Mesenchymal Stem Cell-Laden, Personalized 3D Scaffolds with Controlled Structure and Fiber Alignment Promote Diabetic Wound Healing

Abstract

The management of diabetic wounds remains a major therapeutic challenge in clinics. Herein, we report a personalized treatment using 3D scaffolds consisting of radially or vertically aligned nanofibers in combination with bone marrow mesenchymal stem cells (BMSCs). The 3D scaffolds have customizable sizes, depths, and shapes, enabling them to fit a variety of type 2 diabetic wounds. In addition, the 3D scaffolds are shape-recoverable in atmosphere and water following compression. The BMSCs-laden 3D scaffolds are capable of enhancing the formation of granulation tissue, promoting angiogenesis, and facilitating collagen deposition. Further, such scaffolds inhibit the formation of M1-type macrophages and the expression of pro-inflammatory cytokines IL-6 and TNF-α and promote the formation of M2-type macrophages and the expression of anti-inflammatory cytokines IL-4 and IL-10. Taken together, BMSCs-laden, 3D nanofiber scaffolds with controlled structure and alignment hold great promise for the treatment of diabetic wounds.

Graphical Abstract

1. Introduction

In the United States, chronic wounds affect about 6.5 million patients and carry a fiscal burden estimated to exceed $25 billion per year [1]. Diabetic foot ulcers (DFU), a major complication of diabetes mellitus and type of chronic wound, occurs in 15% of diabetics and contributes a staggering $9 to $13 billion in healthcare costs [2]. Of patients with DFU, a subsequent 14–24% experience a lower extremity amputation, with the mortality rate from amputation approaching 50–59% five years post-procedure [3]. The sustained pro-inflammatory responses [4], impaired cell migration and extracellular matrix (ECM) deposition [5], decreased angiogenesis and granulation tissue formation [6] and difficulty in re-epithelialization often cause delayed healing of DFU [7]. Despite intense basic and clinical research aimed at improving DFU healing, clinical manifestation and prognosis of DFU remain unsatisfactory in most cases, resulting in considerable patient suffering wound persistence [8,9]. Therefore, there is a significant need for an affordable and efficacious treatment strategy.

Given their ability to adequately mimic native ECM architectures, electrospun nanofibers hold great promise for wound healing and regenerative medicine [10]. The incorporation of small molecules or macromolecules to electrospun nanofiber membranes was commonly used for healing diabetic wounds [11,12]. Delivery of small molecules is often associated with limited efficacy. Utilization of growth factors is not practically feasible due to the complex regulatory procedures and potential safety issues. To date, platelet-derived growth factor-BB (PDGF-BB) is the only U.S. Food and Drug Administration (FDA) approved growth factor for use in diabetic wound healing [13]. The employed electrospun nanofiber membranes were composed of densely packed fibers, which severely limited cell infiltration and growth throughout nanofiber scaffolds [10]. In addition, those membranes lacked desired nanotopographic cues such as fiber alignment and thus failed to direct and promote cell migration either from the bottom to the top for granulation tissue formation or from the surrounding to the center for re-epithelialization. Therefore, engineering the desired structure and fiber alignment of 3D nanofiber scaffolds could be an ideal approach to enhance granulation tissue formation and re-epithelialization.

Bone marrow mesenchymal stem cells (BMSCs)-based therapies have warranted enormous attention in wound healing and regenerative medicine (e.g., skin wound healing [14], bone regeneration [15], cardiac repair [16]) given their capabilities in secretion of paracrine growth factors and cytokines [17], differentiation to effector cells [18] and immunomodulation [19]. During skin regeneration, BMSCs could contribute to wound healing by transdifferentiating into effector cells (e.g., keratinocytes, myofibroblasts, pericytes, endothelial cells), suggesting their direct participation in the processes of wound healing [20]. BMSCs were also able to secrete a series of cytokines and chemokines, such as vascular endothelial growth factor (VEGF), insulin-like growth factor 1 (IGF-1), epidermal growth factor (EGF), keratinocyte growth factor (KGF), angiopoietin-1, stromal derived factor-1 (SDF-1), macrophage inflammatory protein-1 alpha and beta (MIP1-α and MIP1-β), to regulate the responses of cells within wounds and contribute to the wound healing in an indirect manner. Despite the many therapeutic benefits, one of the major limitations of using BMSCs with tissue scaffolds is the lack of 3D niche microenvironments to support their survival [14]. Stem cell fate is known to be regulated by signals from these niches, one of the environmental parameters is the extracellular matrix (ECM) that stem cells adhere [21]. Because BMSCs are adult stem cells, their unique functions are retained only in intimate contact with an instructive microenvironment [22]. Artificial stem cell niches such as hydrogels [23], aerogels [24], and microspheres [25], are developed to mimic a real microenvironment for seeding stem cells. However, these artificial niches fail to sustain the phenotype and multipotency of stem cells in the long term, resulting in weakened therapeutic effects. Nanotopographical features of substrate not only influence stem cell fate [26], but also retain stem-cell phenotype and maintain stem cell growth over eight weeks [27]. Considering the importance of nanotopographic cues, utilizing 3D nanofiber scaffolds with controlled structure and alignment could serve as an ideal niche for growth and transplantation of BMSCs.

In this study, we aimed to engineer 3D personalized nanofiber scaffolds with controlled structure and alignment in conjunction with BMSCs to promote diabetic wound healing, which is framed on our recent findings. To overcome spatial limitations, an innovative gas-foaming technique was established to transform 2D nanofiber mats into 3D expanded, highly porous scaffolds [10]. However, this membrane expansion technique was associated with limitations in both shapes (e.g., rectangle) and the direction of fiber alignment (e.g., uniaxial alignment). In a more recent study, we reported an approach inspired by solids-of-revolution that transforms 2D nanofiber mats into once-inaccessible 3D scaffolds with predesigned complex shapes [28]. These 3D scaffolds consisted of either radially-aligned or vertically-aligned nanofibers. Importantly, these scaffolds enabled rapid cell penetration, collagen deposition, and angiogenesis following subcutaneous implantation in rats other than uniform cell proliferation within scaffolds in vitro. Furthermore, nanofiber scaffolds with large pore size are able to increase expression of the M2 macrophages, along with decreased expression of the M1 macrophages [29, 30], suggesting nanofiber scaffolds with large pore size are capable of regulating the local inflammation to pro-regenerative status. Thus, the developed radially aligned scaffolds (RASs) and vertically aligned scaffolds (VASs) also have potentials to control the local inflammation in pro-regenerative status during wound healing because of the inherent large pore size.

