Adipose tissue macrophages promote myelopoiesis and monocytosis in obesity

Abstract

Obesity is associated with infiltration of macrophages into adipose tissue (AT), contributing to insulin resistance and diabetes. However, relatively little is known regarding the origin of AT macrophages (ATMs). We discovered that murine models of obesity have prominent monocytosis and neutrophilia, associated with proliferation and expansion of bone marrow (BM) myeloid progenitors. AT transplantation conferred myeloid progenitor proliferation in lean recipients, while weight loss in both mice and humans (via gastric bypass) was associated with a reversal of monocytosis and neutrophilia. Adipose S100A8/A9 induced ATM TLR4/MyD88 and NLRP3 inflammasome-dependent IL-1β production. IL-1β interacted with the IL-1 receptor on BM myeloid progenitors to stimulate the production of monocytes and neutrophils. These studies uncover a positive feedback loop between ATMs and BM myeloid progenitors and suggest that inhibition of TLR4 ligands or the NLRP3-IL-1β signaling axis could reduce AT inflammation and insulin resistance in obesity.

Figure 1. Leukocytosis in obese mice is due to enhanced myelopoiesis. Ob/Ob or lean controls on a chow diet

A) Representative flow cytometry plots indicating monocyte and neutrophil populations and B) quantified in the accompanying graph. N: neutrophil, M: monocyte, Lo: Ly6-C^lo, Hi: Ly6-C^hi. C) Representative flow cytometry plots depicting the BM progenitor cells and D) quantified. MPC: myeloid progenitor cell. E) Cell cycle analysis of the hematopoietic cells using DAPI. All data mean ± SEM, n=5/group. *P<0.05. See also Figure S1.

Figure 2. Adiposity drives leukocytosis

A) Experimental overview: 1. WT or 2. Ob/Ob BM was transplanted into WT recipients. 3. WT or 4. Ob/Ob BM was transplanted into Ob/Ob recipients. BM was allowed to reconstitute for 5 weeks. B) Monocyte levels, C) BM progenitors and. D) cell cycle analysis of BM progenitors. All data mean ± SEM, n=5/group. *P<0.05. E, F) WT and Ob/Ob mice were treated with a SGLT2i. E) Blood glucose levels. F) Blood leukocytes. All data mean ± SEM, n=5/group. *P<0.05. G) Conditioned media from cultured lean or Ob/Ob VAT was incubated (10% v/v) with BM progenitor cells and GMP proliferation was measured by EdU incorporation. n=6 (independent conditioned media treatments from the VAT of 6 mice/group). H) Experimental overview: FPT: Equal portions of VAT from lean or Ob/Ob mice were transplanted into lean WT recipients. I) Monocyte levels were monitored weekly for 4 weeks. J, K) After 4 weeks the J) abundance and K) cell cycle of the BM progenitor cells were determined by flow cytometry. All data mean ± SEM, n=5/group. *P<0.05. See also Figure S2 and Table S1.

Figure 3. Adipose tissue of HFD-fed mice promotes monocytosis

A–E) WT mice were fed a LFD (Lean) or a HFD (DIO) for 20wks. A) Representative flow cytometry plots depicting blood leukocyte populations and B) quantified. C) Representative flow cytometry plots depicting BM progenitor cells and D) the abundance quantified. E) Cell cycle analysis on the BM progenitor cells. All data mean ± SEM, n=6 lean and 12 DIO mice/group. *P<0.05. F) Experimental overview: FPT: Equal portions of VAT was transplanted into lean WT recipients for 4 weeks. G) Blood leukocyte levels and H) BM progenitors. All data mean ± SEM, n=6/group. *P<0.05. I–L) 16wk HFD-fed DIO obese mice were randomized into 2 groups; 1) DIO-control (DIO-C): continued on a HFD or 2) DIO-weight loss (WL): switched to a chow diet to promote weight loss for 3 weeks. I) Changes in body weight and J) VAT mass after 3 weeks of weight-loss. Circulating K) monocyte and L) neutrophil levels. All data mean ± SEM, n=6/group. *P<0.05. See also Figure S2 and Table S2.

Figure 4. Inflammatory gene expression in the visceral adipose tissue

Lean or Ob/Ob mice were fed a chow diet for 12 wks. Gene expression from A) Total VAT, B) VAT-SVCs and C) VAT-adipocytes was quantified by qPCR. D) Plasma levels of S100A8/A9 in lean and obese mice. E) S100a8 and S100a9 mRNA expression in FACS isolated blood neutrophils. All data mean ± SEM, n=6/group. *P<0.05. See also Figure S3.

