miR-10b-5p expression in Huntington's disease brain relates to age of onset and the extent of striatal involvement

Abstract

Background: MicroRNAs (miRNAs) are small non-coding RNAs that recognize sites of complementarity of target messenger RNAs, resulting in transcriptional regulation and translational repression of target genes. In Huntington's disease (HD), a neurodegenerative disease caused by a trinucleotide repeat expansion, miRNA dyregulation has been reported, which may impact gene expression and modify the progression and severity of HD.

Methods: We performed next-generation miRNA sequence analysis in prefrontal cortex (Brodmann Area 9) from 26 HD, 2 HD gene positive, and 36 control brains. Neuropathological information was available for all HD brains, including age at disease onset, CAG-repeat size, Vonsattel grade, and Hadzi-Vonsattel striatal and cortical scores, a continuous measure of the extent of neurodegeneration. Linear models were performed to examine the relationship of miRNA expression to these clinical features, and messenger RNA targets of associated miRNAs were tested for gene ontology term enrichment.

Results: We identified 75 miRNAs differentially expressed in HD brain (FDR q-value <0.05). Among the HD brains, nine miRNAs were significantly associated with Vonsattel grade of neuropathological involvement and three of these, miR-10b-5p, miR-10b-3p, and miR-302a-3p, significantly related to the Hadzi-Vonsattel striatal score (a continuous measure of striatal involvement) after adjustment for CAG length. Five miRNAs (miR-10b-5p, miR-196a-5p, miR-196b-5p, miR-10b-3p, and miR-106a-5p) were identified as having a significant relationship to CAG length-adjusted age of onset including miR-10b-5p, the mostly strongly over-expressed miRNA in HD cases. Although prefrontal cortex was the source of tissue profiled in these studies, the relationship of miR-10b-5p expression to striatal involvement in the disease was independent of cortical involvement. Correlation of miRNAs to the clinical features clustered by direction of effect and the gene targets of the observed miRNAs showed association to processes relating to nervous system development and transcriptional regulation.

Conclusions: These results demonstrate that miRNA expression in cortical BA9 provides insight into striatal involvement and support a role for these miRNAs, particularly miR-10b-5p, in HD pathogenicity. The miRNAs identified in our studies of postmortem brain tissue may be detectable in peripheral fluids and thus warrant consideration as accessible biomarkers for disease stage, rate of progression, and other important clinical characteristics of HD.

Figure 1

Characterization of miRNA in Huntington’s disease brain. Volcano plot of 75 significantly differentially expressed miRNA after FDR-adjustment for 938 comparisons. Points labeled red were up-regulated in HD and points labeled as blue were down-regulated in HD. Hox-related miRNA points are labeled and represent the top differentially expressed miRNA in HD.

Figure 2

Nine miRNAs are associated with Vonsattel grade. In HD brains, expression of differentially expressed miRNA was compared across Vonsattel grades 0–4. Boxplots represent nine FDR-significant miRNAs (A. miR-10b-5p, B. miR-196a-5p, C. miR-10b-3p, D. miR-196b-5p, E. miR-302a-3p, F. miR-200c-3p, G. miR-4488, H. miR-4449, I. miR-663b) (FDR q < 0.05, adjusted for 75 contrasts) associated with Vonsattel grade by analysis of variance (ANOVA). X-axes represent Vonsattel grade, classified 0–4 in order of the severity of striatal involvement and Y-axes show the VST expression values after batch correction. Significant differences across grades and controls are denoted by letters in the grey banner above the boxplot, labeled a-d. Groups with different letters are significantly different from one another while those with the same letter are not, after correcting for multiple comparisons. For example, group “a” would be significantly different from group “b” and “c.” Conditions represented by multiple letters indicate no significant difference among those groups. For example, group “ab” would not be significantly different than groups “a” and “b,” but would be different group “c.”

Figure 3

miR-10b is associated with age of onset and striatal involvement. In 26 Vonsattel grade 2, 3 and 4 HD brains, both mature miR-10b sequences (−5p and −3p) have FDR-significant relationships to CAG-adjusted Hadzi-Vonsattel striatal score (A and B) and CAG-adjusted onset age (C and D). Y-axes show the variance stabilizing transformation expression values after batch correction and shows that miR-10b-5p is expressed at much higher levels than miR-10b-3p. Grade 0 cases are not included, as they have neither onset age nor H-V striatal score.

Figure 4

CAG-adjusted clinical features of HD show patterns of association with miRNA expression. CAG-adjusted measures of onset age, disease duration, death age, Hadzi-Vonsattel (H-V) striatal and cortical score were correlated with differentially expressed miRNAs in HD brains. miRNAs with at least one nominal p-value < 0.05 are shown. Pearson correlation coefficients and features were independently hierarchically clustered. Red boxes indicate positive correlations and blue boxes indicate negative correlations. Seven miRNAs in the left section are down-regulated in HD and the ten miRNAs in the right section are up-regulated. Unsupervised clustering separated miRNA by their direction of fold change.

Figure 5

Gene ontology terms are similar for mRNA targets of clinically relevant de-regulated miRNAs. (A) Illustrates the overlap in GO Biological Processes between targets of increased miRNA (in orange) and decreased miRNA (in blue) in HD. The x-axis shows the number of gene ontology terms that fall within a given semantic term set, and the y-axis lists the top twenty enriched terms for each set of miRNA targets. Darker colored points represent terms with higher significance and the size of the points represents the union of all genes that fall within a given the term. A number of terms, including “nervous system development” as well as terms relating to transcriptional regulation are shared across up- and down-regulated miRNA target groups. The similarity targets of up-regulated miRNA (in orange) and down-regulated miRNA (in blue) for GO Molecular Function are seen in (B) and for GO Cellular Component in (C).