Herein, we hypothesized that 3D scaffolds consisting of radially or perpendicularly aligned nanofibers could provide a desirable niche for BMSCs growth and promote re-epithelialization or granulation tissue formation. And we tested BMSCs-laden, personalized 3D nanofiber scaffolds with controlled structure and alignment for healing full-thickness wounds in type 2 diabetic mice. In this study, the anti-contraction wound healing model was established in type 2 diabetic mice. This model was able to prevent wound closure caused by skin contraction, and allowed wounds to heal through granulation and re-epithelialization that can closely mimic the process of human skin wound healing [31]. The results obtained from the wounds without contraction were processed to evaluate the wound healing effects of 3D nanofiber scaffolds and BMSCs loaded 3D nanofiber scaffolds.

2. Materials and Methods

2.1. Materials

PCL (Mw=80 kDa), pluronic-F-127, gelatin, sodium borohydride, and Triton X-100 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Dichloromethane (DCM) and N, N-dimethylformamide (DMF) were purchased from BDH Chemicals (Dawsonville, GA, USA). Dulbecco’s modified eagle medium (DMEM), fetal bovine serum (FBS), and penicillin-streptomycin were obtained from Invitrogen (Carlsbad, CA, USA). CD 31, Ki67, TNF-α, IL-4, IL-6, IL-10 primary antibodies and Goat Anti-Rabbit IgG H&L (FITC), Goat Anti-Mouse IgG H&L (Alexa Fluor® 647), Goat Anti-Rat IgG H&L (Cy5®), Donkey Anti-Goat IgG H&L (Cy5®) secondary antibodies, phalloidin-iFluor 488 reagent (ab176753) were purchased from Abcam (Cambridge, MA, USA). CD206 and CCR7 primary antibodies were purchased from NovusBio (Centennial, CO, USA). The Cell Counting Kit-8 (CCK-8) was purchased from Dojindo Molecular Technologies, Inc (Rockville, MD, USA).

2.2. Fabrication of 3D scaffolds consisting of radially or vertically aligned nanofibers

The fabrication of 3D scaffolds consisting of radially or vertically aligned nanofibers were following our previous study [28]. Briefly, the PCL/0.5% F-127 nanofiber mats were cut into rectangle shape in the direction of nanofiber alignment or perpendicular to the direction of nanofiber alignment (5 mm (width) × 1.5 mm (length)). Then, the long side of these PCL/F-127 nanofiber mats were fixed by a thermal treatment (85 °C for 1 s). Subsequently, these PCL/F-127 nanofiber mats with one side fixed were immersed in 1 M NaBH₄ solution which was gently shaken for 30 min. After expansion, the 3D scaffolds with cylindrical shape consisting of radially or vertically aligned nanofibers were transferred into the distilled water and exposed to a vacuum (~200 Pa) for 10 s. This process was repeated for 3 times. Finally, the distilled water was removed, and the 3D scaffolds were exposed to a vacuum until it froze and then they were freeze-dried. In order to enhance the mechanical property, 3D scaffolds were immersed in 0.5% gelatin solution for 10 min. The residual gelatin solution was then removed. Subsequently, these 3D scaffolds were exposed to a vacuum until it froze and then they were freeze-dried. No crosslinking was performed for the gelatin coating.

2.3. SEM characterization

Briefly, the RASs or VASs were immersed into water and frozen under liquid nitrogen. Then the surface of radially or vertically aligned scaffolds were removed by cryo-cutting. Following, Pt sputter coating was performed on these scaffolds. Finally, the cross sections of 3D scaffolds consisting of radially or vertically aligned nanofibers (X-Y, Y-Z, X-Z planes) were characterized by SEM (FEI, Quanta 200, Oregon, USA).

2.4. Mechanical Tests.

In order to test the mechanical properties of 0.5% gelatin-coated, radially and vertically aligned nanofiber scaffolds, the cylindrical 0.5% gelatin-coated radially and vertically aligned nanofiber scaffolds were cut to cuboid-shaped scaffolds (1×1×1 cm³) respectively. The mechanical properties were recorded with an Instron 5640 universal test machine, 10 N load cell was used to perform the mechanical tests. The compressive strain of radially aligned nanofiber scaffold was set as 70% and 76% (Maximum compressive strain of radially aligned nanofiber scaffolds), the compressive strain of vertically aligned nanofiber scaffold was set as 70% and 90%. The superelastic property of 0.5% gelatin-coated radially aligned and vertically aligned nanofiber scaffolds was run 20 cycles with the speed of 18 mm/min under 76% and 90% compressive strain respectively.

2.5. BMSCs culturing on 3D scaffolds

The C57BL/6 mice mesenchymal stem cells were purchased from Cyagen (Santa Clara, CA, US). Cells were maintained in DMEM (low glucose, GlutaMAX™ Supplement) plus with 10% FBS, 1% penicillin-streptomycin in an incubator at 37 °C and 5% CO₂, and the medium was changed every two days. The 3D scaffolds consisting of radially or vertically aligned nanofibers were sterilized with ethylene oxide for 12 h prior to cell culture. The BMSCs suspension with a concentration of 1×10⁷ cells/mL was first prepared. 3D scaffolds were immersed in the cell suspension solution and treated with vacuum for 10 s. Then, the scaffolds were removed from the cell solution and placed into the 0.1% agar pretreated 24-well plate. One mL DMEM plus with 10% FBS, 1% penicillin-streptomycin was added and the medium was changed every two days. At each indicated time point, the scaffolds were collected and washed with PBS for 3 times. Then, they were fixed with 4% paraformaldehyde for 15 min and washed with PBS for 3 times.

