Injectable cinnamon-based gel that promotes new blood vessel growth and skin healing in diabetic wounds

Zinc acetate (Zn(CH₃COO)₂, ≥99%) was purchased from Aladdin Reagent Co., Ltd. (Shanghai, China). Cinnamaldehyde (CA, ≥98%) was obtained from Sigma-Aldrich (St. Louis, MO, United States). GA, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC), and N-hydroxysuccinimide (NHS) were also purchased from Sigma-Aldrich. N,N-Dimethylformamide (DMF) was supplied by Macklin Biochemical Co., Ltd. (Shanghai, China). Microbial transglutaminase (TG, 200 U/g), derived from Streptoverticillium mobaraense, was obtained from Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China).

ZIF-8 nanoparticles were synthesized by a simple precipitation method (Ding et al., 2023). Briefly, 100 μL of zinc acetate dihydrate solution (90 mg/mL) was mixed with 900 μL of 2-methylimidazole solution (311 mg/mL), followed by the addition of 200 μL of reverse osmosis (RO) water. The mixture was stirred vigorously for 5 min at room temperature. The resulting suspension was centrifuged at 12,000 rpm for 10 min, and the precipitate was washed twice with RO water. The final product was dried at 37 °C to obtain ZIF-8 nanoparticles. To prepare CA-loaded ZIF-8 (CA@ZIF-8), CA was introduced at different concentrations (150, 300, 600, and 1,200 μg/mL) into the reaction mixture during ZIF-8 synthesis, following the same procedure as described above.

The morphology and size of ZIF-8 and CA@ZIF-8 nanoparticles were examined by transmission electron microscopy (TEM, JEOL JEM-2100, Japan) and scanning electron microscopy (SEM, JSM-7041F, JEOL, Tokyo, Japan). For TEM analysis, samples were dispersed in ethanol, dropped onto carbon-coated copper grids, and dried at room temperature. For SEM imaging, dried powder samples were mounted on metal stubs using conductive carbon tape and sputter-coated with a thin layer of gold prior to observation. Hydrogel pore size from SEM. Pore openings on the hydrogel surface were measured from top-view SEM micrographs using ImageJ (images calibrated to the embedded scale bar). Diameters were obtained by manual delineation of individual pores and reported as representative values. Fourier-transform infrared (FTIR) spectra were recorded using a Nicolet 5700 (Thermo Fisher Scientific, Waltham, United States) in the range of 4,000–500 cm⁻¹. Dried powder samples were mixed with KBr and pressed into pellets for measurement. The crystalline structures of the nanoparticles were analyzed by X-ray diffraction (XRD, X'pert PRO, PANalytical, Almelo, Netherlands) using Cu Kα radiation with a scanning range of 5°–70° (2θ). X-ray photoelectron spectroscopy (XPS) was performed using an AXIS Ultra DLD spectrometer (Kratos Analytical, United Kingdom) equipped with a monochromatic Al Kα X-ray source to analyze the surface elemental composition and chemical states. High-resolution spectra were recorded for C 1s, N 1s, O 1s, and Zn 2p, and peak deconvolution was carried out using XPS peak 4.1 software.

GGA was synthesized following a previously reported protocol with minor modifications (Li G. et al., 2023). Briefly, gelatin (5 g, from porcine skin, Type A, CAS No. 9000-70-8, Sigma-Aldrich VETEC) was dissolved in 150 mL of deionized water at 40 °C. Separately, GA (1.7 g, 10 mmol) was dissolved in a 3:2 (v/v) mixture of deionized water and DMF (125 mL total volume). EDC (1.91 g, 10 mmol) and NHS (1.6 g, 13.9 mmol) were added to activate the carboxyl groups of GA under stirring at 25 °C for 1 h, adjusting the pH to 4.5. The activated GA solution was then added to the gelatin solution and stirred overnight at 40 °C. The resulting mixture was dialyzed against deionized water for 7 days and subsequently lyophilized to obtain the GGA polymer.

