Coordinated and interactive expression of genes of lipid metabolism and inflammation in adipose tissue and liver during metabolic overload

Abstract

Background: Chronic metabolic overload results in lipid accumulation and subsequent inflammation in white adipose tissue (WAT), often accompanied by non-alcoholic fatty liver disease (NAFLD). In response to metabolic overload, the expression of genes involved in lipid metabolism and inflammatory processes is adapted. However, it still remains unknown how these adaptations in gene expression in expanding WAT and liver are orchestrated and whether they are interrelated.

Methodology/principal findings: ApoE*3Leiden mice were fed HFD or chow for different periods up to 12 weeks. Gene expression in WAT and liver over time was evaluated by micro-array analysis. WAT hypertrophy and inflammation were analyzed histologically. Bayesian hierarchical cluster analysis of dynamic WAT gene expression identified groups of genes ('clusters') with comparable expression patterns over time. HFD evoked an immediate response of five clusters of 'lipid metabolism' genes in WAT, which did not further change thereafter. At a later time point (>6 weeks), inflammatory clusters were induced. Promoter analysis of clustered genes resulted in specific key regulators which may orchestrate the metabolic and inflammatory responses in WAT. Some master regulators played a dual role in control of metabolism and inflammation. When WAT inflammation developed (>6 weeks), genes of lipid metabolism and inflammation were also affected in corresponding livers. These hepatic gene expression changes and the underlying transcriptional responses in particular, were remarkably similar to those detected in WAT.

Conclusion: In WAT, metabolic overload induced an immediate, stable response on clusters of lipid metabolism genes and induced inflammatory genes later in time. Both processes may be controlled and interlinked by specific transcriptional regulators. When WAT inflammation began, the hepatic response to HFD resembled that in WAT. In all, WAT and liver respond to metabolic overload by adaptations in expression of gene clusters that control lipid metabolism and inflammatory processes in an orchestrated and interrelated manner.

Figure 1. HFD feeding leads to obesity and onset of adipose tissue inflammation.

APOE*3Leiden transgenic mice were fed a HFD for 12 weeks and sacrificed at the time points indicated. The average body weight at the start (t=0) of HFD feeding was 29.2 g. A, Body weight gain over time. B, Mass of the epididymal adipose tissue depot during obesity development. Data are presented as mean ± SEM. C, Histological analysis of adipose tissue at start (t=0) and after 12 weeks of HFD or chow feeding (reference for the effect of aging). D, HFD feeding results in adipocyte hypertrophy. Computer-assisted quantification of average adipocytes size (P<0.05). E, Marked accumulation of CCR2 positive cells (arrows) in the HFD fed group.

Figure 2. Cluster analysis of genes of lipid metabolism.

Bayesian cluster analysis of genes with ‘lipid metabolism’ gene ontology resulted in 6 clusters (A, B, C, D, E, and F) with distinct time profiles. Individual gene expression profiles are shown as dotted lines. The bold line represents the cluster average.

Figure 3. Cluster analysis of inflammatory genes.

Bayesian cluster analysis of genes with ‘inflammation’ gene ontology resulted in 4 clusters A, B, C and D with distinct time profiles. Individual gene expression profiles are shown as dotted lines. The bold line represents the cluster average.

Figure 4. Histological analysis of livers.

Hallmarks of non-alcholic fatty liver disease were scored in the livers of the mice that were used for WAT analysis. A, Representative photomicrographs of liver cross-sections after 12 weeks of HFD shows pronounced liver steatosis characterized by micro- and macrovacuolization and hepatocellular hypertrophy. B, Quantitative analysis of total vacuolization and hypertrophy. Data are presented as mean ± SEM (P<0.05).

Figure 5. Comparison of gene expression in liver and WAT over time and analysis of transcriptional regulators.

Venn diagrams of genes with A, ‘lipid metabolism’ gene ontology or B, ‘inflammation’ gene ontology. Time course analysis of the genes that were differentially expressed genes at a particular time point. The intersection represents the number of ‘overlapping genes’, ie. genes that were affected in both tissues. C, Quantitative analysis of the transcriptional activity of HNF4α by TransAM analysis at t=0 and t=12 weeks of HFD feeding relative to reference mice on chow to correct for aging. *P<0.05. D, Differentially expressed target genes of Srebf2 in WAT and liver. Srebf2 is significantly involved in the control of target genes (P<0.05 for both WAT and liver). In both tissues, the calculated Z-score was positive (3.7 for liver and 4.3 for WAT) indicating that Srebf2 is activated. Genes colored in red (green) are upregulated (downregulated).