Pressure-Driven Spreadable Deferoxamine -laden Hydrogels for Vascularized Skin Flaps

Abstract

Hydrogels that support vascularization have been developed to improve the survival of skin flaps, yet establishing homogeneous angiogenic niches without compromising ease of use in a surgical setting remains a challenge. Here, pressure-driven spreadable hydrogels were developed utilizing beta-sheet rich silk nanofiber materials. These silk nanofiber-based hydrogels exhibited excellent spreading under mild pressure to form a thin coating to cover all regions of skin flaps. Deferoxamine (DFO) was loaded onto the silk nanofibers to support vascularization and these DFO-laden hydrogels were implanted under skin flaps in rats to fill the interface between the wound bed and flap using applied pressure. The thickness of the spread hydrogels was below 200 μm, minimizing physical barrier effects from the hydrogels. The distribution of the hydrogels provided homogeneous angiogenic stimulation, accelerating rapid blood vessel network formation and significantly improving the survival of the skin flaps. The hydrogels also modulated immune reactions, further facilitating regeneration of the skin flaps. Considering the homogeneous distribution at wound sites, improved vascularization, reduced barrier effects and low inflammation, these hydrogels appear to be promising candidates for use in tissue repair where a high blood supply is in demand. The pressure-driven spreading property should simplify the use of the hydrogels in a surgical setting to facilitate clinical translation.

Graphical Abstract

Pressure-driven spreadable DFO-laden silk nanofiber hydrogels were developed to form the coatings with a thickness of below 200 μm under the skin flap, providing homogeneous angiogenic stimulation without physical barriers.

1. Introduction

Different scaffolds and hydrogels with physical and chemical cues have been developed to fill tissue defects and to stimulate regeneration.^1-3 However, vascularization remains a key challenge for tissue regeneration.^4-6 Growth factors, small molecule drugs, and ions have been loaded into the matrices, providing niches for improved vascularization and tissue restoration.^7-9 Multiple physical regulators, such as micro-nano hierarchical structures, aligned topographies and mechanical cues have also been fabricated to facilitate angiogenesis in vivo, for more rapid tissue recovery.¹⁰ Although these vascularization strategies are effective for bone and dermis where blood vessels induce the ingrowth of new tissue growth without necrosis, other tissues with a high oxygen demand usually require faster vascularization to prevent necrosis under hypoxic conditions.^11-13

Skin flaps are widely used to cover skin defects in reconstructive surgeries.^14-16 Unlike the regeneration of the dermis, faster angiogenesis is required for skin flaps to reconnect with the wound bed in order to avoid necrosis.^{11, 17} Growth factor-laden hydrogels and scaffolds have been used to stimulate angiogenesis of tissue flaps through multiple injections, which increases time and cost in terms of clinical applications.^18-20 Recently, self-healing and injectable hydrogels were developed to promote neovascularization of skin flaps, yet multiple injections of the hydrogels were still necessary for the skin regeneration.^21-23 These hydrogels were also physical barriers for the wound beds and flaps, partly restricting angiogenesis of the flaps. Homogeneous angiogenic stimulation remains a challenge for these hydrogels. Spreadable hydrogel seems as a preferable option to overcome the issue, but little spreadable hydrogel without the unwanted flowability was developed in regenerative medicine. To overcome these problems, hydrogels with reduced barrier issues and improved vascularization capacity and clinical applicability are preferable. Cell survival is less of an issue when blood vessels and cells are generally below 200 μm apart.^24-26 Therefore, hydrogels that spread to form thin coatings are a good choice to help alleviate physical hindrance. Considering ease of use, pressure-driven spreading of hydrogels can facilitate clinical applications without a negative inefficient impact on workload in the clinic. However, to the best of our knowledge, the hydrogels with these anticipated performances are never reported in previous studies.

