Altered miRNA expression network in locus coeruleus of depressed suicide subjects

Abstract

Norepinephrine (NE) is produced primarily by neurons in the locus coeruleus (LC). Retrograde and ultrastructural examinations reveal that the core of the LC and its surrounding region receives afferent projections from several brain areas which provide multiple neurochemical inputs to the LC with changes in LC neuronal firing, making it a highly coordinated event. Although NE and mediated signaling systems have been studied in relation to suicide and psychiatric disorders that increase the risk of suicide including depression, less is known about the corresponding changes in molecular network within LC. In this study, we examined miRNA networks in the LC of depressed suicide completers and healthy controls. Expression array revealed differential regulation of 13 miRNAs. Interaction between altered miRNAs and target genes showed dense interconnected molecular network. Functional clustering of predicated target genes yielded stress induced disorders that collectively showed the complex nature of suicidal behavior. In addition, 25 miRNAs were pairwise correlated specifically in the depressed suicide group, but not in the control group. Altogether, our study revealed for the first time the involvement of LC based dysregulated miRNA network in disrupting cellular pathways associated with suicidal behavior.

Figure 1

Hierarchical clustering of miRNAs based on median normalized expression data. The construction of dendrogram was based on hierarchical clustering of median normalized miRNA expression data (∆Ct values) in post-mortem locus coeruleus samples which includes both MDD-suicide and healthy control groups. The clustering was prepared following complete linkage method and Euclidean distance. The MDD-suicide group is represented as ‘Sui’ and control is represented as ‘Ctrl’ in this miRNA expression heat map.

Figure 2

Integrated target gene network based on functional clustering of differentially regulated miRNAs. (a) The intense molecular crosstalk (represented as solid lines for direct relationship) is indicative of target enrichment of the differentially regulated miRNAs as found in MDD-suicide group compared to healthy control subjects. In the functional molecular network analysis, shapes of individual molecules are representative of their function and genes are represented as nodes. (b) Gene regulatory network of significantly upregulated miRNAs based on validated targets. Using CyTargetLinker plugin from Cytoscape software (version 3.0), functional gene network was constructed based on validated targets of upregulated miRNAs from MDD-suicide group.

Figure 3

miRNA co-expression network analysis. (a) A core set of 25 miRNAs were selectively and highly co-expressed within the suicide group but not in the healthy group. (b) A core set of 30 microRNAs were strongly co-expressed and interconnected in the healthy control, but not in the suicide group.

Figure 4

Gene regulatory network of top 3 upregulated miRNAs based on TargetScan target prediction algorithm. Construction of an integrated target gene regulatory network is shown, which was based on the TargetScan derived predicted target genes for top three significantly upregulated miRNAs (miR-541-3p, miR-550-5p and miR-1179). The selection of miRNAs was based on the magnitude of upregulated expression in the suicide group (≥1.5 cutoff in relative fold change).

Figure 5

Hierarchical clustering based on predicted target genes of upregulated miRNAs (a) Ontology-based hierarchical clustering of genes identified as predicted targets of upregulated miRNAs. Using mirPath (version 3) from DIANA tools, 10 upregulated miRNAs in the suicide group was used to prepare a heat map based on the gene ontology of predicted target genes. A color-coded representation indicated the functional enrichment of genes as targets of significantly upregulated miRNAs from MDD-suicide group. (b) Cellular pathway based hierarchical clustering of predicted target genes as part of upregulated miRNAs. Kyoto Encyclopedia of Genes and Genomes (KEGG) based mapping of putative target genes in various cellular pathways as part of the deregulated miRNAs demonstrated the possible involvement of 10 up regulated miRNAs and their canonical target genes.

Figure 6

Mapping of predicted target genes in pathways related to various synaptic neurotransmission. (a) Mapping of genes were identified as predicted targets of upregulated miRNAs. Elaboration of canonical pathway related to GABAergic synapse function enriched with predicted target genes of upregulated miRNAs in the MDD-suicide group is shown. (b) Mapping of genes in dopaminergic synapse identified as predicted targets of upregulated miRNAs. Elaboration of canonical pathway related to dopaminergic synapse function enriched with predicted target genes of upregulated miRNAs in the MDD-suicide group is shown. (c) Mapping of genes in glutamatergic synapse identified as predicted targets of downregulated miRNAs. Elaboration of canonical pathway related to dopaminergic synapse function enriched with predicted target genes of downregulated miRNAs in the MDD-suicide group is shown. The pathway images are obtained from Kyoto Encyclopedia of Genes and Genomes (KEGG).

Figure 7

Enriched gene regulatory network of altered miRNAs and their targets with functional relationship to neuropsychiatric disorders. Overlapping target gene network affected by altered miRNAs were mapped. Representing hubs on the network were pathophysiological footprints of associated neuropsychiatric disorders. (p < 0.05, Fisher’s Exact Test).

Figure 8

Relating functional disorders with dysregulated miRNAs based on predicted target genes and their validation. (a) Affected canonical pathways mediated by deregulated miRNA network. Canonical biological pathways associated with genes that were predicted to be targets of significantly altered miRNAs in MDD-suicide subjects are shown (p < 0.05, Fisher’s Exact Test). The ratio is calculated as the number of genes in a given pathway divided by the number of genes that make up the pathway. The p-value for a given process annotation is calculated by considering the number of focus genes that participate in that process and the total number of genes that are known to be associated with that process in the selected reference set. The more focus genes involved, the more likely the association is not due to a random chance. (b) Expression status of few selected genes identified as targets of altered miRNAs. Transcript levels of RELN, GSK-3β, MAOA, CHRM1, PLCB1 and GRIK1 were analyzed in locus coeruleus of MDD-suicide and healthy control subjects by qPCR using primers mentioned in the Methods section. GAPDH normalized relative expression level of RELN, GSK-3β, MAOA, CHRM1, PLCB1 and GRIK1 mRNA transcripts were analyzed in locus coeruleus of MDD-suicide as compared with the normal healthy control group. All data are the mean ± SEM (for RELN, GSK-3β and GRIK1, n = 10 in control and n = 9 in MDD-suicide group; for MAOA, n = 9/group). The level of significance was determined using independent-sample ‘t’ test. ‘*’ denotes significant difference between MDD-suicide and control groups (RELN p = 0.038; GSK-3β p = 0.009; MAOA p = 0.032; GRIK1 p = 0.046).