2.6. Confocal imaging

The distribution and growth of BMSCs in the 3D scaffolds were characterized by confocal laser scanning microscopy (Zeiss 880, Oberkochen, Germany). Briefly, the scaffold seeded with BMSCs were stained with phalloidin-iFluor 488 reagent. Briefly, Five hundred microliter of tracking dye working solution was added into each scaffold, and continuously incubated the cells in a 37°C, 5% CO₂ incubator for 30 minutes. Then the scaffolds were washed with PBS buffer for 3 times. Finally, the cells were fixed with 4% paraformaldehyde and then were placed on cover slides. During the imaging process, 50 μL distilled water was added to each scaffold to avoid drying. The first image and last image of the z-stack range from 0 μm to 150 μm were selected, and the interval was set as 5 μm. Finally, the tile scan was set as 4×4. The depth coding of Zeiss blue software was used to indicate the cells at different depths with different colors.

2.7. CCK-8 assay

At each indicated time point, the old culture medium was removed and replaced with 2 mL fresh culture medium. Then, 20 μL CCK-8 solution was added to each well of 24-well culture plate, and continuously cultured for 3 h at 37 °C and 5% CO_2. Then, 2 mL culture medium with scaffold in each well was transferred to a 15-mL centrifuge tube. The tube was centrifuged for 4 min at 1000 rpm. Finally, 100 μL supernatant of each centrifuge tube was transferred to a 96-well culture plate, and the absorbance at 450 nm was measured using a microplate reader.

2.8. RT-PCR

For detecting the relative expression of VEGF, PDGF-BB, bFGF, TGF-β1, CXCR4 and type 1 collagen genes, BMSCs were seeded on both radially aligned scaffolds and vertically aligned scaffolds for 5 and 9 days. Cells seeded on a 6-well culture plate for 5 and 9 days as control. At each indicated time point, the total RNA was extracted using RNeasy Mini Kit (QIAGEN, Germany) according to the manufacturer’s instructions. One μg of RNA was used to synthesize cDNA using a reverse transcription reagents kit (Quantabio, USA). The quantitative real-time PCR was carried out following the protocol and conducted with an Applied Biosystems 7500 Fast Real-time PCR system (Applied Biosystems, USA) with the following temperature profile: at 95 °C for 3 min, then 40 cycles at 95 °C for 3 s and at 60 °C for 30 s. All primer sequences were as following:

• VEGF forward primer: GCAGCGACAAGGCAGACTAT, reverse primer: AACCTCCTCAAACCGTTGGC;

• PDGF-BB forward primer: TGCCCAAACACTAAGGTGCT, reverse primer: TCGGGAGAAGGGATGCTTTG;

• bFGF forward primer: GTTGGCCAAGGACGGAAGAA, reverse primer: TTTCCCTCCGAAGGTTCGTT;

• TGF-β1 forward primer: AGGGCTACCATGCCAACTTC, reverse primer: CCACGTAGTAGACGATGGGC;

• CXCR4 forward primer: TGGAACCGATCAGTGTGAGT, reverse primer: GCCGACTATGCCAGTCAAGA;

• Type 1 collagen forward primer: CAATGGGGGATAGAGGAGCAT, reverse primer: ATCCCCAGTGCAGTAAGCC.

2.9. Wound model

Briefly, the TALLYHO type 2 diabetic mice (Male, 10–12 weeks, The Jackson Laboratory, https://www.jax.org/strain/005314) were anesthetized using 4% isoflurane in oxygen for approximately 2 min. Mice were placed on a heating pad to maintain their body temperature. An area of 4 × 4 cm² on the back of each animal was shaved, and the povidone-iodine solution was applied three times on the exposed skin. Next, a skin punch was used to make a circular wound (8 mm in diameter). After the wound model was established, the control wounds were left untreated (n=12). The wounds treated with 3D scaffolds consisting of radially aligned nanofibers were set as RAS group (n=12). The wounds treated with 3D scaffolds consisting of vertically aligned nanofibers were set as VAS group (n=12). The wounds treated with BMSCs-laden, 3D scaffolds consisting of radially aligned nanofibers were set as RAS+BMSCs group (n=12). The wounds treated with BMSCs-laden, 3D scaffolds consisting of vertically aligned nanofibers were set as VAS+BMSCs group (n=12). There were three mice for each time point of each group, and totally 6 wounds for each time point. These scaffolds were implanted one time. At days 0, 3, 7, and 10 after treatment, a digital camera was used to record the wounds of each mouse. Then, wound and surrounding tissues were excised and fixed in 4% paraformaldehyde. The animal study was approved by IACUC at the University of Nebraska Medical Center (UNMC).

2.10. Histological and immunohistochemical analysis

Fixed samples were dehydrated in a graded ethanol series (70%−100%), embedded in paraffin, and then sectioned (5 μm thick). Samples were stained with either hematoxylin and eosin (H&E) or Masson’s trichrome. For immunohistochemical staining, slides were deparaffinized and rehydrated followed by antigen retrieval in heated citrate buffer for 5 min (citrate buffer solution, pH 6.0 at 100 °C). Non-specific antibody binding was prevented with a 5% BSA solution. The sections were incubated with primary antibodies overnight at 4 °C. Then the corresponding secondary antibodies were added and incubated for 1 h at room temperature, followed by staining with DAPI for 5 min. Ten randomly selected fields were examined for each group at each time point and used to assess the average positive cells per unit area. The calculation of re-epithelialization rate was based on H&E staining results, which is defined as the re-epithelialization rate (%) = (the total width of the two epidermis tongue / the width of the wound) × 100%.

2.11. Statistical analysis

The data are presented as the mean ± S.D. and the statistical analysis was performed using GraphPad Prism 8.0 software. Differences among groups were assessed using one-way ANOVA followed by post hoc tests. The values of p < 0.05 were considered statistically significant. The values of p < 0.01 were considered statistically very significant.