Silk fibroin (hereafter referred to as silk) has been used in skin engineering as a promising matrix due to its biocompatibility, biodegradability, low inflammation and tunable mechanical properties.^27-29 Silk scaffolds and hydrogels with nanofibrous-microporous hierarchical structures were developed to improve utility in skin regeneration, with the stiffness tuned to 4-7 kPa, and with capacity for vascularization.^{30, 31} Both vascular endothelial growth factor (VEGF) and small molecule deferoxamine (DFO) were loaded in the matrices and released slowly to stimulate the angiogenesis in vivo.^32-35 Recently, beta-sheet rich silk nanofibers (BSNF) were assembled to form shear-thinning hydrogels that offered improve loading capacity for bioactive molecules. Both hydrophobic and hydrophilic cargoes could be loaded on the nanofibers and exhibited better sustained release behaviors than previous silk-based systems.^36-39 Unlike collagen and chitosan hydrogels, these BSNF materials also exhibited a solution-hydrogel transition following increased silk concentration.^{1, 29, 40} The properties suggested a possible balance, where BSNF exhibits outstanding spreading capacity but remains in the hydrogel state. Considering its above advantages as wound matrices, the BSNF hydrogels are suitable choice for the skin flap. Unfortunately, the injectable BSNF hydrogels with different silk concentrations failed to achieve the spreadability without the compromise of flowability. Further treatment is required to tune the BSNF hydrogels, endowing them with pressure-driven spreading behaviors. Thus, the goal here was to design spreadable BSNF hydrogels with vascularization capacity, to accommodate the needs identified earlier, with stimulus of angiogenesis without the unwanted physical barriers for skin flaps.

To simplify clinical applications, pressure-driven spreading is a preferable option for hydrogels involved in skin flap surgery. The spreadable hydrogels with a thickness about 200 μm should also minimize physical barrier between the wound bed and the skin flap. Here, BSNFs with different silk concentrations were tune in terms of their rheological properties for spreadable behavior with applied pressure, forming thin coatings with tunable thickness. Previous studies suggested that the enhancement of vascularization and anti-oxidation is effective to promote wound healing and accelerate the repair of chronic wounds.^41-43 Deferoxamine (DFO) exhibited angiogenic and antioxidative capacities and could be released slowly from the BSNF-based hydrogels. Therefore, the DFO-laden hydrogels were optimized to improve the vascularization capacity.³² The DFO-laden hydrogels were then pressed to cover wound defects, and compared to other hydrogel systems.^{7, 18} Besides the cytocompatibility and vascularization capacity, the hydrogels were also implanted into skin flap models in rats and exhibited significantly more rapid angiogenesis and reduced necrosis versus controls without these features (Scheme 1).^{7, 32, 44, 45}

Scheme 1.

Overview of preparation and utility of DFO-laden BSNF hydrogels: after injection, the hydrogel spreads homogeneously under pressure, forming a thin coating with angiogenic capacity.

2. Materials and Methods

2.1. Preparation of BSNF hydrogels.

The BSNF hydrogels were assembled through a concentration-dilution-thermal incubation process.³² Raw silk fibers (70 g) was boiled in sodium carbonate solution (0.2%) at 100°C for 30 min to remove sericin proteins. After drying overnight, the degummed fibers were dissolved in LiBr solution (9.3M) at 60oC, dialyzed against water for 72 h, and centrifuged at 9,000 rpm for 20 min, obtaining aqueous SF solution with a concentration of about 6 wt%. The silk solution was concentrated for 3-5 d at 40°C to above 20 wt% and then diluted to 0.3-2 wt%. The diluted solutions were incubated at 60°C until hydrogel formation to transform into beta-sheet rich silk nanofibers.^{30, 46} The BSNF hydrogels with different concentrations were then stirred under 500 rpm for 4 h to tune the rheological behaviors.

2.2. Spreading property of BSNF hydrogels with various concentrations.

The pressure exerted by three surgeons on the skin during surgery was measured through a simulating surgery process. Silicone film having similar softness with skin was placed on the tray of the electronic balance. Then the surgeons pressed the silicone film vertically with similar pressure in the real operation. The pressure was measured and kept in the range of 490 g to 530 g. Therefore, we chose 500 g to simulate the pressure of the surgeon’s finger to assess the spreading properties of BSNF hydrogels. To visualize the spreading of BSNF hydrogels, the samples were stained with methylene blue. After the BSNF hydrogels (0.5 ml) were placed under the flaps or between two glass slides, the hydrogels were pressed at 500 g to measure the thickness and spreading area. The spread hydrogels were also imaged after the removal of the pressure to evaluate spreading stability. In order to test the volume change of the BSNF hydrogel after contact with the tissue fluid, the swelling test was also carried out in PBS solutions at 37 °C. Samples were soaked in PBS (pH7.4) at 37 °C for 0.5, 1, 2, 4, 8, 12 and 24h. At indicated time points, the samples were collected from the PBS solution, wiped the remaining solution on the surface of the samples with clean tissue paper, and then weighed. The swelling ratio (%) was obtained using the following equation.

$$\text{Swelling ratio}(\%)=(W_t\text{-}{W}_0)∕W_0$$

where W₀ is the initial weight of the hydrogel, W_t is the weight of the hydrogels in PBS at a certain time point.