The synthesized CA@ZIF-8 nanoparticles were incorporated into the GGA hydrogel matrix to fabricate nanocomposite hydrogels. Briefly, GGA (10 wt%) was dissolved in deionized water at 40 °C to form a uniform hydrogel precursor solution. CA@ZIF-8 nanoparticles were dispersed in RO water via mild sonication and added to the GGA solution under continuous stirring to achieve final nanoparticle concentrations of 2 mg/mL. TG (10 U/mL) was then introduced to initiate enzymatic crosslinking. The mixture was poured into molds and incubated at 37 °C for 5 min to form hydrogels.

The sol-gel transition behavior of hydrogels was evaluated using the vial inversion method. For the torsion test, hydrogel samples (10 × 100 × 1 mm³) were clamped at both ends with tweezers and twisted 360° for 5 min. Sample morphology and integrity were recorded photographically. Tensile testing was conducted on rectangular hydrogel strips (10 × 100 × 1 mm³) using a universal testing machine (AG-IS, Shimadzu, Japan) at a stretching rate of 5 mm/min. For compression tests, cylindrical samples (8 mm diameter, 10 mm height) were compressed at a rate of 5 mm/min using the same machine. The compressive modulus was determined from the slope of the linear portion of the stress-strain curve, and the compressive strength was defined as the maximum stress value. Each test was repeated on four samples, and representative curves were shown. The gelation kinetics and viscoelastic properties of the hydrogels were evaluated using a HAAKE rheometer (Malvern, United Kingdom) at 37 °C. GGA precursor solutions were mixed with 0.5 wt% TG and CA@ZIF-8 nanoparticles. The mixtures were gently stirred and immediately loaded onto a 20 mm quartz plate with a 1 mm gap. Time sweep tests were conducted at 1% strain to monitor the sol–gel transition. Frequency sweeps were subsequently performed on fully crosslinked hydrogels over the range of 0.1–10 Hz to characterize their viscoelastic behavior.

The internal morphology of freeze-dried hydrogels was examined by SEM using the same instrument and sample preparation protocol as described for nanoparticle analysis. FTIR spectra were recorded using the same Nicolet 5700 spectrometer to analyze the functional groups in different hydrogel formulations. XRD patterns were collected on the same X'pert PRO diffractometer to evaluate the crystalline structure of the composites. XPS analysis was performed using the same AXIS Ultra DLD spectrometer, including high-resolution scans for C 1s, N 1s, O 1s, and Zn 2p to investigate surface elemental composition and chemical bonding states.

Cylindrical hydrogels (8 mm diameter, 10 mm height) were immersed in PBS (pH 7.4) at 37 °C for 21 days to assess degradation behavior. At days 1, 4, 7, and 14, samples were lyophilized and weighed. Mass retention was calculated as (W₁/W₀) × 100%, where W₀ and W₁ represent the dry weights before and after immersion, respectively.

For ion and drug release analysis, four hydrogels were incubated in 3 mL PBS, with the medium refreshed at each time point. Zn²⁺ concentrations were measured by ICP-OES (Agilent 730, United States), and CA levels were determined using UV–vis spectrophotometry (UV-2600, Shimadzu, Japan) at 290 nm, based on a PBS-derived standard curve. All experiments were conducted in triplicate.

Human umbilical vein endothelial cells (HUVECs) were obtained from Procell Life Science & Technology Co., Ltd. (Wuhan, China; catalog no. CP-H082) and cultured in endothelial cell medium (ECM; Procell) supplemented with 5% fetal bovine serum (FBS), 1% endothelial cell growth supplement (ECGS), and 1% penicillin-streptomycin. Cells were maintained in a humidified incubator at 37 °C with 5% CO₂. HUVECs at passages 3–6 were used for all subsequent experiments. Mouse fibroblast L929 cells were purchased from Procell Life Science & Technology Co., Ltd. (Wuhan, China; catalog no. CL‐0137) and cultured in Dulbecco's Modified Eagle Medium (DMEM; Gibco, USA) supplemented with 10% FBS and 1% penicillin‐streptomycin. Cells were maintained at 37 °C in a humidified incubator with 5% CO₂, and passages 3‐8 were used for the experiments.

