Granulocyte/Macrophage Colony-stimulating Factor-dependent Dendritic Cells Restrain Lean Adipose Tissue Expansion

Abstract

The physiological roles of macrophages and dendritic cells (DCs) in lean white adipose tissue homeostasis have received little attention. Because DCs are generated from bone marrow progenitors in the presence of granulocyte/macrophage colony-stimulating factor (GM-CSF), we used GM-CSF-deficient (Csf2(-/-)) mice fed a low fat diet to test the hypothesis that adipose tissue DCs regulate the development of adipose tissue. At 4 weeks of age, Csf2(-/-) mice had 75% fewer CD45(+)Cd11b(+)Cd11c(+)MHCII(+) F4/80(-) DCs in white adipose tissue than did wild-type controls. Furthermore, the Csf2(-/-) mice showed a 30% increase in whole body adiposity, which persisted to adulthood. Adipocytes from Csf2(-/-) mice were 50% larger by volume and contained higher levels of adipogenesis gene transcripts, indicating enhanced adipocyte differentiation. In contrast, adipogenesis/adipocyte lipid accumulation was inhibited when preadipocytes were co-cultured with CD45(+)Cd11b(+)Cd11c(+)MHCII(+)F4/80(-) DCs. Medium conditioned by DCs, but not by macrophages, also inhibited adipocyte lipid accumulation. Proteomic analysis revealed that matrix metalloproteinase 12 and fibronectin 1 were greatly enriched in the medium conditioned by DCs compared with that conditioned by macrophages. Silencing fibronectin or genetic deletion of matrix metalloproteinase 12 in DCs partially reversed the inhibition of adipocyte lipid accumulation. Our observations indicate that DCs residing in adipose tissue play a critical role in suppressing normal adipose tissue expansion.

Keywords: mice, GM-CSF, dendritic cells, adipose tissue.

FIGURE 1.

Csf2^−/− mice exhibit increased adiposity when fed a low fat diet. A and B, body weight. Body composition is presented as lean mass (C and D) and fat mass (E and F) normalized to body weight. Adipose tissue morphology mean adipocyte area (G) and images showing the morphology of the adipose tissue by H&E staining (H).

FIGURE 2.

Characterization of glucose metabolism and energy homeostasis. Glucose tolerance (A) and insulin tolerance (B) tests were performed with adult male mice (n = 12–17; *, p < 0.01; **, p < 0.001). Plasma insulin levels were measured following a 4-h fast (C). Energy homeostasis was assessed by indirect calorimetry for five adult male mice. Ambulatory activity (D), oxygen consumption VO₂ (E), and heat production (F) are shown.

FIGURE 3.

Characterization of adipose tissue myeloid populations by surface markers, gene expression, and antigen presentation. Representative FACS plots showing CD11b- and CD11c-expressing CD45⁺ cells and subpopulations in 4-week-old mice (A, pooled from 10 mice) and 25-week-old mice (D, n = 12). F4/80 and MHCII expression by CD45+CD11b-CD11c+ (B⁻C⁺) and CD45⁺CD11b⁺CD11c⁺ (B⁺C⁺) subpopulations was measured in 4-week-old mice (B and C) and 25-week-old mice (E and F). The myeloid subpopulations were quantified for 25-week-old mice as a percentage of CD45⁺ cells (G) and per g of adipose tissue (H). CD11b and CD11c gene expression in the SVF of epididymal adipose depot was quantified for 25-week-old mice (I, n = 12). A T cell proliferation assay was performed to assess the antigen presentation capability of B⁻C⁺ and B⁺C⁺ myeloid subpopulations (J). Myeloid subpopulations were isolated from pooled SVF from 10 adult wild-type mice.

FIGURE 4.

Characterization and comparison of BM-DCs and in vivo CD45⁺CD11b⁺CD11c⁺ counterparts. A, gene expression profile of expressed normalized BM-DM. Dotted line, BM-DMs; black bars, BM-DCs differentiated in vitro (from 6 sets of experiments). B, secreted proteome determined by proteomic analysis of conditioned medium from BM-DMs and BM-DCs (n = 6). C, DC-specific gene expression in SVF (n = 6). D–H, gene expression profiles for the isolated myeloid cell populations generated in vivo (n = 12).

FIGURE 5.

Inhibition of adipocyte differentiation by BM-DCs. A, visualization of inhibition of adipocyte differentiation. B, quantification of dose-dependent inhibition of adipocyte differentiation. C, quantification of adipocyte differentiation after GM-CSF treatment (mean of 4 experiments is presented). D, adipogenesis gene expression profile for preadipocytes treated with BM-DM-conditioned medium (dotted line) and BM-DC-conditioned medium (white bars). E, adipogenesis gene expression profile in epididymal fat pad of wild-type (dotted line) and Csf2^−/− mice (n = 12).

FIGURE 6.

Inhibition of adipocyte differentiation by CD45⁺CD11b⁺CD11c⁺ myeloid cells. A, increased adipogenesis in preadipocytes treated with BM-DC-conditioned medium from Mmp12^−/− mice. B, increased adipogenesis in preadipocytes treated with conditioned medium from BM-DCs in which Fn1 had been silenced. Both experiments were repeated four times, and representative experiments are presented. C, inhibition of adipogenesis when preadipocytes are co-cultured with CD45⁺CD11b⁺CD11c⁺ cells. The experiment was repeated three times. D–F, panel of adipogenesis genes for the CD45⁺CD11b⁺CD11c⁺ co-culture experiments.