Synaptic microRNAs Coordinately Regulate Synaptic mRNAs: Perturbation by Chronic Alcohol Consumption

Abstract

Local translation of mRNAs in the synapse has a major role in synaptic structure and function. Chronic alcohol use causes persistent changes in synaptic mRNA expression, possibly mediated by microRNAs localized in the synapse. We profiled the transcriptome of synaptoneurosomes (SN) obtained from the amygdala of mice that consumed 20% ethanol (alcohol) in a 30-day continuous two-bottle choice test to identify the microRNAs that target alcohol-induced mRNAs. SN are membrane vesicles containing pre- and post-synaptic compartments of neurons and astroglia and are a unique model for studying the synaptic transcriptome. We previously showed that chronic alcohol regulates mRNA expression in a coordinated manner. Here, we examine microRNAs and mRNAs from the same samples to define alcohol-responsive synaptic microRNAs and their predicted interactions with targeted mRNAs. The aim of the study was to identify the microRNA-mRNA synaptic interactions that are altered by alcohol. This was accomplished by comparing the effect of alcohol in SN and total homogenate preparations from the same samples. We used a combination of unbiased bioinformatic methods (differential expression, correlation, co-expression, microRNA-mRNA target prediction, co-targeting, and cell type-specific analyses) to identify key alcohol-sensitive microRNAs. Prediction analysis showed that a subset of alcohol-responsive microRNAs was predicted to target many alcohol-responsive mRNAs, providing a bidirectional analysis for identifying microRNA-mRNA interactions. We found microRNAs and mRNAs with overlapping patterns of expression that correlated with alcohol consumption. Cell type-specific analysis revealed that a significant number of alcohol-responsive mRNAs and microRNAs were unique to glutamate neurons and were predicted to target each other. Chronic alcohol consumption appears to perturb the coordinated microRNA regulation of mRNAs in SN, a mechanism that may explain the aberrations in synaptic plasticity affecting the alcoholic brain.

Figure 1

Difference in expression profiles in SN and TH preparations. (a) Principal component analysis of paired SN (green) and TH samples (blue). All microRNAs on the array (including those from different species) were used for this analysis (25 119 probes). The primary purpose for this analysis is quality control to show the overall detection of transcripts on the array and to facilitate comparison of the preparations. Only the mouse mature microRNAs will be discussed and presented in subsequent tables and figures. (b) Expression levels of microRNAs in SN and TH preparations (control group only). For comparison of SN and TH expression levels, fold changes were calculated as the ratio of SN to TH. MicroRNAs below the diagonal are enriched in SN relative to TH and have a fold change greater than 1 (referred to as ‘SN-enriched' shown in green). MicroRNAs above the diagonal are depleted in the SN relative to the TH and have a fold-change less than 1 (referred to as ‘SN-depleted' shown in blue). (c) Venn diagram showing the number of differentially expressed microRNAs from the SN-control/TH-control analysis and the SN-alcohol/TH-alcohol analysis, and the overlap between them. P-values <0.05 were considered significant.

Figure 2

Alcohol-induced microRNAs are different in SN and TH. (a) The number of alcohol-responsive microRNAs in SN and TH. Alcohol induced fold changes are shown on the y-axis for microRNAs. For comparison of alcohol and control expression levels in each of the preparations, fold changes were calculated as the ratio of alcohol to control expression levels (SN-alcohol/SN-control and TH-alcohol/TH-control). Fold changes greater than 1 are referred to as ‘upregulated' and fold-changes less than 1 are referred to as ‘downregulated'. (b) Volcano plot (scatter plot) of fold changes and P-values of the effects of alcohol on microRNAs in SN and TH.

Figure 3

Alcohol induces coordinated expression of microRNAs that are correlated with alcohol consumption. (a) Hierarchical clustering of microRNAs from SN, including both alcohol and control data. The microRNAs are arranged by covariance similarity; thus, microRNAs under the same branch have greater expression similarity than those outside the branch. The dissimilarity among microRNAs is represented in the y-axis. The six different modules are shown in boxes. The microRNAs represent the co-expressed microRNAs overlapping with alcohol-responsive microRNAs. The width of the box represents the number of microRNAs co-expressed in that module. The correlation of each module with alcohol consumption is shown as a heat map (red represents positive correlation with consumption and green represents negative). The microRNAs in the gap between the modules are ones that did not pass the co-expression threshold as defined by the WGCNA and were not included in any module. (b) Examples of three co-expressed microRNAs and the number of overlapping predicted alcohol-responsive mRNAs. (c) Examples of microRNA predicted interactions. The greater the color intensity, the greater the fold-change magnitude (red is upregulated and green is downregulated). The unmarked circles represent mRNAs. The dotted circle emphasizes mRNAs that are co-targeted by the illustrated microRNAs.

Figure 4

MicroRNA–mRNA interactions are coordinately expressed in response to alcohol and are associated with specific cell types. Shown are the alcohol-responsive mRNA modules and their correlation to individual microRNAs found in SN. The 10 alcohol-responsive mRNA modules are shown. The modules' correlation with consumption for the six alcohol-responsive modules is shown as r-values (Pearson's correlation coefficient). Alcohol-responsive mRNA module correlation to individual microRNAs is represented as a heatmap, with red representing a positive correlation and green representing a negative. Cell type mRNA enrichment is also shown. Modules 1–6 were enriched with astrocytic/microglial mRNAs, whereas modules 7–10 were enriched with neuronal mRNAs. The 16 microRNAs that were significantly correlated with at least one mRNA module are shown (mRNA data are from Most et al, 2014).