The material poly (lactic-co-glycolic acid) (PLGA) (Resomer RG 503, Boehringer, Germany) with a lactide:glycolide ratio of 50:50 and a mean molecular weight of 33,000 Da was utilized herein. Insulin glargine, a sterile solution of insulin (Lantus), was obtained from Sanofi-Aventis Inc (Frankfurt, Germany). Vildagliptin (C₁₇H₂₅N₃O₂) and 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP) were obtained from Sigma-Aldrich (Saint Louis, MO, United States).

Two nanofibrous core-shell scaffolds were fabricated using a special well-designed coaxial electrospinning device that concurrently delivered two solutions from two separate feeding tubes to an aluminum plate (Hsu et al., 2016). Based on our preliminary results, a 1:3 ratio of vildagliptin and PLGA shell design has sufficient core-shell scaffold performance. A mixture of vildagliptin (210 mg) and PLGA (630 mg) from group A (Vildagliptin 25%) or PLGA (840 mg) from group B (Vildagliptin 0%) was dissolved in 3,000 µL HFIP to prepare shell solution. Use 1,000 µL of insulin glargine (equivalent to 3.64 mg) (100%) as the core solution for the feeding syringes. During electrospinning, two independently controlled pumps delivered both liquids onto the aluminum plate at volumetric flow rates of 5 μL/min for the core insulin solution and 15 μL/min for the shell vildagliptin/PLGA solution. The electrospinning studies were conducted at 25°C. Electrospun nanofibers were put in a laminar flow hood for 72 h and then were maintained at 4°C.

The density of the two core-shell electrospun nanofibers were measured using mass divided by volume. The porosity of both core-shell nanofibrous membranes was assessed using the calculation as followed (Lee et al., 2022a). P o r e   % = 1 − σ d r u g s / P L G A   / σ P L G A .   σ = d e n s i t y . (1)

The SEM images of 100 randomly selected areas of fiber (Lee et al., 2015) were used to assess the size distribution and pore area of the electrospun nanofibers. The analysis was conducted using ImageJ software from the National Institutes of Health in the United States. Additionally, the morphology of the core-shell nanofibers was examined using TEM (JEOL JEM-2000EXII, Japan) in three separate instances.

Mechanical properties were evaluated via a Lloyd tensiometer (AMETEK, United States) - tensile strength (the ratio of breaking force (N) to the sample area of cross section (mm²), and is measured in MPa) and elongation at break (the percentage increase in length at the break point (mm) divided by the initial length (mm) and multiplied by 100) of the nanofibrous test samples (n = 3).

To measure the wetting angle of the nanofibers, a water contact angle analyzer featuring a video monitor (First Ten Angstroms, United States) was used. Three samples, each with dimensions of 10 mm × 10 mm, were taken from both sets of membranes and placed on the test plate. Pure water was then carefully released onto the surface of each sample.

The two nanofibers at 0.5, 1, 2, 3, 8, and 24 h were gotten by the water content formula. The water absorption capacity (%) was analyzed as follows. W a t e r   a b s o r p t i o n   c a p a c i t y   % = W   – W 0 / W 0 × 100 (2) where W₀ and W is the sample weight before and after placing in PBS for 3 min after removing surface water with absorbent paper (n = 3).

A release pattern of vildagliptin or insulin from the membranes were calculated using an elution method. Membranes (10 mm × 10 mm) were analyzed (n = 5) in 1,000 µL of phosphate-buffered saline (PBS) and were stored at 37°C before the eluent was collected for analysis. One mL PBS was replaced once a day for 30 days. Hitachi L-2200 Multisolvent Delivery System, an HPLC assay (Japan), with an XBridge C₁₈ 5 μm, 4.6 × 150 mm HPLC column (Waters) for vildagliptin separation, was used to evaluate the profile of vildagliptin release from the fabricated nanofibers. The bioactivity of insulin was evaluated via an human insulin solid-phase sandwich ELISA (enzyme-linked immunosorbent assay) (Thermo Fisher Scientific, United States).

