Overexpression of S100A9 in obesity impairs macrophage differentiation via TLR4-NFkB-signaling worsening inflammation and wound healing

Abstract

Rationale: In obesity the fine-tuned balance of macrophage phenotypes is disturbed towards a dominance of pro-inflammatory macrophages resulting in exacerbation and persistence of inflammation and impaired tissue repair. However, the underlying mechanisms are still poorly understood. Methods: Impact of obesity on macrophage differentiation was studied in high fat diet induced obese and db/db mice during skin inflammation and wound repair, respectively. Mechanisms of S100A9-mediated effects on macrophage differentiation was studied on in vitro generated macrophages by genomic and proteomic approaches. The role of S100A9 on macrophage differentiation was investigated by pharmacological inhibition of S100A9 during skin inflammation and wound repair in obese and db/db mice. Results: We demonstrate an overexpression of S100A9 in conditions of obesity-associated disturbed macrophage differentiation in the skin. We show that saturated free fatty acids (SFA), which are increased in obesity, together with S100A9 induce TLR4 and inflammasome-dependent IL-1β release in macrophages which in turn amplifies S100A9 expression initiating a vicious cycle of sustained S100A9 overexpression in skin inflammation in obesity. We reveal a yet unrecognized impact of obesity-associated S100A9 overexpression on macrophage differentiation. S100A9 binding to TLR4 and activation of NFkB attenuates development of M2-like macrophages and induces pro-inflammatory functions in these cells. Consequently, inhibition of S100A9 restores disturbed M2-like macrophage differentiation in mouse models of obesity-associated skin inflammation and wound repair. Similarly, breaking the vicious cycle of S100A9 overexpression by dietary reduction of SFA restored M2-like macrophage activation. Improvement of skin inflammation and wound repair upon reduction of S100A9 by pharmacological inhibition or by reduction of SFA uncovers the pathogenic role of S100A9 overexpression in obesity. Conclusion: This study identifies S100A9 as a previously unrecognized vital component in obesity-associated disturbed macrophage differentiation and subsequent impaired regulation of inflammation and wound repair. The findings open new opportunities for therapeutic implications for inflammatory diseases and wound repair in obesity.

Keywords: S100A9; macrophage differentiation; obesity; skin inflammation; wound healing.

Figure 1

Disturbed M2-like macrophage differentiation is associated with increased S100A9 expression. A-E) Male C57BL/6 mice were fed with HFD for 15 weeks (obese, red) or chow diet (blue). Skin inflammation was induced by one topical application of imiquimod on the shaved back. A) Relative S100A8 and S100A9 gene expression in the skin over time course detected by quantitative PCR. B) Quantification of densitometric evaluation of S100A9 protein expression in lesional skin at day 4 by Western Blot. GAPDH was used as loading control. C) Relative gene expression of relmα in the skin over time course detected by quantitative PCR. D) Quantification of CD206+ and CD301+ expressing CD45+/CD11b+/Ly6G- cells in lesional skin (d4) by flow cytometry. E) Relative gene expression of M2-like markers in CD11b+ cells isolated from lesional skin at d4 analyzed by quantitative PCR. F-I) 13 weeks old female db/db mice or female C57BL/6 mice were wounded with a 6 mm punch biopsy. F) Relative gene expression of S100A8 and S100A9 in wound tissue analyzed by quantitative PCR analyzed at indicated time points. G) Detection of S100A9 in wounds at day 12 by Western Blot. GAPDH was used as loading control. Quantification of densitometric evaluation of S100A9 protein expression. H) Relative gene expression of relmα in wound tissue at indicated time points analyzed by quantitative PCR. I) Relative gene expression of indicated genes in CD11b+ cells isolated from wounds at day 9 analyzed by quantitative PCR. J-L) Male C57BL/6 mice were fed with HFD for 6 weeks (HFD, red) or chow diet (blue). Skin inflammation was induced by one topical application of imiquimod on the shaved back and followed over time. J) S100A8 and S100A9 expression in the inflamed skin detected by quantitative PCR. K) Quantification of CD206+ and CD301+ expressing CD45+/CD11b+/Ly6G- cells in lesional skin at day 4 by flow cytometry. L) Relative gene expression of M2-like markers in CD11b+ cells isolated from lesional skin analyzed at day 5 by quantitative PCR. Each symbol represents one mouse. Mean is indicated. Unpaired t-test or Mann-Whitney test: *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 2

