Bioprinted Amniotic Fluid-Derived Stem Cells Accelerate Healing of Large Skin Wounds

Hydrogel Preparation

To generate solutions that when printed formed robust gels, the components of fibrin were combined with collagen. Fibrinogen (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) was dissolved in phosphate-buffered saline (PBS) to make a 50 mg/ml solution. Thrombin (Sigma-Aldrich) was dissolved in PBS to make a 20 IU/ml solution. Type I rat tail collagen (BD Biosciences, Bedford, MA, http://www.bdbiosciences.com) was diluted in ice-cold PBS to give a 2.2 mg/ml solution, after which the pH was adjusted to 7.0 using 1 N NaOH. All solutions were then sterile filtered with a 0.4-μm syringe filter. To form gels, the fibrinogen and collagen solutions were mixed together in a 1:1 ratio by volume. That mixed solution was then in turn mixed with the thrombin solution, after which the gel began to cross-link within 30 seconds.

Cell Culture and Cell Labeling

Multipotent stem cells were isolated from human amniotic fluid as previously described [14]. To achieve a relatively homogenous subpopulation, the cells were expanded in culture from a single clone. AFS cells can proliferate rapidly in culture without feeder cells for many passages, while maintaining chromosomal stability.

AFS cells (H1 clone, passage 16) and bone marrow-derived MSCs (passage 5; Lonza, Hopkinton, MA, http://www.lonza.com) were expanded in tissue culture plastic using 150 mm diameter dishes (37°C, 5% CO₂) until 75% confluence with Chang's media (α-minimal essential medium [α-MEM] with 18% Chang B [Irvine Scientific, Santa Ana, CA, http://www.irvinesci.com], 15% embryonic stem cells screened-fetal bovine serum [ES-FBS] [HyClone, Logan, UT, http://www.hyclone.com], 2% Chang C [Irvine Scientific]) and α-MEM with 10% ES-FBS, respectively. Cells were detached from the substrate with Accutase (Innovative Cell Technologies, San Diego, CA, http://www.innovativecelltech.com) and counted prior to centrifugation. Cells were resuspended in the fibrinogen/collagen solution at a density of 16.6 million cells per milliliter.

For visualization and tracking of the printed cells, AFS cells and MSCs were genetically labeled with green fluorescent protein (GFP), using a lentivirus vector carrying the GFP gene (Lenti-GFP), with a standard procedure. Labeling efficiency (70%–80%) was verified by fluorescence-activated cell sorting. The cells were then expanded as needed to reach sufficient numbers for bioprinting.

Bioprinting

In preparation for printing, AFS cells and MSCs were separately suspended in the fibrinogen/collagen solution at a density of 16.6 million cells per milliliter and placed on ice. While under anesthesia, a single full-thickness skin wound (2.0 × 2.0 cm) was surgically created with scissors on the mid-dorsal region of nu/nu mice. A bioprinting device developed in-house [16] was used for cell deposition using fibrin/collagen gels. The device consists of a carriage with three-axis movement capability in which is housed the main print head. The print head is made up of a set of pressure-driven nozzles through which the hydrogel solutions, with or without cells, are printed, and an additional set of high-pressure nozzles through which secondary solutions, such as cross-linking solutions, can be printed. The printable hydrogel solutions are housed in swappable cartridges in line with the back-pressure and the print nozzles. The printing process is then controlled using software also developed in-house.

For treatment, the bioprinter was used to deposit two layers of a fibrin-collagen gel by depositing a layer of thrombin, a layer of fibrinogen/collagen, a second layer of thrombin, a second layer of fibrinogen/collagen, and a final layer of thrombin. Equal amounts of the two solutions were used (0.3 ml), resulting in 5 × 10⁶ AFS cells or MSCs deposited per wound. Additionally, gel-only treatments, containing no cells, were administered. Triple antibiotic (bacitracin zinc, neomycin sulfate, polymyxin-B sulfate; Medique Products, Fort Myers, FL, http://www.mediqueproducts.com) was applied over each wound after gelation, followed by a Tegaderm bandage (3M, St. Paul, MN, http://www.3m.com). Finally, a custom-made bandage was sutured in place in order prevent the Tegaderm from being removed. Wound size was documented by taking photographs immediately after surgery and again at 7 and 14 days (n = 5). Animals were euthanized at 7 and 14 days, and the regenerated skin was harvested for histological analysis. Additionally, several mice were euthanized at 1 and 4 days for cell tracking purposes and to evaluate the consistency and distribution of bioprinted cells by confocal microscopy.

