Diabetes induces stable intrinsic changes to myeloid cells that contribute to chronic inflammation during wound healing in mice

Abstract

Acute inflammation in response to injury is a tightly regulated process by which subsets of leukocytes are recruited to the injured tissue and undergo behavioural changes that are essential for effective tissue repair and regeneration. The diabetic wound environment is characterised by excessive and prolonged inflammation that is linked to poor progression of healing and, in humans, the development of diabetic foot ulcers. However, the underlying mechanisms contributing to excessive inflammation remain poorly understood. Here we show in a murine model that the diabetic environment induces stable intrinsic changes in haematopoietic cells. These changes lead to a hyper-responsive phenotype to both pro-inflammatory and anti-inflammatory stimuli, producing extreme M1 and M2 polarised cells. During early wound healing, myeloid cells in diabetic mice show hyperpolarisation towards both M1 and M2 phenotypes, whereas, at late stages of healing, when non-diabetic macrophages have transitioned to an M2 phenotype, diabetic wound macrophages continue to display an M1 phenotype. Intriguingly, we show that this population predominantly consists of Gr-1(+) CD11b(+) CD14(+) cells that have been previously reported as 'inflammatory macrophages' recruited to injured tissue in the early stages of wound healing. Finally, we show that this phenomenon is directly relevant to human diabetic ulcers, for which M2 polarisation predicts healing outcome. Thus, treatments focused at targeting this inflammatory cell subset could prove beneficial for pathological tissue repair.

Fig. 1.

Analysis of inflammatory cell polarisation in wounds of non-diabetic and diabetic mice over a healing time-course. (A) Overview of a representative whole-wound section from a day-7 diabetic wound with boxes indicating where images were captured for analyses (scale bar: 1 mm). (B) Immunofluorescent detection of CD45, Nos2 and Arg1 in day-4 (i) and day-7 (ii) wounds of non-db and db mice. Merged images show individual CD45⁺ cells that are also positive for Nos2 but not Arg1 (M1 phenotype, yellow arrow), Arg1 but not Nos2 (M2 phenotype, blue arrow), and for both Nos2 and Arg1 (‘mixed’ phenotype, white arrows) (scale bar: 10 μm). (C,D) Quantification of macrophage phenotypes as shown in B at (C) day 4 following wounding and (D) day 7 following wounding, from six non-db and six db mice at each time point. Non-db, non-diabetic; db, diabetic; **P<0.01, ***P<0.001, unless otherwise indicated.

Fig. 2.

Analysis of myeloid cell subsets in day-5 and day-10 non-db and db wounds. (A) Representative flow plots from tissue-dissociated single-cell suspensions of whole wounds gated on CD11b⁺ cells and showing Gr-1 and CD14 expression in (i) isotype control, (ii) non-db day-5, (iii) db day-5, (iv) non-db day-10 and (v) db day-10 wounds. Upper left region of plot (P4) shows cells counted as positive for Gr-1, but negative for CD14. Upper right region of plot (P3) shows cells double positive for Gr-1 and CD14. Lower right region of plot (P5) shows cells negative for Gr-1 but positive for CD14. (B,C) Graphs of flow cytometry data shown in A for day-5 (B) and day-10 (C) wounds displaying percent of total wound cells positive for the listed marker combinations (grey bars, non-db; black bars, db; n=6, *P<0.05, **P<0.01, ***P<0.001). (D) Representative images of non-db and db day-5 wound-derived sorted cells shown in A following cytospin and Giemsa staining: Gr-1⁺ CD14⁻ cells (left-most two panels), Gr-1⁺ CD14⁺ cells (centre panels) and Gr-1⁻ CD14⁺ cells (rightmost two panels). Scale bar: 5 μm.

Fig. 3.

Analysis of differentiation and polarisation in day-5 and day-10 wound-derived myeloid cells. (A,B) qRT-PCR analysis of (A) general macrophage maturation marker expression in wound-derived CD14⁺ (all) cells, or (B) general granulocyte maturation marker expression in wound-derived Gr-1⁺ CD14⁻ cells, pooled from three to six non-db or three to six db samples (three pools each) 5 days post-wounding. Bars indicate gene expression in db-derived CD14⁺ (all) cells relative to the non-db sample (indicated by dashed line) and reference genes Hist2h2aa1 and Hsp90ab1. Means and s.e.m. of three biological replicates are shown (*P<0.05). Nm to ref and ndb, normalised to reference genes and non-diabetic samples. (C) Analysis of M1 and M2 polarisation markers in CD14⁺ (all) cells (CD14+) and Gr-1⁺ CD14⁻ cells (Gr-1+) isolated from wounds of non-diabetic (grey bars) or diabetic (black bars) mice 5 or 10 days after wounding (time point indicated, *P<0.05, **P<0.01, ***P<0.001). (D) Analysis of H₂O₂ production from wound-derived CD11b⁺ myeloid cells from three non-db and three db mice after 24-hour culture (mean ± s.e.m., **P<0.01).