S100A9 impairs M2-like macrophage differentiation. A/B) Peritoneal cells were incubated with IL-4 or IL-4+S100A8 homodimer, IL4+S100A9 homodimer or IL-4+S1008/A9 heterodimer. A) Gene expression analysis by quantitative PCR. B) Cytokine levels in the supernatant detected by ELISA. Each symbol represents macrophages from one mouse. Mean is indicated. ANOVA with multiple comparisons *P < 0.05, **P < 0.01, ***P < 0.001, **** P < 0.0001. C/D) Peritoneal macrophages were incubated with LPS / INFγ or IL-4 to induce a M1 and M2-like macrophage phenotype, respectively. S100A9 was added during M2-like macrophage differentiation. Genome wide expression analysis. Display of the 60 most upregulated gene in (C) M1- versus M2-like macrophages and (D) M2- versus M1-like macrophages. E) Pathway analysis using the KEGG database of differentially expressed proteins and phosphoproteins with a log fold change of > 2 or < 0.5 identified by proteome and phospho-proteome microarray. Macrophages differentiated with IL-4 and IL-4+S100A9 of two different mice were compared. F/G) Detection and quantification of Stat6, p38, p65NFkB, erk1/2, and Akt, phosphorylation in peritoneal macrophages incubated with LPS / INFγ, IL-4 or IL-4 and S100A9 by Western Blot. Rpl-26 served as loading control. Representative Western Blots are shown (F). Quantification of phosphorylation (G). Each symbol represents macrophages from one mouse. Mean is indicated. Unpaired t-test or Mann-Whitney test between IL-4 and IL-4+S100A9 treated cells: *P < 0.05, **P < 0.01, ***P < 0.001, ns - not significant. H) Peritoneal macrophages were incubated with IL-4 or IL-4+S100A9 in the presence or absence of TLR4 signaling inhibitor (CLI-095, 1 µM), RAGE antagonistic peptide (ELKVLMEKEL, 10 µM), NFkB inhibitor (BMS-345541, 4 µM), PI3K inhibitor (Apitolisib, 5 µM), or MAPK inhibitor (Cobimitinib, 10 nM). Data represent the mean of ≥ 4 independent experiments. Numbers indicate the mean of AU (RS36-normalized) detected by quantitative PCR and protein detected by ELISA.

Figure 3

Inhibition of S100A9 restores M2-like macrophage differentiation and prevents high fat diet- and obesity induced exacerbation of skin inflammation. Male C57BL/6 mice were fed with chow diet (blue) or high fat diet (red) for 15 weeks (obese, A-D) or 6 weeks (HFD, E-J). Skin inflammation was induced by one topical application of imiquimod on the shaved back and ears. Paquinimod was applied 10 h after imiquimod (HFD+Paq; obese+Paq; green). Skin inflammation was analyzed at day 4. A/E) Quantification of CD206+ and CD301+ macrophages within CD45+/CD11b+/Ly6G- cells in the lesional skin by flow cytometry. B/F) Relative expression of indicated genes detected by quantitative PCR in CD11b+ myeloid cells isolated from lesional skin. C/H) Severity score, D/I) Ear swelling, G) Macroscopic appearance of lesional skin area. Images of mice (one representative example of each group with at least 8 mice). J) Epidermal thickness within skin lesions quantified in hematoxylin and eosin stainings of skin tissue sections. Each dot represents one mouse. Mean is indicated. ANOVA with multiple comparisons *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4