3. Results

3.1. Personalized 3D Scaffolds

Based on our previous studies, we fabricated 3D scaffolds consisting of radially or vertically aligned nanofibers (Fig. 1). The sizes, thicknesses and shapes were readily customized according to the sizes, depths and shapes of diabetic patients’ wounds. Fig. 1A and Fig. 1B show the photographs of a series of cylinders consisting of radially or vertically aligned nanofibers with diameters of 12, 10, 8 and 6 mm. Fig. 1C and Fig. 1D show the photographs of a series of cylinders consisting of radially or vertically aligned nanofibers with thicknesses of 1, 3, 5 and 7 mm. Fig. 1E and Fig. 1F show various shapes of RASs or VASs including triangular, square, pentagonal and heart-shape. Further morphological characterizations using SEM reveal that the top view of the RAS displayed numerous thin layers of radially aligned nanofibers, while the side view exhibited a highly porous structure (Fig. 1G). The average distance between two layers of nanofibers in RAS was 376 ± 53 μm. The top view of the VAS shows a highly porous structure, and the side view reveals that the pore walls were composed of numerous vertically aligned nanofibers (Fig. 1H). The average distance between two layers of nanofibers in VAS was 450 ± 87 μm.

Fig. 1.

Customized 3D scaffolds consisting of radially or vertically aligned nanofibers with tailored sizes, thicknesses and shapes. (A, B) 3D scaffolds consisting of radially or vertically aligned nanofibers with different sizes. (C, D) 3D scaffolds consisting of radially or vertically aligned nanofibers with different thicknesses. (E, F) 3D scaffolds consisting of radially or vertically aligned nanofibers with different shapes. (G) Top view and side view of 1-mm thick, radially aligned scaffold. (H) Top view and side view of 1-mm thick, vertically aligned scaffold. The double-headed arrows indicate the alignment of nanofibers.

3.2. Mechanical Properties

To fully characterize the mechanical properties of the scaffolds, compression tests were carried out. Intriguingly, 3D RASs and VASs were shape re-coverable after compression in the air (Fig. 2A and Fig. 2B). Moreover, they were capable of completely returning to its original shape in the water following compression (Fig. 2C and Fig. 2D). The maximum compressive strain and stress of 3D RASs were 76% and 46 ± 10 kPa (Fig. 2E). Cyclic compression tests indicate that the RAS was able to recover (95 ± 3) % of its original shape in the air after 20 compressive cycles (Fig. 2F). The maximum compressive strain and stress of 3D VASs were 90% and 33 ± 4 kPa (Fig. 2G). Cyclic compression tests show that the VAS can recover 100% of its original shape in the air after 20 compressions (Fig. 2H). Required mechanical properties of scaffolds frequently vary depending on wound locations. For example, scaffolds for plantar ulcers may require higher compressive strength. Mechanical properties can be tuned by increasing the concentration of coated gelatin solutions (Fig. S1). For example, the compressive stress of 1% gelatin-coated RASs reached up to 102 ± 7 kPa, which was significantly higher than that achieved by 0.5% gelatin-coated RASs.

Fig. 2.

Shape recoverable property of 3D RAS and VAS. Photographs of 3D RAS (A) and VAS (B) before, under and after complete compression in the air. Photographs of 3D RAS (C) and VAS (D) under and after complete compression in the water. The RAS and VAS could completely recover the original shape in the air and water. (E) Compression tests of 3D RAS under 70% and 76% (maximum) compressive strain (n=5). (F) Cyclic compression tests of 3D radially aligned scaffolds under 76% compressive strain (20 cycles) (n=5). RAS was able to recover (95 ± 3) % of its original shape in the air after 20 cycles of compressive stress. (G) Compression tests of 3D vertically, aligned nanofiber scaffolds under 70% and 90% compressive strain (n=5). (H) Cyclic compression tests of 3D VAS under 90% compressive strain (20 cycles) (n=5). VAS was able to recover 100 % of its original shape in the air after 20 compressive cycles. RAS: Radially aligned scaffolds, VAS: Vertically aligned scaffolds. The data are presented as the mean ± S.D.

3.3. BMSCs Culture

To explore the therapeutic effects of incorporating BMSCs in expanded scaffolds, BMSCs were cultured on both RASs and VASs. Fig. 3A shows a laser scanning confocal microscopy image indicating the distribution of BMSCs on both RASs and VASs after 9 days of culture. Uniform cell distribution was observed throughout the entire scaffolds, with cells displaying a radially-aligned and porous pattern, recapitulating the structures of RASs and VASs, respectively. Due to the limited imaging thickness of confocal microscopy, a maximum depth of about 140 μm from the surface to the inside of the scaffolds was imaged. Fig. 3B shows the quantification of BMSC proliferation on the scaffolds. BMSCs significantly proliferated from day 1 to day 9 on RASs and VASs, and the proliferation reached a maximum value on day 9. There was no significant difference in the proliferation of BMSCs between day 9 and day 14 on both types of scaffolds. In addition, there was also no significant difference in the proliferation of BMSCs on both types of scaffolds at each indicated time point. Moreover, the living BMSCs in the scaffolds after proliferation for 9 days were about (95 ± 3)% (Fig. S2).

Fig. 3.

Mouse BMSCs proliferated on 3D RAS and VAS. (A) The distribution of mouse BMSCs in the center and the edge of 1-mm thick, RAS or VAS after culturing for 9 days. (B) The rapid proliferation of mouse BMSCs on 1-mm thick, RAS and VAS after culturing for 1, 5, 9 and 14 days, respectively (n=6). (C) The increased expression of growth factors (VEGF, PDGF-BB, bFGF), migration related genes (TGF-β1, CXCR4), and decreased expression of type 1 collagen of BMSCs cultured on 1-mm thick, RAS or VAS for 5 and 9 days (n=9). RAS: Radially aligned scaffolds, VAS: Vertically aligned scaffolds. The data are presented as the mean ± S.D. *p<0.05, **p<0.01.