HUVECs were seeded onto hydrogel-coated 24-well plates at a density of 2 × 10⁴ cells/well and cultured in endothelial cell medium (ECM) supplemented with 10% FBS for 24 h. Live and dead cells were stained using Calcein-AM (2 μM) and propidium iodide (PI, 4 μM) (Beyotime, China) according to the manufacturer's instructions. Fluorescence images were captured with an inverted fluorescence microscope (Nikon Ti2, Japan).

For cytoskeletal staining, HUVECs cultured on different hydrogels for 24 h were fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 for 10 min, and stained with rhodamine–phalloidin (50 μg/mL, Sigma) for 30 min at room temperature. Nuclei were counterstained with DAPI (5 μg/mL, Sigma) for 5 min. Fluorescence images were acquired using a confocal laser scanning microscope (Zeiss LSM 880, Germany) under identical exposure settings across all groups.

After 24 h incubation with hydrogel extracts, HUVECs were fixed with 4% paraformaldehyde and blocked with 1% BSA for 1 h. Cells were incubated overnight at 4 °C with primary antibodies against CD31 (1:200, Abcam, ab28364) or VEGF (1:200, Abcam, ab46154), followed by Alexa Fluor–conjugated secondary antibodies (1:500, Invitrogen, A-11001/A-11008) for 1 h at room temperature. Nuclei were counterstained with DAPI (5 μg/mL, Sigma, D9542). Images were captured using a confocal microscope (Zeiss LSM 880, Germany) with identical exposure times for all groups. Fluorescence intensity was quantified using ImageJ software, and values were normalized to the control group (set as 1.0).

A Transwell assay was conducted using 8 μm pore-size inserts (Corning, United States). HUVECs (1 × 10⁵) suspended in serum-free medium were seeded in the upper chamber, and 600 μL of hydrogel extract medium was added to the lower chamber. After 12 h, migrated cells on the lower surface were fixed, stained with 0.1% crystal violet, and counted under a microscope in five random fields.

Matrigel (Corning, United States) was added to 96-well plates (50 μL/well) and incubated at 37 °C for 30 min. HUVECs (1 × 10⁴/well) were seeded in hydrogel extract–containing medium and incubated for 6 h. Tubular structures were imaged and quantified using ImageJ (Angiogenesis Analyzer plugin) for branch number and total tube length.

All animal experiments were approved by the Institutional Animal Care and Use Committee of Shenzhen Peking University–The Hong Kong University of Science and Technology Medical Center (Approval No. 2025-525). Male Sprague–Dawley (SD) rats (8 weeks old, 220–250 g) were rendered diabetic by intraperitoneal injection of streptozotocin (STZ, 50 mg/kg) for five consecutive days. Rats with fasting blood glucose levels exceeding 16.7 mmol/L for three consecutive days were considered diabetic and included in subsequent experiments. Full-thickness wounds were created 4 weeks after STZ injection, by which time the diabetic state was stable. Blood glucose levels were monitored weekly throughout the wound healing observation period, and remained consistently above 16.7 mmol/L without significant fluctuations among groups.

Under anesthesia, a full-thickness circular skin wound (2 cm in diameter) was created on the dorsal area of each rat using a sterile biopsy punch. The animals were randomly assigned to four groups (n = 5): Control (PBS), GGA, ZIF-8/GGA, and CA(0.6)@ZIF-8/GGA. Each wound was topically treated with 200 μL of the respective hydrogel and covered with a sterile occlusive dressing (Tegaderm™).

Wound healing was documented photographically on days 0, 4, 8, and 12 under standardized bright-field conditions (fixed camera–wound distance, identical illumination intensity, and locked exposure/white balance; a 1-cm scale bar was included in every image). Wound area was measured in ImageJ by manual tracing of the wound boundary and expressed as A_t/A₀ ×100%; data are reported as mean ± SD (n = 6 per group). A standardized healing score (0–4) was assigned based on macroscopic appearance, re-epithelialization, and hair regrowth, using the following rubric: 0 = no closure, moist surface/exudate; 1 = ≤25% closure with persistent exudate/necrosis; 2 = 26–50% closure with partial re-epithelialization; 3 = 51–75% closure with obvious re-epithelialization/granulation and minimal exudate; 4 = ≥90% closure with continuous epithelium, dry surface, minimal/no exudate, and initial hair regrowth. Full criteria are provided in Supplementary Table S1.

