. Med. 3, 925?937 (2008). 205. Food and Drug Administration. Minimal Manipulation of Human Cells, Tissues, and Cellular and Tissue-Based Products, Draft Guidance for Industry and Food and Drug Administration Staff. Guidance for Industry and Food and Drug Administration Staff (2017). 206. World Health Organization. Second Global Consultation on Regulatory Requirements for Human Cells and Tissues for Transplantation: Towards Global Harmonization through Graduated Standards. World Health Organization 53, (2006). 207. Food and Drug Administration. Same Surgical Procedure Exception under 21 CFR 1271.15(b): Questions and Answers Regarding the Scope of the Exception Guidance for Industry. (2017). 208. FDA/CBER. Guidance for Industry Preclinical Assessment of Investigational Cellular and Gene Therapy Products. FDA (2013). 209. Bailey, A. M., Mendicino, M. & Au, P. An FDA perspective on preclinical development of cell-based regenerative medicine products. Nat. Biotechnol. 32, 721?723 (2014). 210. Food and Drug Administration. Fast Track, Breakthrough Therapy, Accelerated Approval, Priority Review. (2018). 211. Food and Drug Administration. INTERACT Meetings (Initial Targeted Engagement for Regulatory Advice on CBER products). (2018). 212. FDA 5. Expedited Programs for Serious Conditions ? Drugs and Biologics. Draft Guidance for Industry (2014). doi:U.S. Food & Drug Administration 213. Cervelli, V. et al. Treatment of stable vitiligo by ReCell system. Acta Dermatovenerologica Croat. 17, 273?278 (2009). 214. Prendergast, M. E., Solorzano, R. D. & Cabrera, D. Bioinks for biofabrication: current state and future perspectives. J. 3D Print. Med. 3dp?2016?0002 (2016). doi:10.2217/3dp-2016-0002 215. Hull, C. W. Apparatus for production of three-dimensional objects by stereolithography. US Pat. 4,575,330 1?16 (1986). at 216. Jose, R. R., Rodriguez, M. J., Dixon, T. A., Omenetto, F. & Kaplan, D. L. Evolution of Bioinks and Additive Manufacturing Technologies for 3D Bioprinting. ACS Biomater. Sci. Eng. acsbiomaterials.6b00088 (2016). doi:10.1021/acsbiomaterials.6b00088 281 217. Ferris, C. J., Gilmore, K. G., Wallace, G. G. & In Het Panhuis, M. Biofabrication: An overview of the approaches used for printing of living cells. Appl. Microbiol. Biotechnol. 97, 4243?4258 (2013). 218. Sant, S., Hancock, M. J., Donnelly, J. P., Iyer, D. & Khademhosseini, A. Biomimetic gradient hydrogels for tissue engineering. Can. J. Chem. Eng. 88, 899?911 (2010). 219. Fedorovich, N. E. et al. Biofabrication of osteochondral tissue equivalents by printing topologically defined, cell-laden hydrogel scaffolds. Tissue Eng. Part C. Methods 18, 33?44 (2012). 220. Shim, J.-H. et al. Bioprinting of a mechanically enhanced three-dimensional dual cell-laden construct for osteochondral tissue engineering using a multi- head tissue/organ building system. J. Micromechanics Microengineering 22, 085014 (2012). 221. Gurkan, U. A. et al. Engineering Anisotropic Biomimetic Fibrocartilage Microenvironment by Bioprinting Mesenchymal Stem Cells in Nanoliter Gel Droplets. Mol. Pharm. 11, 2151?2159 (2014). 222. Bell, A., Kofron, M. & Nistor, V. Multiphoton crosslinking for biocompatible 3D printing of type I collagen. Biofabrication 7, 035007 (2015). 223. Ferlin, K. M., Prendergast, M. E., Miller, M. L., Kaplan, D. S. & Fisher, J. P. Influence of 3D printed porous architecture on mesenchymal stem cell enrichment and differentiation. Acta Biomater. 32, 161?169 (2016). 224. Miller, J. S. et al. Rapid casting of patterned vascular networks for perfusable engineered three-dimensional tissues. Nat. Mater. 11, 768?774 (2012). 225. Kolesky, D. B., Homan, K. a, Skylar-Scott, M. a & Lewis, J. a. Three- dimensional bioprinting of thick vascularized tissues. Proc. Natl. Acad. Sci. U. S. A. 113, 3179?84 (2016). 226. Lee, Y. B. et al. Bio-printing of collagen and VEGF-releasing fibrin gel scaffolds for neural stem cell culture. Exp. Neurol. 223, 645?652 (2010). 227. Kundu, J., Shim, J.-H., Jang, J., Kim, S.-W. & Cho, D.-W. An additive manufacturing-based PCL-alginate-chondrocyte bioprinted scaffold for cartilage tissue engineering. J. Tissue Eng. Regen. Med. 9, 1286?1297 (2015). 228. Lee, W. et al. Multi-layered culture of human skin fibroblasts and keratinocytes through three-dimensional freeform fabrication. Biomaterials 30, 1587?1595 (2009). 229. Guo, S.-Z., Heuzey, M.-C. & Therriault, D. Properties of Polylactide Inks for Solvent-Cast Printing of Three-Dimensional Freeform Microstructures. Langmuir 30, 1142?1150 (2014). 230. Gaetani, R. et al. Cardiac tissue engineering using tissue printing technology and human cardiac progenitor cells. Biomaterials 33, 1782?1790 (2012). 231. Janmey, P. A. et al. Fibrin gels and their clinical and bioengineering 282 applications. J. R. Soc. Interface 6, 1?10 (2009). 232. Hutmacher, D. W. et al. Mechanical properties and cell cultural response of polycaprolactone scaffolds designed and fabricated via fused deposition modeling. J. Biomed. Mater. Res. 55, 203?16 (2001). 233. Zein, I., Hutmacher, D. W., Tan, K. C. & Teoh, S. H. Fused deposition modeling of novel scaffold architectures for tissue engineering applications. Biomaterials 23, 1169?1185 (2002). 234. Trachtenberg, J. E. et al. Extrusion-Based 3D Printing of Poly(propylene fumarate) in a Full-Factorial Design. ACS Biomater. Sci. Eng. 2, 1771?1780 (2016). 235. Hinton, T. J. et al. Three-dimensional printing of complex biological structures by freeform reversible embedding of suspended hydrogels. Sci. Adv. 1, (2015). 236. Beaman, J. J. & Deckard, C. R. Selective laser sintering with assisted powder handling. (1990). 237. Goodridge, R. D., Wood, D. J., Ohtsuki, C. & Dalgarno, K. W. Biological evaluation of an apatite?mullite glass-ceramic produced via selective laser sintering. Acta Biomater. 3, 221?231 (2007). 238. Hao, L. et al. Characterization of selective laser-sintered hydroxyapatite-based biocomposite structures for bone replacement. Proc. R. Soc. A Math. Phys. Eng. Sci. 463, 1857?1869 (2007). 239. Salmoria, G. V., Hotza, D., Klauss, P., Kanis, L. A. & Roesler, C. R. M. Manufacturing of Porous Polycaprolactone Prepared with Different Particle Sizes and Infrared Laser Sintering Conditions: Microstructure and Mechanical Properties. Adv. Mech. Eng. 6, 640496?640496 (2015). 240. Xia, Y. et al. Selective laser sintering fabrication of nano-hydroxyapatite/poly- ?-caprolactone scaffolds for bone tissue engineering applications. Int. J. Nanomedicine 8, 4197?213 (2013). 241. Kinstlinger, I. S. et al. Open-Source Selective Laser Sintering (OpenSLS) of Nylon and Biocompatible Polycaprolactone. PLoS One 11, e0147399 (2016). 242. Murphy, S. V & Atala, A. 3D bioprinting of tissues and organs. Nat. Biotechnol. 32, 773?785 (2014). 243. Boland. Ink-jet printing of viable cells. Biomaterials 26, 93?99 (2006). 244. Xu, T. et al. Characterization of Cell Constructs Generated With Inkjet Printing Technology Using In Vivo Magnetic Resonance Imaging. J. Manuf. Sci. Eng. 130, 021013 (2008). 245. Xu, T., Kincaid, H., Atala, A. & Yoo, J. J. High-Throughput Production of Single-Cell Microparticles Using an Inkjet Printing Technology. J. Manuf. Sci. Eng. 130, 021017 (2008). 