Furthermore, the effects of 3D RASs and VASs on the biological functions of BMSCs were examined. As shown in Fig. 3C, there was no difference in the relative gene expression of VEGF among the control, RAS and VAS groups after culture for 5 and 9 days. While, the relative gene expression of PDGF-BB on RAS and VAS groups were higher than the control group on both day 5 (p<0.01) and day 9 (p<0.01). Similarly, the relative gene expression of bFGF on RAS and VAS groups were also increased compared to the control group on both day 5 (p<0.05) and day 9 (p<0.01). Interestingly, the fold change of the relative expression of migration related genes (TGF-β1, CXCR4) on RAS and VAS groups were dramatically increased compared to the control group at day 5 (p<0.01) and day 9 (p<0.01). The relative gene expression of type 1 collagen on RAS and VAS groups were lower than the control group after culture for 5 days. And there was no difference in type 1 collagen expression among the control, RAS and VAS groups after culture for 9 days.

3.4. Granulation Tissue Formation

In order to examine how our BMSCs-laden personalized 3D scaffolds affect healing tissues in diabetic ulcers, two full-thickness, splinted excisional wounds were created on the back side of type 2 diabetic mice. The blood glucose level of mice ranged from 323 mg/dl to 512 mg/dl, which was higher than the standard 200 mg/dl. And the bodyweight of mice ranged from 32.5 g to 43.5 g. There was no significant change in body weight of each mouse before and after the operation. In this study, we evaluated wounds at days 3 and 7, as the mice started to scratch off the splints after 7 days posttreatment. Subsequent wound healing after splint removal would affect the accuracy of comparison among the treatment groups as wounds contract in the absence of splints (Fig. S3). A greater number of cells positively stained with Ki67 (a proliferative marker) in wounds treated by RAS+BMSCs and VAS+BMSCs groups compared to the control, RAS, VAS groups on both day 3 and day 7 (p<0.01) (Fig. S4). Masson’s trichrome staining shows that the wounds treated by RAS, VAS, RAS+BMSCs and VAS+BMSCs groups had more collagen deposition (indicated by blue arrows) and new tissue formation relative to the control group on both days 3 and 7. Wounds treated by RAS+BMSCs and VAS+BMSCs groups also showed more collagen deposition than RAS and VAS groups (Fig. 4A). Moreover, more blood vessels (indicated by green arrows) formed in wounds treated by RAS, VAS, RAS+BMSCs and VAS+BMSCs groups compared to the control group on both days 3 and 7 (p < 0.01). In addition, the numbers of newly formed blood vessels within wounds treated by RAS+BMSCs and VAS+BMSCs groups was significantly higher than the RAS and VAS groups on day 7 (p < 0.01) (Fig. 4A and Fig. 4C). The wounds treated by RAS+BMSCs and VAS+BMSCs groups had higher expression of CD31, an endothelial cell marker, than the RAS and VAS groups on both days 3 and 7 (p<0.01). In contrast, there was only marginal CD31 expression in the wounds treated by the control group (Fig. 4B and Fig. 4D). More visible granulation tissues were found in the Control, RAS, VAS, RAS+BMSCs and VAS+BMSCs groups after surgery for 10 days (Fig. S5). The wounds in the RAS, VAS, RAS+BMSCs and VAS+BMSCs groups were fully filled with a mass of dense granulation tissues, while the structure of granulation tissue in control group was still loose (Fig. S5).

Fig. 4.

The enhanced granulation tissue formation of wound area of RAS, VAS, RAS+BMSCs, VAS+BMSCs groups compared to control group after 3 and 7 days postsurgery. (A)The trichrome staining of Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after 3 days postsurgery. Green arrows indicate newly formed blood vessels. Blue arrows indicate newly formed collagen fibers. (B) The CD31 immunohistochemical staining of Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after 7 days postsurgery. (C, D) The quantification of newly formed blood vessels and CD 31 positive cells of Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after 3 and 7 days postsurgery (n=5). RAS: Radially aligned scaffolds, VAS: Vertically aligned scaffolds, RAS+BMSCs: BMSCs loaded, radially aligned scaffolds, VAS+BMSCs: BMSCs loaded, vertically aligned scaffolds, S: scaffold area, W: wound area. The data are presented as the mean ± S.D. *p<0.05, **p<0.01.

3.5. Re-epithelialization

Fig. 5 shows the re-epithelialization of the wounds treated by RAS, VAS, RAS+BMSCs and VAS+BMSCs groups on day 3 and day 7. No visible epithelial tongues were found at two edges of the wounds in the control group three days after surgery. However, keratinocytes migrated out of two edges of wounds in RAS, VAS, RAS+BMSCs and VAS+BMSCs treatment groups. The re-epithelialization rates for the RAS, VAS, RAS+BMSCs and VAS+BMSCs treatment groups were (12 ± 5) %, (3 ± 2) %, (10 ± 4) %, (6 ± 2) % (Fig. 5A and Fig. S6), while only (1 ± 1) % in the control group. After surgery for 7 days, the re-epithelialization rate for the wounds treated by the RAS group was (46 ± 13) %, which was dramatically higher than those in the control ((9 ± 3) %), VAS ((8 ± 2) %), RAS+BMSCs ((20 ± 6) %), VAS+BMSCs ((19 ± 8) %) groups. Additionally, the re-epithelialization rates of RAS+BMSCs and VAS+BMSCs groups were faster than the VAS group (Fig. 5B and Fig. S6). The wounds treated by RAS and RAS+BMSCs completed re-epithelialization after surgery for 10 days. The re-epithelialization in the wounds of VAS and VAS+BMSCs groups were also complete. However, there was still a large area of dermis without epidermis coverage in the wounds of control group after surgery for 10 days (Fig. S5).

Fig. 5.

3D RAS guide and accelerate the re-epithelialization of diabetic wounds. (A) H & E staining of wounds treated by Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after surgery for 3 days. (B) H & E Staining of wounds treated by Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after surgery for 7 days. The bottom two images are higher magnification of the left and right edge of wound, respectively. Green dot lines indicate the migrated keratinocytes. RAS: Radially aligned scaffolds, VAS: Vertically aligned scaffolds, RAS+BMSCs: BMSCs loaded, radially aligned scaffolds, VAS+BMSCs: BMSCs loaded, vertically aligned scaffolds, S: scaffold area, W: wound area, N: normal tissue.

