Supplementary Figure 2: AML reduces adipocyte numbers more profoundly in BM versus non-BM sites.

(a) Absolute PLN⁺ adipocyte counts and DAPI⁺ nucleated cell counts per xenograft femur section. n = 9 (healthy) and n = 8 (AML) pooled from 5 experiments (values correspond to frequencies reported in Fig. 2h). (b) Representative images of xenograft femur sections showing that AML engraftment reduces the number of adipocytes per unit area, whereas nucleated cell counts remain stable. Scale bar, 70 μm. (c) Body weights as a function of human leukemic burden in BM. Values reflect n = 7, 2, 7, 7 independent xenografted mice (healthy human chimerism, 20% AML, 20–80% AML, and >80% AML) pooled from 3 experiments (1 healthy sample, AML Patients #2 and #6). (d,e) Nonfasted serum glucose levels in mice reconstituted with healthy human hematopoieisis or human AML (Patient #1). n = 5 biologically independent xenografted mice/group from 1 experiment (1 mouse with <20% AML, 3 mice with 20–80% AML, and 1 mouse with >80% AML). (f) Representative immunofluorescence images of inguinal fat pad tissue collected from mice engrafted with human AML (Patient #2) or matched non-transplanted controls. PLN, perilipin. Low magnification images (left) represent full tissue montage images of up to 347 fields (scale bar, 1.3 mm), while high magnification images (right), represent individual fields (scale bar, 70 μm). BM leukemic chimerism was >90%. (g,h) Adipocyte number (g) and cross-sectional area measurements (h) from fat pad sections shown in f. Area measurements reflect n = 73422 (control) and n = 46242 (AML) independent adipocytes from individual tissue sections. Analysis is representative of 3 independent mice/group from separate experiments. All data are means ± s.e.m. Statistical significance was assessed by unpaired t-tests (a,d), Kruskal–Wallis test (c), or Mann–Whitney U test (h). No statistical analyses were performed in e. Source data can be found in Supplementary Table 8.