The rheological behavior of BSNF hydrogels was measured with a rheometer (AR2000, TA Instruments, New Castle, DE).

2.3. Preparation and characterization of DFO-laden BSNF hydrogels.

Different amounts of DFO powders (Sigma Aldrich, Shanghai, China) were added to BSNF hydrogels (1 wt%) directly and stirred at 500 rpm for 4h. The DFO content in the hydrogels were tuned in the range of 30uM-180uM to optimize vascularization outcomes. All DFO-laden hydrogels were sterilized (60Coγ, 25 kGy) before cell culture in vitro and wound healing in vivo. Raman spectra were used to confirm the loading of DFO in the hydrogels, measured with a Confocal Raman spectrometer (LabRam HR800, HORIBA, France) with 633 nm excitation wavelength. The samples were assessed on a 20 mm cone plate (Ti, 20/1°) over a frequency range from 1 to 1000 rad s⁻¹ at 37oC. The viscosity of BSNF hydrogels and DFO-laden BSNF hydrogels against shear rates were analyzed for dynamic viscosity.^{32, 39} BSNF hydrogels were pushed through a needle (22 G) at 1.5 N to assess injectability.

2.4. Release behavior of DFO from hydrogels.

To investigate the release behavior of the DFO from hydrogels, 2 ml of DFO-laden BSNF hydrogels were transferred into a dialysis tube (Thermal Scientific, MWCO 3500) and soaked in 10 ml of phosphate buffered saline (PBS) solution for 40 days. At designed time points, 1 mL was collected and immediately replaced with an equal amount of fresh PBS. The collected supernatants were combined with ferric chloride (6 mM) at volume ratio of 1:1. The DFO release was obtained by detecting absorbance at 485nm with multiscan spectrometer (Biotek, USA) as previously reported.⁴⁷

2.5. Cell culture.

Endothelial cells were cultured in vitro to assess the cytocompatibility of the hydrogels. Human umbilical vein endothelial cells (HUVECs) were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The CCK-8 assay was used to evaluate the proliferation of HUVECs on the hydrogels. The cells were seeded on the surface of DFO-laden BSNF hydrogels with different amounts of DFO (DFO concentrations, 30uM-180uM) at 5×103 cells using 96-well plates. As a control, the cells were also cultured in the medium with an equal amount of free DFO and on the surface of DFO-free BSNF hydrogels. After culturing for 1, 2 and 3 days, the cells were incubated with CCK-8 solution in a cell incubator for 2 h and the absorbance at 450 nm was measured with multiscan spectrometer.^{48, 49}

2.6. Tube formation on the DFO-laden BSNF hydrogels.

The vascularization capacity of DFO-laden BSNFs was evaluated in vitro by a tube formation assay. Matrigel (500uL/well, BD Biosciences, USA) was coated on the BSNF hydrogels in a 24-well plate according to reported work⁴⁷ HUVECs were cocultured on the surface with cell density of 3×10⁴ cells/each well. After culturing for 12h, tube formation of HUVECs was observed with an inverted microscope (Axio Vert.A1, Carl Zeiss, German) and analyzed with Image J.