3.6. Immune Responses

3.6.1. Macrophage Formation

To examine the influence of BMSC-laden 3D scaffolds on immune responses, we first performed immunohistochemical analysis of the mouse macrophages in M1 phase and M2 phase within the wound area. Fig. 6A and Fig. 6C show CD206 (a marker for M2 macrophages) positive cells stained in red, indicating no difference in the formation of M2 macrophages among the control and all the treatment groups at day 3. However, the numbers of CD206 positive cells in RAS, VAS, RAS+BMSCs, and VAS+BMSCs groups were obviously higher than that in the control group at day 7 (p < 0.01). The numbers of CD206 positive cells in RAS+BMSCs and VAS+BMSCs groups were significantly larger than those in RAS and VAS groups at day 7 (p < 0.01). Fig. 6B and Fig. 6D shows that no difference in the numbers of CCR7 (a marker for M1 macrophage) positive cells among the control and all the treatment groups at day 3 exist. The numbers of CCR7 positive cells in RAS, VAS, RAS+BMSCs, and VAS+BMSCs groups were obviously lower than that in the control group at day 7 (p < 0.01). The numbers of CCR7 positive cells in RAS+BMSCs and VAS+BMSCs groups were significantly lower when compared to the RAS and VAS groups at day 7 (p<0.01). In addition, the M2/M1 ratio in the control group was 0.87 at day 3, and it further reduced to 0.27 at day 7. The M2/M1 ratios in RAS (1.51) and VAS (1.28) groups were higher than 1 at day 3, and M2/M1 ratios in RAS (1.37) and VAS (1.51) groups at day 7 were similar as the ones at day 3. In comparison, the M2/M1 ratios in RAS+BMSCs (1.06) and VAS+BMSCs (1.05) groups were higher than 1 at day 3, and M2/M1 ratios increased dramatically in the RAS+BMSCs (6.51) and VAS+BMSCs (8.45) groups at day 7 (Fig. 6E).

Fig. 6.

BMSCs loaded RAS and VAS provide a regenerative microenvironment. (A, B) The M2 type macrophage (CD206) and M1 type macrophage (CCR7) immunohistochemical staining of wound area treated by Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after 3 and 7 days postsurgery. (C, D) The quantification of CD206 and CCR7 positive cells within the wound area of Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after 3 and 7 days postsurgery (n=5). (E) The ratio between the number of CD206 positive cells (M2) and the number of CCR7 positive cells (M1) within the wound area treated by Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after 3 and 7 days postsurgery. RAS: Radially aligned scaffolds, VAS: Vertically aligned scaffolds, RAS+BMSCs: BMSCs loaded, radially aligned scaffolds, VAS+BMSCs: BMSCs loaded, vertically aligned scaffolds, S: scaffold area, W: wound area. The data are presented as the mean ± S.D. *p < 0.05, **p < 0.01.

3.6.2. Pro-inflammatory Cytokine Expression

Given their important role in both innate and adaptive immune response, an investigation into the expression of pro-inflammatory cytokines IL-6 and TNF-α at the wound site was carried out using immunohistochemistry. Fig. 7A and Fig. 7D shows IL-6 positive cells stained in green, suggesting the expression of IL-6 in RAS+BMSCs and VAS+BMSCs groups were lower than the control, RAS and VAS groups at day 3 (p < 0.05). The expression of IL-6 in the RAS, VAS, RAS+BMSCs, and VAS+BMSCs groups was significantly reduced compared to the control group at day 7 (p < 0.01). Similarly, Fig. 7B and Fig. 7C shows that the expression of TNF-α in the RAS, VAS, RAS+BMSCs, and VAS+BMSCs groups was lower than the control group at day 3 (p < 0.01) and day 7 (p < 0.05). The expression of TNF-α in the RAS+BMSCs and VAS+BMSCs groups were decreased compared to RAS and VAS groups (p < 0.05) at day 3.

Fig. 7.

BMSCs loaded RAS and VAS downregulate the expression of inflammatory cytokines. (A, B) The IL-6 and TNF-α immunohistochemical staining of wound area treated by Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after surgery for 3 and 7 days. (C, D) The quantification of TNF-α and IL-6 positive cells within the wound area of Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after surgery for 3 and 7 days (n=5). RAS: Radially aligned scaffolds, VAS: Vertically aligned scaffolds, RAS+BMSCs: BMSCs loaded, radially aligned scaffolds, VAS+BMSCs: BMSCs loaded, vertically aligned scaffolds, S: scaffold area, W: wound area. The data are presented as the mean ± S.D. *p < 0.05, **p < 0.01.

3.6.3. Anti-inflammatory Cytokine Expression

To study the expression of anti-inflammatory cytokines within wounds, IL-4 and IL-10 were detected using immunohistochemistry. Fig. 8A and Fig. 8B shows IL-4 and IL-6 positive cells stained in red. The expression of IL-4 in the RAS, VAS, RAS+BMSCs, and VAS+BMSCs groups was significantly higher than the control group at day 3 (p < 0.01) and day 7 (p < 0.01). In addition, the expression of IL-4 in the RAS+BMSCs and VAS+BMSCs groups were higher than the RAS and VAS groups on day 3 (p < 0.01), and there was no difference in the expression of IL-4 among the four groups at day 7 (Fig. 8C). Similarly, the expression of IL-10 in the RAS, VAS, RAS+BMSCs, and VAS+BMSCs groups was significantly increased than the control group at day 3 (p < 0.01) and day 7 (p < 0.01). The expression of IL-4 in the RAS+BMSCs and VAS+BMSCs groups was higher than the RAS and VAS groups at day 3 (p < 0.01) and day 7 (p < 0.01) (Fig. 8D).

Fig. 8.