To evaluate systemic biocompatibility, major organs including the lung, liver, heart, brain, spleen, and kidney were harvested from each group on day 12 post-treatment. Tissues were fixed in 4% paraformaldehyde, embedded in paraffin, sectioned at 5 µm thickness, and stained with hematoxylin and eosin (H&E) using standard protocols. Histopathological examination was performed under a light microscope (20× and 40× magnification) in a blinded manner by a trained pathologist. The assessment focused on structural integrity, inflammatory infiltration, and potential tissue damage. No additional blood biochemical tests were conducted in this study.

Figure 1 displays the materials characterization of ZIF-8 and CA-loaded ZIF-8 nanoparticles. As illustrated in Figure 1A, the TEM and SEM images revealed that pristine ZIF-8 nanoparticles exhibited a uniform polyhedral morphology with an average size of approximately 89.8 ± 7.1 nm. After loading CA, the particle size gradually increased with CA content, reaching up to 130.6 ± 8.1 nm for CA(1.2)@ZIF-8, while maintaining structural integrity. FTIR spectra (Figure 1B) confirmed the presence of characteristic peaks from both ZIF-8 and CA. Peaks at ∼1,578 cm⁻¹ (C=N) and 425 cm⁻¹ (Zn–N) are attributed to the imidazole ring and the coordination between Zn²⁺ ion and imidazole groups, respectively (Qian et al., 2022; Qian et al., 2023), indicating the successful formation of the ZIF-8 framework. After CA loading, new bands appeared at ∼1,678 cm⁻¹ (–CHO) and ∼1,625 cm⁻¹ (C=C, aromatic ring), confirming the incorporation of CA without disrupting the ZIF-8 structure. XRD patterns (Figure 1C) showed that all CA@ZIF-8 samples retained the crystalline structure of ZIF-8. XPS confirmed the elemental composition and surface states of ZIF-8 before and after CA loading. In the survey spectra (Figure 1D), both samples display C 1s (284.8 eV), N 1s (399.1 eV), Zn 2p (1,021.8 eV), and O 1s (531.5 eV) signals. The O 1s intensity is slightly higher for CA(1.2)@ZIF-8, consistent with the presence of oxygen-containing species introduced during loading and/or ubiquitous surface adsorbates. High-resolution C 1s spectra (Figure 1E) are well deconvoluted into four components located at 284.8 eV (C–C/C–H), ≈284.3 eV (C=C, aromatic/conjugated), ≈285.5 eV (C–N), and ≈286.3 eV (C=N), in agreement with the imidazolate linker and the aromatic framework. The N 1s spectra (Figure 1F) comprise N–Zn (≈398.3 eV), N=C (≈399.1 eV) and N–C (≈400.0 eV) components; the peak positions remain essentially unchanged after CA loading, indicating that the Zn–N coordination environment of ZIF-8 is preserved. The Zn 2p spectrum (Figure 1G) shows a Zn 2p_3/2 peak centered at 1,021.8 eV, characteristic of Zn²⁺. Overall, the XPS results verify successful CA loading without disrupting the ZIF-8 coordination framework.

The sol–gel transition of GGA, ZIF-8/GGA, and CA(0.15, 0.3, 0.6, 1.2)@ZIF-8/GGA hydrogels was assessed by vial inversion, indicating successful gel formation across all formulations (Figure 2A). The resulting hydrogels exhibited good shape integrity after molding. Macroscopic deformation tests demonstrated the hydrogels' ability to withstand torsion, stretching, and compression without fragmentation (Figure 2B). Rheological analysis (Figure 2C) showed that all hydrogel samples exhibited typical viscoelastic solid behavior, with storage modulus (G′) consistently higher than loss modulus (G″) across the tested frequency range. Notably, both G′ and G″ values increased with higher CA@ZIF-8 content, indicating enhanced gel stiffness and elasticity. In tensile tests (Figure 2D), the stress-strain curves demonstrated improved tensile strength and extensibility as CA@ZIF-8 concentration increased. Similarly, the compressive stress-strain curves (Figure 2E) revealed a marked increase in compressive resistance with higher filler content, suggesting that CA@ZIF-8 incorporation effectively reinforced the hydrogel network. Ion release studies showed a time-dependent and concentration-dependent release of both Zn²⁺ and CA (Figures 3A,B). Higher CA@ZIF-8 loading led to faster and more sustained release of Zn²⁺, while CA release exhibited an initial burst phase followed by a slower, sustained profile, with release rates positively correlated with CA content. Degradation analysis (Figure 3C) demonstrated a gradual decrease in hydrogel mass over 12 days, with faster degradation observed in samples containing higher CA@ZIF-8 content, suggesting that filler concentration influences hydrogel stability in PBS.