2.7. In vivo flap model.

All animal procedures were performed in accordance with the guidelines for care and use of laboratory animals of Soochow University and approved by the animal ethics committee of Soochow University. Rats used were maintained under specific pathogen-free conditions (SPF). Twenty-four Sprague-Dawley (SD) adult male rats (weight 200 ± 50 g) were randomly divided into three groups (n=8): Control group, DFO-free BSNF hydrogel group and DFO-laden BSNF hydrogel group (DFO, 60 uM). The rats were anesthetized with 4% chloral hydrate (intraperitoneal injection,1mL/100g), followed by having the hair shaved from the back with an electric clipper. The random flap model was also used in our spreadable hydrogels, but excellent flap survival (near 100%) appeared for both of DFO-free and DFO-laden hydrogels (data not shown). Different flap model with less blood supply is required for our present study. The iliolumbar vessels are often attached in the random flap, which then influences the necrotic ratio. Recently, a pingpong-shape flap with narrower width and longer length was designed to avoid vascular penetration. The repeatable distal necrosis is obtained for the pingpong-shape flap, suggesting its advantages in evaluating the survival of skin flap.^{50, 51} Therefore, a pingpong-shape flap model was used in the present study to reveal the function of the hydrogels in stimulating the blood vessel formation and reducing the necrosis. Intraoperative observation of the pingpong-shape flap revealed the blood vessel free state of the pedicle.⁵¹ A caudally-based dorsal flap with racket shape (pedicle: 1.0 × 3.0 cm, flap: d=3.0 cm) was incised with scalpel. The full-thickness skin was raised to separate from the deeper muscular fascia, leaving an un-incised pedicle (width, 1 cm). No axial vessel passed through the pedicle. The hydrogels were pushed under the flap with a syringe, following by suturing the flap closed. The sutured flaps were pressed by hand to spread the hydrogel, ensuring homogeneous distribution of the hydrogels under the flaps (Fig. S1). Finally, the flaps were covered with sterile gauze. The rats were housed individually without antibiotics.

2.8. Evaluation of flap survival and blood perfusion.

At 1, 3 and 7 days post-operation, photographs of the flaps were taken and analyzed with a digital camera (Canon 5D Mark IV, Japan). The necrotic areas were evaluated through the color, temperature, capillary reaction and elasticity of the flap using standard methods, evaluated individually by two dermatologic pathologists using a double-blind method.^{11, 44, 52} The percentage of skin necrosis was calculated through dividing the necrotic area by the round area. Blood perfusion of the flap was measured with the infrared thermal imager (Ti450, Fluke, China) after 3 and 7 days post-operation.

2.9. Evaluation of neovascularization and inflammation.

After 7 days post-surgery, the rats were euthanized for flap sample collection. The flaps and adjacent tissues were harvested and fixed with 4% paraformaldehyde. The samples were stained with Hematoxylin and Eosin (H&E) to evaluate survival and angiogenesis. Immunofluorescence staining for CD31 (1:100 dilution, Ab28364, Abcam) was used to evaluate the blood vessel formation and CD68 and CD206 immunofluorescence staining was performed to investigate inflammatory responses.

2.10. Statistical analysis.

The data were statistically analyzed by SPSS v.16.0 software. One-way AVOVA was used to compare the mean values of the data. p < 0.05 was considered statistically significant.

3. Results and Discussion

Multiple strategies have been developed to design hydrogels with vascularization capacity for reducing necrosis.^{17, 21, 53, 54} The use of hydrogels are often incompatible with flap surgery, thereby, limiting clinical applications. Although previous hydrogels containing vascularization stimulators promoted angiogenesis and reduced necrosis, these hydrogels were also physical barriers for the wound beds and flaps, partly restricting angiogenesis of the flaps.^{42, 53} Multiple injections of hydrogels also resulted in heterogeneous outcomes in terms of skin flap regeneration.^{20, 21, 54} Ideally, hydrogels with vascularization cues should distribute homogeneously at the interface of the wound bed and the flap, forming a thin coating with a thickness below 200 μm to minimize barrier influences.