BMSCs loaded RAS and VAS upregulate the expression of anti-inflammatory cytokines. (A, B) The IL-4 and IL-10 immunohistochemical staining of wound area treated by Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after surgery for 3 and 7 days. (C, D) The quantification of IL-4 and IL-10 positive cells within the wound area treated by Control, RAS, VAS, RAS+BMSCs, VAS+BMSCs groups after surgery for 3 and 7 days (n=5). RAS: Radially aligned scaffolds, VAS: Vertically aligned scaffolds, RAS+BMSCs: BMSCs-loaded, radially aligned scaffolds, VAS+BMSCs: BMSCs-loaded, vertically aligned scaffolds, S: scaffold area, W: wound area. The data are presented as the mean ± S.D. *p < 0.05, **p < 0.01.

4. Discussion

The roles of electrospun nanofiber membranes in healing diabetic wounds were investigated [32,33]. However, these studies focused on 2D electrospun nanofiber membrane system in combination with unique morphological patterns [32,34]. antibacterial agents [35,36], different types of skin cells and stem cells [37], growth factors and cytokines [38], and plant extracts [39,40]. Despite the variety of approaches, the densely-packed fibers limit cell infiltration, greatly reducing the efficacy of incorporated cell therapeutics and re-epithelialization. In this study, we demonstrated 3D scaffolds consisting of radially or vertically aligned nanofibers with controlled structure, porosity, and shape-recoverable properties. This class of scaffolds not only overcomes the limitations of traditional 2D nanofiber membranes, but also provide a perfect temporary niche for BMSCs growth and transplantation for treatment of diabetic wounds.

To our knowledge, the current study reports, for the first time, the design of customized 3D scaffolds that can completely match the size, depth, and shape of diabetic wounds with unique porous nanofiber structure when compared to the current commercial dressing of DFU treatment. The personalized scaffolds can be readily obtained by cutting, which is attributed to the gelatin coating. Our previous work demonstrated that the 0.5% gelatin coating was able to enhance the mechanical properties of expanded PCL nanofiber scaffolds, making them superelastic and shape-recoverable [41]. In addition, it was also able to retain its original shape after wetting. With regard to the 3D scaffolds consisting of radially aligned nanofibers, when a stress is applied, the straight nanofibers bend and fold to form numerous spring structures which serve as elastic units. During compression, the formed spring unites store the energy (stress). As the compression releases, the stored stress is slowly released and the compressed scaffolds return to its original shape (Fig. S7A and Fig. S7C). The shape recovery mechanism is different for the 3D scaffolds consisting of vertically aligned nanofibers. The shape recovery is most likely attributed to the layered arch structures which serve as elastic units. During compression, because of the densification of pores, the pore walls start to contact each other, providing strong fatigue resistance (Fig. S7B and Fig. S7D) [41]. Therefore, the 3D scaffolds consisting of radially and vertically aligned nanofibers can completely recover their shape and maintain structural integrity after repeated compressive strain release. Such properties could make these scaffolds unaffected during walking, bending, or other load-bearing movements.

The advantages of BMSCs transplantation on skin wound healing are more widely appreciated [14, 42]. Here, we found the 3D RASs and VASs were capable of enhancing the biological functions of BMSCs. These 3D nanofiber scaffolds could accelerate the relative expression of growth factors (PDGF-BB and bFGF), migration related genes (TGF-β1 and CXCR4). PDGF-BB and bFGF are the major growth factors involved in the process of wound healing, which contribute to angiogenesis, granulation tissue formation and ECM secretion during wound healing [43]. TGF-β1 is a very strong chemokine which belongs to the superfamily of transforming growth factor β. It was demonstrated that TGF-β1 was able to promote cancer cells [44], endothelial cells [45], epidermal cell [46] and BMSCs [47] migration and invasion. CXCR-4 also known as fusin or CD184, is a membrane receptor of stromal-derived-factor-1 (SDF-1), which plays significant roles in the homing and migration of multiple stem cell types, and it is highly expressed in migrated BMSCs [48]. We speculate that these enhanced expression of growth factors and migration-related genes of BMSCs cultured on radially or vertically aligned PCL nanofiber scaffolds could be mainly due to the nanotopographic cues rendered by aligned PCL nanofibers. The higher expressed TGF-β1 and CXCR4 could be due to the induction of aligned nanofibers which are capable of guiding and promoting the migration of BMSCs. Here, we also found that the relative expression of type 1 collagen of BMSCs on both RAS and VAS groups were lower than the ones cultured in culture plate. We speculate that BMSCs seeded on RASs and VASs showed higher cell migration activity compared to BMSCs cultured in plates, likely resulting in the decreased type 1 collagen expression.

The wound healing process of humans differs greatly from that of rodents, where human wound healing relies on granulation tissue formation and re-epithelialization, rather than wound contraction as found in rodents [31]. Granulation tissue consists of new blood vessels, fibroblasts, immune cells, and provisional ECM [49]. Newly formed blood vessels are capable of affording enough oxygen and nutrients to regenerating tissues, an essential elemental in complete wound closure [50]. Here, the angiogenesis in the RAS and VAS groups was significantly improved as compared to the control group. Further, BMSCs loading enhanced revascularization in the RAS+BMSCs and VAS+BMSCs groups through a variety of mechanisms including multipotential differentiation and paracrine growth factor effects. Oswald et al. and Perin et al. demonstrated that human BMSCs could be differentiated into endothelial cells in vitro and in vivo [51, 52]. In addition, Phillips et al. revealed that autologous mesenchymal stem cell transplantation could definitely enhance the expression of VEGF, resulting in increased vascular density and regional blood flow in human [53]. Other growth factors (e.g., FGF and PDGF) secreted by BMSCs could also contribute to the promotion of new blood vessel regeneration [54]. Dermal fibroblasts begin to migrate to and proliferate in the wound site along during angiogenesis. Simultaneously, the fibroblasts produced type I collagen, a main component of ECM. In this study, the RAS+BMSCs and VAS+BMSCs groups showed a dense collagen deposition, while only loose collagen deposition was detected in the RAS and VAS groups, thus confirming the BMSCs effect on ECM deposition. Tredget et al. reported that the profibrotic characteristics of dermal fibroblasts could be enhanced by BMSCs [55]. Fibroblasts co-cultured with BMSCs had higher expression of type 1 collagen and lower expression of inhibitor of metalloproteinase 1 and collagenase, suggesting that BMSCs encourage fibroblasts to deposit more collagen while suppressing ECM degradation.