Transwell filters (Costar, United States) with pores (8.0 µm in diameter) were applied for EPC migration assay, as reported elsewhere (Lee et al., 2021; Lee et al., 2022b). The cells were acquired from Molecular Pharmacology Laboratory (Chang-Gung University, Taiwan). Data were gotten from five randomly selected regions using the eluent released from insulin/vildagliptin-eluting PLGA or insulin/PLGA scaffolds at different time point (n = 4).

All operations related to diabetic model were carried out following the confirmation of Institutional Animal Experiment Committee of Chang Gung University (CGU 14–045). Induction of diabetes was performed using fourteen Sprague-Dawley rats with streptozotocin (STZ) (Sigma, United States). Seven days following intraperitoneal injections of 70 mg/kg of STZ, diabetic status was determined by identifying hyperglycemia (≧300 mg/dL) before diabetic wound assessment was performed.

Seven animals were prepared with insulin/vildagliptin-eluting PLGA nanofibrous scaffolds as group A, and the other seven were prepared with insulin/PLGA scaffolds as group B.

Following being anesthetized, a sterile 8.0 mm plastic template was put on mid-back of the diabetic rats, and an excisional wound with full-thickness was created by taking away the skin based on the template. Wound edges were examined using glass microscope slides and assessed for size by planimetry using ImageJ. The entire wound was excised down to the fascia on day 14 along with a margin (5 mm) of uninjured skin. Prior to performing the frozen section with cryostats, the sample was placed in a microtome using an optimal cutting temperature compound.

Samples were washed extensively in PBS buffer containing 0.05% (v/v) Tween-20 (PBST) and 2% bovine serum albumin blocking at room temperature for 30 min for immunofluorescence staining. The samples were then got along with primary antibodies against α-smooth muscle actin (α-SMA) (1/500, ab5694, Abcam) at least 8 h at 4°C. After being treated with AF 546 goat anti-rabbit secondary antibodies (1/500, Life Technologies), the samples were left to stain for a minimum of 8 h in a cool room at 4°C. On the following day, 4,6-Diamidino-2-phenylindole (DAPI) (nuclear stain, 1/2000, Sigma) was added and the samples were then made ready for visualization using a Leica confocal microscope (n = 4).

Data were analyzed in SPSS software (version 17.0 for Windows; SPSS Inc., United States). Data are represented as mean ± standard deviation (SD). For continuous variables with a normal distribution, means were compared using an unpaired Student’s t-test. Otherwise, a Mann-Whitney U test was performed. Statistical significance was determined by a p-value less than 0.05.

Figures 1A, B present the morphologies and diameter distributions of the coaxially electrospun nanofibers (magnification ×3,000) that had been loaded with insulin/vildagliptin-eluting PLGA (Group A) or insulin/PLGA nanofibers (Group B) (Supplementary Figure S1). The pictures can be clearly noticed that the vildagliptin-loaded core-shell nanofiber membranes clearly had a larger diameter (662 ± 162 nm) than the insulin/PLGA nanofibers (478 ± 99 nm) (p = 0.007) (Figure 1C). The pore area of the coaxial nanofibrous insulin/vildagliptin-eluting PLGA membranes [1965 ± 238 × 10³ nm (Menke et al., 2015)] significantly exceeded that of the insulin/PLGA membranes [1,128 ± 71 × 10³ nm (Menke et al., 2015)] (p < 0.001) (Figure 1D). The value of the core-shell group A membranes porosity (88.9 %± 2.0%) was higher than that of the group B scaffolds (78.7 %± 2.0%) (p = 0.004).

The mixture of vildagliptin with PLGA might have increased the stability of bending and stretching properties and the jet during electrospinning, increasing the diameter of the fabricated fibers over that of the crude PLGA fibers, and yielding the smooth morphology and structure of the consequently formed nanofibrous scaffolds. Nanofibrous scaffolds that were prepared by the electrospinning method mimicked the biological functions or structure of the natural extracellular matrix (ECM), offering mechanical strength to the structure of cells and guiding cellular migration (Ingavle and Leach, 2014). Additionally, the nanofibrous structures had an elevated ratio of surface area-to-volume and high porosity, favoring cellular proliferation, migration, and adhesion (Khorshidi et al., 2016).

