Three genetically diverse mouse strains gained different amounts of body weight and fat when fed a high fat high sugar diet

(a) Metabolic phenotyping schedule (n = at least 4 independent mice for each condition, except for glucose tolerance tests n = at least 12 independent mice). (b-c) Weekly (b) and final (c) body weight measurements in CAST (green), B6 (grey), and NZO (blue) mice starting at six weeks of age (week 0) continuing for 9 weeks on HFHS, H, (dashed) or control, C, (solid) diets. (d) Adiposity (fat mass/total mass x 100) percent after 5 weeks on HFHS (H) or control (C) diet. Box plots show median interquartile range, whiskers indicate 1.5 times interquartile range. Statistical analyses were performed using a two-way ANOVA followed by a Tukey post-hoc test. *P<0.05; **P<0.01. Abbreviations: HFHS or H, high-fat high-sugar; C, control diet; NMR, nuclear magnetic resonance; IC, indirect calorimetry; GTT, glucose tolerance test; B, B6; C, CAST; N, NZO.

The HFHS diet worsened glucose intolerance and hyperinsulinemia in NZO mice

(a) Intraperitoneal glucose tolerance test (GTT) performed after 7 weeks on H or C diet. (b) Fasting blood glucose levels after 7 weeks on H or C diet. (c) Blood glucose levels two hours after intraperitoneal glucose bolus. (d) Adjusted glucose AUC. (e-g) Serum insulin (picomole per liter) (e) c-peptide (picomole per liter) (f) and molar ratio of insulin to c-peptide (g) measured after four-hour fast in B6 (grey), CAST (green) or NZO (blue) mice fed HFHS (H) or control (C) diet for eight weeks (n = 4 independent mice for each condition). (h) WPIC (nanogram per milliliter per gram of tissue) measured from mice fed H or C diets for nine weeks. (i) Representative immunohistochemical stain for insulin on fixed paraffin embedded pancreas samples after nine weeks on H or C diet. Scale bar = 50 µM. (j) Representative hematoxylin and eosin staining on NZO mice fed C or H for nine weeks showing islets (dashed circles), lipid droplets (asterisk), and immune infiltration (arrowheads). Scale bar = 100 µM. GTT timepoints are plotted as mean ± s.e.m. Box plots depict interquartile range and median, whiskers indicate 1.5 times interquartile range. Statistical analyses were performed using a two-way ANOVA followed by a Tukey post-hoc test. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. Abbreviations: AUC, area under the curve; WPIC, whole pancreas insulin content; C, control diet; H, high-fat high-sugar diet.

Strain-specific energy expenditure is independent of diet

(a) Bodyweight adjusted energy expenditure (metabolic rate) is shown over a 54-hour period with dark/light cycle indicated by shading. Feeding status (top bar) and strain indicated by B6 (grey), CAST (green) and NZO (blue) mice. (b) Distance traveled on wheel (meters per hour) over 54-hour period. (c) Respiratory exchange ratio (volume of CO2 per volume of O2) over 54-hour period with solid line indicating the food quotient for the control diet and the dotted line indicating the food quotient for the HFHS diet. Error bars indicate means ± s.e.m. Abbreviations: HFHS, high-fat high-sugar; EE, Energy expenditure; RER, respiratory exchange ratio; FQ, food quotient.

Transcriptional profiling of metabolic tissues reveals genetically driven heterogeneity in the HFHS diet response

(a) Tissue collection summary. (b) UMAP plot based on the expression levels of 9,114 genes expressed in all tissues and among the top 3,000 most variable genes for at least one tissue. B6, CAST, and NZO mice showed large differences in gene expression in each tissue independent of diet. Colors of the outline ellipses indicate tissues denoted in (a). (c) Numbers of genes identified by differential gene expression analyses that modeled the effects of strain, sex, diet and their interactions, or compared HFHS vs Control by strain. (d) Venn diagrams illustrate uniqueness of the HFHS vs Control diet differential expressed genes for each strain. (e) Overrepresentation analyses showed Hallmark MSigDB pathways significantly affected by the diet response. (f-g) Volcano plots depic the response of adipose tissue to HFHS diet in B7, CAST, and NZO mice and differences among genes with the largest changes without (f) and with (g) strain effects. (h-i) Enrichment scores derived from GSEA based on diet responsive (no strain effect) (h) and strain x diet genes (with strain effect) (i) showed that specific Hallmark MSigDB pathways were affected in a common and strain-specific response to HFHS diet.

Differential Immune Sensitivity of Adipose Tissue to HFHS Diet Across Mouse Strains

(a) Heatmap depicting hierarchical clustering of scaled levels of serum cytokine levels showed large differences between B6, CAST, and NZO on HFHS or Control diets. (b-d) Serum levels of TNF, MCP-1 and IP-10 varied between B6, CAST, and NZO mice. (e-g) Gene expression of Tnf, Ccl2, and Cxcl10 in adipose tissue. (h-j) Gene expression of Tnf, Ccl2, and Cxcl10 in pancreatic islets. (k) Serum levels of EPO in B6, CAST and NZO mice. (l) Gene expression of Epo, in kidney. Gene expression: **FDR< 0.01; Serum cytokines: *P<0.05, **P<0.01.

CAST haplotype on Chr19 is associated with reduced monocyte content in DO adipose tissue, shows human-mouse correlation with lower BMI

(a) Deconvolution of gene expression data was used to estimate the monocyte fraction in adipose tissue and predicted differences between B6, CAST, and NZO mice. (b) The estimated monocyte fraction in DO mice showed variability similar to B6, CAST, and NZO mice. (c) Positive correlation of serum leptin, body weight and fat pad weight with monocyte fraction of DO mice. (d) LOD scores varied across the genome and a significant genetic association was identified on chromosome 19. (e) Allele effects for the QTL on chromosome 19 in A/J (yellow), B6 (grey), 129S1/SvlmJ (pink), NOD/ShiLtJ (dark blue), NZO (light blue), CAST (green), PWK/PhJ (red), and WSB/EiJ (purple) mouse strains. (f) Mediation analyses identified eight candidate genes that decreased the LOD score by 4 or more. (g) The human region orthologous to the Chr19 locus had very strong genetic evidence for an association with BMI. (h-i) Expression of Chr19 locus candidate genes, Cfl1 (h) and 2010003K11Rik (i), showed expression patterns between strains that matched susceptibility to HFHS diet induced obesity. Dashed line indicates LOD=7.4. *FDR<0.05; **FDR< 0.01. Abbreviations: DO, Diversity Outbred; QTL, quantitative trait loci; LOD, logarithm of the odds.

MHC class I gene expression are associated with heightened risk of obesity caused by HFHS diet

(a) Heatmap showing correlation between gene set enrichment scores for hallmark pathways using strain x diet genes and various anthropometric and glycemic traits in DO mice. (b) Top results of QTL mapping and mediation based on strain x diet gene expression levels in adipose tissue of DO mice fed a HFHS diet. QTL with significant LOD scores were identified for five pathways. Mediation analysis was used to determine whether significance was dependent on the expression of genes within each loci. (c-e) Adipose tissue expression of candidate genes in the Chr17 locus, H2-K1, H2-Q7, and Tap1, showed large differences in B6, CAST, and NZO mice, consistent with differences in body weight and obesity. (f) Summary of human orthologues and genetic evidence for their role obesity and diabetes related phenotypes. Genetic evidence was determined using the HuGE calculator49. Abbreviations: Tg, triglyceride; WPIC, whole pancreas insulin content; Gcg, glucagon.