246. Derby, B. et al. Bioprinting: inkjet printing proteins and hybrid cell-containing materials and structures. J. Mater. Chem. 18, 5717 (2008). 283 247. Xu, T., Jin, J., Gregory, C., Hickman, J. J. J. J. & Boland, T. Inkjet printing of viable mammalian cells. Biomaterials 26, 93?9 (2005). 248. Xu, T. et al. Viability and electrophysiology of neural cell structures generated by the inkjet printing method. Biomaterials 27, 3580?8 (2006). 249. Murphy, S. V, Skardal, A. & Atala, A. Evaluation of hydrogels for bio-printing applications. J. Biomed. Mater. Res. A 101, 272?84 (2013). 250. Khalil, S. & Sun, W. Biopolymer deposition for freeform fabrication of hydrogel tissue constructs. Mater. Sci. Eng. C 27, 469?478 (2007). 251. Hennink, W. . & van Nostrum, C. . Novel crosslinking methods to design hydrogels. Adv. Drug Deliv. Rev. 54, 13?36 (2002). 252. Kim, J. D., Choi, J. S., Kim, B. S., Chan Choi, Y. & Cho, Y. W. Piezoelectric inkjet printing of polymers: Stem cell patterning on polymer substrates. Polymer (Guildf). 51, 2147?2154 (2010). 253. Lin, H. et al. Application of visible light-based projection stereolithography for live cell-scaffold fabrication with designed architecture. Biomaterials 34, 331? 9 (2013). 254. Lu, Y. & Chen, S. Projection printing of 3-dimensional tissue scaffolds. Methods Mol. Biol. 868, 289?302 (2012). 255. Nichol, J. W. et al. Cell-laden microengineered gelatin methacrylate hydrogels. Biomaterials 31, 5536?44 (2010). 256. Ma, X. et al. Deterministically patterned biomimetic human iPSC-derived hepatic model via rapid 3D bioprinting. Proc. Natl. Acad. Sci. U. S. A. 113, 2206?2211 (2016). 257. Placone, J. K. et al. Development and Characterization of a 3D Printed, Keratin-Based Hydrogel. Ann. Biomed. Eng. 1?12 (2016). doi:10.1007/s10439- 016-1621-7 258. Koroleva, A. et al. Fabrication of fibrin scaffolds with controlled microscale architecture by a two-photon polymerization?micromolding technique. Biofabrication 4, 015001 (2012). 259. Zhu, W. et al. 3D printing of functional biomaterials for tissue engineering. Curr. Opin. Biotechnol. 40, 103?112 (2016). 260. Miller, J. S. et al. Bioactive hydrogels made from step-growth derived PEG- peptide macromers. Biomaterials 31, 3736?43 (2010). 261. Yue, K. et al. Synthesis, properties, and biomedical applications of gelatin methacryloyl (GelMA) hydrogels. Biomaterials 73, 254?271 (2015). 262. Highley, C. B., Rodell, C. B. & Burdick, J. A. Direct 3D Printing of Shear- Thinning Hydrogels into Self-Healing Hydrogels. Adv. Mater. 27, 5075?5079 (2015). 263. Wei Xu et al. Rapid Prototyping Three-Dimensional Cell/Gelatin/Fibrinogen Constructs for Medical Regeneration. J. Bioact. Compat. Polym. 22, 363?377 284 (2007). 264. Cui, X. & Boland, T. Human microvasculature fabrication using thermal inkjet printing technology. Biomaterials 30, 6221?7 (2009). 265. Xu, T. et al. Hybrid printing of mechanically and biologically improved constructs for cartilage tissue engineering applications. Biofabrication 5, 015001 (2013). 266. Smith, C. M. et al. Three-dimensional bioassembly tool for generating viable tissue-engineered constructs. Tissue Eng. 10, 1566?76 (2004). 267. Rhee, S., Puetzer, J. L., Mason, B. N., Reinhart-King, C. A. & Bonassar, L. J. 3D Bioprinting of Spatially Heterogeneous Collagen Constructs for Cartilage Tissue Engineering. ACS Biomater. Sci. Eng. 2, 1800?1805 (2016). 268. Moon, S. et al. Layer by Layer Three-dimensional Tissue Epitaxy by Cell- Laden Hydrogel Droplets. Tissue Eng. Part C Methods 16, 157?166 (2010). 269. Tamayol, A. et al. Hydrogel Templates for Rapid Manufacturing of Bioactive Fibers and 3D Constructs. Adv. Healthc. Mater. 4, 2146?2153 (2015). 270. Duan, B., Hockaday, L. A., Kang, K. H. & Butcher, J. T. 3D bioprinting of heterogeneous aortic valve conduits with alginate/gelatin hydrogels. J. Biomed. Mater. Res. A 101, 1255?64 (2013). 271. Williams, S. K., Touroo, J. S., Church, K. H. & Hoying, J. B. Encapsulation of Adipose Stromal Vascular Fraction Cells in Alginate Hydrogel Spheroids Using a Direct-Write Three-Dimensional Printing System. doi:10.1089/biores.2013.0046 272. Merceron, T. K. et al. A 3D bioprinted complex structure for engineering the muscle?tendon unit. Biofabrication 7, 035003 (2015). 273. Christensen, K. et al. Freeform inkjet printing of cellular structures with bifurcations. Biotechnol. Bioeng. 112, 1047?55 (2015). 274. Gaetani, R. et al. Epicardial application of cardiac progenitor cells in a 3D- printed gelatin/hyaluronic acid patch preserves cardiac function after myocardial infarction. Biomaterials 61, 339?348 (2015). 275. Skardal, A. et al. Bioprinting Cellularized Constructs Using a Tissue-specific Hydrogel Bioink. J. Vis. Exp. 1?8 (2016). doi:10.3791/53606 276. Wang, X. et al. Generation of Three-Dimensional Hepatocyte/Gelatin Structures with Rapid Prototyping System. Tissue Eng. 12, 83?90 (2006). 277. Bhise, N. S. et al. A liver-on-a-chip platform with bioprinted hepatic spheroids. Biofabrication 8, 014101 (2016). 278. Lee, V. K. et al. Creating perfused functional vascular channels using 3D bio- printing technology. Biomaterials 35, 8092?8102 (2014). 279. Bertassoni, L. E. et al. Hydrogel bioprinted microchannel networks for vascularization of tissue engineering constructs. Lab Chip 14, 2202?11 (2014). 280. Ang, T. . et al. Fabrication of 3D chitosan?hydroxyapatite scaffolds using a 285 robotic dispensing system. Mater. Sci. Eng. C 20, 35?42 (2002). 281. Kuo, C.-Y. et al. Development of a 3D Printed, Bioengineered Placenta Model to Evaluate the Role of Trophoblast Migration in Preeclampsia. ACS Biomater. Sci. Eng. 2, 1817?1826 (2016). 282. Melchels, F. P. W. et al. Development and characterisation of a new bioink for additive tissue manufacturing. J. Mater. Chem. B 2, 2282 (2014). 283. Bertassoni, L. E. et al. Direct-write bioprinting of cell-laden methacrylated gelatin hydrogels. Biofabrication 6, 24105?11 (2014). 284. Soman, P., Chung, P. H., Zhang, A. P. & Chen, S. Digital microfabrication of user-defined 3D microstructures in cell-laden hydrogels. Biotechnol. Bioeng. 110, 3038?3047 (2013). 285. Skardal, A. et al. Photocrosslinkable Hyaluronan-Gelatin Hydrogels for Two- Step Bioprinting. Tissue Eng. Part A 16, 2675?2685 (2010). 286. Suri, S. et al. Solid freeform fabrication of designer scaffolds of hyaluronic acid for nerve tissue engineering. Biomed. Microdevices 13, 983?93 (2011). 287. Hung, B. P. et al. Three-Dimensional Printing of Bone Extracellular Matrix for Craniofacial Regeneration. ACS Biomater. Sci. Eng. 2, 1806?1816 (2016). 288. Reichert, J. C. et al. A Tissue Engineering Solution for Segmental Defect Regeneration in Load-Bearing Long Bones. Sci. Transl. Med. 4, (2012). 289. Ali Akbari Ghavimi, S., Ebrahimzadeh, M. H., Solati-Hashjin, M. & Abu Osman, N. A. Polycaprolactone/starch composite: Fabrication, structure, properties, and applications. J. Biomed. Mater. Res. Part A 103, 2482?2498 (2015). 290. Jakus, A. E. et al. Hyperelastic ?bone?: A highly versatile, growth factor?free, osteoregenerative, scalable, and surgically friendly biomaterial. Sci. Transl. Med. 8, (2016). 