BSNFs composed of beta-sheet rich conformation had a high negative charge density, providing shear-thinning properties and solution-hydrogel transitions at different silk concentrations.^{27, 28, 32, 46, 56} Spreadability of the hydrogel could facilitate its homogeneous distribution in skin flap, but also results in the flowability that worsens the function. It remains a great challenge to develop spreadable hydrogels without the unwanted flowability. To reveal the balance of spreading and adhesion of the BSNF materials, the BSNFs assembled under different silk concentrations in the range of 0.3-2%. Although the injectable DFO-laden silk hydrogels have been developed in our previous studies to accelerate the wound healing,³² BSNF hydrogels with different silk concentrations failed to achieve the spreadability without the compromise of flowability. The negative charge repulsion could also rearrange the aggregation of the BSNFs after the stirring treatment, which further tuned the flowability and spreading capacity of the hydrogels. Therefore, solid BSNF hydrogels were treated with stirring process to change the rheological behaviors. After the treatment, the pressure-driven spreadable hydrogels without the unwanted flowability were prepared, which would facilitate the application in the skin flap (Fig. S2). The pressure of the surgeon’s finger during surgery is usually in the range of 490-530 g (Fig. S3), the BSNF hydrogels were pressed at 500 g to evaluate spreading and adhesion (Fig. 1a,b and c). For the BSNFs with concentrations below 1%, flowability resulted in the loss of BSNFs after pressing. When the SF concentration was increased to 1%, the BSNF hydrogels easily spread and formed thin coatings (Fig. 1d). The coatings remained stable without shrinking or contraction after the removal of the pressure, suggesting that the hydrogels would distribute at the interface of the wound bed and flap in a homogeneous fashion. The thickness of the coating was investigated to evaluate the minimization as a physical barrier, and hydrogels with silk concentration of 1% formed a coating with about 178.6±11.3μm (below 200μm) (Table 1). Swelling behaviors of the hydrogels were studied to evaluate the volume changes of BSNF hydrogel in vitro. The results showed that BSNF hydrogel remained stable without significant volume changes (< 10%) in the PBS solution (Fig. S4). Similar to the BSNF hydrogels without stirring treatment, the treated hydrogels could be injected with needles (22 G) easily, suggesting its injectability (Fig. S5).Therefore, the hydrogels with a silk concentration of 1 wt% were used to load DFO to promote angiogenesis of the flaps.

Fig. 1.

(a) Schematic diagram of pressure driven hydrogels under 500 g pressure in vitro; (b) 0.5ml BSNF hydrogels were spread to form uniform thin coatings at 500g pressure between glass plates; (c) Schematic diagram of 10 cm² full-thickness skin defect on the back of rats. After 0.5ml BSNF materials with different concentrations were injected on the wound, the skin was reapplied to cover the wound bed and pressed to test the spreading of the hydrogels; (d) Spreading and adhesion properties of the BSNF materials with different concentrations. Silk nanofiber hydrogels were marked with methylene blue. (→ means the hydrogels overflowed from the edges of the skin, Δ means a high degree of fluidity of the hydrogels, ☆ means shrinking of the hydrogels after the removal of pressure)

Table 1.

The area and thickness of coating formed between two glass slides under 500 g pressure. Every sample was measured five times, the difference between every two groups were significant ( p< 0.05).

Silk (%)	Coating (cm2 ± SD)	Thickness (μm ± SD)
0.3	78.5±2.5	63.7±16.7
0.5	63.6±1.9	78.6±12.9
0.7	36.2±2.3	138.1±15.2
1.0	28.0±1.8	178.6±11.3
1.5	13.8±1.7	361.3±9.2
2.0	9.6±1.2	519.8±5.7

To optimize the cytocompatibility and vascularization capacity of the hydrogels, different amounts of DFO were loaded in the BSNFs directly for in vitro cell culture. Raman spectra of DFO-laden hydrogels had typical peaks of DFO and silk ( 584 cm⁻¹, 1050 cm⁻¹, 1350 cm⁻¹, 1450 cm⁻¹ for DFO and 1064 cm⁻¹, 1232 cm⁻¹, 1665 cm⁻¹ for silk), confirming the successful loading of DFO (Fig. 2a). Similar to pure BSNF hydrogels, the DFO-laden hydrogels retained the shear-thinning behavior and exhibited a slight increase of viscosity. The hydrogels with different amounts of DFO had similar viscoelasticity without significant differences (Fig. 2b). The results suggested that the introduction of DFO did not influence the spreading. Different amounts of DFO were released slowly from the hydrogels over 40 days, suggesting sustained vascularization stimulation could be achieved with the hydrogels (Fig. 2c and d). Previous study revealed that the sustained release of DFO from BSNF hydrogels stimulated the quick angiogenesis at dermal wound sites, superior to other DFO-laden biomaterial systems.^{32, 47} Therefore, these BSNF hydrogels with vascularization potential were developed without compromising the original, beneficial mechanical properties.

Fig. 2.

