Age-Associated Salivary MicroRNA Biomarkers for Oculopharyngeal Muscular Dystrophy

Abstract

Small non-coding microRNAs (miRNAs) are involved in the regulation of mRNA stability. Their features, including high stability and secretion to biofluids, make them attractive as potential biomarkers for diverse pathologies. This is the first study reporting miRNA as potential biomarkers for oculopharyngeal muscular dystrophy (OPMD), an adult-onset myopathy. We hypothesized that miRNA that is differentially expressed in affected muscles from OPMD patients is secreted to biofluids and those miRNAs could be used as biomarkers for OPMD. We first identified candidate miRNAs from OPMD-affected muscles and from muscles from an OPMD mouse model using RNA sequencing. We then compared the OPMD-deregulated miRNAs to the literature and, subsequently, we selected a few candidates for expression studies in serum and saliva biofluids using qRT-PCR. We identified 126 miRNAs OPMD-deregulated in human muscles, but 36 deregulated miRNAs in mice only (pFDR < 0.05). Only 15 OPMD-deregulated miRNAs overlapped between the in humans and mouse studies. The majority of the OPMD-deregulated miRNAs showed opposite deregulation direction compared with known muscular dystrophies miRNAs (myoMirs), which are associated. In contrast, similar dysregulation direction was found for 13 miRNAs that are common between OPMD and aging muscles. A significant age-association (p < 0.05) was found for 17 OPMD-deregulated miRNAs (13.4%), whereas in controls, only six miRNAs (1.4%) showed a significant age-association, suggesting that miRNA expression in OPMD is highly age-associated. miRNA expression in biofluids revealed that OPMD-associated deregulation in saliva was similar to that in muscles, but not in serum. The same as in muscle, miRNA expression levels in saliva were also found to be associated with age (p < 0.05). Moreover, the majority of OPMD-miRNAs were found to be associated with dysphagia as an initial symptom. We suggest that levels of specific miRNAs in saliva can mark muscle degeneration in general and dysphagia in OPMD.

Keywords: OPMD; circulating miRNA; miRNA; muscle aging; muscle atrophy.

Figure 1

Workflow of our study. In step 1, the oculopharyngeal muscular dystrophy (OPMD)-miRNA candidates were identified in three RNAseq studies. (I) human control and OPMD, vastus lateralis (VL), (II) mouse OPMD model A17.1 and Friend Virus B laboratory mouse (FVB) control, tibialis anterior (TA). From both studies, differentially expressed (DE) miRNAs were identified. (III) A small RNAseq in muscle and saliva from OPMD patients was carried out in order to identify the highly correlated miRNAs between muscle and saliva. miRNA candidates were selected from the three studies. In step 2, assessment of expression was carried out using qRT-PCR in both serum and saliva from control and OPMD.

Figure 2

OPMD and A17.1 differentially expressed miRNAs. (A) and (B) Volcano plots show the average fold-change (log2) against p-value (-log10) in OPMD vastus lateralis (VL; A) or in A17.1 mouse tibialis anterior (TA; B). The miRNAs with pFDR < 0.05 are depicted in red. (C) Bar chart shows the fraction of deregulated miRNA in humans or mice. The upregulated miRNAs are depicted in red, and downregulated in blue. (D) Bar chart shows the number of total and overlapping miRNAs between mouse and human (depicted in black or grey, respectively). The proportion of overlapping miRNAs is from the total. The overlapping deregulated miRNAs are depicted in red, and the proportion of the deregulated miRNAs is from the overlapping miRNA pull. (E) Scatter plot of the average fold-change of the 252-overlapping miRNA. Pearson correlation coefficient (r) and regression line are depicted. Similar fold-change directions between human and mouse are depicted in cyan and the miRNA with opposite direction are in grey. The overlapping miRNAs with pFDR < 0.05 are encircled in red. The miRNAs with the highest or lowest fold-changes in both humans and mice are indicated. (F) Bar chart of the overlapping miRNAs with a similar fold-change direction in mouse and human. The number of miRNAs with a fold-change (FC) > 2 or < 0.5 are depicted in red and blue, respectively.