Inhibition of S100A9 restores M2-like macrophage differentiation and improves wound healing in db/db mice. Db/db mice were wounded with 6 mm punch biopsies. Paquinimod was applied 2 d after wounding (db/db+Paq; green). Wounds were analyzed 9 days after wounding. A) Relative expression of S100A9 in wounds detected by quantitative PCR. B) Quantification of CD206+ and CD301+ macrophages within CD45+/CD11b+/Ly6G- cells in the wounds by flow cytometry. C) Relative expression of indicated genes detected by quantitative PCR in CD11b+ myeloid cells isolated from wounds. D) Masson's trichrome staining of tissue sections of wounds. One representative example out of 5 db/db and 5 paquinimod treated db/db mice is shown. E/F) Area of granulation tissue and epidermal thickness was quantified in Masson trichrome stainings of tissue sections. Each dot represents one mouse. Mean is indicated. Unpaired t-test or Mann-Whitney test: *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5

S100A9 is expressed in keratinocytes and dermal white adipose tissue during inflammation and tissue repair. A-D) Detection of S100A9 (turquoise) by immunofluorescence staining. Nuclei were stained with DAPI (violet). Bar 50 µm. A) Staining with isotype control antibody. B) S100A9 in healthy skin of lean wildtype mice and of db/db mice. C) S100A9 in imiquimod-induced inflamed skin of wildtype mice at d4. D) S100A9 in wounds of db/db mice at d9. E) Male C57BL/6 mice were fed with high fat diet or chow diet for 15 (obese) weeks. Skin inflammation was induced by one topical application of imiquimod (IMQ). Relative gene expression of S100A9 in epidermal cells of chow and obese mice and of lesional skin of chow (chow-IMQ, blue) and obese mice (obese-IMQ, red) at day 4 of inflammation. F) Detection of S100A9 relative gene expression by quantitative PCR in cultured primary murine keratinocytes stimulated with 10 ng/ml TNFα, 10 ng/ml IL-1β, 10 ng/ml TNFα/IL-1β, 1.5 µg/ml S100A9, 1 µg/ml insulin, 25 mM glucose or 500 µM palmitic acid (PA) complexed with BSA or BSA alone. G) Detection of S100A9 in dWAT areas in inflamed skin of obese mice and in wounds of db/db mice by immunofluorescence staining (turquoise). Nuclei were stained with DAPI (violet). H) Relative S100A9 gene expression in dWAT from chow (blue) and obese (red) and in dWAT, adipocytes and stromal vascular fraction (SVF) isolated from lesional skin of imiquimod treated chow and obese mice at day 6. I) Ex vivo skin cultures with biopsies of wildtype mice were stimulated with 10 ng/ml TNFα, 10 ng/ml IL-1β, 10 ng/ml TNFα/IL-1β, 1.5 µg/ml S100A9, 1 µg/ml insulin, 25 mM glucose or 500 µM palmitic acid (PA) complexed with BSA for 24h. Relative gene expression of S100A9 was detected in dWAT isolated from the skin. J/K) Skin biopsies from lean chow-fed mice were stimulated with IL-1β in the presence of NFkB inhibitor (BMS-345541, 4 µM) or MAPK inhibitor (Cobimitinib, 0.2 µM). Relative S100A9 gene expression in the epidermis (J) and dWAT (K) was detected by quantitative PCR. Each dot represents one mouse. Mean is indicated. ANOVA with multiple comparisons or unpaired t-test: *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 6

S100A9 induces IL-1β secretion in macrophages in the presence of saturated fatty acids. A) Myeloid peritoneal cells from lean chow-fed mice were incubated with 500 µM BSA-complexed PA, stearic acid (SA), or oleic acid (OA), BSA alone or 25 mM glucose (hGlc). IL-1β and TNFα levels within the supernatants were detected by ELISA. B/C) Myeloid peritoneal cells from lean chow-fed mice were incubated with S100A9 in the presence or absence of TLR4 signaling inhibitor (CLI-095, 1 µM), RAGE antagonistic peptide (ELKVLMEKEL, 10 µM), NFkB inhibitor (BMS-345541, 4 µM). B) Relative IL-1β gene expression detected by quantitative PCR; C) IL-1β levels in the supernatant detected by ELISA. D) Myeloid peritoneal cells from lean chow-fed mice were incubated with 500 µM BSA-complexed PA, stearic acid (SA), or oleic acid (OA), or 25 mM glucose followed by stimulation with S100A9. IL-1β and TNFα levels within the supernatants were detected by ELISA. E/F) Myeloid peritoneal cells from lean chow-fed mice were incubated with 500 µM PA-BSA complexes or BSA alone followed by stimulation with S100A9. Inflammasome activation was inhibited by addition of E) 25 µg/ml casapase-1 inhibitor (Ac-YVAD-cmk) or F) 20 mM N-acetyl-L-cysteine (NAC), 10 µM diphenyliodoniumchloride (DPI) or 5 µM Ca074 Me. IL-1β was detected after 24 h by ELISA. Each dot represents one mouse. Mean is indicated. ANOVA with multiple comparisons: *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 7