Gross Histology: Wound Closure, Contraction, and Re-Epithelialization

Wound closure, contraction, and re-epithelialization percentages were calculated using photographs and histological examination of the wounds, taken at the time of surgery, day 7, and day 14. Using ImageJ software, the original wound area was defined as A, the clearly re-epithelialized (but thinner) skin area was defined as B, and the remaining unclosed wound was defined as C. Percentage of wound closure was defined as C/A × 100%, percentage of contraction of the wound was defined as (A − B)/A × 100%, and percentage of re-epithelialization was defined as (B − C)/B × 100%.

Histology

Harvested skin tissues were first rolled around a syringe needle prior to being fixed overnight in 4% paraformaldehyde. Samples were then washed in PBS three times for 30 minutes per wash, after which the samples were transferred to 30% sucrose for an overnight incubation at 4°C. The rolled tissues were then sliced in half and flash frozen in Tissue-Tek OCT Compound (Sakura Finetek, Torrance, CA, http://www.sakura.com) blocks in liquid nitrogen. A cryotome (Leica, Heerbrugg, Switzerland, http://www.leica.com) was used to generate 6-μm sections made up of the entire cross-sections of the regenerating wounds. These slides were stored at −20°C until histological procedures were performed.

Sections were stained with hematoxylin and eosin for histology, and slides were imaged under light microscopy. Attention was paid to the presence of blood vessels in the regenerated tissue and re-epithelialization of keratinocytes across the surface of the wound area.

Immunohistochemistry

Immunohistochemical (IHC) staining with smooth muscle actin (SMA) was used to visualize smooth muscle cells (SMCs) in mature blood vessels in the regenerating skin, or with von Willebrand factor (vWF) to visualize endothelial cells in newly formed capillaries and lining the mature blood vessels. For IHC, all incubations were carried out at room temperature unless otherwise stated. Slides were warmed at 60°C for 1 hour to increase bonding to the slides. Antigen retrieval was performed on all slides and achieved with incubation in proteinase K (Dako, Carpinteria, CA, http://www.dako.com) for 5 minutes.

For SMA staining, sections were permeabilized with 0.05% Triton X-100 in PBS (PBST) for 5 minutes. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide solution in methanol for 30 minutes. Nonspecific antibody binding was minimized by incubating sections for 10 minutes in Protein Block Solution (Abcam, Cambridge, U.K., http://www.abcam.com). Sections were incubated for 90 minutes in a humidified chamber with primary anti-α-SMA antibodies (catalog no. ab5694; Abcam) at a 1:200 dilution. Following primary incubation, slides were washed three times in PBS for 5 minutes. Sections were then incubated for 30 minutes with biotinylated secondary antibody solution at a 1:400 dilution. After washing, sections were incubated with horseradish peroxidase-streptavidin RTU (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) for 30 minutes. Visualization of immunoreactivity was achieved by incubating sections in the 3,3′-diaminobenzidine peroxidase substrate kit (Vector Laboratories) for 1–2 minutes. The sections were washed in PBS, counterstained with hematoxylin, dehydrated, and coverslipped. Native skin samples were present as positive controls and were used for comparison. Negative controls were set up at the same time as the primary antibody incubations and included incubation with PBS in place of the primary antibody. No immunoreactivity was observed in these negative control sections.

For vWF staining, immunofluorescent staining was used. Nonspecific antibody binding was blocked in 10% normal donkey serum (NDS) with 50 μg/ml anti-mouse IgG Fab fragments (Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com) for 60 minutes. Sections were incubated for 90 minutes in a humidified chamber with the primary anti-vWF antibodies (catalog no. AB7356; Millipore, Billerica, MA, http://www.millipore.com) at a 1:200 dilution in 3% NDS. Following primary incubation, slides were washed three times in PBS for 5 minutes. Sections were then incubated for 60 minutes with DyLight 594-conjugated AffiniPure donkey anti-rabbit IgG secondary antibodies in a 1:200 dilution in 3% NDS. The sections were washed in PBS, counterstained with 4′,6-diamidino-2-phenylindole (DAPI), and coverslipped with Prolong Gold Antifade (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Positive and negative controls were performed in the same manner as described above.

