Interleukin-8/Matrix Metalloproteinase-9 Axis Impairs Wound Healing in Type 2 Diabetes through Neutrophil Extracellular Traps-Fibroblast Crosstalk

Abstract

Neutrophils interact with and activate fibroblasts through the release of neutrophil extracellular traps (NETs). We investigated the role of NETs-fibroblast crosstalk in the cutaneous wound healing of type 2 diabetes (T2D). Neutrophils/NETs, serum, and primary human skin fibroblasts (HSFs) were obtained from individuals with T2D and age/sex-matched controls. NET-stimulation studies were performed on neutrophils/HSFs, with and without specific inhibitors, while HSF healing capacity was assessed using a scratch wound healing assay. T2D HSFs display a profibrotic phenotype, showing increased CCN2/CTGF, α-smooth muscle actin, and collagen release, albeit with impaired healing capacity, elevated type I collagen C-terminal telopeptide, and collagen degradation associated with increased (∼3.5-fold) matrix metalloproteinase-9 (MMP-9) in T2D neutrophils/NETs. IL-8 induced the expression of MMP-9 in neutrophils/NETs. Moreover, T2D neutrophils/NETs exhibited increased IL-8 content, which acted in an autocrine/paracrine fashion to further augment its production by neutrophils/HSFs. The findings were validated in normoglycemic individuals during a hyperglycemic clamp with concomitant lipid infusion and further corroborated immunohistochemically in diabetic plantar ulcer biopsies. This novel, vicious circle of NETs/interleukin-8/MMP-9/HSFs was hindered by IL-8 or MMP-9 blockade via specific inhibitors or by dismantling the NET-scaffold with DNase I, suggesting candidate therapeutic targets in wound healing impairment of T2D.

Keywords: interleukin‐8; matrix metalloproteinase‐9; neutrophil extracellular traps; type 2 diabetes mellitus; wound healing.

FIGURE 1

T2D fibroblasts are characterized by a profibrotic but functionally impaired phenotype. (A–D) Immunostainings in HSFs isolated by HI (A), (B), and T2D patients (C), (D). (A), (C) blue: DAPI, green: aSMA, red: CCN2; (B), (D) blue: DAPI, green: Vimentin, red: CCN2. (E) Corresponding mean fluorescence intensity (MFI) quantification of aSMA and CCN2. CCN2 protein (F) and mRNA (G) levels were assessed by cytoblot and RT‐qPCR, respectively, in HI and T2D HSFs. (H) Sircol collagen release assay. (I) Wound healing assay. (A–D), (I) One representative example out of four independent experiments is presented. (A–D) Confocal microscopy. Magnification: 40×, scale bar 10 µm. (I) Optical microscopy. Magnification: 5×. (E–G) Mann–Whitney U‐test was applied, n = 4, *p < 0.05. aSMA, alpha smooth muscle actin; CCN2, cellular communication network factor 2; HI, healthy individuals; HSFs, human skin fibroblasts; T2D, type 2 diabetes mellitus.

FIGURE 2

Healthy fibroblasts, during their crosstalk with T2D NETs, acquire a functionally defective phenotype, similar to primary T2D fibroblasts. (A–E) Stimulations of HI HSFs with either HI or T2D NETs. Assessment of CCN2 by (A) immunostaining (blue: DAPI, green: Vimentin, red: CCN2), (B) Cytoblot and (C) RT‐qPCR. (D) Sircol collagen release assay. (E) Wound healing assay. (A), (E) One representative example out of four independent experiments is presented. (A) Confocal microscopy. Magnification: 40×, scale bar 10 µm. (E) Optical microscopy. Magnification: 5×. (B–D) Kruskal–Wallis followed by Dunn's test for multiple comparisons, was applied, n = 4, ***p < 0.001, ns, not significant. CCN2, cellular communication network factor 2; HI, healthy individuals; HSFs, human skin fibroblasts; T2D, type 2 diabetes mellitus.

FIGURE 3

Hyperglycemic T2D NETs overexpress MMP‐9, which impairs the wound healing process. (A, B) MMP‐9 concentration in HI and T2D NETs assessed by ELISA (A) and immunostaining (B) (blue: DAPI, green: MMP‐9, red: NE). (C) Immunostaining (blue: DAPI, green: CTX1, red: COL1) and (D) wound healing assay of HI HSFs stimulated with HI or T2D NETs in the absence/presence of DNAse I or MMP‐9‐IN‐1. (E) Collagenase activity assay of HI HSFs treated under the aforementioned conditions. (F) MMP‐9 ELISA measured in NETs obtained from patients with good (HbA1c<7%) and poor (HbA1c>7%) glycemic control. (G) Correlation between MMP‐9 concentration in NETs and HbA1c levels. (A), (E), (F) Kruskal–Wallis followed by Dunn's test for multiple comparisons was applied, *p < 0.05, **p < 0.01, ****p < 0.0001, ns, not significant. (A), (F) HI, n = 12; T2D, n = 24; (E) n = 6. (B–D) One representative example out of four independent experiments is presented. (B), (C) Confocal microscopy. (B) Magnification: 60×, scale bar 10 µm. White arrowheads depict NETs. (C) Magnification: 40×, scale bar 10 µm. (D) Optical microscopy. Magnification: 5×. (G) Spearman's rank correlation coefficient was used. CTX1, C‐terminal cross‐linked telopeptide of type I collagen; COL1, collagen type I; HbA1c, hemoglobin A1c; HI, healthy individuals; HSFs, human skin fibroblasts; MMP‐9, matrix metalloproteinase‐9; MMP‐9‐IN‐1; matrix metalloproteinase‐9 inhibitor‐1; NE, neutrophil elastase; T2D, type 2 diabetes mellitus.