Figure 5. MyD88 dependent TLR4 signaling is required for leukocytosis in obesity

A) Experimental overview: Ob/Ob mice were transplanted with BM from donor mice (1–5) and allowed to reconstitute for 5 week before blood leukocytes were assessed. VAT was harvested from Ob/Ob mice after BMT (1, 2) and equal portions of VAT were transplanted into lean WT mice. B–C) Ob/Ob mice BMT with WT, Tlr4^−/− or Rage^−/− BM. B) Blood leukocyte levels and C) OGTT (1g/kg of glucose). D) Blood leukocyte levels in Ob/Ob BMT with WT, Myd88^−/− or Cd14^−/− BM. E) Abundance of hematopoietic progenitor cells in the BM and F) cell cycle analysis. G) Macrophage abundance in the VAT was assessed by immunohistochemistry using anti-Mac3 antibody. X20 objective. H) Expression of inflammatory genes in the SVCs of the VAT from Ob/Ob BMT mice (1, 2). I–K) FPT of VAT from Ob/Ob mice (1, 2). I) Blood leukocyte levels, J) BM progenitor abundance and K) cell cycle analysis. All data mean ± SEM, n=6/group. *P<0.05. See also Figure S4.

Figure 6. TLR4/MyD88 signaling in CD11c⁺ ATMs drives leukocytosis in obesity

A) Experimental outline: Ob/Ob mice were transplanted with BM from Myd88^flx/flx, Myd88^flx/flx LysM Cre or Myd88^flx/flx CD11c Cre and allowed to reconstitute for 5 weeks. B) Schematic overview of the LysM and CD11c Cre induced deletion of Myd88 in the respective myeloid cells and the overlapping CD11c⁺ macrophage. C) Blood leukocyte levels, D) BM progenitor abundance and E) cell cycle analysis. All data mean ± SEM, n=6/group. *P<0.05. F) Experimental overview: Ob/Ob mice were injected intraperitoneally with PBS or clodronate (CLO) liposomes (to deplete ATMs). The VAT was harvested washed and transplanted into lean WT mice for 4 weeks. G) Flow cytometry plots of the ATM subsets (parent gate CD45⁺), H) F4/80 staining of the VAT and I) F4/80 and Cd68 mRNA expression in the VAT. J) Circulating blood monocyte levels, K) BM progenitor abundance and L) proliferation after transplantation. All data mean ± SEM, n=5/group. *P<0.05. See also Figure S5A.

Figure 7. IL-1β signaling contributes to obesity-driven leukocytosis

A–C) Ob/Ob mice were transplanted with WT or Nlrp3^−/− BM and allowed to reconstitute for 5 weeks. A) Representative flow cytometry plots depicting blood leukocytes and quantified. B) Abundance of BM progenitor cells and C) cell cycle analysis. All data mean ± SEM, n=5/group. *P<0.05. D–F) WT and IL-1r^−/− mice were fed a HFD for 6 months. A) Blood leukocyte levels. B) Representative flow cytometry plots depicting the BM progenitors and quantified. MCP: myeloid progenitor cells. C) Representative cell cycle histograms and quantification. All data mean ± SEM, n=6/group. *P<0.05. G) IL-1R expression on CMPs and GMPs in the BM from lean or DIO mice. Expression was normalized to an isotype control. All data mean ± SEM, n=6/group. *P<0.05. H–I) Competitive BMT studies. Ob/Ob mice were transplanted with equal amounts of CD45.1 WT/CD45.2 WT or CD45.1 WT/CD45.2 IL-1r^−/− BM and allowed to repopulate for 6 weeks. Ratios of CD45.1:CD45.2 for H) BM progenitor cells and I) blood leukocytes. All data mean ± SEM, n=5/group. *P<0.05. J) BM derived macrophages from WT, Tlr4^−/− of Myd88^−/− mice were stimulated with S100A8/A9 for 4hrs and gene expression was assessed using qPCR. Dashed red line indicates control; unstimulated WT macrophages. All data mean ± SEM, n=5/group. *P<0.05. K–L) WT mice were fed a HFD for 16 weeks and then given daily i.p injections of Anakinra (IL-1ra; 7.5mg/kg) or saline for 10 days. K) Blood leukocytes and L) OGTT. All data mean ± SEM, n=6/group. *P<0.05. M) Schematic overview of the proposed mechanism of how obesity promotes enhanced myelopoiesis. See also Figure S5.