Saturated fatty acids in combination with S100A9 drive S100A9 overexpression in skin inflammation via induction of IL-1β secretion in macrophages. A) Skin biopsies from lean chow-fed mice were cultured with supernatants of PA-, BSA-, PA/S100A9- and BSA/S100A9-stimulated myeloid cells (Ma) or medium. IL-1β and TNFα were neutralized by pre-incubation with 200 ng/ml IL-1 receptor antagonist (IL1RA) and 1 µg/ml anti-TNFα antibody, respectively. Relative expression of S100A9 in the skin was detected by quantitative PCR. B-D) Male C57BL/6 mice were fed with high fat diet (HFD, red) or chow diet (blue) for 6 weeks. Skin inflammation was induced by one topical application of imiquimod on the shaved back. B) Detection of IL-1β in untreated skin of lean, HFD and obese mice and in lesional skin (day 4) of chow and HFD by ELISA. IL-1β concentration related to protein content is shown. C) Detection of mature caspase in lesional skin (day 4) by Western Blot. GAPDH was used as loading control. D) Detection of S100A9 in lesional skin at day 4 by Western Blot. GAPDH was used as loading control. E-G) Male C57BL/6 mice were fed with chow diet (blue) or high fat diet (HFD, red) for 6 weeks. Skin inflammation was induced as in B-D. Paquinimod was applied 10h after imiquimod (HFD+Paq; green). Mice were analyzed at day 4. E) Detection of IL-1β in lesional skin by ELISA. IL-1β concentration related to protein content is shown. F) Relative S100A9 gene expression in the epidermis detected by quantitative PCR. G) Detection and densitometric evaluation of S100A9 by Western Blot. GAPDH was used as loading control. H-K) Male C57BL/6 mice received either HFD for 6 weeks (red), HFD for 5 weeks followed by chow diet for one week (green), or chow diet for 6 weeks (blue). Skin inflammation was induced as in B-D. H) Concentrations of saturated fatty acids in mM, C14:0 - myristic acid, C16:0 - palmitic acid, C18:0 - stearic acid. I) Detection of IL-1β in lesional skin by ELISA. IL-1β concentration related to protein content. J) Relative gene expression of S100A9 in the epidermis detected by quantitative PCR. K) Detection of S100A9 in lesional skin by Western Blot. GAPDH was used as loading control. Each dot represents one mouse. Mean is indicated. ANOVA with multiple comparisons: *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 8

Overexpression of S100A9 in obesity impairs M2-like macrophage differentiation resulting in amplified and sustained inflammation as well as impaired tissue repair. A) Obesity and high fat diet result in elevation of saturated fatty acids (SFA), which have no impact on keratinocytes (KC) in homeostatic conditions. B) S100A9 release upon injury/inflammation in the absence of elevated SFA does not promote inflammasome activation in resting macrophages/monocytes. C) Upon injury/inflammation S100A9 is released by keratinocytes and dermal white adipose tissue (dWAT). S100A9 binding to TLR4 and activation of NFkB induces gene expression of IL-1β in macrophages. In the presence of high SFA, elevated by obesity or HFD, inflammasome is activated in the cells resulting in IL-1β release. Macrophage-derived IL-1β in turn amplifies S100A9 expression in epidermis and dWAT via NFkB and MAPK signaling pathways, initiating a perpetuating S100A9 response in inflammation in obese skin. S100A9 impairs an appropriate macrophage activation and polarization via TLR4-NFkB signaling preventing resolution of skin inflammation that finally contributes to a sustained and amplified skin inflammation and impaired tissue repair.