SU52 - UTILITY OF PERIPHERAL miRNA EXPRESSION PROFILES AS BIOMARKERS AND PATHOLOGICAL INDICATORS OF NEUROPSYCHIATRIC DISEASE

Abstract

Background Given the inaccessibility of brain tissue for large cohort-based and longitudinal studies, analysis of peripheral blood as a surrogate tissue has become a mainstay within the neuropsychiatric field for analysis of gene and miRNA expression and DNA methylation. Although the importance of peripheral biomarkers as diagnostic and prognostic indicators is well recognized, translation of findings into meaningful pathology is more difficult to achieve. Lack of validation of peripheral biomarkers in brain tissue may be confounded by smaller sample sizes, investigation of inappropriate brain regions, or post-mortem complications. Methods To shed some light on the utility of blood-based biomarkers to inform upon pathology of disease, we investigated correlations between miRNA expression levels in blood and brain tissue from 14 cortical and subcortical regions within two groups of baboons (n=4, n=6; 7 brain regions each). To better understand how blood-brain correlations might relate to human biology and disease, we investigated a subset of 212 miRNAs whose peripheral blood expression was at least moderately correlated (r≥0.5) with 50% or more of the brain regions examined in a family-based cohort of Mexican Americans (n=896) for whom we have extensive neurocognitive and neuroanatomical phenotypes as well as peripheral blood miRNA profiles. We tested 145 human peripheral blood-expressed miRNAs for association against 34 cortical and 10 subcortical structures (using magnetic resonance imaging data), and 24 neurocognitive phenotypes. Due to high inter-relatedness and potentially unknown correlations across many of the phenotypes assessed, we corrected only for the number of miRNAs tested, setting the significance threshold at p Results The overall correlation (r) of miRNA expression profiles between blood and 14 different brain regions in baboons ranged from 0.62–0.66. We identified 444 miRNAs that were expressed in 90% of all samples analyzed and of these, 212 showed at least moderate correlation (r≥0.5) between blood and 50% or more of the brain regions examined. Within this subset of baboon-defined correlated miRNAs, we identified 145 that were also expressed in at least 90% of our human cohort and tested these for association with neurological traits. We identified 34 significant associations between miRNAs and neurological traits. Of these, one of the most interesting findings was association of peripheral miR-144-3p expression with amygdala (p=8.50×10-5) and hippocampus (p=5.86×10-5) volumes, as well as thickness of the fusiform gyrus (p=1.98×10-5). In our baboon study, peripheral expression of this miRNA was highly correlated with expression in the amygdala (r=1.00) and hippocampus (r=0.66), expression was not measured in the fusiform gyrus. Discussion Prior studies have shown that overexpression of miR-144-3p in the basolateral amygdala of mice rescued impaired fear extinction, and also that expression of miR-144-3p is increased in the hippocampus of rats following chronic treatment of both lithium and valproate mood stabilizers. Our studies highlight the utility of using peripheral blood as a surrogate tissue for miRNA profiling and indicate that disease associations identified using peripheral blood may point toward pathological processes within the brain, particularly for miRNAs whose expression is correlated between blood and brain.