Characterization of BSNF hydrogels: (a) Raman spectra; (b) viscosity of the hydrogels before and after loading DFO; (c, d) DFO release properties in vitro of the 1% BSNF hydrogels in 3 days and 42 days; (e) cell viability of HUVECs for 3 days; (f) tube formation of HUVECs at 12 h, scale bars are 300 μm; (g, h) tube length of HUVECs. Statistically significant *P ≤ 0.05 compared with control group, #P ≤ 0.05 compared with cells cultured in free DFO solutions of the same concentration.

DFO-laden hydrogels have been used to stimulate dermal regeneration, considering vascularization and cytocompatibility, BSNF hydrogels with 120 μm DFO were considered the better matrices for the dermal defect repair.³² Unlike the dermis application in our prior studies, the skin flap model requires hydrogels with improved biocompatibility and vascularization for faster re-connection with the wound bed. Human umbilical vein endothelial cells (HUVECs) were cultured on different DFO-laden hydrogels to assess the cytocompatibility and vascularization capacity in vitro. Free DFO exhibited significant toxicity when the concentration of DFO was above 30 μM. When the cells were cultured for 3 days, the viability of HUVECs was below 40%. These results suggested that direct injection of DFO was inappropriate for skin flap regeneration. The loading of DFO in BSNF hydrogels significantly attenuated the toxicity of DFO, achieving higher cell viability on the DFO-laden hydrogels (Fig. 2e). However, significant cell cytotoxicity remained for the DFO-laden hydrogels when the DFO amounts were above 60μM where the viability of HUVECs were below 60% after cultured for 3 days. More than 75% of the HUVECs remained viable on the hydrogels at DFO concentrations of 60 μM, suggesting the acceptable cytocompatibility of the DFO-laden hydrogels. DFO as an effective angiogenic drug has been used to facilitate the blood vessel formation in various tissue regeneration studies in vitro and in vivo.^{47, 57-59} Since the vascularization capacity of DFO-laden hydrogels has been investigated extensively in our recent study,³² the in vitro vascularization of the hydrogels with different amounts of DFO was evaluated through tube-formation assay. Although the longest tubes from the HUVEC appeared in the hydrogels with DFO concentrations of 120 μM, many tubes also formed in the hydrogels with the DFO concentration of 60 μM (Fig. 3f, g and h), showing slight decrease of vascularization (within 10%). Considering acceptable cytocompatibility and comparable vascularization capacity, the BSNF hydrogels containing 60 μM DFO were selected as the better matrices to stimulate vascularization of skin flaps and were subsequently used in the in vivo rat models.

Fig. 3.

Skin flap regeneration treated with various samples: (a) Digital photographs of the flaps at days 1, 3, and 7 postoperation; (b) blood perfusion of the flaps measured with infrared thermal imager laser at days 0, 3, and 7 post-operation; (c) quantitative analysis of flap necrotic via ImageJ software at day 7. Statistically significant *P ≤ 0.05 compared with control group, #P ≤ 0.05 compared with SF group. (d) the detached flaps at day 7 from different group; (e) Blood vessels in the base and H&E stained of the flaps at 7 days, scale bars are 100 μm. (Δ means inflammation of the flap, ☆ means hair follicles, sweat glands and other skin appendages, → means microvessels)

The random flap model in rats is usually used to evaluate the influence of the hydrogels with vascularization capacity for flap regeneration.^{11, 54, 60} Necrosis appears at the distal end of the flap because of inferior blood supply. Although the quality of the regenerated flap could be evaluated based on angiogenesis density and necrosis area, more objective evaluation remains a challenge since little necrosis formed in the random flap model without hydrogel treatment. ^{41, 61} The random flap model was also used in our spreadable hydrogels, but excellent flap survival (near 100%) appeared for both of DFO-free and DFO-laden hydrogels (data not shown). Different flap model with less blood supply is required for our present study. The iliolumbar vessels are often attached in the random flap, which then influences the necrotic ratio. Recently, a pingpong-shape flap with narrower width and longer length was designed to avoid vascular penetration. The repeatable distal necrosis is obtained for the pingpong-shape flap, suggesting its advantages in evaluating the survival of skin flap.^50,51 Therefore, a pingpong-shape flap model was used in the present study to reveal the function of the hydrogels in stimulating the blood vessel formation and reducing the necrosis. Intraoperative observation of the pingpong-shape flap revealed the blood vessel free state of the pedicle (Fig S6). Developing a modified flap model with less blood supply is necessary to reveal the function of the hydrogels in stimulating the blood vessel formation and reducing the necrosis. Therefore, a pingpong-shape flap model in rats was used in the present study. ^{50, 51} Through tuning the length of the rectangle area, more necrosis formed and occupied most of the round area when the length of the rectangle was 5 cm (Fig. S6). Since the function of hydrogels used could be assessed easily through the ratio of necrosis in the round zone, the pingpong-shape flap is a better model to study the stimulating function of the hydrogels in skin flap regeneration. Therefore, the DFO-laden hydrogels were implanted under the modified pingpong-shape flap (10 cm²) and spread by pressing the flap.