Figure 3

Analysis of miRNAs in human OPMD muscles. (A) Scatter plot shows average count per million (CPM, log 2) versus average fold-change (log2) in OPMD versus control muscles. The OPMD upregulated miRNAs are depicted in red, and the downregulated in blue, and those with fold-change > 2 or < 0.5 with a larger circle. The name of most abundant miRNAs and those with the highest fold-change is depicted. (B) Scatter plot shows average CPM versus average fold-change in OPMD for selected miRNAs that have been reported to be associated with several muscle pathologies and physiological conditions: The myomiRs are reported in muscular dystrophies; myositis miRs are reported in pathologies with muscle inflammation; AgingmiRs have been reported in aging muscles, and muscle wasting miRs have been reported in models with muscle atrophy and wasting. The miRNAs whose dysregulation direction differs between OPMD and the reference condition are marked with an arrow, and its color refers to the condition. Differences in dysregulation directions were found for muscular dystrophies and muscle wasting. (C) Scatter plot of the Pearson correlation coefficient (r) of miRNA expression levels with age in control and OPMD subjects. Analysis was carried out for the OPMD-deregulated miRNAs (pFDR < 0.05). The significantly correlated (p < 0.05) miRNAs in OPMD are depicted in red, and in control in black. The reported aging-associated miRNAs, miR-133a/b, miR-146b, miR-26b and let-7a are blue encircled. (D,F). Scatter plots of expression values versus age in three significantly age-correlated and OPMD-deregulated miRNAs. Controls are depicted in grey and OPMD in red. A dotted line shows a linear regression line, the correlation coefficient (r) is denoted. Pearson p-value < 0.05 is denoted *, respectively, or non-specific (NS). The slope (=s) of the regression line was determined with a linear regression model.

Figure 4

A graphical summary of predicted targets analysis. The OPMD upregulated miRNAs and those that are also age-associated are listed. The OPMD upregulated miRNAs inclusion criteria: average fold-change (FC) > 2 and average CPM > 2.5. In bold are the four miRNAs that are common between the two conditions. The miRNAs without predicted gene targets are in grey. Venn diagram shows the overlap between the number of OPMD-downregulated predicted targets to the OPMD upregulated miRNAs that are age-associated (yellow) and the OPMD upregulated miRNAs (red). The percentage of predicted targets from the OPMD downregulated gene list is depicted.

Figure 5

miRNA candidates for OPMD in serum and saliva. (A) Scatter plot shows average expression levels in saliva vs. serum. The dysregulation direction (up or down in OPMD muscles) is denoted in red or blue, respectively. miRNAs with unchanged levels in muscles are denoted in grey. The diagonal (1:1 ratio) is denoted with a dashed line. (B) A summary qRT-PCR for 14 miRNAs in serum or saliva. p-value (p < 0.05) indicates a significant difference between control and OPMD. miRNAs that are higher in OPMD muscles are in red, those with lower expression levels in blue, and unchanged in grey. p-values were determined with an unpaired t-test, Welch corrected. ND, undetected; NS, not significant. miR-133b was not detected in serum from controls. (C,D) Box plots show expression levels in saliva (C) and serum (D) in control and OPMD groups of the significantly differentially expressed miRNAs. Fold-change (FC) is depicted. Every subject is denoted with a dot. The fold-change of all the differentially expressed miRNAs in serum has an opposite direction than in OPMD muscles. (E,F) An age-associated expression level is visualized with a scatter plot. Significant correlation (Pearson, p < 0.05, denoted with *) was found only for in miR-200c and miR-451a in saliva, and only for the OPMD samples. NS = not significant. Correlation coefficient (r) is denoted next to the regression line (dashed lines). A thin dashed line denoted the regression line of the controls and the bold line denotes the OPMD samples. (G,H) A summary of miRNA differential expression analysis per initial symptom, ptosis (n > 12) or dysphagia (n > 16). The p-value (unpaired t-test) and fold-change (FC) are denoted for the significantly differentially expressed miRNAs. miRNAs that are higher in OPMD muscles are in red, those with lower expression levels in blue, and unchanged in grey.