To verify the core-shell structure, nanofibers containing recombinant green fluorescence protein (reGFP) at the core and pure PLGA as the shell were fabricated through co-electrospinning. The reGFP image in Figure 2A observed using a fiuorescence microscope shows a continuous and clear fluorescent signal (green string-like filaments), indicating the presence of a highly bioactive protein at the core. The TEM image in Figure 2B indicates that the fabricated nanofibers had a distinct core-shell structure and a density that differed between the two-white dashed. These images confirm the core–shell structure of the spun nanofibers and the encapsulation of insulin within the polymer PLGA shell.

As a dressing for wound repair, a scaffold should be able to support cell growth and resist stress and movement. Figure 2C displays the mechanical properties of both electrospun core-shell nanofibrous scaffolds. The mean tensile strengths for nanofibrous membranes with insulin/vildagliptin-eluting (Group A) and insulin/PLGA (Group B) were about 1.23 ± 0.19 MPa and 2.90 ± 0.05 MPa, respectively (p < 0.001). The nanofibrous scaffolds in Group A had less elongation at breakage (29.9 %± 3.0% vs. 180.4 % ± 15.1%, p < 0.001) than those in group B. In a drug-loaded polymeric composite, the polymer is the component that resists the external force and supports plastic extension. The mixture with vildagliptin decreased the content of PLGA in the composite herein. The maximum elongation and strength at breakage of PLGA nanofibers were reduced accordingly (Wang et al., 2021a; Wang et al., 2021b).

Figure 3 represents the contact angles of water on core-shell insulin/Vildagliptin PLGA nanofibers with vildagliptin (Figure 3A) or insulin/PLGA nanofibers with no vildagliptin (Figure 3B). The results of the core-shell insulin/vildagliptin-loaded and the insulin/PLGA nanofibrous scaffolds were 68.3 ± 8.5° and 121.4 ± 2.0°, respectively. Mixing with vildagliptin significantly increased the hydrophilicity of the core-shell insulin/PLGA nanofibrous membranes (Figure 3C) (p = 0.006). Figure 3D shows the percentages of water absorption capacity in the both core-shell nanofibers following PBS immersion for 0.5, 1, 2, 3, 8, and 24 h. Group A nanofibrous membranes had the highest water content (379 %± 9%) at 2 h and maintained a higher water content (>272%) during the first 24 h. Without vildagliptin, the capacity of group B nanofibers for storing water was always lower than that of group A and group B nanofibrous scaffolds also had their highest water content at 2 hours (283 %± 24%) (Supplementary Table S1, water absorption capacity of both groups) Due to the capillary effect of the numerous pores, the water uptake of nanofibers increased with time after immersion in an aqueous environment. The uptake reached peak values at 2 h, and gradually diminished and leveled off thereafter. The modification of the wetting properties of a polymer surface can be achieved by the integration of hydrophilic moieties. This is attributed to the increased polarity of the polymer and the subsequent reduction in the angle of contact with water (Kiss et al., 2002; Yilgor et al., 2012). Consequently, the inclusion of the water-soluble hydrophilic drug, vildagliptin, facilitated the water absorption of nanofibers (Hess et al., 2011; Lee et al., 2022c). Nanofibrous scaffolds in Group A thus exhibited a greater water uptake capability than those in Group B. As a scaffold for skin dressings, surface conditions significantly affect the response of cells to the water content of biomaterials. The core-shell nanofibrous scaffolds demonstrated improved wettability, which in turn would promote cellular activities such as proliferation, differentiation, and attachment, ultimately leading to tissue regeneration (Wu et al., 2015; Wang et al., 2016).