Granulation tissue formation may promote re-epithelialization, where keratinocytes migrate from two edges of the wound and toward the center of the wound to form a barrier between the wound and the environment [56]. Though granulation tissue formation was achieved in the RAS+BMSCs and VAS+BMSCs groups, re-epithelialization in both groups was unsatisfactory. We speculate that the tight connections between wound beds and scaffolds could prevent sufficient migration of keratinocytes. However, re-epithelialization may be boosted after complete escharosis exfoliation. By comparison, the RAS group exhibited a faster re-epithelialization. The connections between wound beds and scaffolds are loose, and keratinocytes can migration along the bottom of RASs, even without a large amount of granulation tissue support. Although the connections between wound beds and scaffolds are loose in the VAS group, the migrating direction of keratinocytes is perpendicular to the direction of fiber alignment, which inhibits the migration of keratinocytes, resulting in a slow re-epithelialization.

The sustained pro-inflammatory response is one of the dominant factors restricting diabetic wound healing. M1 type macrophages play a decisive role in the sustained pro-inflammatory response and lead to elevated levels of pro-inflammatory cytokines and proteases, ultimately reducing levels of various growth factors [57]. Pro-inflammatory or pro-regenerative immune response largely depends on the ratio of M2/M1 [58]. In this study, the ratio of M2/M1 in the control group was less than 1 at days 3 and 7, which means the wounds were under a pro-inflammatory microenvironment with higher expression of pro-inflammatory cytokines (TNF-α and IL-6) and lower expression of anti-inflammatory cytokines (IL-4 and IL-10). The ratios of M2/M1 in the RAS and VAS groups were slightly above 1 at days 3 and 7. However, the fold change of the ratios of M2/M1 in the RAS+BMSCs and VAS+BMSCs groups was about 6 from day 3 to day 7, suggesting a pro-regenerative microenvironment with lower expression of pro-inflammatory cytokines (TNF-α and IL-6) and higher expression of anti-inflammatory cytokines (IL-4 and IL-10).

According to the University of Texas DFU Classification System, diabetic ulcers are classfied into 4 stages based on the depth of ulcers and affected tissues: Stage 0, pre-or postulcerative site that has healed; Stage 1, superficial ulceration; Stage 2, wound penetrating to tendon or capsule; and Stage 3, wound penetrating to bone or joint [59]. Considering the difference of diabetic wounds from person to person, a personalized 3D scaffolds consisting of hierarchically assembled nanofibers with controlled size, thickness, and shape can perfectly fit various diabetic wounds. The RASs could be suitable for stage 0 and stage 1 DFU, which aim to accelerate the re-epithelialization of superficial wounds. And the VASs could be suitable for stage 2 and stage 3 DFU, which aim to promote the formation of granulation tissues of deep wounds. Fig. 9A shows the schematic illustrating how to apply customized, RASs for healing stage 0 and stage 1 DFU, which could enhance angiogenesis, granulation tissue formation, ECM deposition and re-epithelialization (Fig. 9C). Fig. 9B shows the schematic illustrating how to apply customized, VASs for the management of stage 2 and stage 3 DFU, which could promote angiogenesis, granulation tissue formation, and ECM deposition (Fig. 9D). Besides stem cells, growth factors (e.g., PDGF-BB [60], FGF-2 [61], VEGF [62] and EGF [63]), short peptides [64] and siRNA [65] can be loaded to these scaffolds to further promote diabetic wound healing. Moreover, silver nanoparticles [66], antibiotics [67], antimicrobial peptides [36] and other antibacterial agents can also be incorporated to such scaffolds for the treatment of infected diabetic wounds.

Fig. 9.

Schematic illustrating the application of 3D scaffolds consisting of radially or vertically aligned nanofibers in diabetic wound healing and their potential mechanisms. (A) Schematic illustrating the application of 3D scaffolds consisting of radially aligned nanofibers in stages 1 & 2 DFU healing, which aim to accelerate the re-epithelialization of superficial wounds. (B) Schematic illustrating the application of 3D scaffolds consisting of vertically aligned nanofibers in stages 3 & 4 DFU healing, which aim to promote the formation of granulation tissues of deep wounds. (C) The potential mechanism of 3D scaffolds consisting of radially aligned nanofibers for diabetic wound healing, including enhancing angiogenesis, granulation tissue formation, ECM deposition and re-epithelialization. (D) The potential mechanism of 3D scaffolds consisting of vertically aligned nanofibers for diabetic wound healing, including enhancing angiogenesis, granulation tissue formation, and ECM deposition.

5. Conclusions

In summary, we have demonstrated 3D scaffolds consisting of radially or vertically aligned nanofibers in combination with BMSCs offers a robust, customizable platform to dramatically improve diabetic wound healing. These scaffolds can provide a personalized treatment by tailoring the size, depth, and shape of diabetic wounds. The BMSCs laden, 3D scaffolds significantly enhanced granulation tissue formation, angiogenesis and ECM deposition, and simultaneously elicited a pro-regenerative response to accelerate wound healing. Furthermore, given the relatively simple mechanical features, such scaffolds can be further enhanced with a variety of therapies to help achieve proper treatment of diabetic chronic wounds.

Statement of Significance.

In this study, we developed 3D radially and vertically aligned nanofiber scaffolds to transplant bone marrow mesenchymal stem cells (BMSCs). We personalized 3D scaffolds that could completely match the size, depth, and shape of diabetic wounds. Moreover, both the radially and vertically aligned nanofiber scaffolds could completely recover their shape and maintain structural integrity after repeated loads with compressive stresses. Furthermore, the BMSCs-laden 3D scaffolds are able to promote granulation tissue formation, angiogenesis, and collagen deposition, and switch the immune responses to the pro-regenerative direction. These 3D scaffolds consisting of radially or vertically aligned nanofibers in combination with BMSCs offer a robust, customizable platform potentially for a significant improvement of managing diabetic wounds.