291. Mironov, A. V. et al. 3D printing of PLGA scaffolds for tissue engineering. J. Biomed. Mater. Res. Part A (2016). doi:10.1002/jbm.a.35871 292. Hockaday, L. A. et al. Rapid 3D printing of anatomically accurate and mechanically heterogeneous aortic valve hydrogel scaffolds. Biofabrication 4, 035005 (2012). 293. Rutz, A. L., Hyland, K. E., Jakus, A. E., Burghardt, W. R. & Shah, R. N. A multimaterial bioink method for 3D printing tunable, cell-compatible hydrogels. Adv. Mater. 27, 1607?1614 (2015). 294. Arcaute, K., Mann, B. K. & Wicker, R. B. Stereolithography of three- dimensional bioactive poly(ethylene glycol) constructs with encapsulated cells. Ann. Biomed. Eng. 34, 1429?41 (2006). 295. Warner, J., Soman, P., Zhu, W., Tom, M. & Chen, S. Design and 3D Printing of Hydrogel Scaffolds with Fractal Geometries. ACS Biomater. Sci. Eng. 2, 1763?1770 (2016). 286 296. Han, L.-H., Mapili, G., Chen, S. & Roy, K. Projection Microfabrication of Three-Dimensional Scaffolds for Tissue Engineering. J. Manuf. Sci. Eng. 130, 021005 (2008). 297. Cui, X., Breitenkamp, K., Finn, M. G., Lotz, M. & D?Lima, D. D. Direct Human Cartilage Repair Using Three-Dimensional Bioprinting Technology. Tissue Eng. Part A 18, 1304?1312 (2012). 298. Homan, K. A. et al. Bioprinting of 3D Convoluted Renal Proximal Tubules on Perfusable Chips. Sci. Rep. 6, 34845 (2016). 299. Lan, P. X., Lee, J. W., Seol, Y.-J. & Cho, D.-W. Development of 3D PPF/DEF scaffolds using micro-stereolithography and surface modification. J. Mater. Sci. Mater. Med. 20, 271?9 (2009). 300. Melchiorri, A. J. et al. 3D-Printed Biodegradable Polymeric Vascular Grafts. Adv. Healthc. Mater. 5, 319?325 (2016). 301. Bhattacharjee, T. et al. Writing in the granular gel medium. (2015). doi:10.1126/sciadv.1500655 302. Wu, C., Luo, Y., Cuniberti, G., Xiao, Y. & Gelinsky, M. Three-dimensional printing of hierarchical and tough mesoporous bioactive glass scaffolds with a controllable pore architecture, excellent mechanical strength and mineralization ability. Acta Biomater. 7, 2644?2650 (2011). 303. K?pf, M., Campos, D. F. D., Blaeser, A., Sen, K. S. & Fischer, H. A tailored three-dimensionally printable agarose-collagen blend allows encapsulation, spreading, and attachment of human umbilical artery smooth muscle cells. Biofabrication 8, 025011 (2016). 304. Kang, D. et al. Endothelial monolayers on collagen-coated nanofibrous membranes: cell-cell and cell-ECM interactions. Biofabrication 8, 025008 (2016). 305. Kim, M. & Kim, G. H. Electrohydrodynamic direct printing of PCL/collagen fibrous scaffolds with a core/shell structure for tissue engineering applications. Chem. Eng. J. 279, 317?326 (2015). 306. Skardal, A. & Atala, A. Biomaterials for Integration with 3-D Bioprinting. Ann. Biomed. Eng. 43, 730?746 (2015). 307. Wang, X., Yan, Y. & Zhang, R. Rapid prototyping as a tool for manufacturing bioartificial livers. Trends Biotechnol. 25, 505?513 (2007). 308. Carrow, J. K., Kerativitayanan, P., Jaiswal, M. K., Lokhande, G. & Gaharwar, A. K. Polymers for bioprinting. Essentials 3D Biofabrication Transl. 229?248 (2015). doi:10.1016/B978-0-12-800972-7.00013-X 309. Guvendiren, M., Molde, J., Soares, R. M. D. & Kohn, J. Designing Biomaterials for 3D Printing. ACS Biomater. Sci. Eng. 2, 1679?1693 (2016). 310. Miranda-Nieves, D. & Chaikof, E. L. Collagen and Elastin Biomaterials for the Fabrication of Engineered Living Tissues. ACS Biomater. Sci. Eng. 287 acsbiomaterials.6b00250 (2016). doi:10.1021/acsbiomaterials.6b00250 311. Lee, A. et al. Featuring work from the labs of As featured in: Engineering anastomosis between living capillary networks and endothelial cell-lined microfluidic channels. 16, (2014). 312. Pati, F. et al. Printing three-dimensional tissue analogues with decellularized extracellular matrix bioink. Nat. Commun. 5, 3935 (2014). 313. Chang, R., Nam, J. & Sun, W. Direct cell writing of 3D microorgan for in vitro pharmacokinetic model. Tissue Eng. Part C. Methods 14, 157?66 (2008). 314. Narayanan, L. K. et al. 3D-Bioprinting of Polylactic Acid (PLA) Nanofiber? Alginate Hydrogel Bioink Containing Human Adipose-Derived Stem Cells. ACS Biomater. Sci. Eng. acsbiomaterials.6b00196 (2016). doi:10.1021/acsbiomaterials.6b00196 315. Kolesky, D. B. et al. 3D Bioprinting of Vascularized, Heterogeneous Cell- Laden Tissue Constructs. Adv. Mater. 26, 3124?3130 (2014). 316. Poldervaart, M. T. et al. Sustained Release of BMP-2 in Bioprinted Alginate for Osteogenicity in Mice and Rats. PLoS One 8, e72610 (2013). 317. Bellan, L. M. et al. Fabrication of an artificial 3-dimensional vascular network using sacrificial sugar structures. Soft Matter 5, 1354 (2009). 318. Antonov, E. N. et al. Fabrication of polymer scaffolds for tissue engineering using surface selective laser sintering. Laser Phys. 16, 774?787 (2006). 319. Williams, J. M. et al. Bone tissue engineering using polycaprolactone scaffolds fabricated via selective laser sintering. Biomaterials 26, 4817?4827 (2005). 320. Saggiomo, V. & Velders, A. H. Simple 3D Printed Scaffold-Removal Method for the Fabrication of Intricate Microfluidic Devices. Adv. Sci. 2, n/a-n/a (2015). 321. Hinton, T. J., Hudson, A., Pusch, K., Lee, A. & Feinberg, A. W. 3D Printing PDMS Elastomer in a Hydrophilic Support Bath via Freeform Reversible Embedding. ACS Biomater. Sci. Eng. acsbiomaterials.6b00170 (2016). doi:10.1021/acsbiomaterials.6b00170 322. Bhattacharjee, T. et al. Liquid-like Solids Support Cells in 3D. ACS Biomater. Sci. Eng. 2, 1787?1795 (2016). 323. Albritton, J. L. et al. Ultrahigh-throughput generation and characterization of cellular aggregates in laser-ablated microwells of poly(dimethylsiloxane). RSC Adv. 6, 8980?8991 (2016). 324. Roudsari, L. C. et al. A 3D Poly(ethylene glycol)-based Tumor Angiogenesis Model to Study the Influence of Vascular Cells on Lung Tumor Cell Behavior. Sci. Rep. 6, 32726 (2016). 325. Kang, H.-W. et al. A 3D bioprinting system to produce human-scale tissue constructs with structural integrity. Nat. Biotechnol. 34, 312?319 (2016). 326. Sooppan, R. et al. In Vivo Anastomosis and Perfusion of a Three- 288 Dimensionally-Printed Construct Containing Microchannel Networks. Tissue Eng. Part C. Methods 22, 1?7 (2016). 327. Zhang, B. et al. Biodegradable scaffold with built-in vasculature for organ-on- a-chip engineering and direct surgical anastomosis. Nat. Mater. advance on, (2016). 328. Huh, D. et al. Reconstituting Organ-Level Lung Functions on a Chip. Science (80-. ). 328, 1662?1668 (2010). 329. Huh, D. et al. A Human Disease Model of Drug Toxicity-Induced Pulmonary Edema in a Lung-on-a-Chip Microdevice. Sci. Transl. Med. 4, 159ra147- 159ra147 (2012). 330. Zietek, T., Rath, E., Haller, D. & Daniel, H. Intestinal organoids for assessing nutrient transport, sensing and incretin secretion. Sci. Rep. 5, 16831 (2015). 331. Fatehullah, A., Tan, S. H. & Barker, N. Organoids as an in vitro model of human development and disease. Nat. Cell Biol. 18, 246?54 (2016). 332. Devriese, S. et al. T84 monolayers are superior to Caco-2 as a model system of colonocytes. Histochem. Cell Biol. 148, 85?93 (2017). 333. Sakolish, C. M., Esch, M. B., Hickman, J. J., Shuler, M. L. & Mahler, G. J. Modeling Barrier Tissues In Vitro: Methods, Achievements, and Challenges. EBioMedicine 5, 30?39 (2016). 