Investigation into the underlying molecular mechanisms of white adipose tissue through comparative transcriptome analysis of multiple tissues

Abstract

Adipose tissue has a primary role in lipid and glucose metabolism as a storage site for fatty acids, and also functions as an endocrine organ, producing large numbers of hormones and cytokines. Adipose dysfunction triggers a number of obesity‑associated health problems. The aim of the present study was, therefore, to investigate the molecular mechanisms of white adipose tissue (WAT). The GSE9954 microarray data were downloaded from the Gene Expression Omnibus. Adipose‑specific genes were identified through limma package analysis, based on samples of WAT and 17 other types of non‑adipose tissue obtained from mice. Process and pathway enrichment analyses were performed for these genes. Finally, protein‑protein interaction (PPI) and co‑expression networks were constructed and analyzed. In total, 202 adipose‑specific genes were identified, which were involved in key biological processes (including fat cell differentiation and lipid metabolic process) and one key pathway [namely, the adenine monophosphate‑activated protein kinase (AMPK) signaling pathway]. Construction of the PPI network and further molecular complex detection revealed the presence of 17 key genes, including acetyl‑CoA carboxylase α, peroxisome proliferator‑activated receptor (PPAR) γ and leptin, that were involved in the AMPK, PPAR and insulin signaling pathways. In addition, amine oxidase copper containing 3 and adrenoceptor beta 3 were communication hubs in the co‑expression network of adipose‑specific genes. In conclusion, the present study promotes our understanding of the underlying molecular mechanisms of WAT, and may offer an insight into the prevention and treatment of obesity‑associated diseases caused by adipose dysfunction.

Keywords: white adipose tissue; Gene Expression Omnibus; microarray data; differentially expressed genes.

Figure 1.

Hierarchical clustering analysis of adipose-specific genes in 58 samples from 18 tissue types included in the GSE9954 data series. Each row represents one gene, with a total of 202 adipose-specific genes. Each column represents one tissue sample (GEO sample number), including 3 epididymal adipose, 3 liver, 3 diaphragm, 3 salivary gland, 3 spleen, 4 muscle, 3 brain, 3 lung, 3 kidney, 3 adrenal gland, 4 bone marrow, 5 pituitary gland, 3 seminal vesicle, 3 thymus, 3 testis, 3 heart, 3 small intestine and 3 eye tissue samples. Relative gene expression is indicated according to the color scale, where red indicates elevated expression and blue indicates reduced expression. Gene expression values were transformed into log₂-based values (distribution range, 6 and 14).

Figure 2.

GO biological processes and KEGG pathways analyses of adipose-specific genes. (A) GO annotations of the adipose-specific genes with the top 20 enrichment scores for biological processes. (B) Interactions and overlap of associated molecules among the top 20 biological processes. (C) KEGG pathway enrichment analysis of the adipose-specific genes with the top 20 enrichment scores. (D) Interaction and overlap of associated molecules among the top 20 pathways. Ellipses represent the most significantly enriched biological processes or pathways, while diamonds indicate the genes that act as a link hinge between biological processes or pathways. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 3.

PPI network complex and MCODE module analysis of adipose-specific genes. (A) The PPI network was visualized using Cytoscape software (isolated nodes were not included). Node size represents the node degree (a larger size indicates a higher degree). The width and transparency of the edge represent the combined score of the edge (a wider or more opaque edge indicates a higher combined score). (B) Top 30 genes with the highest node degrees. (C) MCODE module screening for the top 30 genes and pathway enrichment analysis for the modules. PPI, protein-protein interaction; MCODE, molecular complex detection; AMPK, adenine monophosphate-activated protein kinase; PPAR, peroxisome proliferator-activated receptor.

Figure 4.

Construction of a co-expression network and identification of key candidate genes. (A) Co-expression network of adipose-specific genes. Node size represents the node degree (a larger size indicates a higher degree). The width and transparency of the edge represent the weight score of the edge (a wider or more opaque edge indicates a higher weight score). (B) Aoc3 and Adrb3 were strongly positively correlated with the key genes Lep, Retn and Cidec in 58 tissue samples from 18 types of mouse tissues. Aoc3, amine oxidase copper containing 3; Adrb3, adrenoceptor beta 3; Lep, leptin; Retn, resistin; Cidec, cell death-inducing DFFA-like effector c.