Quantification of neovascularization was done by determining the microvessel density (MVD) and average vessel diameter (AVD) in hematoxylin and eosin (H&E)-stained wound cross-sections. To do this, ImageJ software was used to perform measurements on histological images. For MVD, first the cross-sectional area of tissue in the image was determined. Within that area the number of vessels was counted. MVD values were calculated by dividing the area by the number of vessels, after which the average MVD per group was determined. For AVD, vessels were selected at random from histological images taken of each animal. To calculate each individual vessel diameter, the perimeter of the vessel was traced and ImageJ was used to calculate the cross-sectional vessel area. The diameter, D, was then derived by the equations Area = πr² and D = 2r, after which the AVD was determined.

Fluorescence Imaging of Cells

To investigate whether the deposited cells remain in the regenerating skin long-term after the bioprinting, GFP-transfected AFS cells and MSCs cells were used. Animals were euthanized on days 1, 4, 7, and 14 after cell bioprinting, and skin samples were harvested and prepared for histology as described above. Samples were then washed three times in PBST, counterstained with DAPI, and washed three times before mounting with Prolong Gold Antifade reagent (Invitrogen). Sections were imaged using a fluorescence microscope, and representative images were recorded.

Cell-Conditioned Media

AFS cells and MSCs were cultured as described above in 150 mm diameter dishes until 75% confluence. At this point, the medium was aspirated and the cells were washed with PBS to remove any remaining serum. Next, 10 ml of serum-free α-MEM was added to each dish. The cells were returned to the incubator for 3 additional days, after which the conditioned medium was collected. The conditioned medium samples were concentrated by lyophilization, and analyzed for growth factor and cytokine content using a Quantibody Human Growth Factor Array (RayBiotech, Norcross, GA, http://www.raybiotech.com).

In Vitro AFS Cell-Induced Transwell Migration Assays of Human Umbilical Vein Endothelial Cells

To investigate the effect of AFS cells to induce migration of endothelial cells for angiogenesis, Transwell (Corning Enterprises, Corning, NY, http://www.corning.com) migration assays were performed using human umbilical vein endothelial cells (HUVECs) as the migrating populations. H1 AFS cells were expanded as described above, before being plated into collagen-coated soft substrates (2 kPa stiffness, Excellness Biotech, Lausanne, Switzerland, http://www.excelleness.com) or plastic 24-well plates at a density of 100,000 cells per well. The aforementioned soft substrates were implemented to provide in vitro substrates that might better mimic the mechanical properties experienced by the AFS cells in the in vivo portion of this study. HUVECs were expanded on tissue culture plastic using 150 mm diameter dishes coated with fibronectin (37°C, 5% CO₂) until 75% confluence with endothelial basal medium containing 2% FBS (Lonza). After 24 hours of H1 AFS culture in the wells, HUVECs were detached from the substrate with Accutase, counted prior to centrifugation, and seeded into Transwell inserts that had been coated in fibronectin at a density of 50,000 cells per insert. The inserts were then placed in the previously prepared H1-seeded soft 2-kPa and plastic wells, as well as cell-free control wells with Chang's media.

After 4 hours, inserts were fixed with 4% paraformaldehyde for 15 minutes. Cells still attached to the interior side of the inserts (nonmigratory) were removed with cotton swab applicators. Cells attached to the outside of the insert membranes were stained by placing each insert in 0.05% crystal violet solution for 20 minutes. Inserts were then washed two times in distilled water to remove excess dye. The insert membranes were then cut out of the inserts and mounted on glass slides for photography and analysis of migrated cells.

Bioprinting of Cells onto Skin Wounds

The bioprinting procedure, depicted in Figure 1A, consistently yielded a fibrin/collagen gel that provided 100% coverage over the wound area and formed a tight seal with the skin at the edges of the excisional wound (not shown). To assess the distribution of cells, delivered to the wound by the bioprinting procedure, confocal microscopy of gel/tissue at day 1 was used to visualize the fluorescently labeled AFS cells. Figure 1B shows a uniform distribution of cells in a representative image spanning an approximately 900 × 900 μm area of the harvested tissue. Figure 1C shows a cross-sectional view (z axis) of the same sample, in which the vertical distribution of cells is consistent across the sample.

