Thiazolidinediones regulate adipose lineage dynamics

Abstract

White adipose tissue regulates metabolism; the importance of this control is highlighted by the ongoing pandemic of obesity and associated complications such as diabetes, atherosclerosis, and cancer. White adipose tissue maintenance is a dynamic process, yet very little is known about how pharmacologic stimuli affect such plasticity. Combining in vivo lineage marking and BrdU labeling strategies, we found that rosiglitazone, a member of the thiazolidinedione class of glucose-lowering medicines, markedly increases the evolution of adipose progenitors into adipocytes. Notably, chronic rosiglitazone administration disrupts the adipogenic and self-renewal capacities of the stem cell compartment and alters its molecular characteristics. These data unravel unknown aspects of adipose dynamics and provide a basis to manipulate the adipose lineage for therapeutic ends.

Figure 1. TZDs Stimulate New Adipocyte Formation

(A) Two-month old AdipoTrak mice were treated with or without TZD while receiving BrdU-water for 2 month, and adipocyte nuclei were probed for BrdU uptake with flow cytometry. Top diagram is the experimental design. X-axis of the flow profiles is GFP intensity, and Y-axis is BrdU (bottom left panels). The percent change of +TZD BrdU+GFP+ events compared to −TZD is displayed in the graph on the bottom right. n = 4 per cohort (B–E) Two-month old AdipoTrak mice were treated as indicated for 2 months and examined for reporter expression. (B) IWAT and RWAT were stained (left panels) or quantified (the graph on the right) for β-galactosidase activity. n = 4 per cohort (C) Adipocyte nuclei were analyzed for GFP with flow cytometry. The relative percentages of GFP+ nuclei are graphed on the right. See also Figure S1. n = 4 to 6 per cohort (D) IWAT and RWAT were stained (left panels) or quantified (blue, the graph on the right) for β-galactosidase activity. AdipoTrak mice carrying a R26R^RFP reporter were similarly treated, and total RFP from all adipocytes isolated from each experimental group were quantified and displayed as a relative value in the graph on the right (red). (E) Adipose tissues were stained with X-Gal, sectioned and counterstained with nuclear fast red (top panels). Black arrows point to small adipocytes rarely seen in non-treated controls. Adipose tissues were examined for GFP, perilipin (lipid droplets) and caveolin-1 (CAV1, cytoplasmic membrane) (bottom panels). White arrows point to adipocytes that contain multiple lipid droplets rarely seen in non-treated controls. *p < 0.05 by two-tailed Student’s t-Test. Error bars indicate SEM. Scale bars: 2mm (B left, D left); 50 μm (E)

Figure 2. TZDs stimulate in vivo progenitor proliferation and differentiation

(A–B) Adult AdipoTrak mice were treated as indicated for 2 months, and adipose SV cells were analyzed for GFP (A) or the apoptotic marker Annexin V (B) with flow cytometry. (A) X-axis is GFP intensity; Y-axis is PE channel to illustrate the distribution of GFP+ cells. The percentage of GFP+ cells are as indicated at top of the flow profiles, and displayed as percent changes in the graph on the right. See also Figure S2. n = 4~6 per cohort (B) X-axis is GFP intensity; Y-axis is Annexin V. The relative percentages of GFP+/Annexin V+ cells compared to −TZD are shown on the right. n = 4 per cohort (C) AdipoTrak mice were injected with BrdU for 5 consecutive days, and then treated with or without TZD for 1 month. Top diagram shows the experimental design. Isolated SV cells (bottom left) and adipocyte nuclei (bottom right) were analyzed for BrdU incorporation with flow cytometry. The relative percentages of GFP+/BrdU+ cells are displayed at the bottom. n = 3~4 per cohort (D) Adipose tissues from mice treated as in (C) were sectioned and stained for GFP, isolectin (red) and BrdU (blue). White arrows indicate label-retaining progenitors located in the vascular niche. The percent of GFP+BrdU+ cells in GFP+ cells on small (1–2 cells in diameter) or relatively larger (3–5 cells in diameter) blood vessels (as stained positive for isolectin) were quantified and displayed on the right. *p < 0.05 by two-tailed Student’s t-Test. Error bars indicate SEM. Scale bars: 50μm (D left)

Figure 3. TZDs exhaust the progenitor pool

(A–C) Adult AdipoTrak mice were treated with or without TZD for 2 months, and the adipose SV cells were analyzed for GFP (A) or sorted for GFP+ cells, plated in equal numbers, and cultured in insulin (B–C). (A) The percentages of GFP+ cells are as indicated at top of the flow profiles. X-axis is GFP intensity; Y-axis is APC channel to illustrate cell distribution. The relative percentages of GFP+ cells, compared to −TZD, are shown on the graph on the right. See also Figure S3. n = 4 per cohort (B–C) Adipogenesis was assayed with Oil Red O staining (left two panels (B), fat stains red), triglyceride quantification (graph on the right, (B)), and qPCR analyses of a panel of adipogenic markers (C). ADPN: adiponectin. (D) SV cells from untreated AdipoTrak mice were isolated, cultured in insulin with or without TZD. Adipogenesis was assayed by Oil Red O staining (left two panels) and triglyceride quantification (right panel). (E) AdipoTrak R26R^lacZ or R26R^RFP reporter mice were treated with or without TZD for 2 months, and adipose SV cells were examined for β-galactosidase by X-gal staining (left panels) and cell counting (2,500~5,500 cells, blue in the graph) or RFP with flow cytometry (red in the graph). *p < 0.05 by two-tailed Student’s t-Test. Error bars indicate SEM. Scale bars: 50 μm (B, D, E)

Figure 4. In vivo TZD treatment alters the molecular characteristics of adipose progenitors

(A–D) Adult AdipoTrak mice were treated with or without TZD for 2 months, and adipose GFP+ SV cells were analyzed with a panel of cell surface markers (A) or sorted and interrogated with gene expression profiling (B) or qPCR (C–D). (A) Some markers reduced by TZD are displayed as percentage of GFP+ cells that co-expressed the marker of interest. (B) The heat map illustrates 133 genes in which TZD induced differential expression. The color bar on the right indicates gene expression level in log 2 scale. n = 3–4 mice per cohort, 3 cohorts for each experimental group (C–D) qPCR analyses of selected genes identified from microarray that are involved in immune response (C) or have established roles in fat formation or obesigenic responses (D). (E) GFP+ SV cells were stained with an anti-p75 NTR antibody, and sorted into p75− and p75+ fractions; equal numbers of cells from each fraction were plated, incubated in insulin, and stained with Oil Red O. *p < 0.05 by two-tailed Student’s t-Test. Error bars indicate SEM. Scale bars: 50 μm (E)