Deregulated immune cell recruitment orchestrated by FOXM1 impairs human diabetic wound healing

Abstract

Diabetic foot ulcers (DFUs) are a life-threatening disease that often result in lower limb amputations and a shortened lifespan. However, molecular mechanisms contributing to the pathogenesis of DFUs remain poorly understood. We use next-generation sequencing to generate a human dataset of pathogenic DFUs to compare to transcriptional profiles of human skin and oral acute wounds, oral as a model of "ideal" adult tissue repair due to accelerated closure without scarring. Here we identify major transcriptional networks deregulated in DFUs that result in decreased neutrophils and macrophages recruitment and overall poorly controlled inflammatory response. Transcription factors FOXM1 and STAT3, which function to activate and promote survival of immune cells, are inhibited in DFUs. Moreover, inhibition of FOXM1 in diabetic mouse models (STZ-induced and db/db) results in delayed wound healing and decreased neutrophil and macrophage recruitment in diabetic wounds in vivo. Our data underscore the role of a perturbed, ineffective inflammatory response as a major contributor to the pathogenesis of DFUs, which is facilitated by FOXM1-mediated deregulation of recruitment of neutrophils and macrophages, revealing a potential therapeutic strategy.

Fig. 1. Wound-activated gene signature found at baseline of oral mucosa is activated in DFUs.

a Representative image of a diabetic foot ulcer (DFU). b Heatmap of genes regulated in DFUs. c Volcano plot indicating differentially regulated genes in DFUs. Dotted line region is magnified on the right panel highlighting some of the most significantly upregulated genes in DFUs. d Wound-activated gene signature of a subset of genes involved in epidermal differentiation, intermediate filaments and inflammatory cytokines shows similarities between baseline oral mucosa and DFUs.

Fig. 2. Pathways and functions involved in inflammatory response and cellular movement are inhibited in DFUs but activated in oral and skin wounds.

a Venn diagram of significantly regulated genes in oral day 3 (D3)/day 1 (D1), skin D3/D1, and DFU/DFS (diabetic foot skin). b Top canonical pathways and c diseases and functions found to be enriched in oral D3/D1, skin D3/D1, and DFU/DFS showing downregulation of cellular movement and inflammatory response in DFUs.

Fig. 3. Upstream regulators involved in promoting proliferation and cell survival of immune cells are inhibited in DFUs and activated in oral and skin acute wounds.

a Top upstream regulators enriched in oral D3/D1 vs. skin D3/D1 vs. DFU/DFS. b Upstream regulators found to be activated in oral and skin wounds that are suppressed or partially regulated in DFUs involved in proliferation, inflammatory response, leukocyte migration and proliferation of leukocytes. NA non applicable. c FOXM1 predicted network shows activation of proliferation and inflammatory response in oral and skin wounds compared to suppression in DFUs. d qPCR validations of upstream regulators FOXM1, STAT3, and TNFα confirms suppression in DFUs compared to activation in oral and skin human wounds. n = 2 biologically independent samples for oral and skin wounds; n = 3 biologically independent samples for DFUs. Data presented as mean ± SD. FOXM1 P-values oral D3/D1 vs DFU/DFS: **P = 0.0095; FOXM1 P-values skin D3/D1 vs DFU/DFS: **P = 0.008; STAT3 P-values oral D3/D1 vs DFU/DFS: **P = 0.003; STAT3 P-values skin D3/D1 vs DFU/DFS: **P = 0.0034; TNF P-values oral D3/D1 vs DFU/DFS: **P = 0.0041; TNF P-values skin D3/D1 vs DFU/DFS: *P = 0.0436 (two-way ANOVA followed by Tukey’s post hoc test). e Immunostaining of FOXM1 (green immunofluorescence signal) and keratin 5 (K5, in red) in oral and skin day 3 wounds and DFUs. Robust staining of FOXM1 found in oral and acute skin wounds but is absent in DFUs. Stainings were performed once with three biologically independent patient samples per group. Scale bar = 100 µm.

Fig. 4. Inflammatory signature is inhibited in DFUs.

a Venn diagram of significantly regulated genes from oral, skin, and DFUs. b Top enriched GO processes from commonly regulated genes in oral and skin wounds compared to DFU specific genes (circled in yellow in part a) demonstrates processes involved in cellular proliferation and inflammation to be deregulated in DFUs compared to acute wounds (oral and skin). c Top canonical pathways involved in inflammation and d IPA-predicted network of inflammation in oral and skin compared to DFUs (n = 8 biologically independent samples) shows inhibition of cell proliferation and inflammation in DFUs compared to acute wounds (oral and skin).

