Development of a Novel Pre-Vascularized Three-Dimensional Skin Substitute Using Blood Plasma Gel

Culture of Endothelial Progenitor Cells

The isolation and culture of EPCs were performed as previously described with modifications ¹² . In brief, human UCB was layered on the top of Ficoll-Paque^TM PLUS (GE Healthcare, Chandler, Arizona, USA) followed by centrifugation at 900 g for 30 min. The white layer including mononuclear EPCs was transferred to a new centrifuge tube and subsequently incubated with RBC lysis buffer (Gibco, Paisley, Scotland, UK) for 5 min followed by centrifugation. The lower layer of cells was harvested and suspended in 10 ml of HBSS (Gibco). Cells were washed and incubated with 6 ml of EBM-2 culture medium supplemented with EGM-2 (Lonza, Basel, Switzerland) and 1% antibiotic-mycotic (Sigma-Aldrich, St. Louis, Missouri, USA). The cell solution was pipetted into the fibronectin (Millipore, Billerica, Massachusetts, USA)-coated plate for 2 days to obtain primary EPCs. Primary EPCs between the second and fifth passage were used in the experiment. The expression of specific endothelial progenitor markers including CD31, C-Kit, Tie-2 and VE-Cadherin (Abcam, Cambridge, Massachusetts, USA) was examined to confirm the phenotype of EPCs prior to the study. The study protocol was reviewed and approved by the Institutional Review Board (IRB) in the Tri-Service General Hospital, R.O.C. (TSGHIRB No. 100-05-251). A written informed consent according to the IRB guidelines was obtained from each donor.

The Preparation and Biocompatibility of Plasma Gel

The schematic experimental procedures of cell-contained plasma gel scaffold are presented in Fig. 1. Human blood samples were centrifuged at 800 g for 30 min to separate the plasma from blood cells. The plasma was then filtered through a 0.22 μm filter (Millipore) to remove impurities. To generate the bottom layer of the plasma gel scaffold, 0.5 ml of plasma was mixed with 0.5 ml of culture medium in a 6 cm culture plate, and then treated with various concentrations (2.5, 5, 10, 20 and 40 mM) of CaCl₂ (5 μl, Sigma-Aldrich) at 37°C to determine the optimal agglutination condition for the plasma gel scaffold. To form the top layer of the plasma gel scaffold, the plasma was mixed with culture medium containing 1 × 10⁵ cells of human dermal fibroblasts from foreskin (passage 3–5) and then treated with different concentrations of CaCl₂. Cell growth in the plasma gel scaffold was measured at 6, 12, 24, and 48 h after cell seeding.

OPEN IN VIEWER

Fig. 1.

The schematic diagram for the preparation of cell-contained 3D plasma gel scaffold. After centrifugation, the endothelial progenitor cells were isolated from whole blood and mixed with fixed concentrations of CaCl₂ for solidification. Endothelial progenitor cells proliferated and differentiated for 7 days to form microvascular-like structures, and keratinocytes were then seeded onto the surface of the plasma gel for 7 days to produce the epidermal layer. The cell-containing 3D plasma gel was prepared and then tested using a wound repair model on the dorsal skin of mice.

Preparation of Pre-Vascularized 3D-Gel Scaffold and Cell Proliferation Assay

EPCs and fibroblasts at various ratios were mixed with 250 μl of plasma gel in a 24-well culture plate and incubated in co-culture medium to form pre-vascularized 3D-skin substitutes. Typically, 2% FBS containing EBM-2 medium is used to induce the differentiation of EPCs to endothelial cells. In this experiment, EPCs and fibroblasts were mixed prior to seeding into the 3D-plasma gel scaffolds and co-cultured for 7 days to construct a pre-vascularized skin substitute. The co-culture medium was made up of equal amounts of DMEM-High glucose/F12 medium (3:1) and EBM-2 medium containing 1% PSA and 2% (or 20%) FBS. After 7 days, keratinocytes were seeded at the density of 1×10⁶ cells/cm² onto the external surface of plasma gel scaffolds and incubated for another 7 days. The co-culture medium was prepared with equal amounts of DMEM/F-12 medium (3:1, containing 5% FBS, 1% penicillin/streptomycin, 5 mg/ml insulin, 0.4 mg/ml hydrocortisone, and 1 mM isoproterenol) and EBM-2 medium (supplemented with 1% PSA and 20% FBS). In this study, the pre-vascularized skin substitute contained populations of EPCs, fibroblasts and keratinocytes that were seeded on the plasma gel. Non-vascularized skin substitutes, however, only contained fibroblasts and keratinocytes that were seeded on the plasma gel.