Wound Closure

Wound closure was determined using photographs taken at week 0 (the time of surgery), week 1, and week 2 (Fig. 2A). ImageJ software was used to quantify wound closure (Fig. 2B). Because of variations in the starting area of individual wounds, all measurements were converted to percentages of the corresponding wound sizes from the wound size at time of surgery. At week 1, AFS-treated mice showed an average of 42% unclosed wounds and MSC-treated mice showed an average of 44% unclosed wounds. These values were not significantly different from each other but were significantly lower than those of mice treated with gel only, which had an average of 77% unclosed wounds (p < .01). At week 2, AFS-treated mice showed an average of 3% unclosed wounds and the MSC-treated mice showed an average of 2% unclosed wounds. These values were significantly lower than those of mice treated with gel only, which had an average of 13% unclosed wounds (p < .05). At all time points, wound closure in the AFS cell- and MSC-treated groups was not significantly different.

Contraction and Keratinocyte Re-Epithelialization

Wound contraction and re-epithelialization was determined by comparing the gross histology of the wounds at the time of surgery with that of 1 and 2 weeks after the surgery (Fig. 3A, 3B). At week 1, AFS cell- and MSC-treated wounds had contracted 44% and 43%, respectively, which was significantly higher than the 17% contraction observed in gel-only-treated wounds (p < .01). At week 2, only the MSC-treated wounds were significantly more contracted than gel-only-treated wounds (p < .05), at 83% compared with 73%. AFS cell- and MSC-treated groups had greater re-epithelialization at both time points. At week 1, AFS cell- and MSC-treated wounds had 25% and 21% re-epithelialization, respectively, compared with 7% re-epithelialization in gel-only-treated wounds (p < .01). At week 2, AFS cell- and MSC-treated wounds had 84% and 89% re-epithelialization, respectively, compared with 51% re-epithelialization in gel-only-treated wounds (p < .01 and p < .05, respectively).

Additionally, H&E-stained histological sections of skin samples, harvested at week 2, showed clear differences in the quality of the epidermal layers near the center of the wound areas in each group. In gel-only-treated wounds, poorly defined epidermal layering was observed (Fig. 3C). In fact, the structural integrity of the regenerated skin was poor, resulting in crumbling during cryosectioning. In contrast, MSC-treated wounds (Fig. 3D) and AFS cell-treated wounds (Fig. 3E) showed well-defined, organized epidermal layers of the regenerated skin.

Histological examination of the wound site at multiple times after deposition of human AFS cells and MSCs showed no indications of immune cell activity. Although the nu/nu (NU/J) mice used in this study lack mature T cells and therefore suffer from a lack of cell-mediated immunity and show only a partial defect in B-cell development (probably due to the absence of functional T-cells), their innate immunity is intact, with normal circulating levels of natural killer cells, dendritic cells, monocytes, and macrophages, as well as normal complement activity. This result may be due to the immunomodulatory activity of both AFS cells and MSCs but could be a result of the immune-compromised mouse model used. As such, further examination of these cells in an immune-competent animal is necessary.

Neovascularization of Regenerating Skin

H&E-stained histological sections of regenerating skin tissues illustrate the presence of blood vessels (Fig. 4A–4F). At week 1, gel-only-treated wounds remained relatively thin, with blood vessels primarily only visible in the subcutaneous adipose tissue (Fig. 4A). In contrast, MSC-treated (Fig. 4B) and AFS cell-treated (Fig. 4C) wounds showed noticeably thicker regenerating tissue. Furthermore, a greater number of blood vessels was observed in the regenerating skin in these groups, especially in the AFS cell-treated one. At week 2, the difference in the number of visible blood vessels between groups remained. Gel-only-treated wounds showed the fewest vessels (Fig. 4D), followed by MSC-treated wounds (Fig. 4E), whereas AFS cell-treated wounds showed the largest number of blood vessels (Fig. 4F). Additionally, vessels in the AFS cell-treated tissues appeared noticeably larger.

