MMP-9 responsive hydrogel promotes diabetic wound healing by suppressing ferroptosis of endothelial cells

Abstract

Ferroptosis plays a crucial role in the progression of diabetic wounds, suggesting potential therapeutic strategies to target ferroptosis. Transient receptor potential ankyrin 1 (TRPA1) is a non-selective calcium channel that acts as a receptor for a variety of physical or chemical stimuli. Cinnamaldehyde (CA) is a specific TRPA1 agonist. In in vitro experiments, we observed that high glucose (HG) treatment induced endothelial cell ferroptosis, impairing cell function. CA successfully inhibited endothelial cell ferroptosis, improving migration, proliferation, and tube formation. Further mechanistic studies showed that CA-activated TRPA1-induced Ca²⁺ influx promoted the phosphorylation of calmodulin-dependent protein kinase II (CaMKII) and nuclear factor-E 2-related factor 2 (Nrf2) translocation, which contributed to the elevation of glutathione peroxidase 4 (GPX4), leading to the inhibition of endothelial cell ferroptosis. In addition, CA was incorporated into an MMP-9-responsive injectable duplex hybrid hydrogel (CA@HA-Gel), allowing its efficient sustained release into diabetic wounds in an inflammation-responsive manner. The results showed that CA@HA-Gel inhibited wound endothelial cell ferroptosis and significantly promoted diabetic wound healing. In summary, the results presented in this study emphasize the potential therapeutic application of CA@HA-Gel in the treatment of diseases associated with ferroptosis.

Keywords: Diabetes mellitus; Ferroptosis; Hydrogel; TRPA1; Wound healing.

Scheme 1. Schematic illustration of therapeutic process for diabetic wound healing by using CA-loaded hydrogel.

Fig. 1

The diabetic microenvironment induces endothelial cell ferroptosis in vivo. (A) A diagram illustrating the diabetic mellitus model created using a combination of high-fat diet (HFD) feeding and low-dose streptozotocin (STZ) injection. The red squares in the figure indicate the duration of processing. (B) Representative macroscopic pictures of wound healing and closure traces. Scale bar: 2.5 mm. (C) Quantitative analysis of wound closure on days 0, 3, 7, 10 and 14 of control and DM groups. (D) Representative photographs of Masson's trichrome staining and (G) statistical evaluation of collagen-occupied area. Scale bar: 50 μm. (E) A representative panoramic view of H&E staining and (H) quantitative analysis of scar width was used in each group. Scale bar: 1 mm. (F) Representative immunohistochemical images of MMP-9, CD31, GPX4. Scale bar: 50 μm. (I) Western blotting (WB) images of wound tissues with (J) statistical evaluation of the expression level of GPX4. (K–M) Quantitative analysis of immunohistochemistry of MMP-9, CD31, GPX4. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Each group contained three mice.

Fig. 2

HG treatment activates ferroptosis and suppresses TRPA1 expression in endothelial cells. (A) Heatmap showing differential gene expression between control and DM groups of HUVCEs. (B) KEGG pathway enrichment analysis. (C) GSEA enrichment score revealed a major reduction in TRP channels in the HG group. (D) GSEA enrichment score revealed significant improvements in the ferroptosis pathway in the HG group. (E) RT-qPCR analysis of TRP channels genes in HUVECs treated with PBS or HG for 24 h. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

Fig. 3

TRPA1 alleviates HG-induced ferroptosis in endothelial cells. (A) Representative confocal images of C11-BODIPY staining and (J) its quantitative analysis. Scale bar: 20 μm. (B) TEM images of mitochondrial morphology after different interventions. Scale bar: 200 nm. (C) Intracellular MDA levels of HUVECs. (D) Intracellular GSH levels of HUVECs. (E) Depolarized cells labeled by JC-1 staining were observed by flow cytometer. (F) Quantitative analysis of depolarized cells. (G–I) Western-blot images and evaluation of GPX4 and TRPA1 expression in HUVECs. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

