Blood-derived microRNAs are related to cognitive domains in the general population

Abstract

Introduction: Blood-derived microRNAs (miRNAs) are potential candidates for detecting and preventing subclinical cognitive dysfunction. However, replication of previous findings and identification of novel miRNAs associated with cognitive domains, including their relation to brain structure and the pathways they regulate, are still lacking.

Methods: We examined blood-derived miRNAs and miRNA co-expression clusters in relation to cognitive domains, structural magnetic resonance imaging measures, target gene expression, and genetic variants in 2869 participants of a population-based cohort.

Results: Five previously identified and 14 novel miRNAs were associated with cognitive domains. Eleven of these were also associated with cortical thickness and two with hippocampal volume. Multi-omics analysis showed that certain identified miRNAs were genetically influenced and regulated genes in pathways like neurogenesis and synapse assembly.

Discussion: We identified miRNAs associated with cognitive domains, brain regions, and neuronal processes affected by aging and neurodegeneration, making them promising candidate blood-based biomarkers or therapeutic targets of subclinical cognitive dysfunction.

Highlights: We investigated the association of blood-derived microRNAs with cognitive domains. Five previously identified and 14 novel microRNAs were associated with cognition. Eleven cognition-related microRNAs were also associated with cortical thickness. Identified microRNAs were linked to genes associated with neuronal functions. Results provide putative biomarkers or therapeutic targets of cognitive aging.

Keywords: biomarker; cognition; cortical thickness; miR‐10401‐3p; miR‐125b‐5p; miR‐128‐3p; miR‐134‐5p; miR‐192‐5p; miR‐215‐5p; miR‐4677‐5p; miR‐4732‐3p; miR‐92a‐3p; miR‐92b‐3p; microRNA; population‐based.

FIGURE 1

Replication of previous findings on miRNAs associated with cognition, AD, or MCI. Replication of miRNAs that were identified as associated with cognition in previous cross‐sectional population‐based studies, , , , (A), one previous longitudinal population‐based study (B), and a previous meta‐analysis of miRNAs dysregulated in AD and MCI (C). Details on the studies included can be found in Table S1 in supporting information. The sample sizes of included studies are in parentheses next to the study names. Statistical significance is indicated by the asterisks. MiRNAs that appeared in multiple included studies are marked with “†.” MiRNAs are marked with “+” in parentheses when their higher expression was associated with better cognitive function in included studies, and with “–” in the opposite case. For the AD and MCI meta‐analysis, “+” indicates higher expression of the miRNA in controls compared to patients. MiRNAs that were significantly associated with cognition with a similar direction in the included studies and in our data have been annotated with blue color and are outlined with a blue box. Annotated with red color are miRNAs that were significantly associated with cognition in our study, but in an opposite direction compared to included studies. AD, Alzheimer's disease; MCI, mild cognitive impairment; miRNA, microRNA; SD, standard deviation.

FIGURE 2

Hypothesis‐free association of miRNA and module expression with cognition. A–C, Linear regression coefficients and significance levels for the association between co‐expression modules and cognitive scores. Shown are Model 1 (A) which was adjusted for age and sex; Model 2 (B) which was adjusted for age, sex, first (native) language, and educational level; and Model 3, which was adjusted for age, sex, and blood cell counts. Modules are named after colors, and numbers in parentheses indicate the number of miRNAs within each module. Significance was determined after correcting for multiple comparisons using the FDR method. (D) Volcano plots of the per miRNA association of all miRNAs measured in our study with global cognition, executive function, and episodic verbal memory scores. The p value shown in the y axis has not been corrected for multiple testing. Top miRNAs for each score are annotated and co‐expression modules are shown in parentheses. MiRNAs that reached statistical significance for at least one phenotype, after multiple testing correction (FDR ≤ 0.05), are shown with triangles and have been marked with an asterisk (*). Hub miRNAs are annotated with blue color. FDR, false discovery rate; miRNA, microRNA; SD, standard deviation.

FIGURE 3

Association of miRNA expression with cognitive scores and brain imaging measures. The heatmap (A) shows the association of miRNA expression with cognitive scores and brain imaging measures. Included were miRNAs identified as associated with cognition in our literature replication, co‐expression, and per miRNA analyses. The association of some miRNAs with cognitive domain scores was modified by sex and age. This is illustrated in (B) for age and (C) for sex, which show the standardized regression betas and 95% confidence intervals for the age‐ and sex‐stratified association of these miRNAs with each cognitive score. Linear regression analyses were based on Model 1, which included as covariates age and sex, except for the sex‐stratified analysis which was controlled only for age. CI, confidence interval; FDR, false discovery rate; HCV, hippocampal volume; miRNA, microRNA; SD, standard deviation.

FIGURE 4

Region‐specific association between miRNA expression and brain cortical thickness. Regions of interest have been mapped based on the Desikan–Killiany cortical atlas. The color indicates the standardized β for the regions in which miRNA expression was significantly associated with cortical thickness. The analysis was performed only for miRNAs that were significantly associated with cognition and mean cortical thickness across both hemispheres. miRNA, microRNA; SD, standard deviation.

FIGURE 5

Tissue expression analysis of cognition‐related miRNAs. Boxplots of tissue expression of miRNAs associated with cognition. Data were obtained from the miRNA Tissue Atlas, based on post mortem samples taken from six human donors (here indicated by the colored dots). The module each miRNA belongs to is shown in parentheses. The top five tissues for each miRNA, based on median miRNA expression, are shown. Note that, to allow for better comparison of relative expression in tissues for each miRNA, the scale of the x axis (expression z score) varies. miRNA, microRNA; TSI, tissue specificity index.

FIGURE 6

Function genomics analysis of cognition‐related miRNAs. A, Results of Gene Ontology: Biological Processes enrichment analysis of negatively associated target genes of the miRNAs in each co‐expression module. The top eight most enriched terms (based on p values) are shown. The numbers in parentheses indicate the number of target genes in each enriched term. p values have been corrected for multiple testing using the FDR method, with the dashed line indicating the FDR = 0.05 threshold. Note that, to allow for better comparison of relative enrichment in each module, the scale of the x axis (negative logarithm of FDR) varies. B, Overlap and overrepresentation (determined with a hypergeometric test) of negatively associated target genes of miRNA in a GWAS of global cognitive function. FDR, false discovery rate; GWAS, genome‐wide association study; miRNA, microRNA.