334. Duval, K. et al. Modeling Physiological Events in 2D vs. 3D Cell Culture. Physiology 32, 266?277 (2017). 335. Little, M. H. Organoids: a Special Issue. Development 144, 935?937 (2017). 336. Huch, M., Knoblich, J. A., Lutolf, M. P. & Martinez-Arias, A. The hope and the hype of organoid research. Development 144, 938?941 (2017). 337. Braakhuis, H. M. et al. Progress and future of in vitro models to study translocation of nanoparticles. Archives of Toxicology 89, 1469?1495 (2015). 338. Bannasch, H., Momeni, A., Knam, F., Stark, G. B. & F?hn, M. Tissue engineering of skin substitutes. Panminerva Med. 47, 53?60 (2005). 339. Mortensen, L. J. et al. Quantification of quantum dot murine skin penetration with UVR barrier impairment. Nanotoxicology 7, 1386?1398 (2013). 340. Huong, S. P., Bun, H., Fourneron, J. D., Reynier, J. P. & Andrieu, V. Use of various models for in vitro percutaneous absorption studies of ultraviolet filters. Ski. Res. Technol. 15, 253?261 (2009). 341. Abd, E. et al. Skin models for the testing of transdermal drugs. Clin. Pharmacol. Adv. Appl. 8, 163?176 (2016). 342. Bond, J. R. & Barry, B. W. Limitations of hairless mouse skin as a model for in vitro permeation studies through human skin: hydration damage. J. Invest. Dermatol. 90, 486?9 (1988). 343. Rissmann, R., Oudshoorn, M. H. M., Hennink, W. E., Ponec, M. & Bouwstra, J. a. Skin barrier disruption by acetone: Observations in a hairless mouse skin 289 model. Arch. Dermatol. Res. 301, 609?613 (2009). 344. Scheuplein, R. J. Mechanism of percutaneous absorption. II. Transient diffusion and the relative importance of various routes of skin penetration. J. Invest. Dermatol. 48, 79?88 (1967). 345. Teodorescu, F. et al. Transdermal skin patch based on reduced graphene oxide: A new approach for photothermal triggered permeation of ondansetron across porcine skin. J. Control. Release 245, 137?146 (2017). 346. Van Bocxlaer, K., Yardley, V., Murdan, S. & Croft, S. L. Drug permeation and barrier damage in Leishmania-infected mouse skin. J. Antimicrob. Chemother. 71, 1578?1585 (2016). TM 347. Mun, G. C. et al. Further development of the EpiDerm 3D reconstructed human skin micronucleus (RSMN) assay. Mutat. Res. Toxicol. Environ. Mutagen. 673, 92?99 (2009). 348. Garland, M. J. et al. Influence of skin model on in vitro performance of drug- loaded soluble microneedle arrays. Int. J. Pharm. 434, 80?89 (2012). 349. Oshima, S. et al. The Use of an Artificial Skin Model to Study Transdermal Absorption of Drugs in Inflamed Skin. Biol. Pharm. Bull. 35, 203?209 (2012). 350. Flamand, N. et al. Development of genotoxicity test procedures with Episkin?, a reconstructed human skin model: Towards new tools for in vitro risk assessment of dermally applied compounds? Mutat. Res. - Genet. Toxicol. Environ. Mutagen. 606, 39?51 (2006). 351. Gr?goire, S. et al. Prediction of chemical absorption into and through the skin from cosmetic and dermatological formulations. Br. J. Dermatol. 160, 80?91 (2009). 352. Carreras, N., Alonso, C., Mart?, M. & Lis, M. J. Mass transport model through the skin by microencapsulation system. J. Microencapsul. 32, 358?363 (2015). 353. Davies, D. J., Heylings, J. R., Gayes, H., McCarthy, T. J. & Mack, M. C. Further development of an in vitro model for studying the penetration of chemicals through compromised skin. Toxicol. Vitr. 38, 101?107 (2017). 354. Ali, S. & Rytting, E. Influences of nanomaterials on the barrier function of epithelial cells. Adv. Exp. Med. Biol. 811, 45?54 (2014). 355. Ananta, M., Brown, R. A. & Mudera, V. A Rapid Fabricated Living Dermal Equivalent for Skin Tissue Engineering: An In Vivo Evaluation in an Acute Wound Model. Tissue Eng. Part A 18, 353?361 (2012). 356. Guo, Z. et al. Building a microphysiological skin model from induced pluripotent stem cells. Stem Cell Res Ther 4 Suppl 1, S2 (2013). 357. Reijnders, C. M. A. et al. Development of a Full-Thickness Human Skin Equivalent In Vitro Model Derived from TERT-Immortalized Keratinocytes and Fibroblasts. Tissue Eng. Part A 21, 2448?2459 (2015). 358. Kim, B. S., Lee, J. S., Gao, G. & Cho, D. W. Direct 3D cell-printing of human 290 skin with functional transwell system. Biofabrication 9, (2017). 359. Koch, L. et al. Skin tissue generation by laser cell printing. Biotechnol. Bioeng. 109, 1855?1863 (2012). 360. Ata?, B. et al. Skin and hair on-a-chip: in vitro skin models versus ex vivo tissue maintenance with dynamic perfusion. Lab Chip 13, 3555 (2013). 361. Lee, S. et al. Construction of 3D multicellular microfluidic chip for an in vitro skin model. Biomed. Microdevices 19, 22 (2017). 362. van den Broek, L. J., Bergers, L. I. J. C., Reijnders, C. M. A. & Gibbs, S. Progress and Future Prospectives in Skin-on-Chip Development with Emphasis on the use of Different Cell Types and Technical Challenges. Stem Cell Rev. Reports 418?429 (2017). doi:10.1007/s12015-017-9737-1 363. Li, Y. et al. Evaluation of the Effect of the Structure of Bacterial Cellulose on Full Thickness Skin Wound Repair on a Microfluidic Chip. Biomacromolecules 16, 780?789 (2015). 364. Mori, N., Morimoto, Y. & Takeuchi, S. Skin integrated with perfusable vascular channels on a chip. Biomaterials 116, 48?56 (2017). 365. Wufuer, M. et al. Skin-on-a-chip model simulating inflammation, edema and drug-based treatment. Sci. Rep. 6, 1?12 (2016). 366. Abaci, H. E., Guo, Z., Doucet, Y., Jack?w, J. & Christiano, A. Next generation human skin constructs as advanced tools for drug development. Experimental Biology and Medicine 242, 1657?1668 (2017). 367. Podola, B., Li, T. & Melkonian, M. Porous Substrate Bioreactors: A Paradigm Shift in Microalgal Biotechnology? Trends Biotechnol. 35, 121?132 (2017). 368. Scheper, T. Bioreactor Systems for Tissue Engineering. 112, (Springer Berlin Heidelberg, 2009). 369. Kim, J., Kennedy, K. & Vunjak-Novakovic, G. in Principles of Regenerative Medicine 787?803 (Elsevier, 2019). doi:10.1016/B978-0-12-809880-6.00045- X 370. Stoffel, M. et al. Towards bioreactor development with physiological motion control and its applications. Med. Eng. Phys. 39, 106?112 (2017). 371. Toume, S., Gefen, A. & Weihs, D. Printable low-cost, sustained and dynamic cell stretching apparatus. J. Biomech. 49, 1336?1339 (2016). 372. D?marteau, O. et al. Dynamic compression of cartilage constructs engineered from expanded human articular chondrocytes. Biochem. Biophys. Res. Commun. 310, 580?588 (2003). 373. Nguyen, B.-N. B., Ko, H. & Fisher, J. P. Tunable osteogenic differentiation of hMPCs in tubular perfusion system bioreactor. Biotechnol. Bioeng. 113, 1805? 1813 (2016). 374. Niu, M., Hammond, P. & Coger, R. N. The Effectiveness of a Novel Cartridge- Based Bioreactor Design in Supporting Liver Cells. Tissue Eng. Part A (2009). 291 doi:10.1089/ten.tea.2008.0279 375. Usuludin, S. B. M., Cao, X. & Lim, M. Co-culture of stromal and erythroleukemia cells in a perfused hollow fiber bioreactor system as an in vitro bone marrow model for myeloid leukemia. Biotechnol. Bioeng. (2012). doi:10.1002/bit.24400 376. Lin, H., Lozito, T. P., Alexander, P. G., Gottardi, R. & Tuan, R. S. Stem Cell- Based Microphysiological Osteochondral System to Model Tissue Response to Interleukin-1?. Mol. Pharm. 