The in vitro pressure of the hydrogels with glass panels revealed that 0.3 ml of hydrogels (1 wt%) could occupy 10 cm² area, forming a thin coating with thickness of about 178 μm. Various volumes of hydrogels were implanted under the flap to determine a suitable amount of the hydrogel for the flap model in vivo and above 0.5 ml was identified (Fig. S7). H&E images further revealed that the thickness of the hydrogel at the interface of the wound beds and flaps were 120 μm,160 μm, 240μm, and 470 μm after 0, 0.5 ml, 1 ml and 2 ml of the hydrogels were implanted under the flap, respectively (Fig. S7). The necrosis area increased with the higher amounts of hydrogel, reflecting the impact as a physical barrier effect (above 200 μm). ^{24-26, 62} Therefore, 0.5 ml of hydrogel could cover the flap area with suitable thickness and was used to induce angiogenesis in vivo.

DFO-laden hydrogels with different amounts of DFO (0.5 ml) were also implanted under the flap of the rats, and the best viability of the flaps was achieved for the groups with DFO concentration of 60 μM (Fig. S8). Therefore, the DFO-laden hydrogels with DFO concentration of 60 μM were used in the in vivo studies. To reveal the stimulating effect of the DFO-laden hydrogels on angiogenesis, the same amount of normal saline and DFO-free BSNF hydrogels (0.5 ml) were used as controls and implanted under the flaps. Compared to the blank control, the laser speckle contrast images showed significantly lighter color in the flap area for both the DFO-laden and DFO-free BSNF hydrogel groups after the surgery (Fig. 3a and b). The results confirmed that the hydrogels occupied the interface between the wound bed and the flap, in contrast to the punctate distribution in previous hydrogel systems.^{9, 21, 23} At day 3 after the surgery, almost all of the round area of the flap in the blank group showed a lighter color than the surrounding regions, while about 25% of the round part in the DFO-free hydrogel group remained same color with normal tissue.The results indicated that BSNF itself stimulated angiogenesis, resulting in improved blood flow under the flap. The vascularization capacity was further improved after the introduction of DFO. All of the flap region retained a similar color with the surrounding tissues, indicating the best angiogenesis in the DFO-laden hydrogel group. After 7 days, the color difference become more evident where the best result was achieved in the DFO-laden hydrogel group (Fig. 3a). Necrosis and scabbing were consistent with the laser speckle contrast images. At day 7 after operation, the necrosis area in the blank group was 75% and then decreased to 60% in the DFO-free hydrogel group. The vascularization capacity of the DFO-laden hydrogels reduced necrosis. Only 30% of the round region was occupied with necrosis and 70% of the flap successfully survived and had same appearance with the normal surrounding tissues (Fig. 3c). Although it is difficult to compare with previous hydrogel systems due to different flap models,^18-20 the DFO-free hydrogel has achieved comparable flap survival in the random flap model to that of previous reported works, but was inferior to the DFO-laden hydrogel group in the pingpong-shape flap model, suggesting possible better performances of the DFO-laden hydrogels. The results suggested that the DFO-laden hydrogels stimulated blood vessel formation and improved flap survival.