Data from the EPC transwell migration assay (Figures 4A–F), more cells migrated after treatment with the eluate on day 2 (Figures 4A–C) and day 14 (Figures 4D–F) in group A (day 2 eluate: 847 ± 33 cells/mm²; day 14 eluate: 762 ± 77 cells/mm²) than after treatment in group B (day 2 eluate: 469 ± 20 cells/mm²; day 14 eluate: 424 ± 23 cells/mm²) or DPBS (day 2 eluate: 340 ± 24 cells/mm²; day 14 eluate: 389 ± 36 cells/mm²) (all p < 0.001) (group B vs. control, p < 0.001 for day 2 eluate, and p = 0.005 for day 14 eluate). Figure 4G shows that cell migration with vildagliptin and insulin was superior to insulin alone or DPBS alone (all p < 0.001). Figure 4H plots release curves in vitro of vildagliptin every day. The core-shell insulin/vildagliptin-eluting PLGA nanofibers delivered vildagliptin constantly for 30 days, with an rapid release of the beginning on day one (188 ± 20 μg/mL), and then continuously release (at least 16 μg/mL) until day 30. Figure 4I plots the daily release profile of insulin in vitro. Both core-shell nanofibers delivered stable and sustained insulin concentrations from day 1 (6.1 ± 1.2 vs. 5.7 ± 1.7mU/mL, p = 0.789) to day 14 (6.2 ± 1.0 vs. 6.8 ± 0.9mU/mL, p = 0.469). (See Supplementary Table S2). The augmented hydrophilicity of vildagliptin-loaded nanofibers may enhance insulin release concentration from day 1 to day 3 when compared with the non-vildagliptin-loaded group.

EPCs contribute to angiogenesis and vascularization during wound recovery stage but impaired EPCs recruitment in cases of diabetes (Loomans et al., 2004; Ambasta et al., 2017). Under hyperglycemic conditions, endothelial progenitor cells (EPCs) are functionally impaired in their ability to migrate, mobilize, and integrate/differentiate into existing vasculature (Saboo et al., 2016; Hu et al., 2018). DPP4 inhibitors, such as vildagliptin and sitagliptin, induce pro-angiogenic and anti-apoptotic effect in EPCs by improving glucose tolerance and restoring normal biologic functions (Fiordaliso et al., 2016; Liu et al., 2016; Lee et al., 2022c). Insulin treatment has also been shown to enhance EPC proliferation and their angiogenic activity (Zhao et al., 2011). Therefore, delivery of insulin and vildagliptin using core-shell nanofibrous membranes to promote diabetic wound healing may be due to increased EPC activity rather than increased circulating EPCs and their homing to the wound area. The internal structure and composition of nanofibers critically determine whether certain drug-release profiles can be achieved. The high porosity of the fibrous scaffolds and its large contact area for dissolution make nanofibers promising candidates for drug release (Braghirolli et al., 2014; Potrč et al., 2015). To prolong drug release, either core-shell nanofibers, sandwich-type nanofibrous meshes, or multi-drug-loaded layers can be used (Lee et al., 2020). Supplementary Figure S2 displays the results of FTIR measurements of the vildagliptin-loaded PLGA and pure PLGA core-shell nanofibers. The FTIR data of vildagliptin/PLGA nanofibers elucidated important wide spikes between 3,110 and 3,700 cm⁻¹, demonstrating N–H stretching and OH vibrations. The other vibrational peak at 2,250 cm⁻¹ was indicated to nitrile stretching vibrations (Harsha et al., 2015). An overlap of typical peaks was noticed, illustrating reasonable stability and incorporation of the vildagliptin during core-shell electrospinning of nanofibers using HFP solvent.

In summary, the biocompatible core-shell nanofibers that are prepared in this work had effective concentrations of vildagliptin and insulin for 30 and 14 days, respectively.

Figure 5 presents representative images of a healing wound on a diabetic animal from both treatment groups using different nanofibers on days zero, seven, and 14 after surgery with core-shell nanofibrous membranes. The wound regions in the two treatment groups were similar on the days zero (52.1 ± 2.2 vs. 52.4 ± 2.2 mm (Menke et al., 2015), p = 0.881). However, wounds in group A (insulin/vildagliptin-eluting PLGA core-shell membranes) were significantly smaller than those in group B (insulin/PLGA) on days seven and 14 (Day 7: 11.6 ± 2.2 vs. 24.4 ± 1.6 mm (Menke et al., 2015), p = 0.002; Day 14: 1.9 ± 0.3 vs. 8.0 ± 1.4 mm (Menke et al., 2015), p = 0.002) (Figure 5C). On the 0th, seventh, and 14th days, the average blood glucose values of group A were consistent with those of group B (all p > 0.05) (Supplementary Table S3).