Fig. 5. Deregulated activation of immune cells in DFUs.

a Top functions enriched in the inflammatory response demonstrating inhibition of immune-cell activation in DFUs. b Prediction of estimated proportions of a subset of leukocytes (macrophages and granulocytes) in oral, skin, and DFUs based on gene expression demonstrates decreased macrophage activation and neutrophil recruitment with increased eosinophils in DFUs. c Subset of genes involved in activation of phagocytes showing downregulation of several genes in DFUs (n = 8 biologically independent samples). d Representative pictures of oral and skin day 3 wounds and DFUs show basal keratin marker K5, macrophage marker, and neutrophil marker MPO. Quantification of n = 5 biologically independent samples demonstrates decreased macrophage activation (**P = 0.0038, ****P ≤ 0.0001) and neutrophils (**P = 0.0048, ****P ≤ 0.0001) in DFUs compared to oral and skin wounds. Data presented as mean ± SD (two-tailed unpaired Student’s t-test). Scale bar = 100 µm.

Fig. 6. Inhibition of FOXM1 suppresses immune-cell response and inhibits wound healing in vivo.

a Schematic of in vivo wound-healing assay. CD1 (non-diabetic) mice were wounded and treated topically with either the FOXM1 inhibitor FDI-6 or vehicle every other day for 8 days. b Representative images of wounded skin after topical treatment with either vehicle or FDI-6 at 0, 2, 4, 6, and 8 days after wounding. c Percent of wound area at each time following vehicle or FDI-6 treatment relative to the original wound area. Quantification of wound areas in n = 6 (Veh) and 8 (FDI-6) wounds per group were performed with Fiji software. Data presented as mean ± SEM. *P = 0.042 (two-tailed unpaired Student’s t test). d H&E staining of day 4 wounds demonstrating decreased immune cell infiltrates in FDI-6 treated wounds compared to vehicle control wounds. e Representative pictures of vehicle and FDI-6 treated wounds at day 4 show basal keratin marker K5, and neutrophil marker MPO. Treatment of wounds with FDI-6 resulted in decreased neutrophils compared to vehicle treated wounds. n = 3 animals per group examined over two independent experiments. Data presented as mean ± SD. ***P = 0.0005 (two-tailed unpaired Student’s t test). White arrows indicate the wound edge of the migrating epithelial tongue. Scale bar = 100 µm.

Fig. 7. Inhibition of FOXM1 further impairs wound healing and decreases frequency of macrophages and neutrophils in the wounds of diabetic mice.

a. Schematic of in vivo wound-healing assay in STZ-induced diabetic mice. CD1 mice were i.p. injected with STZ to induce diabetes and were maintained for 6 weeks for effects of diabetes on wound healing to occur. Mice were wounded and treated topically with either the FOXM1 inhibitor FDI-6 or vehicle every other day for 8 days. b Representative images of wounded skin after topical treatment with either vehicle or FDI-6 at 0, 2, 4, 6, and 8 days after wounding. c Percent of wound area at each time following vehicle or FDI-6 treatment relative to the original wound area. Quantification of wound areas in n = 10 (diabetic) and n = 16 (diabetic+FDI-6) wounds per group were performed with Fiji software. Data presented as mean ± SEM. **P = 0.0053 (two-way ANOVA followed by Tukey’s post hoc test). d Wound edge skin at day 4 was collected and frequencies of macrophages (F4/80 + Ly6G−) and neutrophils (F4/80-Ly6G+) within gated myeloid cells, CD11b+ cells were quantified by flow cytometry. Data represent 4 wounds per group. Data presented as mean ± SEM. **P = 0.0028 for control group, *P = 0.023 for diabetic group for macrophages; *P = 0.027 for diabetic group for neutrophils, as calculated using one-way ANOVA with Tukey’s multiple comparisons test.

Fig. 8. The diagram summarizes our findings and shows a model of unwounded oral mucosa, skin, and DFUs.

It demonstrates similar wound-activated signature of genes involved in differentiation, cytokines, and intermediate filaments. Inhibition of FOXM1, STAT3, and TNFα regulators results in lack of immune-cell activation, proliferation and survival in the DFU environment contributing to deregulated inflammatory response and inhibition of wound healing.