A history of obesity leaves an inflammatory fingerprint in liver and adipose tissue

Abstract

Background/objectives: Dieting is a popular yet often ineffective way to lower body weight, as the majority of people regain most of their pre-dieting weights in a relatively short time. The underlying molecular mechanisms driving weight regain and the increased risk for metabolic disease are still incompletely understood. Here we investigate the molecular alterations inherited from a history of obesity.

Methods: In our model, male high-fat diet (HFD)-fed obese C57BL/6J mice were switched to a low caloric chow diet, resulting in a decline of body weight to that of lean mice. We measured body composition, as well as metrics of glucose, insulin and lipid homeostasis. This was accompanied by histological and gene expression analysis of adipose tissue and liver to assess adipose tissue inflammation and hepatosteatosis. Moreover, acute hypothalamic response to (re-) exposure to HFD was assessed by qPCR.

Results & conclusions: Within 7 weeks after diet switch, most obesity-associated phenotypes, such as body mass, glucose intolerance and blood metabolite levels were reversed. However, hepatic inflammation, hepatic steatosis as well as hypertrophy and inflammation of perigonadal, but not subcutaneous, adipocytes persisted in formerly obese mice. Transcriptional profiling of liver and perigonadal fat revealed an upregulation of pathways associated with immune function and cellularity. Thus, we show that weight reduction leaves signs of inflammation in liver and perigonadal fat, indicating that persisting proinflammatory signals in liver and adipose tissue could contribute to an increased risk of formerly obese subjects to develop the metabolic syndrome upon recurring weight gain.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure 1

Ad libitum switch to low-fat diet reverses diet-induced obesity phenotype of male C57BL/6J mice. (a) Body weight curve of lean (n=23), obese (n=24) and formerly obese mice (n=23) Statistical difference was determined via two-way ANOVA and reached significance with P< 0.05 (*) Significance is indicated for formerly obese against lean mice. (b) Lean mass and fat mass was determined in a subgroup of lean, obese and formerly obese mice (n=10–11), after 27 weeks of feeding and 35 weeks of age. (c) Intraperitoneal glucose tolerance (GTT) test (glucose 2 g per kg body weight) in a subgroup of formerly obese, lean and obese mice (n=11–12) at 35 weeks of age. (d) Area under the curve (AUC) of performed GTT (n=11–12). Data are presented as mean±standard error of the mean (s.e.m.).

Figure 2

Depot-specific adipose tissue heterogeneity after a history of obesity. (a) Subcutaneous (scWAT) and perigonadal (gWAT) tissue wet weight (g) of lean, obese and formerly obese mice after 27 weeks of experimental diet (n=12). (b) Frequency distribution of adipocyte cell sizes (μm2) from scWAT and gWAT of lean, obese and formerly obese mice (n=4–5). Significant differences are indicated with+(formerly obese vs lean mice), # (obese vs formerly obese) and *=P<0.05 (obese vs lean) and were determined via two-way ANOVA followed by Tukey’s multiple comparison test. (c) % of counted crown-like structures (CLS) per number of counted adipocytes of scWAT and gWAT (n=4–5). (d) qPCR analysis of macrophage markers in scWAT (left panel) and gWAT (right panel). Expression levels are normalized to housekeeping gene TBP and shown as fold-change compared to the lean group (n=7–11). (e) Representative sections from H&E-stained scWAT (left panel) and gWAT (right panel) of lean, obese and formerly obese mice (× 200 magnification, scale bar=50 μm). Arrows indicate CLS.

Figure 3

A history of diet-induced obesity covers a proinflammatory adipogenic transcription profile. (a) Heat map of 309 significantly (P<0.01) regulated genes in gWAT of formerly obese vs lean mice (n=6). Top up- and downregulated genes are presented separately. A detailed list of the 309 differentially expressed genes can be found in Supplementary Table 2. Statistical analysis was performed with the limma t-test and FC>1.3 ×, P<0.01 was considered as statistical significant. (b) Top five statistically significant (P<0.001) enriched canonical pathways associated with the 309 differentially expressed genes of formerly obese vs lean mice. A complete list of significantly enriched canonical pathways is attached in Supplementary Table 3. (c) Venn-diagram of (overlapping) significantly differentially expressed genes (FC>1.3, FDR<10%) in three pairwise comparisons ‘formerly obese vs lean’ (left panel), ‘obese vs lean’ (upper panel) and ‘formerly obese vs obese’ (right panel). A detailed list of the differentially expressed genes regarding each comparison can be found in Supplementary Tables 2, 4 and 5.

Figure 4

Ad libitum switch to low-fat diet partly reverses hepatic steatosis. (a) Liver weights of lean, obese and formerly obese mice, after 27 weeks of feeding the experimental diets (n=10–12). (b) Triglyceride contents (μg per mg of liver tissue) from livers of lean, obese and formerly obese mice (n=7–8). (c) Representative Masson’s trichrome-stained liver sections of a lean, obese and formerly obese mouse (× 200 magnification, scale 100 μm). Arrows indicate fibrotic lesions around the portal area and portal-to-portal bridging. (d) Histological scoring of H&E- and Masson’s Trichrome-stained liver sections (n=5–8). Data are given as boxplots indicating mean, minimum and maximum. (e) QPcr analysis of hepatic genes. Expression levels are normalized to housekeeping gene HPRT and shown as fold-change compared to the lean group. Data are presented as mean±standard error of the mean (s.e.m.) and analyzed using one-way ANOVA followed by Tukey’s multiple comparison test (*P<0.05, **P<0.01).

Figure 5

A history of diet-induced obesity covers a proinflammatory hepatic transcription profile. (a) Heat map of 322 significantly (P<0.01) regulated hepatic genes in the comparison of formerly obese and lean mice (n=6). Top up- and downregulated genes are presented separately. A detailed list of the 322 differentially expressed genes is attached in Supplementary Table 6. (b) Top five statistically significant (by P-value; P<0.01) activated ‘Diseases and Bio Functions’ associated with the 322 regulated genes. A detailed list of annotated terms can be found in Supplementary Table 7. (c) Significantly activated predicted upstream regulators of the 322 regulated hepatic genes. A list of all predicted activated upstream regulators of the 322 differentially expressed genes can be found in Supplementary Table 8. (d) Venn-diagram of (overlapping) differentially expressed genes (FC>1.2 P<0.01) in three pairwise comparisons ‘formerly obese vs lean’ (left panel), ‘obese vs lean’ (upper panel) and ‘formerly obese vs obese’ (right panel). A detailed list of the differentially expressed genes regarding each comparison can be found in Supplementary Tables 6, 9–13.

Figure 6

A history of diet-induced obesity does not facilitate hyperphagia within 48 h of re-feeding. (a) Body weight of lean and formerly obese mice, (re-) fed with HFD for 48 h (n=12). (b) Cumulative energy intake from lean and formerly obese mice, (re-) fed with HFD for 48 h (n=6) calculated as kcal/48 h/2 mice. (c) Plasma metabolites of lean and formerly obese mice (re-) fed with HFD for 48 h (n=8). (d) mRNA expression of hypothalamic orexigenic and inflammatory genes lean and formerly obese mice (re-) fed with HFD for 48 h and obese mice (n=6). Data are given as mean±standard error of the mean (s.e.m.). Statistical significance was indicated with *=P<0.05.