11, 2203?2212 (2014). 377. Nakagawa, Y. et al. Two differential flows in a bioreactor promoted platelet generation from human pluripotent stem cell-derived megakaryocytes. Exp. Hematol. 41, 742?748 (2013). 378. Spitters, T. W. G. M. et al. A Dual Flow Bioreactor with Controlled Mechanical Stimulation for Cartilage Tissue Engineering. Tissue Eng. Part C Methods 19, 774?783 (2013). 379. Chang, C. H., Lin, F. H., Lin, C. C., Chou, C. H. & Liu, H. C. Cartilage tissue engineering on the surface of a novel gelatin-calcium- phosphate biphasic scaffold in a double-chamber bioreactor. J. Biomed. Mater. Res. - Part B Appl. Biomater. 71, 313?321 (2004). 380. Giusti, S. et al. A novel dual-flow bioreactor simulates increased fluorescein permeability in epithelial tissue barriers. Biotechnol. J. 9, 1175?1184 (2014). 381. Mori, N., Morimoto, Y. & Takeuchi, S. Skin integrated with perfusable vascular channels on a chip. Biomaterials 116, 48?56 (2017). 382. Melchiorri, A. J., Bracaglia, L. G., Kimerer, L. K., Hibino, N. & Fisher, J. P. In Vitro Endothelialization of Biodegradable Vascular Grafts Via Endothelial Progenitor Cell Seeding and Maturation in a Tubular Perfusion System Bioreactor. Tissue Eng. Part C Methods (2016). doi:10.1089/ten.tec.2015.0562 383. Skardal, A. et al. Bioprinted Amniotic Fluid-Derived Stem Cells Accelerate Healing of Large Skin Wounds. Stem Cells Transl. Med. 1, 792?802 (2012). 384. Ladd, M. R., Lee, S. J., Atala, A. & Yoo, J. J. Bioreactor maintained living skin matrix. Tissue Eng. Part A 15, 861?8 (2009). 385. Jeong, C. et al. Applicability and Safety of in Vitro Skin Expansion Using a Skin Bioreactor: A Clinical Trial. Arch. Plast. Surg. 41, 661 (2014). 386. Tokuyama, E., Nagai, Y., Takahashi, K., Kimata, Y. & Naruse, K. Mechanical stretch on human skin equivalents increases the epidermal thickness and develops the basement membrane. PLoS One (2015). doi:10.1371/journal.pone.0141989 387. Wendt, D., Riboldi, S. A., Cioffi, M. & Martin, I. Bioreactors in tissue engineering: Scientific challenges and clinical perspectives. Advances in Biochemical Engineering/Biotechnology (2009). doi:10.1007/978-3-540- 69357-4_1 292 388. Purdue, G. F. et al. A multicenter clinical trial of a biosynthetic skin replacement, dermagraft-tc, compared with cryopreserved human cadaver skin for temporary coverage of excised burn wounds. J. Burn Care Rehabil. (1997). doi:10.1097/00004630-199701000-00009 389. Halberstadt, C. R. et al. The in vitro growth of a three? dimensional human dermal replacement using a single? pass perfusion system. Biotechnol. Bioeng. (1994). doi:10.1002/bit.260430808 390. Nicodemus, G. D. & Bryant, S. J. Cell Encapsulation in Biodegradable Hydrogels for Tissue Engineering Applications. Tissue Eng. Part B Rev. 14, 149?165 (2008). 391. Young, S., Wong, M., Tabata, Y. & Mikos, A. G. Gelatin as a delivery vehicle for the controlled release of bioactive molecules. in Journal of Controlled Release (2005). doi:10.1016/j.jconrel.2005.09.023 392. Boateng, J. S., Matthews, K. H., Stevens, H. N. & Eccleston, G. M. Wound Healing Dressings and Drug Delivery Systems: A Review. J. Pharm. Sci. 97, 2892?2923 (2008). 393. Malafaya, P. B., Silva, G. A. & Reis, R. L. Natural-origin polymers as carriers and scaffolds for biomolecules and cell delivery in tissue engineering applications. Advanced Drug Delivery Reviews (2007). doi:10.1016/j.addr.2007.03.012 394. Morais, J. M., Papadimitrakopoulos, F. & Burgess, D. J. Biomaterials/Tissue Interactions: Possible Solutions to Overcome Foreign Body Response. AAPS J. 12, 188?196 (2010). 395. Fuchs, E. Keratins and the Skin. Annu. Rev. Cell Dev. Biol. 11, 123?154 (1995). 396. Popescu, C. & H?cker, H. Hair - The most sophisticated biological composite material. Chem. Soc. Rev. (2007). doi:10.1039/b604537p 397. Ham, T. R. et al. Tunable keratin hydrogels for controlled erosion and growth factor delivery. Biomacromolecules 17, 225?236 (2016). 398. Layman, H. et al. Enhanced angiogenic efficacy through controlled and sustained delivery of FGF-2 and G-CSF from fibrin hydrogels containing ionic-albumin microspheres. J. Biomater. Sci. Polym. Ed. (2012). doi:10.1163/092050610X546417 399. Sierpinski, P. et al. The use of keratin biomaterials derived from human hair for the promotion of rapid regeneration of peripheral nerves. Biomaterials 29, 118?128 (2008). 400. Kirfel, J., Magin, T. M. & Reichelt, J. Keratins?: a structural scaffold with emerging functions. Cell. Mol. Life Sci. 60, 56?71 (2003). 401. Steinert, P. M. & Gullino, M. I. Bovine epidermal keratin filament assembly in vitro. Biochem. Biophys. Res. Commun. (1976). doi:10.1016/0006- 293 291X(76)91131-1 402. Thomas, H., Conrads, A., Phan, K. H., van de L?cht, M. & Zahn, H. In vitro reconstitution of wool intermediate filaments. Int. J. Biol. Macromol. (1986). doi:10.1016/0141-8130(86)90038-3 403. Ikkai, F. & Naito, S. Dynamic light scattering and circular dichroism studies on heat-induced gelation of hard-keratin protein aqueous solutions. Biomacromolecules (2002). doi:10.1021/bm010160i 404. de Guzman, R. C. et al. Mechanical and biological properties of keratose biomaterials. Biomaterials 32, 8205?8217 (2011). 405. Hill, P., Brantley, H. & Van Dyke, M. Some properties of keratin biomaterials: Kerateines. Biomaterials 31, 585?593 (2010). 406. Richter, J. R., De Guzman, R. C., Greengauz-Roberts, O. K. & Van Dyke, M. Structure-property relationships of meta-kerateine biomaterials derived from human hair. Acta Biomater. 8, 274?281 (2012). 407. Marshall, R. C., Orwin, D. F. G. & Gillespie, J. M. Structure and biochemistry of mammalian hard keratin. Electron Microscopy Reviews (1991). doi:10.1016/0892-0354(91)90016-6 408. Buchanan, J. H. A cystine-rich protein fraction from oxidized ?-keratin. Biochem. J. 167, 489?491 (1977). 409. Hill, P. S. et al. Repair of Peripheral Nerve Defects in Rabbits Using Keratin Hydrogel Scaffolds. Tissue Eng. Part A 17, 1499?1505 (2011). 410. Thilagar, S., Arul Jothi, N., Sheikh Omar, A. R., Kamaruddin, M. Y. & Ganabadi, S. Effect of keratin-gelatin and bFGF-gelatin composite film as a sandwich layer for full-thickness skin mesh graft in experimental dogs. J. Biomed. Mater. Res. - Part B Appl. Biomater. (2009). doi:10.1002/jbm.b.31024 411. Wang, Y., Zhang, W., Yuan, J. & Shen, J. Differences in cytocompatibility between collagen, gelatin and keratin. Mater. Sci. Eng. C (2016). doi:10.1016/j.msec.2015.09.093 412. Rouse, J. G. & Van Dyke, M. E. A review of keratin-based biomaterials for biomedical applications. Materials (Basel). 3, 999?1014 (2010). 413. Humphries, J. D. Integrin ligands at a glance. J. Cell Sci. 119, 3901?3903 (2006). 414. Singh, R. et al. Evaluation of hydrogel matrices for vessel bioplotting: Vascular cell growth and viability. J. Biomed. Mater. Res. - Part A (2016). doi:10.1002/jbm.a.35590 415. Elvin, C. M. et al. A highly elastic tissue sealant based on photopolymerised gelatin. Biomaterials 31, 8323?8331 (2010). 416. Sando, L. et al. Photochemical crosslinking of soluble wool keratins produces a mechanically stable biomaterial that supports cell adhesion and proliferation. J. Biomed. Mater. Res. Part A 95A, 901?911 (2010). 