The internal microstructure of the freeze-dried hydrogels was observed by SEM (Figure 4A), showing that pore size decreased and the network became denser with increasing CA@ZIF-8 content, indicating the influence of nanoparticle incorporation on hydrogel architecture. XRD patterns (Figure 4B) revealed that all composite hydrogels maintained the characteristic peaks of ZIF-8, with intensities gradually increasing as the CA@ZIF-8 content increased, suggesting enhanced crystalline phase retention. FTIR spectra (Figure 4C) showed the characteristic absorption bands of GGA, ZIF-8, and CA, confirming the coexistence of the three components in the composite hydrogels. XPS analysis further confirmed compositional changes: the C 1s spectra (Figure 4D) showed increased intensity of C=O and C-N components; Zn 2p peaks (Figure 4E) were present in ZIF-8-containing samples but absent in GGA; the N 1s spectrum (Figure 4F) showed the emergence of N-Zn peak; and the O 1s spectrum (Figure 4G) exhibited enhanced O=C and O-Zn components with increasing CA@ZIF-8 incorporation, confirming the successful integration of the nano-fillers and their interaction with the hydrogel matrix.

The cytocompatibility of the hydrogels was further verified using L929 fibroblasts (). Live/dead staining () showed that most cells remained viable in the GGA and ZIF-8/GGA groups, with a progressive increase in live cell density upon CA incorporation. In particular, CA(0.6)@ZIF-8/GGA exhibited the highest proportion of live cells and minimal dead cells, whereas CA(1.2)@ZIF-8/GGA induced marked cytotoxicity with a large number of dead cells. Consistently, quantitative analysis by CCK-8 assay () confirmed that CA(0.6)@ZIF-8/GGA significantly promoted fibroblast viability, while excessive CA loading led to a sharp decline. These results demonstrate that CA(0.6)@ZIF-8/GGA provides an optimal balance of cell compatibility and bioactivity in fibroblasts, supporting its potential application in skin tissue repair. Supplementary Figure S2 Supplementary Figure S2A Supplementary Figure S2B

Live/dead staining of HUVECs (Figure 5A) showed that the GGA hydrogel alone did not enhance cell survival compared to the control group. Incorporation of ZIF-8 modestly improved viability, which was further enhanced when CA was loaded into ZIF-8. A concentration-dependent effect was observed, with the CA(0.6)@ZIF-8/GGA group displaying the highest number of live cells and the fewest dead cells, whereas excessive CA in CA(1.2)@ZIF-8/GGA resulted in significant cytotoxicity. Quantitative analysis of fluorescence intensity (Supplementary Figure S1) confirmed these observations: CA(0.6)@ZIF-8/GGA exhibited the strongest calcein-AM (live) signal and reduced PI (dead) signal, while CA(1.2)@ZIF-8/GGA showed the opposite trend.

F-actin/DAPI staining (Figure 5B) further supported these findings, as cells in the CA(0.6)@ZIF-8/GGA group exhibited optimal adhesion and cytoskeletal organization. Quantification of cell adhesion (Figure 5C) showed the highest relative adhesion area in the CA(0.6)@ZIF-8/GGA group, which was consistent with enhanced spreading. The CCK-8 assay (Figure 5D) also confirmed the superior viability of this group, while CA(1.2)@ZIF-8/GGA again showed a sharp reduction in cell survival. Collectively, these results demonstrate that CA(0.6)@ZIF-8/GGA provides the most favorable cytocompatibility, and thus it was selected for subsequent angiogenesis-related experiments.