Cell proliferation of EPCs and fibroblasts in the plasma gel scaffolds were determined by WST-8 assay (Dojindo Laboratories, Kumamoto, Japan) at days 1, 3, 5 and 7. In brief, the medium was replaced with 300 µl of WST-8 reagent solution (90 µl WST-8 solution and 210 µl co-culture medium). After incubating with WST-8 reagent solution for 2 h, the absorbance was measured on a microplate reader (Sjeia Auto Reader II; Sanko Junyaku, Tokyo, Japan) with a test wavelength at 450 nm and a reference wavelength at 630 nm. Cell number was then derived using a standard curve. The differentiation of EPCs to endothelial cells was confirmed using an immunohistochemistry assay. As the ratio of EPCs to fibroblasts is a critical factor associated with the vascularization of EPCs, the cells were co-cultured at EPC:fibroblast ratios of 1:0.5, 1:1 and 1:2 in the plasma gel scaffolds. The morphology of EPC-derived endothelial cells was observed by an immunohistochemistry assay using CD31 antibody (ab9498, Abcam).

Immunohistochemistry Assay

Primary EPCs isolated from human UCB were incubated for 3 weeks before confirming the EPC phenotype using a CD31 marker for immunofluorescent staining. Cells were washed with phosphate buffered saline (PBS) twice and fixed in 4% paraformaldehyde (Sigma-Aldrich) for 24 h. The plasma gel was immersed in 100% alcohol for 1 min, washed with PBS three times, and then blocked with blocking buffer (1× PBS containing 5% BSA and 0.3% Triton X-100) at room temperature for 1 h. The plasma gel was then incubated with primary mouse monoclonal anti-human CD31 antibody (1:200 dilution) overnight at 4°C, and subsequently with a secondary alexa-conjugated 488 donkey anti-mouse IgG (ab150105, 1:100 dilution, Abcam) overnight at 4°C. The cell nuclei were stained with DAPI (Thermo Fisher Scientific, Waltham, Massachusetts, USA).

In Vivo Animal Studies

In vivo studies using a skin wound healing model in nude mice were performed as previously described ¹³ . The protocol was approved by the Institutional Animal Care and Use Committee (IACUC) in National Defense Medical Center, Taiwan, R.O.C. (IACUC-14-286). Eight-week-old nude mice (BALB/c-nu; BioLASCO, Taipei, Taiwan) were anesthetized and a full-thickness cutaneous wound (diameter 1.5 cm) was surgically created on the dorsum of the mice. All the surgical instruments were sterilized and the surgical procedures were performed under laminar flow. The surgical sites were sterilized with Easy Antiseptic Liquid 2% (Panion & BF, Taipei, Taiwan) before surgery. The mice were assigned to one of four treatment groups including blank, plasma gel, non-vascularized skin and pre-vascularized skin. The plasma gel scaffolds were placed on the wounds, sewn into place with 10–12 stitches using NC125 L Nylon 5-0 surgical sutures (UNIK, Taipei, Taiwan), and covered with Tegaderm film (3 M Health Care, St. Paul, Minnesota, USA). The wounds were continuously monitored over a period of 14 days. After 7 and 14 days, the wound tissues were excised, fixed with 10% formalin for at least 24 h at room temperature, embedded in paraffin, and then sectioned in 5-μm increments. The sections were stained with H&E staining and observed under an optical microscope (ZEISS Axio Scan.Z1, Carl Zeiss, Jena, Germany). Masson’s trichrome staining was performed under standard experimental procedures to compare the gross morphology of collagen fibers between the experimental and blank groups. The thickness of the epidermis and dermis was examined using ZEN software. In addition, immunohistochemical analysis was performed on skin sections on day 14 post-surgery with an anti-smooth muscle actin antibody and anti-PECAM/CD31 antibody, to confirm the expression of smooth muscle cells and endothelial cells, which are components of blood vessels. The double stain kit (TADS03, BIOTnA Biotech, Taoyuan, Taiwan) that could detect smooth muscle cells (containing HRP Green Chromogen) and endothelial cells (containing DAB Brown Chromogen) on the same section slide was used following the manufacturer’s protocol.

Tensile Tests

Mechanical properties were determined by using a universal testing machine (HT-8504 Hung Ta Instruments, Taichung, Taiwan) on the tensile mode setting. After 14 days post-surgery, the mice were sacrificed and the skin of the dorsal wound was excised. Each skin tissue was cut into a rectangular section of 1.5 cm × 4 cm. Tissue specimens were stretched at a crosshead speed of 50 mm/min until rupture. Young’s modulus and tensile strengths at breakage were calculated from the stress-strain data ¹⁴ . Six specimens per experimental group were measured to obtain the average values.