294 417. Balaji, S. et al. Characterization of keratin-collagen 3D scaffold for biomedical applications. Polym. Adv. Technol. 23, 500?507 (2012). 418. Tomblyn, S. et al. Keratin hydrogel carrier system for simultaneous delivery of exogenous growth factors and muscle progenitor cells. J. Biomed. Mater. Res. - Part B Appl. Biomater. 104, 864?879 (2016). 419. Wang, S. et al. Culturing fibroblasts in 3D human hair keratin hydrogels. ACS Appl. Mater. Interfaces 7, 5187?5198 (2015). 420. Fancy, D. A. & Kodadek, T. Chemistry for the analysis of protein-protein interactions: Rapid and efficient cross-linking triggered by long wavelength light. Proc. Natl. Acad. Sci. 96, 6020?6024 (1999). 421. O?Brien, F. J. Biomaterials & scaffolds for tissue engineering. Mater. Today 14, 88?95 (2011). 422. Placone, J. K. et al. Development and Characterization of a 3D Printed, Keratin-Based Hydrogel. Ann. Biomed. Eng. 45, 237?248 (2017). 423. de Guzman, R. C. et al. Bone regeneration with BMP-2 delivered from keratose scaffolds. Biomaterials 34, 1644?1656 (2013). 424. Pace, L. A., Plate, J. F., Smith, T. L. & Van Dyke, M. E. The effect of human hair keratin hydrogel on early cellular response tosciatic nerve injury in a rat model. Biomaterials 34, 5907?5914 (2013). 425. Pace, L. A. et al. A Human Hair Keratin Hydrogel Scaffold Enhances Median Nerve Regeneration in Nonhuman Primates: An Electrophysiological and Histological Study. Tissue Eng. Part A 20, 131115063659000 (2013). 426. Poranki, D. R. & Van Dyke, M. E. The effect of gamma keratose on cell viability invitro after thermal stress and the regulation of cell death pathway- specific gene expression. Biomaterials 35, 4646?4655 (2014). 427. Dias, G. J. et al. Keratin-hydroxyapatite composites: Biocompatibility, osseointegration, and physical properties in an ovine model. J. Biomed. Mater. Res. - Part A 95, 1084?1095 (2010). 428. Kowalczewski, C. J. et al. Reduction of ectopic bone growth in critically-sized rat mandible defects by delivery of rhBMP-2 from kerateine biomaterials. Biomaterials 35, 3220?3228 (2014). 429. Tachibana, A., Furuta, Y., Takeshima, H., Tanabe, T. & Yamauchi, K. Fabrication of wool keratin sponge scaffolds for long-term cell cultivation. J. Biotechnol. 93, 165?170 (2002). 430. Verma, V., Verma, P., Ray, P. & Ray, A. R. Preparation of scaffolds from human hair proteins for tissue-engineering applications. Biomed. Mater. 3, (2008). 431. Lin, C.-W. et al. Evaluation of adhesion, proliferation, and differentiation of human adipose-derived stem cells on keratin. J. Polym. Res. 25, 40 (2018). 432. Han, S., Ham, T. R., Haque, S., Sparks, J. L. & Saul, J. M. Alkylation of 295 human hair keratin for tunable hydrogel erosion and drug delivery in tissue engineering applications. Acta Biomater. 23, 201?213 (2015). 433. Richter, J. R., de Guzman, R. C. & Van Dyke, M. E. Mechanisms of hepatocyte attachment to keratin biomaterials. Biomaterials 32, 7555?7561 (2011). 434. Baker, H. B. et al. Cell and Growth Factor-Loaded Keratin Hydrogels for Treatment of Volumetric Muscle Loss in a Mouse Model. Tissue Eng. Part A 23, 572?584 (2017). 435. Passipieri, J. A. et al. Keratin Hydrogel Enhances In Vivo Skeletal Muscle Function in a Rat Model of Volumetric Muscle Loss. Tissue Eng. Part A 23, 556?571 (2017). 436. Saul, J. M., Ellenburg, M. D., De Guzman, R. C. & Dyke, M. Van. Keratin hydrogels support the sustained release of bioactive ciprofloxacin. J. Biomed. Mater. Res. - Part A 98 A, 544?553 (2011). 437. Pines, M. & Nagler, A. Halofuginone: A Novel Antifibrotic Therapy. Gen. Pharmacol. Vasc. Syst. 30, 445?450 (1998). 438. Luo, Y., Xie, X., Luo, D., Wang, Y. & Gao, Y. The role of halofuginone in fibrosis: more to be explored? J. Leukoc. Biol. 102, 1333?1345 (2017). 439. Pines, M., Snyder, D., Yarkoni, S. & Nagler, A. Halofuginone to treat fibrosis in chronic graft-versus-host disease and scleroderma. Biol. Blood Marrow Transplant. 9, 417?425 (2003). 440. Leiba, M. Halofuginone inhibits NF-kB and p38 MAPK in activated T cells. J. Leukoc. Biol. 80, 399?406 (2006). 441. Navarro, J., Swayambunathan, J., Lerman, M., Santoro, M. & Fisher, J. P. Development of keratin-based membranes for potential use in skin repair. Acta Biomater. 83, 177?188 (2019). 442. Navarro, J., Swayambunathan, J., Santoro, M. & Fisher, J. Assessment of the Effects of Energy Density in Crosslinking of Keratin-Based Photo-Sensitive Resin. in 2018 IX International Seminar of Biomedical Engineering (SIB) 1?6 (IEEE, 2018). doi:10.1109/SIB.2018.8467744 443. Kato, Y., Uchida, K. & Kawakishi, S. Aggregation of collagen exposed to UVA in the presence of riboflavin: A plausible role of tyrosine modification. Photochem. Photobiol. 59, 343?349 (1994). 444. Singer, A. J., Thode Jr, H. C. & McClain, S. A. Development of a Histomorphologic Scale to Quantify Cutaneous Scars after Burns. Acad. Emerg. Med. 7, 1083?1088 (2000). 445. Wallace, J. et al. Validating continuous digital light processing (cDLP) additive manufacturing accuracy and tissue engineering utility of a dye- initiator package. Biofabrication 6, (2014). 446. Stansbury, J. W. & Idacavage, M. J. 3D printing with polymers: Challenges 296 among expanding options and opportunities. Dent. Mater. 32, 54?64 (2016). 447. Tyan, Y.-C., Liao, J.-D., Lin, S.-P. & Chen, C.-C. The study of the sterilization effect of gamma ray irradiation of immobilized collagen polypropylene nonwoven fabric surfaces. J. Biomed. Mater. Res. 67A, 1033?1043 (2003). 448. Escudero-Castellanos, A., Ocampo-Garc?a, B. E., Dom?nguez-Garc?a, M. V., Flores-Estrada, J. & Flores-Merino, M. V. Hydrogels based on poly(ethylene glycol) as scaffolds for tissue engineering application: biocompatibility assessment and effect of the sterilization process. J. Mater. Sci. Mater. Med. 27, 176 (2016). 449. Kanjickal, D., Lopina, S., Evancho-Chapman, M. M., Schmidt, S. & Donovan, D. Effects of sterilization on poly(ethylene glycol) hydrogels. J. Biomed. Mater. Res. Part A 87A, 608?617 (2008). 450. Hodder, E. et al. Investigating the effect of sterilisation methods on the physical properties and cytocompatibility of methyl cellulose used in combination with alginate for 3D-bioplotting of chondrocytes. J. Mater. Sci. Mater. Med. 30, 10 (2019). 451. Metcalfe, A. D. & Ferguson, M. W. J. Bioengineering skin using mechanisms of regeneration and repair. Biomaterials 28, 5100?5113 (2007). 452. Zhong, S. P., Zhang, Y. Z. & Lim, C. T. Tissue scaffolds for skin wound healing and dermal reconstruction. Wiley Interdiscip. Rev. Nanomedicine Nanobiotechnology 2, 510?525 (2010). 453. Powell, H. M. & Boyce, S. T. Wound closure with EDC cross-linked cultured skin substitutes grafted to athymic mice. Biomaterials 28, 1084?1092 (2007). 454. POWELL, H., SUPP, D. & BOYCE, S. Influence of electrospun collagen on wound contraction of engineered skin substitutes. Biomaterials 29, 834?843 (2008). 455. Ono, I., Tateshita, T. & Inoue, M. Effects of a collagen matrix containing basic fibroblast growth factor on wound contraction. J. Biomed. Mater. Res. 48, 621?630 (1999). 