FIGURE 4

Overproduced IL‐8 by neutrophils and fibroblasts instigates autocrine and paracrine feedback loops. (A) IL‐8 mRNA levels, (B) IL‐8 ELISA, and (C) IL‐8 immunostaining (blue: DAPI, green: IL‐8, red: NE), assessed in neutrophils/NETs obtained by HI and T2D patients. (D) IL‐8 mRNA levels and (E) IL‐8 immunostaining (blue: DAPI, green: IL‐8, red: Vimentin) assessed in HI HSFs stimulated with HI, T2D, or IL‐8 neutralized T2D NETs. (F) IL‐8 mRNA levels and (G) IL‐8 immunostaining (blue: DAPI, green: IL‐8, red: NE), assessed in HI neutrophils treated with HI or T2D or IL‐8 neutralized T2D NETs or recombinant IL‐8. (H) MMP‐9 ELISA and (I) MMP‐9 immunostaining (blue: DAPI, green: MMP‐9, red: NE), in HI neutrophils stimulated with HI or T2D or neutralized T2D NETs, or recombinant IL‐8. (A), (B) Mann–Whitney U‐test was applied. (A) n = 6, (Β) ΗΙ, n = 12; T2D, n = 24. (C), (E), (G), (I) One representative example out of four independent experiments is presented. Confocal microscopy. Magnification: 40x, scale bar 10 µm. (D), (F), (H) Kruskal–Wallis followed by Dunn's test for multiple comparisons was applied, n = 6. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns, not significant. HI, healthy individuals; HSFs, human skin fibroblasts; IL‐8, interleukin‐8; MMP‐9, matrix metalloproteinase‐9; NE, neutrophil elastase; T2D, type 2 diabetes mellitus.

FIGURE 5

Hyperglycemic clamps with lipid infusion induce phenotypic features in healthy neutrophils, similar to those observed in T2D. (A–C) Immunostainings depicting IL‐8 (green)/NE (red) and MMP‐9 (green)/NE (red) immunoreactivity in neutrophils obtained before HGC administration (timepoint 0’, T0’) (A), 240’ minutes (timepoint 240’, T240’) (B), and 24 h (timepoint 24 h, T24 h) (C), after HGC infusion in healthy controls. (D–F) Wound healing assay showing the migratory capacity of HI HSFs treated with either (D) T0’, (E) T240’, or (F) T24 h NETs. (G) IL‐8 and (H) MMP‐9 ELISA measured in T0’, T240’, and T24h’ NETs. (I) Collagenase activity of HI HSFs treated with T0, T240’ or T24 h NETs. (A–F) One representative example out of three independent experiments is presented. (A–C) Confocal microscopy. Magnification: 40x, scale bar 10 µm. (D–F) Optical microscopy. Magnification: 5×. (G–I) Qualitative representation; orange line: subject #1, purple line: subject #2, green line: subject #3. HGC, hyperglycemic clamp with lipid infusion; HI, healthy individuals; HSFs, human skin fibroblasts; IL‐8, interleukin‐8; MMP‐9, matrix metalloproteinase‐9; NE, neutrophil elastase; T2D, type 2 diabetes mellitus.

FIGURE 6

Chronic diabetic ulcers exhibit an increased presence of MMP‐9/IL‐8 in neutrophils and degraded collagen. (A–C) Tissue immunostainings in unaffected skin and ulcer sections obtained from healthy and diabetic individuals. (A) Blue: DAPI, green: MMP‐9, red: NE, (B) blue: DAPI, green: IL‐8, red: NE, (C) blue: DAPI, green: CTX1, red: COL1. (D) MFI‐based grading scale depicting MMP‐9, NE, IL‐8, CTX1, and COL1 immunoreactivity in the aforementioned conditions. (A–C) One representative example out of four independent experiments is presented. Confocal microscopy. Magnification: 20×, scale bar 20 µm. CTX1, C‐terminal cross‐linked telopeptide of type I collagen; COL1, collagen type I; HI, healthy individuals; IL‐8, interleukin‐8; MMP‐9, matrix metalloproteinase‐9; NE, neutrophil elastase; T2D, type 2 diabetes mellitus.