Angiogenesis Assay

To assess angiogenesis in the untreated control, plasma gel, non-vascularized and pre-vascularized groups 14 days after implantation, animals were intravenously injected with AngioSense 750 EX (Perkin Elmer, Waltham, Massachusetts, USA; 2 nmol/100 μl prepared in sterile PBS) via tail vein injection ¹⁵ . Fluorescence imaging (excitation: 745 nm; emission: 800 nm; high-pass filter cut off: 770 nm; illumination: 30%) in the anesthetized animal was performed 24 h after the injection of the fluorescence dye, using the IVIS Spectrum Imaging System (Perkin Elmer) ¹⁵ .

Statistical Analysis

Results were presented as the mean ± standard deviation of three replicates for each experiment. Statistical analysis was performed using SPSS version 22 for Windows (SPSS Inc., Chicago, Illinois, USA). The statistically significant differences between groups were assessed by one-way analysis of variance (ANOVA) with Tukey’s HSD Post Hoc Test. P < 0.05 was considered statistically significant.

Preparation of the Plasma Gel Scaffolds

The gel scaffolds were prepared by using a series of CaCl₂ concentrations at 2.5, 5, 10, 20 and 40 mM, into a mixture of equal amounts of plasma and culture medium (Fig. 2). Our results demonstrated that 10, 20, and 40 mM CaCl₂ were able to induce agglutination of the plasma gel, but 2.5 and 5 mM CaCl₂ were too low to successfully form the gel (Fig. 2A). Cell proliferation of fibroblasts inside the plasma gel scaffolds was analyzed using WST-8 assay (Fig. 2B). We found that 40 mM CaCl₂ significantly inhibited cell growth from 12 h to 48 h, resulting in cell numbers decreasing from 1×10⁵ cells at 0 h to 0.94×10⁵ cells at 48 h. In contrast, an increase in cell number from 1×10⁵ cells (0 h) to 2.07×10⁵ cells (48 h) was observed in plasma gel scaffolds prepared using 10 mM CaCl₂ (Fig. 2B). Therefore, plasma gel scaffolds prepared with 10 mM CaCl₂ were used in the following experiments. This plasma gel scaffold is translucent and gelatinous, and can be easily handled and detached from a 6 cm dish plate after casting.

OPEN IN VIEWER

Fig. 2.

The preparation of plasma gel with addition of different concentrations of CaCl₂. (A) The effect of plasma agglutination was tested by different concentrations of CaCl₂. (B) The proliferation of fibroblasts for up to 48 h was investigated using plasma gel containing varying specific concentrations of CaCl₂. (C) The morphology of plasma gel containing 10 mM CaCl₂. *: indicates P < 0.05; **: indicates P < 0.01 when compared with plasma gel with 10 mM CaCl₂.

Pre-Vascularization Using EPC Incorporated Plasma Gel Scaffolds

Primary EPCs were isolated and harvested from UCB samples. First, EPC phenotypes were identified using specific markers of endothelial progenitors such as CD31 (Platelet endothelial cell adhesion molecule), C-Kit (stem cell marker that is found in EPCs), Tie-2 (endothelial cell-specific tyrosine kinases 2) and VE-Cadherin (an endothelial-specific cell–cell adhesion molecule) in the present study (see Supplement S1). After 3 weeks of culture, the expression of CD31 surface bio-markers on EPCs was confirmed using immunofluorescence staining (Fig. 3). To determine if EPCs have the potential to differentiate into vascular-like endothelial cells in a 3D environment, EPCs were cultured in DMEM-High glucose: F12 (3:1 v/v) 2% or 20% FBS medium. As shown in Fig. 3A, a greater extent of vascularization was observed when EPCs were cultured in 20% FBS medium compared with 2% FBS culture medium over 7 days. Furthermore, cell proliferation assay results demonstrated that the cell growth of EPCs was significantly inhibited in the medium containing 2% FBS compared with 20% FBS medium (Fig. 3B). In contrast, the cell growth of fibroblasts was not affected by the concentration of FBS, resulting in very similar rates of cell growth from day 1 to day 7 (Fig 3C).

OPEN IN VIEWER

Fig. 3.