To further evaluate the angiogenic potential of the hydrogels, immunofluorescence staining was performed to assess the expression of two key markers: CD31 and VEGF in HUVECs cultured with different hydrogel extracts. As shown in Figure 6A, the expression of CD31 in the control and GGA groups was relatively weak, while ZIF-8/GGA treatment led to moderately enhanced CD31 staining. Notably, CA(0.6)@ZIF-8/GGA treatment resulted in a marked increase in CD31 signal, accompanied by improved cytoskeletal organization and cell spreading. Quantitative analysis (Figure 6C) confirmed a statistically significant upregulation of CD31 in the CA(0.6)@ZIF-8/GGA group compared to the other groups. Figure 6B demonstrates the expression of VEGF, another critical angiogenic factor. A similar trend was observed, with the CA(0.6)@ZIF-8/GGA group showing the highest VEGF fluorescence intensity. Quantification (Figure 6D) revealed a significant increase in VEGF levels, particularly in the CA(0.6)@ZIF-8/GGA group (***p < 0.001), indicating that the composite hydrogel synergistically enhances angiogenic signaling via upregulation of both structural (CD31) and secretory (VEGF) markers.

To further assess the proangiogenic capacity of the CA(0.6)@ZIF-8/GGA hydrogel, we performed in vitro HUVEC migration and tube formation assays. In the Transwell migration assay (Figure 7A), HUVECs treated with CA(0.6)@ZIF-8/GGA extracts exhibited the highest number of migrated cells compared to the other groups. Quantitative analysis (Figure 7C) showed a ∼3.5-fold increase in migrated cell number relative to the control group (**p < 0.01, ***p < 0.001), suggesting that the composite hydrogel strongly enhances endothelial motility. In the Matrigel tube formation assay (Figure 7B), CA(0.6)@ZIF-8/GGA stimulated the formation of well-defined, interconnected tubular networks with significantly more branches and longer overall tube length compared to the GGA and ZIF-8/GGA groups. Quantification revealed marked increases in both the number of branches (Figure 7D) and total tube length (Figure 7E), with CA(0.6)@ZIF-8/GGA outperforming all other treatments (***p < 0.001). These results corroborate the immunofluorescence findings and confirm that CA(0.6)@ZIF-8/GGA hydrogel effectively promotes angiogenesis-related processes through synergistic regulation of endothelial behavior.

To investigate the therapeutic efficacy of the CA(0.6)@ZIF-8/GGA hydrogel in vivo, a full-thickness excisional wound model was established in streptozotocin-induced diabetic mice. Wounds were treated with PBS (control), GGA, ZIF-8/GGA, or CA(0.6)@ZIF-8/GGA, and monitored over 12 days. As shown in Figure 8A, wound healing was notably faster in the CA(0.6)@ZIF-8/GGA group, with almost complete re-epithelialization and hair regrowth by day 12. In contrast, the control and GGA groups displayed delayed wound closure and prominent scab formation. Quantitative analysis of wound area (Figure 8B) revealed that CA(0.6)@ZIF-8/GGA significantly reduced wound size from day 4 onwards, achieving over 90% closure by day 12 (***p < 0.001). The healing score, which integrates wound appearance, epithelialization, and contraction, was also highest in the CA(0.6)@ZIF-8/GGA group (Figure 8C). These results suggest that the multifunctional composite hydrogel markedly promotes tissue regeneration and wound closure in diabetic conditions, likely through coordinated regulation of inflammation and angiogenesis.

To assess systemic biosafety, major organs (lung, liver, heart, brain, spleen, and kidney) were collected at day 12 and subjected to H&E staining (Figure 9). All groups, including the CA(0.6)@ZIF-8/GGA-treated rats, exhibited normal histoarchitecture comparable to the normal and diabetic controls. The lung tissues showed intact alveolar structures without hemorrhage or inflammation; liver and kidney sections maintained normal hepatocyte and glomerular morphology; and no myocardial degeneration, neuronal damage, or splenic abnormalities were observed. These findings confirm that topical application of CA(0.6)@ZIF-8/GGA hydrogel does not induce detectable systemic toxicity, supporting its favorable biosafety profile for diabetic wound healing applications.