To quantify these observations, MVD and AVD values were determined from the histological sections by ImageJ software (Fig. 4G, 4H). Anti-vWF immunostaining confirmed the vascular structures (not shown). At week 1, MVD values in MSC and AFS groups were both significantly greater than that of the gel-only group (p < .05), whereas AVD values were not significantly different. At week 2, MVD values of the AFS group were significantly greater than those of both MSC and gel-only groups (p < .05). In addition, AVD values of the AFS group significantly increased and were significantly greater than those of the MSC group (p < .05), whose AVD values were significantly greater than those of the gel-only group (p < .05). This change in AVD suggested stabilization and maturation of the new vessels over time.

To further investigate vessel maturation, histological sections were stained for SMA (Fig. 4I–4K) and extravasation of red blood cells (RBCs) was examined in these sections (Fig. 4L–4N). In the gel-only group (Fig. 4I) very few SMA-stained SMCs were observed. Although many SMA-stained SMCs were observed in the MSC group (Fig. 4J), they did not form a complete circle around the vascular structures. In contrast, in the AFS group (Fig. 4K), complete circles of strongly stained SMCs were observed around the vascular structures, further suggesting blood vessel maturation in this group. Furthermore, large numbers of RBCs were observed in the regenerating skin tissue in the gel-only (Fig. 4L) and MSC (Fig. 4M) groups, suggesting extravasation from immature, “leaky” vasculature. In contrast most RBCs were contained inside the vasculature in the regenerating skin in the AFS cell group (Fig. 4N).

Fluorescence Imaging of AFS Cells and MSCs

Fluorescence imaging of GFP-expressing AFS cells and MSCs was carried out in order to track the cells in the regenerating skin tissue, following cell bioprinting. Skin samples were harvested on days 1, 4, 7, and 14 after surgery and examined by fluorescent microscopy (Fig. 5). The tissues were stained with DAPI to reveal all cells in the tissue. At day 1 the large number of GFP-labeled cells were visible in the wounds of the AFS cell and MSC groups. In both cell types, this number decreased by roughly 50% at day 4, and only a few GFP-labeled cells were seen at day 7. In tissues harvested at day 14 no GFP-labeled cells could be detected. Furthermore, GFP-labeled cells remained at the wound surface and did not migrate into the underlying tissue. These data indicate that the printed cells remained transiently in the tissue, and did not permanently integrate. These observations suggest that wound closure and neovascularization, observed in groups treated with bioprinting of cells, may be due to secretion of trophic factors from the cells rather than cell integration in the regenerating skin.

Analysis of Growth Factors in Conditioned Media of AFS Cells and MSCs

Proteomics analysis of AFS-conditioned media (CM) and MSC-CM illustrated differences in their growth factor secretory profiles (Table 1). In general, AFS-CM contained more growth factors at higher concentrations. Although MSC-CM had twofold more vascular endothelial growth factor (VEGF) than AFS-CM, it did not contain basic fibroblast growth factor (bFGF) or hepatocyte growth factor (HGF), found in AFS-CM. bFGF levels in AFS-CM were about 500 pg/ml CM, whereas VEGF levels in MSC-CM were about 5–10 ng/ml. In addition to bFGF and HGF, members of the insulin-like growth factor-binding protein superfamily were expressed at much higher levels in AFS-CM compared with MSC-CM.

Endothelial Cell Migration in Response to AFS Cell-Secreted Growth Factors

To test the activity of the angiogenic growth factors secreted by AFS cells, we performed an in vitro migration assay. To mimic the environment in which the AFS cells were bioprinted onto the wounds, the AFS cells were cultured on soft substrates (gels), and compared with AFS cells cultured in standard tissue culture plastic dishes. After 24 hours of culture, HUVECs, seeded inside Transwell inserts, were placed above the AFS cell cultures, as well as above nonseeded culture dishes. The numbers of HUVECs that migrated across the filter support of the insert was determined after 4 hours (Fig. 6A–6D). Significantly more HUVECs migrated in inserts placed over AFS cells that were cultured on soft 2-kPa substrates, compared with those on plastic, or in wells without cells (p < .05).