The differentiation and proliferation of EPCs and fibroblasts in plasma gel in co-culture medium containing 2% or 20% FBS. (A) Immunofluorescent staining for the pre-vascularization of EPCs using a labeled CD31 marker in plasma gel with the different co-culture media at 1, 3, 5 and 7 days. (B) The proliferation of EPCs in co-culture medium was observed at 1, 3, 5 and 7 days. (C) The proliferation of fibroblasts in co-culture medium was investigated at 1, 3, 5 and 7 days. *: indicates P < 0.05; **: indicates P < 0.01; ***: indicates P < 0.001 when compared with the co-culture medium containing 2% FBS.

Pre-Vascularization of EPC on the Plasma Gel Modulated by Fibroblast

The differentiation of EPCs for vascularization was examined in co-culturing conditions with fibroblasts at varying ratios of 1:0.5, 1:1 and 1:2. Positive staining of CD31 was used to assess vascularization of EPCs in the plasma gel (Fig. 4A). The density of vascularized area significantly increased when EPCs and fibroblasts were cultured at a ratio of 1:0.5 over 7 days. On day 3, the vascularization area was calculated at 17.83% and increased to 25.31% by day 7. In contrast, when ECPs:fibroblasts were cultured at either 1:1 or 1:2, the vascularized area was relatively low at approximately 9–12% over 7 days (Fig. 4B).

OPEN IN VIEWER

Fig. 4.

The differentiation of EPCs under different ratios of EPCs to fibroblasts in plasma gel. (A) Immunofluorescent staining presenting a vascular-like tubular structure from EPCs were labeled with CD31 marker and cell nuclei were labeled with DAPI at 3, 5, and 7 days. (B) The calculation of the vessel density by Image J Software. *: indicates P < 0.05; **: indicates P < 0.01 when compared with 0.5:1 ratio of fibroblasts to EPCs group.

The Wound Repair by Neovascularized 3D Skin in Nude Mice

To assess the effect of this pre-vascularized 3D skin substitute on wound healing in vivo, we used a nude mouse wound healing model. On day 0, all gel scaffolds were able to tightly stitch onto the wound edges. The wound areas of mice grafted with the blank or the plasma gel slowly healed over 14 days (Fig. 5A-a, 5A-e, 5A-i). The wound area contained a wet surface, indicating that the epidermis and dermis has been only partially formed. In contrast, wound healing was accelerated in mice grafted with non-vascularized 3D skin substitutes (Fig. 5A–c, 5A–g, 5A–k) or pre-vascularized 3D skin substitutes (Fig. 5A–d, 5A–h, 5A–l) on day 7 and 14 post-surgery. Histological analysis showed that both non-vascularized and pre-vascularized 3D gels were well integrated with surrounding tissue compared with the blank and plasma group 7 days post-surgery (Fig. 6A-day 7). Pre-vascularized 3D gels were found to induce an abundant neovascularized network in the central wound area (Fig. 6/ day7-B-d). Histological analysis further confirmed that all gel scaffolds were well tolerated in the mice, with mild inflammatory responses, and no cell necrosis and apoptosis observed on day 7 and 14 (Fig. 6-day 7 and day 14). The non-vascularized/pre-vascularized 3D skin graft was replaced with the cutaneous construction at the study site, and the neo-vascularized network also achieved a condition of homeostasis. In addition, an epithelial layer formed at the wound site of mice grafted with non-vascularized/pre-vascularized 3D skin substitutes.

OPEN IN VIEWER

Fig. 5.

The in vivo wound repair model and angiogenesis located at the dorsal area of nude mice. The observation of wound repair was performed at 0, 7 and 14 days post-surgery. The blank (a, e, i) was not engrafted with any artificial skin. The plasma gel (b, f, j), non-vascularized (c, g, k) and pre-vascularized (d, h, l) were engrafted into the wound.

OPEN IN VIEWER

Fig. 6.

Histological examination of skin wound healing was observed by H&E staining of the whole skin wound (A) and the center of skin wound (B) wound engrafted by blank (a), plasma gel scaffold (b), non-vascularized (c), and pre-vascularized skin (d) under 20× magnification at 7 and 14 days post-surgery. Green dashed rectangle marks the wound center on the dorsal area of nude mice, and the magnification presented as (B) on the right side of images. The epidermal layer could not be found in the blank group. Scale bars = 100 μm. N indicates normal skin. W indicates wound bed. Black triangles in the images indicate the boundaries between the wound and the surrounding normal skin. Red arrows indicate micro-vessels. n = 6 for each group.