Skin is a complex and dynamic organ, whose integrity following skin injury is restored in a process that depends on the interaction of fibroblasts, neovascularization, and biomacromolecules (Arthur and Luciano, 2000; Barrientos et al., 2008). The exogenous application of insulin has been demonstrated to stimulate cellular migration and diabetic wound recovery (Abdelkader et al., 2016; Lee et al., 2020). DPP-4 inhibitors favor diabetic wound healing by promoting the recruitment of EPCs and supporting the final scaffolding of wounds (Saboo et al., 2016). DPP-4 inhibitors also support angiogenesis and have a wide range of effects, including optimizing the immune response and decreasing the adverse effect of hypoxia in prolonged diabetes wounds (Huang et al., 2012). Therefore, dressing a diabetic rat wound with core-shell insulin/vildagliptin-eluting PLGA nanofibrous scaffolds healed faster than the core-shell insulin/PLGA group after 2 weeks.

The histological photos reveal that nanofibrous core-shell insulin/vildagliptin-eluting PLGA membranes (Figure 6A) increased cell proliferation in the dermis area than the insulin/PLGA nanofibers group (Figure 6B). Treatment with group A scaffold for 2 weeks showed a thicker epidermis in the cross-sectional wound than achieved in group B (56.3 ± 13.9 µm vs. 24.8 ± 4.9 µm, p = 0.005) (Figure 6C). The results of immunostaining for α-SMA expression in both groups (Figure 6D for group A; Figure 6E for group B) revealed a significant difference when normalized to those of DAPI nuclear staining (group A: 0.72 ± 0.23 vs. group A: 2.07 ± 0.37, p < 0.001).

Diabetic wound healing are typically associated with reduced granulation tissue thickness, delayed re-epithelialization, decreased infiltrated cells, reduced matrix density, and impaired angiogenesis (Broadley et al., 1988; El Gazaerly et al., 2013). High DPP4 levels in diabetic wounds impair wound healing by sustaining an inflammatory status (Baticic Pucar et al., 2017). The vildagliptin, one of DPP4 inhibitor, that was slowly and sustainably released from core-shell nanofibrous scaffolds improves wound healing herein may be due to the possible effect of increasing the migration rate of keratinocytes and fibroblasts to diabetic wound regions, promoting re-epithelialization of the wound (Saboo et al., 2016; Adeghate et al., 2017; Lee et al., 2022c).

α-SMA is the actin isoform that dominates in smooth-muscle cells and has an vital role in fibrogenesis (Storch et al., 2007). Myofibroblasts are morphologically and metabolically distinctive fibroblasts that express α-SMA, and their activation has a critical role in the expansion of the fibrotic tissue and the release of large amounts of extracellular component proteins (Langevin et al., 2010). The presence of more myofibroblasts (expressing α-SMA protein) later in wound recovery is a specific hallmark of impaired wound recovery, as most of them disappear during granulation cell apoptosis (Desmouliere et al., 1995; Toriseva et al., 2012). The experimental data in this study suggest that the biomolecule-loaded scaffold can greatly reduce the expression ofα-SMA protein by releasing vildagliptin and insulin. The use of nanofibrous scaffolds, which release vildagliptin and insulin, therefore accelerates wound closure and promotes the healing of diabetic wounds.

Despite its strengths and contributions, this study has certain limitations that warrant further investigation. First, many molecular and cellular response differences between acute and chronic wounds have been demonstrated (George Broughton et al., 2006; Enoch and Price, 2004). For example, chronic wounds often exhibit persistent inflammation, delayed re-epithelialization, suboptimal granulation tissue formation, and poor fibroblast infiltration (Martin and Nunan, 2015). Therefore, diabetic wounds are often used as one of the in vivo models of chronic wounds developed by subjecting acute wounds to the main clinical causes of chronic wounds (Loots et al., 1998; Blakytny and Jude, 2006). Second, when is the optimal period for wound management after diabetes induction? According to the literature, 2–3 days (Kishore et al., 2017; Siracusa et al., 2018), 7 days (Gallagher et al., 2007; Yuan et al., 2018), or 2–3 weeks (Lee et al., 2012; Ramachandran et al., 2012) following the induction have been used. Finally, after 7-day protocol to induce experimental diabetes, skin wounds may not have all the features of chronic lesions, and the imperative to conduct further research is needed in order to attain a more comprehensively understand of the mechanisms associated with the diabetic wound healing.