Supplementary Figure 5: In vivo GW1929 treatment expands adipose tissue and promotes healthy myelo-erythropoiesis.

(a) Schematic showing 4-day Transwell co-culture of healthy human Lin⁻ cells with healthy human BM-MSCs or differentiated human adipocytes (analysis in b,c). (b) FACS quantification of CD15⁺ granulocyte-lineage cells. Data represent group averages from n = 3 biologically independent experiments. (c) Quantification of human erythroid progenitors using methylcellulose CFU assays. n = 7 (MSCs) and n = 14 (adipocytes) independently assayed wells, pooled from 3 experiments. (d) Heatmap of probe sets most preferentially expressed in mature human osteoblasts versus adipocytes (GSE9451²⁹). (e) Normalized PPARγ transcript expression levels in mature human osteoblasts versus adipocytes (n = 3 biologically independent samples/group, GSE9451²⁹). (f) Experimental design to validate in vivo GW1929 dosing in non-transplanted mice (analysis in g–k; all n = 5 (vehicle) and n = 4 (GW1929) independent mice from 1 experiment). (g,h) Effects of GW1929 treatment on previously reported³⁹ PPARγ-sensitive parameters of fat mass (g) and non-fasted serum glucose (h). (i–k) Whole femur dot plots (i) and image-based quantification of BM adipocyte frequency (j) and size (k). PLN, perilipin. (l) Experimental design to evaluate effects of in vivo GW1929 treatment on BM adipocytes and healthy human myelo-erythroid cells (analysis in m,n). (m,n) Correlation of BM adipocyte frequencies versus human erythroid colony number (m) or FACS-quantified healthy human CD45⁺CD15⁺ granulocytes (n). Data points represent n = 9 independently assayed wells (m) or n = 7 independent xenografted mice (n) from 1 experiment. Findings were reproduced in 3 additional experiments with independent healthy human cells (Fig. 6e, g). (o) Schematic showing 4-day in vitro culture of healthy human Lin⁻ cells in 0.1% DMSO (vehicle) or 20 μM GW1929 (analysis in p,q). (p) FACS quantification of CD15⁺ granulocyte-lineage cells. Data represent group averages from n = 6 biologically independent experiments. (q) Quantification of human erythroid progenitors using methylcellulose CFU assays. Data points represent n = 6 independently assayed wells/group, pooled from 2 experiments. All data are means ± s.e.m. Statistical significance was assessed by paired t-tests (b,p), Mann–Whitney U tests (c,q), two-tailed unpaired t-test (e), one-tailed unpaired t-tests (g,h,j,k), or Pearson’s correlation (m,n). Source data can be found in Supplementary Table 8.