Integrative miRNA-mRNA profiling of adipose tissue unravels transcriptional circuits induced by sleep fragmentation

Abstract

Obstructive sleep apnea (OSA) is a prevalent condition and strongly associated with metabolic disorders. Sleep fragmentation (SF) is a major consequence of OSA, but its contribution to OSA-related morbidities is not known. We hypothesized that SF causes specific perturbations in transcriptional networks of visceral fat cells, leading to systemic metabolic disturbances. We simultaneously profiled visceral adipose tissue mRNA and miRNA expression in mice exposed to 6 hours of SF during sleep, and developed a new computational framework based on gene set enrichment and network analyses to merge these data. This approach leverages known gene product interactions and biologic pathways to interrogate large-scale gene expression profiling data. We found that SF induced the activation of several distinct pathways, including those involved in insulin regulation and diabetes. Our integrative methodology identified putative controllers and regulators of the metabolic response during SF. We functionally validated our findings by demonstrating altered glucose and lipid homeostasis in sleep-fragmented mice. This is the first study to link sleep fragmentation with widespread disruptions in visceral adipose tissue transcriptome, and presents a generalizable approach to integrate mRNA-miRNA information for systematic mapping of regulatory networks.

Figure 1. Outline of a general approach for integrating mRNA-miRNA data.

Concurrent miRNA and mRNA profiling of the same tissue under control and exposure conditions is performed. The mRNA information is processed using gene set enrichment analysis with further refinements based on leading edges of enriched biologic modules. Meanwhile, differentially expressed miRNAs as well as their computationally-derived targets are also identified. These data are then merged based on anti-correlated expression of leading edge genes and their corresponding putative miRNAs.

Figure 2. Integrated mRNA-miRNA analysis of visceral fat tissue transcriptome during SF.

Heatmaps depict gene expression profiles of enriched pathways and differentially expressed miRNAs in response to SF. The miRNAs are linked to their respective modules based on whether these pathways harbor target genes. Note that one mRNA microarray experiment was excluded due to poor hybridization (gray column).

Figure 3. Network analysis of the “insulin regulation and diabetes” pathway.

Members of this enriched module were linked together based on verified gene product interactions. Several distinct motifs were identified (shown in different colors). As expected, insulin assumed a central position in this network. One upregulated candidate, MLX (red), was the putative target of multiple downregulated miRNAs (blue) and is postulated to be a critical controller of visceral adipocytes' metabolic response to SF. Note that several other downregulated miRNAs interacted with members of this network, but are not shown in order to improve clarity.

Figure 4. Acute sleep fragmentation causes dysregulation in glucose homeostasis as demonstrated by the development of insulin resistance in SF mice using the homeostatic model assessment (HOMA-IR).

Data are presented as mean ± SEM.

Figure 5. Reduced glucose tolerance in mice exposed to SF compared to control animals.

Mice (n = 8 control, n = 7 SF) were injected with of D-glucose (2 g/kg) intra-peritoneally and glucose levels measured at 0, 15, 30, 60 and 120 min following injection. P-value was based on two-way ANOVA and corresponds to the significance of exposure (glucose) effect. All data are shown as mean ± SEM.

Figure 6. Chronic sleep fragmentation is associated with elevated circulating triglycerides.

Mice were subjected to 3 weeks of SF (n = 5) or used as controls (n = 5). Fasting plasma triglyceride levels were measured and are presented as mean ± SEM.