456. Morris, D. E. et al. Acute and chronic animal models for excessive dermal scarring: Quantitative studies. Plast. Reconstr. Surg. (1997). doi:10.1097/00006534-199709000-00021 457. Gallant, C. L., Olson, M. E. & Hart, D. A. Molecular, histologic, and gross phenotype of skin wound healing in red Duroc pigs reveals an abnormal healing phenotype of hypercontracted, hyperpigmented scarring. Wound Repair Regen. (2004). doi:10.1111/j.1067-1927.2004.012311.x 458. Harunari, N. et al. Histology of the thick scar on the female, red Duroc pig: Final similarities to human hypertrophic scar. Burns (2006). doi:10.1016/j.burns.2006.03.015 459. Jansen, J. a, de Ruijter, J. E., Janssen, P. T. & Paquay, Y. G. Histological 297 evaluation of a biodegradable Polyactive/hydroxyapatite membrane. Biomaterials 16, 819?27 (1995). 460. Nieminen, T. et al. In vivo and in vitro degradation of a novel bioactive guided tissue regeneration membrane. Int. J. Oral Maxillofac. Surg. 35, 727?732 (2006). 461. Retzepi, M. & Donos, N. Guided Bone Regeneration: Biological principle and therapeutic applications. Clin. Oral Implants Res. 21, 567?576 (2010). 462. Thoma, D. S. et al. Evaluation of a new biodegradable membrane to prevent gingival ingrowth into mandibular bone defects in minipigs. Clin. Oral Implants Res. 20, 7?16 (2009). 463. Yang, F., Both, S. K., Yang, X., Walboomers, X. F. & Jansen, J. A. Development of an electrospun nano-apatite/PCL composite membrane for GTR/GBR application. Acta Biomater. 5, 3295?3304 (2009). 464. Cai, Y. et al. Silk fibroin membrane used for guided bone tissue regeneration. Mater. Sci. Eng. C 70, 148?154 (2017). 465. Gillespie, J. M. Proteins rich in glycine and tyrosine from keratins. Comp. Biochem. Physiol. -- Part B Biochem. 41, (1972). 466. Verma, V., Verma, P., Ray, P. & Ray, A. R. Preparation of scaffolds from human hair proteins for tissue-engineering applications. Biomed. Mater. 3, (2008). 467. Nayak, K. K. & Gupta, P. In vitro biocompatibility study of keratin/agar scaffold for tissue engineering. Int. J. Biol. Macromol. 81, 1?10 (2015). 468. Nakata, R., Tachibana, A. & Tanabe, T. Preparation of keratin hydrogel/hydroxyapatite composite and its evaluation as a controlled drug release carrier. Mater. Sci. Eng. C 41, 59?64 (2014). 469. Menges, F. Spectragryph - Software for optical spectroscopy. (2017). at 470. Hirschl, C. et al. Determining the degree of crosslinking of ethylene vinyl acetate photovoltaic module encapsulants - A comparative study. Sol. Energy Mater. Sol. Cells 116, 203?218 (2013). 471. Lazzara, M. J. & Deen, W. M. Effects of concentration on the partitioning of macromolecule mixtures in agarose gels. J. Colloid Interface Sci. 272, 288? 297 (2004). 472. Cussler, E. L. Diffusion, Mass Transfer in Fluid Systems. (Cambridge University Press, 2009). doi:10.1017/CBO9780511805134.010 473. Wexler, S. A. et al. Adult bone marrow is a rich source of human mesenchymal ?stem? cells but umbilical cord and mobilized adult blood are not. Br. J. Haematol. 121, 368?374 (2003). 474. Yang, Z., Schmitt, J. F. & Lee, E. H. in Methods 698, 353?366 (2011). 475. Hunter, K. T. & Ma, T. In vitro evaluation of hydroxyapatite-chitosan-gelatin 298 composite membrane in guided tissue regeneration. J. Biomed. Mater. Res. - Part A 101 A, 1016?1025 (2013). 476. Hurt, A. P., Getti, G. & Coleman, N. J. Bioactivity and biocompatibility of a chitosan-tobermorite composite membrane for guided tissue regeneration. Int. J. Biol. Macromol. 64, 11?16 (2014). 477. Bottino, M. C., Thomas, V. & Janowski, G. M. A novel spatially designed and functionally graded electrospun membrane for periodontal regeneration. Acta Biomater. 7, 216?224 (2011). 478. G?rb?z, S. et al. Multi-layered functional membranes for periodontal regeneration: Preparation and characterization. Mater. Lett. 178, 256?259 (2016). 479. Jamuna-Thevi, K., Saarani, N. N., Abdul Kadir, M. R. & Hermawan, H. Triple-layered PLGA/nanoapatite/lauric acid graded composite membrane for periodontal guided bone regeneration. Mater. Sci. Eng. C 43, 253?263 (2014). 480. Xue, J. et al. Preparation and in vivo efficient anti-infection property of GTR/GBR implant made by metronidazole loaded electrospun polycaprolactone nanofiber membrane. Int. J. Pharm. 475, 566?577 (2014). 481. Xue, J. et al. Fabrication of drug-loaded anti-infective guided tissue regeneration membrane with adjustable biodegradation property. Colloids Surfaces B Biointerfaces 135, 846?854 (2015). 482. Li, H. et al. A quaternary composite fiber membrane for guided tissue regeneration. Polym. Adv. Technol. 27, 178?184 (2016). 483. Vashi, A. V., Werkmeister, J. A., Vuocolo, T., Elvin, C. M. & Ramshaw, J. A. M. Stabilization of collagen tissues by photocrosslinking. J. Biomed. Mater. Res. Part A 100 A, n/a-n/a (2012). 484. Adekogbe, I. & Ghanem, A. Fabrication and characterization of DTBP- crosslinked chitosan scaffolds for skin tissue engineering. Biomaterials 26, 7241?7250 (2005). 485. Zhang, X., Jia, C., Qiao, X., Liu, T. & Sun, K. Silk fibroin microfibers and chitosan modified poly (glycerol sebacate) composite scaffolds for skin tissue engineering. Polym. Test. 62, 88?95 (2017). 486. Malencik, D. A., Sprouse, J. F., Swanson, C. A. & Anderson, S. R. Dityrosine: Preparation, isolation, and analysis. Anal. Biochem. 242, 202?213 (1996). 487. Dijkgraaf, L. C. et al. Crosslinking of fibrinogen and fibronectin by free radicals: A possible initial step in adhesion formation in osteoarthritis of the temporomandibular joint. J. Oral Maxillofac. Surg. 61, 101?111 (2003). 488. Scott McCall, A. et al. Mechanisms of corneal tissue cross-linking in response to treatment with topical riboflavin and Long-Wavelength Ultraviolet Radiation (UVA). Investig. Ophthalmol. Vis. Sci. 51, 129?138 (2010). 489. Lattuada, M. et al. Kinetics of free-radical cross-linking polymerization: 299 Comparative experimental and numerical study. Macromolecules 46, 5831? 5841 (2013). 490. Fuxman, A. M., McAuley, K. B. & Schreiner, L. J. Modeling of Free-radical Crosslinking Copolymerization of Acrylamide and N,N?- Methylenebis(acrylamide) for Radiation Dosimetry. Macromol. Theory Simulations 12, 647?662 (2003). 491. Bennett, J. Measuring UV curing parameters of commercial photopolymers used in additive manufacturing. Addit. Manuf. 18, 203?212 (2017). 492. Steyrer, B., Neubauer, P., Liska, R. & Stampfl, J. Visible light photoinitiator for 3D-printing of tough methacrylate resins. Materials (Basel). 10, 1?11 (2017). 493. Seo, H., Heo, S. G., Lee, H. & Yoon, H. Preparation of PEG materials for constructing complex structures by stereolithographic 3D printing. RSC Adv. 7, 28684?28688 (2017). 494. Sigma. Fluorescein Isothiocyanate-Dextran Product Information. 1?3 (1997). 495. Tdb Consultancy. FITC-Dextran Fluorescein isothiocyanate dextran. 1?5 (2010). at 496. Arumugasaamy, N., Navarro, J., Kent Leach, J., Kim, P. C. W. & Fisher, J. P. In Vitro Models for Studying Transport Across Epithelial Tissue Barriers. Ann. Biomed. Eng. (2018). doi:10.1007/s10439-018-02124-w 497. Shook, B. et al. The Role of Adipocytes in Tissue Regeneration and Stem Cell Niches. Annu. Rev. Cell Dev. Biol. 32, 609?631 (2016). 