Fig. 4.

Analysis of macrophage polarisation markers in BM macrophages after 7 days of culture in M-CSF. (A,B) Analysis of M1 (A) and M2 (B) polarisation markers in non-activated (NA), classically activated (CA) and alternatively activated (AA) non-db (grey bars) and db (black bars) BM macrophages (n=3 independent RNA samples, *P<0.05, **P<0.01). nm to ref, normalised to reference genes. (C) Production of IL-12 and INF-γ in non-db (grey bars) and db (black bars) BM macrophages (n=3 independent samples, *P<0.05, ***P<0.001). (D) Analysis of H₂O₂ production from BM-derived macrophages from non-db and db mice (n=3 independent samples, *P<0.05).

Fig. 5.

Differentiation and behavioural assays of non-db- and db-derived BM macrophages. (A) Representative bright-field images of nondb- and db-derived BM macrophages after 7 days culture in M-CSF (scale bar: 25 μm). (B) Representative flow cytometry plots of non-dband db-derived BM macrophages after 7 days culture in M-CSF, showing expression of macrophage markers CD11b and F4.80. (C) qRTPCR analysis of myeloid cell markers Emr1 (F4.80, left panel), Ly6g (Gr-1, centre panel) and Itgam (CD11b, right panel) at day 1 and day 7 of culture in M-CSF in non-db (grey bars) and db (black bars) BM macrophages from three independent RNA isolations for each condition. (D) qRT-PCR analysis of CCR2 expression in nondb (grey bar) and db (black bar) BM macrophages after 7 days of culture in M-CSF from three independent RNA isolations for each condition. (E) Chemotactic response to MCP-1 in a transwell assay showing mean number of cells per field of view (FoV) of non-db and db BM macrophages from three independent experiments. (F) Scratch wound migration assay of non-db (grey diamonds) and db (black squares) BM macrophages showing mean number of cells migrated into scratch wound at 4-hour intervals over 16 hours from three independent experiments. (G) Adhesion assay of non-db (grey bar) and db (black bar) BM macrophages showing mean number of cells per field of view adhered onto activated endothelial cells after 4 hours from three independent experiments. (H) Adhesion assays comparing classically activated (CA) and alternatively activated (AA) BM macrophages with wound-derived macrophages at days 3 and 7 following wounding (n=3 for each condition).

Fig. 6.

Analysis of the role of the leptin receptor on recruitment and retention of inflammatory cells in non-diabetic and diabetic wound environments. (A) Representative images of total cells (DAPI), transplanted Lepr^+/− (wt) and Lepr^−/− (db) BM-derived cells (GFP) and inflammatory cells in sections of non-diabetic (wt) and diabetic (db) wounds in the following donor >recipient combinations: Lepr^−/− donor into Lepr^−/− recipient (db >db, top row of panels), Lepr^+/− donor into Lepr^−/− recipient (wt >db, second row of panels), Lepr^−/− donor into Lepr^+/− recipient (db >wt, third row of panels) and Lepr^+/− donor into Lepr^+/− recipient (wt >wt, bottom row of panels). Scale bar: 200 μm. (B, left panel) Quantification of recruitment and retention data of total BM-derived cells (GFP positive) in the transplant combinations shown in A (n=6; n.s., not significant). (B, left panel) Quantification of recruitment and retention data of BM-derived inflammatory cells (double positive for GFP and CD45) in the transplant combinations shown in A (n=6; n.s., not significant).

Fig. 7.

Analysis of Arg1⁺ cells in human healing and non-healing diabetic foot ulcers. (A) Cartoon diagram showing punch biopsy site, indicated by blue dotted circle. (B) H&E stain of whole-wound section showing representative areas used for analyses of CD68, Arg1 and Nos2 immunofluorescent staining (scale bar: 600 μm). (C) Representative images of CD68 and Arg1 immunofluorescent staining in sections from healing and non-healing diabetic foot ulcers. The first three panels of each group show fluorescent staining as indicated (scale bar: 100 μm). The fourth panel of each group shows a high-magnification view of the area indicated by the white box in the merge panels (scale bar: 25 μm). (D) Quantification of the number of CD68⁺ cells per mm² in healing and non-healing wounds. (E) Quantification of the percentage of macrophages (Macs) that are Arg1⁺ (double positive for Arg1 and CD68) in healing and non-healing wounds (n=7, *P<0.05).