Masson’s trichrome staining (Fig. 7A) was conducted to assess collagen formation post skin excision and grafting. Wounds grafted with pre-vascularized 3D skin substitutes displayed well-organized collagen fibers with corrugated structure. However, in the blank (Fig. 7A–b) and plasma gel scaffold treatment (Fig. 7A–c) groups, the density of collagen fibers found at the wound site was low. By day 14, more collagen fibers were obviously apparent in the non-vascularized (Fig. 7A–i) and pre-vascularized (Fig. 8A–j) gel groups compared with the other groups. The quantitative analysis of collagen density demonstrated that collagen production in non-vascularized and pre-vascularized groups was significantly higher compared with the blank and plasma gel groups 14 days post-surgery (Fig. 7B). The thickness of the epidermis layer in the central area of the wounds was measured using a ZEN slidescan tool 14 days post-surgery (Fig. 7C). The thickness of the regenerated epidermis was significantly greater in the non-vascularized 3D gel-treated wounds, but the thickness of the epidermis was found to be similar between the wounds grafted with pre-vascularized skin substitute and normal skin. Blank or plasma gel scaffold-treated wounds displayed limited re-epithelialization in 14 days. The thickness of the dermis layer was also examined 14 days post-surgery (Fig. 7D). Dermal thickness was significantly increased in the non-vascularized 3D gel-treated group, but not in the pre-vascularized groups. Thinner dermal thickness was observed in the blank and plasma gel groups.

OPEN IN VIEWER

Fig. 7.

(A) Masson’s trichrome staining for the histological analysis of repaired wound after treatment with blank (b, g), plasma gel scaffold (c, h), non-vascularized skin (d, i), pre-vascularized skin (e, j), and normal skin (a, f) at 7 and 14 days post-surgery. Scale bars = 100 μm. (B) quantification of collagen density after Masson’s trichrome staining (C) the calculation of the thickness of the epidermal layer and (D) thickness of the dermal layer as well as the evaluation of mechanical properties measured by (E) Young’s modulus and (F) tensile strength 14 days post-surgery in the five groups. *: indicated P < 0.05; **: indicates P < 0.01; ***: indicates P < 0.001 when compared with sham group.

OPEN IN VIEWER

Fig. 8.

(A) IVIS imaging of neo-vascular blood perfusion of pre-vascularized 3D gel implants and quantification of fluorescent signals. 14 days after implantation, fluorescence imaging of nude mice were performed using IVIS Spectrum Imaging System as described in the Materials and Methods. *: indicates P < 0.05; **: indicates P < 0.01; ***: indicates P < 0.001. (B) Immunohistochemical staining for normal skin and pre-vascularized skin substitute presenting a vascular-like tubular structure from EPCs on day 14 labeled with CD31 and αSMA markers. CD31 and αSMA were found present in normal skin tissue, while only CD31 labeling was observed in pre-vascularized skin group. CD31: brown, αSMA: green. Black arrows indicate CD31 marker. Red arrows indicate αSMA marker.

Evaluation Skin Wound Restoration by Tensile Test

The mechanical strength of healed skin was determined using a universal testing machine (Fig. 7E and F). In the blank group, the wound area was not healed on day 14, therefore the strength assessment could not be performed for the blank group. The epidermal layer also could not be found in the blank group, as shown in the histological analysis (Fig. 6A–a). Non-vascularized and pre-vascularized healed skin maintained a similar Young’s modulus and tensile strength as normal skin. However, a significant reduction in Young’s modulus of 0.66 MPa and tensile strength of 0.22 MPa was observed in the skin healed following plasma scaffold treatment.

Angiogenesis

In order to assess the angiogenesis in the wound area, fluorescence probes (AngioSense 750 EX) were injected on day 14 and fluorescence imaging was recorded 24 h after the injection (Fig. 8A). The highest accumulation of Angiosense 750 fluorophores was found on the wound areas in the pre-vascularized skin group, indicating the formation of functional neovasculature for the regulation of blood flow (Fig. 8A). In addition, non-vascularized skin substitutes showed secondary high fluorescence signals compared with the blank or plasma groups, suggesting that the co-culture of fibroblast may benefit microvascular formation. Furthermore, wounds grafted with non-vascularized and pre-vascularized substitutes formed an epidermis on the surface of wound area. The differentiation of EPCs into smooth muscle cells was further examined using immunohistological double-staining analysis (Fig. 8B). In a normal control skin section, CD31 signals are shown in brown and αSMA expression was shown in green. However, only CD31 (shown as brown) expression was observed in pre-vascularized skin substitutes on day 14, indicating that EPC did not differentiate into smooth muscle cells in the 3D plasma scaffolds during the wound healing process.