498. Yeatts, A. B. & Fisher, J. P. Tubular Perfusion System for the Long-Term Dynamic Culture of Human Mesenchymal Stem Cells. Tissue Eng. Part C Methods 17, 337?348 (2011). 499. Yeatts, A. B., Gordon, C. N. & Fisher, J. P. Formation of an Aggregated Alginate Construct in a Tubular Perfusion System. Tissue Eng. Part C Methods 17, 1171?1178 (2011). 500. Gleghorn, J. P., Lee, C. S. D., Cabodi, M., Stroock, A. D. & Bonassar, L. J. Adhesive properties of laminated alginate gels for tissue engineering of layered structures. J. Biomed. Mater. Res. Part A 85A, 611?618 (2008). 501. Harley, B. A. et al. Design of a multiphase osteochondral scaffold III: Fabrication of layered scaffolds with continuous interfaces. J. Biomed. Mater. Res. Part A 9999A, NA-NA (2009). 502. Kim, B. S. et al. Human collagen-based multilayer scaffolds for tendon-to- bone interface tissue engineering. J. Biomed. Mater. Res. - Part A 102, 4044? 4054 (2014). 503. Kim, M. & Kim, G. H. 3D multi-layered fibrous cellulose structure using an electrohydrodynamic process for tissue engineering. J. Colloid Interface Sci. 457, 180?187 (2015). 300 504. Levingstone, T. J. et al. Cell-free multi-layered collagen-based scaffolds demonstrate layer specific regeneration of functional osteochondral tissue in caprine joints. Biomaterials 87, 69?81 (2016). 505. Yang, G., Lin, H., Rothrauff, B. B., Yu, S. & Tuan, R. S. Multilayered polycaprolactone/gelatin fiber-hydrogel composite for tendon tissue engineering. Acta Biomater. 35, 68?76 (2016). 506. Pi, Q. et al. Digitally Tunable Microfluidic Bioprinting of Multilayered Cannular Tissues. Adv. Mater. 30, 1?10 (2018). 507. Spalazzi, J. P., Doty, S. B., Moffat, K. L., Levine, W. N. & Lu, H. H. Development of Controlled Matrix Heterogeneity on a Triphasic Scaffold for Orthopedic Interface Tissue Engineering. Tissue Eng. 12, 061120052454001 (2006). 508. Mosher, C. Z., Spalazzi, J. P. & Lu, H. H. Stratified scaffold design for engineering composite tissues. Methods 84, 99?102 (2015). 509. Costa, P. F. et al. Advanced tissue engineering scaffold design for regeneration of the complex hierarchical periodontal structure. J. Clin. Periodontol. 41, 283?294 (2014). 510. Camarero-Espinosa, S., Rothen-Rutishauser, B., Weder, C. & Foster, E. J. Directed cell growth in multi-zonal scaffolds for cartilage tissue engineering. Biomaterials 74, 42?52 (2016). 511. Hadjipanayi, E., Brown, R. A. & Mudera, V. Interface integration of layered collagen scaffolds with defined matrix stiffness: implications for sheet-based tissue engineering. J. Tissue Eng. Regen. Med. 3, 230?241 (2009). 512. Papenburg, B. J. et al. Development and analysis of multi-layer scaffolds for tissue engineering. Biomaterials 30, 6228?6239 (2009). 513. Asano, Y., Shimoda, H., Okano, D., Matsusaki, M. & Akashi, M. Transplantation of three-dimensional artificial human vascular tissues fabricated using an extracellular matrix nanofilm-based cell-accumulation technique. J. Tissue Eng. Regen. Med. 11, 1303?1307 (2017). 514. Matsusaki, M. et al. Development of full-thickness human skin equivalents with blood and lymph-like capillary networks by cell coating technology. J. Biomed. Mater. Res. Part A 103, 3386?3396 (2015). 515. Matsuzawa, A., Matsusaki, M. & Akashi, M. Construction of three- dimensional liver tissue models by cell accumulation technique and maintaining their metabolic functions for long-term culture without medium change. J. Biomed. Mater. Res. - Part A 103, 1554?1564 (2015). 516. Nishiguchi, A., Yoshida, H., Matsusaki, M. & Akashi, M. Rapid construction of three-dimensional multilayered tissues with endothelial tube networks by the cell-accumulation technique. Adv. Mater. 23, 3506?3510 (2011). 517. Sobachkin, A. & Dumnov, G. Numerical Basis of CAD-Embedded CFD. 301 NAFEMS World Congr. 2013 1?20 (2014). doi:10.1007/s10590-014-9157-9 518. Placone, J. K. et al. Development and Characterization of a 3D Printed, Keratin-Based Hydrogel. Ann. Biomed. Eng. 45, (2017). 519. Lembong, J., Lerman, M. J., Kingsbury, T. J., Civin, C. I. & Fisher, J. P. A Fluidic Culture Platform for Spatially Patterned Cell Growth, Differentiation, and Cocultures. Tissue Eng. Part A 24, 1715?1732 (2018). 520. Lozito, T. P. et al. Three-dimensional osteochondral microtissue to model pathogenesis of osteoarthritis. Stem Cell Res. Ther. 4, S6 (2013). 521. Dimitrijevich, D., Dodd, C. W., Anderson, W. J. & Schwarz, R. P. Two Chamber Cell Culture Vessel. 15 (2003). 522. Yourek, G., McCormick, S. M., Mao, J. J. & Reilly, G. C. Shear stress induces osteogenic differentiation of human mesenchymal stem cells. Regen. Med. 5, 713?724 (2010). 523. Hung, B. P., Babalola, O. M. & Bonassar, L. J. Quantitative characterization of mesenchymal stem cell adhesion to the articular cartilage surface. J. Biomed. Mater. Res. - Part A 101, 3592?3598 (2013). 524. Dong, J. De et al. Response of mesenchymal stem cells to shear stress in tissue-engineered vascular grafts. Acta Pharmacol. Sin. 30, 530?536 (2009). 525. Lu, H. H., Subramony, S. D., Boushell, M. K. & Zhang, X. Tissue engineering strategies for the regeneration of orthopedic interfaces. Ann. Biomed. Eng. 38, 2142?2154 (2010). 526. Di Luca, A., Van Blitterswijk, C. & Moroni, L. The osteochondral interface as a gradient tissue: From development to the fabrication of gradient scaffolds for regenerative medicine. Birth Defects Res. Part C - Embryo Today Rev. 105, 34?52 (2015). 527. Wang, F. et al. Collagen/chitosan based two-compartment and bi-functional dermal scaffolds for skin regeneration. Mater. Sci. Eng. C 52, 155?162 (2015). 528. Berry, J. L. et al. Three-Dimensional Prevascularized, Engineered Tissue Constructs, Methods of Making and Methods of using the Tissue Constructs. 29 (2015). 529. Donofrio, L. & Weinkle, S. The third dimension in facial rejuvenation: A review. J. Cosmet. Dermatol. 5, 277?283 (2006). 530. Kim, Y. S. et al. The Application of Three-Dimensional Printing in Animal Model of Augmentation Rhinoplasty. Ann. Biomed. Eng. 43, 2153?2162 (2015). 531. Nyberg, E. L. et al. 3D-Printing Technologies for Craniofacial Rehabilitation, Reconstruction, and Regeneration. Ann. Biomed. Eng. 45, 45?57 (2017). 532. Matsumoto, D. et al. Cell-Assisted Lipotransfer?: Supportive Use of Human Adipose-Derived Cells for Soft Tissue Augmentation. 12, (2006). 533. Dasiou-Plakida, D. Fat injections for facial rejuvenation: 17 years experience 302 in 1720 patients. J. Cosmet. Dermatol. 2, 119?125 (2003). 534. Alhadlaq, A., Tang, M. & Mao, J. J. Engineered Adipose Tissue from Human Mesenchymal Stem Cells Maintains Predefined Shape and Dimension?: Implications in Soft Tissue Augmentation and Reconstruction. Tissue Eng. 11, 556?566 (2005). 535. Martella, E. et al. Secreted adiponectin as a marker to evaluate in vitro the adipogenic differentiation of human mesenchymal stromal cells. Cytotherapy 16, 1476?1485 (2014). 536. Pf?tzner, A. et al. Mesenchymal stem cell differentiation into adipocytes is equally induced by insulin and proinsulin in vitro. Int. J. Stem Cells 10, 154? 159 (2017). 537. Gerhardt, C. C., Romero, I. A., Cancello, R., Camoin, L. & Strosberg, A. D. Chemokines control fat accumulation and leptin secretion by cultured human adipocytes. Mol. Cell. Endocrinol. 175, 81?92 (2001). 303 304