Fig. 4

TRPA1 alleviates HG-induced endothelial cell dysfunction. (A) Representative fluorescence images and (B) statistical evaluation of EdU in HUVECs. Scale bar: 100 μm. (C) Representative WB images of CyclinD1, CyclinD3, and VEGF proteins. (D) Scratching assay and (E) quantitative statistics showing the scratch closure ratio of HUVECs with different treatments. Scale bar: 200 μm. (F) Representative images and (H) corresponding statistical evaluation showing the migrated cells. Scale bar: 100 μm. (G) Results and (I) quantitative analysis of tube formation assays. Scale bar: 100 μm. (J) Flow cytometry (FCM) images of cell cycle distribution, with (K) statistical evaluation. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

Fig. 5

TRPA1 promotes nuclear translocation of Nrf2 and inhibits ferroptosis of endothelial cells. (A) Confocal immunofluorescence images of Nrf2, cytoskeleton, and nucleus. Scale bar: 10 μm. (B) Western blotting (WB) images and (C–D) quantitative statistics of cytoplasmic and nuclear Nrf2 protein levels in HUVECs. (E) Representative confocal images of C11-BODIPY staining and (I) its quantitative analysis. Scale bar: 20 μm. (F–G) Intracellular MDA and GSH levels of HUVECs. (J) Depolarized cells labeled by JC-1 staining observed by flow cytometer and (H) its quantitative statistics. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

Fig. 6

TRPA1-mediated CaMKII activation facilitates the nuclear translocation of Nrf2. (A) Representative confocal immunofluorescence pictures and (B) quantitative statistics from Fluo-4-stained HUVECs. Scale bar: 50 μm. (C)WB images and (D–E) quantitative statistics of ρ-CaMKII, CaMKII, and GPX4 protein levels in HUVECs. (F) Confocal immunofluorescence images of Nrf2 (green fluorescence), cytoskeleton (red fluorescence), and nucleus (blue fluorescence). Scale bar: 10 μm. (G) Western blotting images and (H–I) quantitative statistics of Nrf2 protein levels of nuclear and cytoplasmic in HUVECs. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

Fig. 7

Characterization of HA-Gel hydrogel. (A) The SEM images, (B) pore size, (C) swelling ratio analysis, and (D) rheological time sweep of HA-Gel hydrogels with a fixed concentration of Gel-ADH (16 %) and the different final concentrations of HA-CHO (4 %, 6 % and 8 %). Scale bar: 100 μm. (E) Shear tinning test of HA-Gel hydrogel (Gel-ADH: 16 % and HA-CHO: 4 %). (F) HA-Gel hydrogel degradation at various MMP-9 concentrations in vitro. (G) In vivo biocompatibility testing for HA-Gel hydrogel. Scale bar: 50 μm. (H) Live and dead staining at 1,2,3 days of HUVECs. Scale bar: 100 μm. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001.

Fig. 8

CA@HA-Gel accelerates diabetic wound healing in vivo. (A) Representative gross images of wound healing and (B) wound closure traces. Scale bar: 2.5 mm. (C) Quantitative statistics of wound closure on days 0, 3, 7, 10, and 14 of DM, DM + CA, DM + HA-Gel, DM + CA@HA-Gel groups. (D) Panoramic view of H&E staining and (G) analysis of scar width in each group. Scale bar: 1 mm. (E) Pictures of Masson's trichrome staining and (H) its quantitative statistics of collagen occupied region. Scale bar: 100 μm. (F) Western blotting images of wound tissue and (K) quantitative statistics of the expression level of GPX4 and Nrf2. (I) Representative immunohistochemistry images of CD31 and GPX4. Scale bar: 50 μm. (J) Quantitative analysis of CD31 and GPX4 immunohistochemistry. (L) Immunofluorescence images and (M–N) quantification of α-SMA (red fluorescence) and COL1A1 (red fluorescence) at wound sites. Scale bar: 100 μm. The results are shown as the mean ± standard deviation (SD), with a sample size of n = 3. The symbols ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.001. Each group contained three mice.