The skin flaps were peeled off at day 7 after the implantation to reveal blood vessel formation (Fig. 3d). Few blood vessels appeared under the flap in the blank group, indicating inferior angiogenesis. New blood vessels formed under the flap treated with DFO-free BSNF hydrogels, confirming the promoting effect of silk nanofibers on neovascularization. The loaded DFO further stimulated rapid angiogenesis. Rich blood vessel networks were distributed homogeneously under the flap in the DFO-laden hydrogel group, which further improved the survival of the flap. H&E staining was used to evaluate micro-vessel formation at the interface of the wound bed and the flap (Fig. 3e). Few fragile micro-vessels appeared under the blank control flap, which confirmed the inferior blood supply. No SF aggregates were observed below the flap treated with DFO-free hydrogels, suggesting that the nanofibers were distributed homogeneously without aggregation. Although the number of vessels increased in the DFO-free hydrogel groups than that of control, significantly higher neo-vessels were further regenerated in the DFO-laden hydrogel group (Fig. 4a). Different regions of the round area were investigated and showed similar angiogenesis. The results revealed that the implanted DFO-laden hydrogels provided homogeneous angiogenic stimulating niches for the skin flaps, an improvement over previous hydrogel systems.^{8, 20, 21} CD 31 immunohistochemical staining images further clarified the density and morphology of the formed vessels. The DFO-laden hydrogel group showed 108 vessels per cm2, 3 times higher than in the blank group and 1.5 times higher than in the DFO-free hydrogel group. The diameters of the main vessels in the DFO-laden hydrogel group were about 8.6 μm, 1.8 times larger than in the DFO-free hydrogel group, suggesting better and mature blood vessel networks in the DFO-laden hydrogel group (Fig. 4a and c). Therefore, the DFO-laden hydrogels provided the anticipated angiogenic stimulation.

Fig. 4.

(a) immunofluorescence staining of neovascularization of different treatment groups, CD31 (red) and DAPI (nuclei; blue). (b) Immunoregulation at flap sites when treated with different groups for 7 days. Macrophages are marked with pan-maker CD68 (green), M2 phenotype was marked with marker CD 206 (red); (c) the microvessel density (MVD) and mean vessel diameter of CD31+ vessels; ( d ) CD68+ macrophages number per mm², (e) percentage of M2 phenotype macrophages over total CD68+ macrophages. Data presented as mean ± SD, n = 6, *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001.

Previous hydrogels with vascularization capacity usually induce inflammation due to foreign body reactions, weakening vascularization in the skin flap. ^63-65 Compared to other materials, silk has a lower inflammatory reaction after implantation, strengthening biomedical utility in tissue regeneration.^66-69 DFO regulated inflammatory reactions and angiogenesis to accelerate the healing of diabetic chronic wounds.^{9, 70-73} Thus, DFO-laden SF hydrogels could minimize foreign body reactions and further improve survival of skin flaps. The macrophage density under the flap was measured with immunohistochemical staining (Fig. 4b). Compared with the blank control group, the number of macrophages was significantly decreased following the use of BSNF hydrogels and then the introduction of DFO. Besides the lower macrophage density, the DFO-laden hydrogels modulated the M1-M2 phenotype of macrophages. A significantly higher ratio of M2 macrophages appeared in the DFO-laden hydrogel group compared to both the blank and DFO-free hydrogel groups (Fig. 4d and e). Therefore, the DFO-laden hydrogels helped to tune the inflammatory behavior, which further accelerated skin flap regeneration. Unlike previously reported hydrogels with vascularization capacity^{21, 54}, the present work developed a pressure-driven spreadable hydrogel to provide homogeneous angiogenic stimulation for skin flaps. The hydrogels stimulated angiogenesis and tuned inflammation, significantly improving survival of the flap. The spreadable property minimized the physical barrier effect of the hydrogel on tissue in growth.

4. Conclusions

The spreading and adhesion properties of BSNF hydrogels were tuned by changing SF concentration to achieve spreadable hydrogels suitable for skin flap regeneration. DFO was loaded on the BSNF nanofibers to endow the hydrogels with vascularization capacity. The hydrogels occupies the flap region through simple pressing to provide homogeneous angiogenic stimulation, improving the survival of the flap. The thin hydrogel coatings that formed minimized physical barrier effects and modulated inflammation behavior, further accelerating flap regeneration. The vascularization capacity, inflammation modulation and spreading properties suggest that these DFO-laden hydrogels are promising candidates for flap regeneration and also suitable for other tissues with high blood supply demand.