miRNA cargo in circulating vesicles from neurons is altered in individuals with schizophrenia and associated with severe disease

Abstract

While RNA expression appears to be altered in several brain disorders, the constraints of postmortem analysis make it impractical for well-powered population studies and biomarker development. Given that the unique molecular composition of neurons are reflected in their extracellular vesicles (EVs), we hypothesized that the fractionation of neuron derived EVs provides an opportunity to specifically profile their encapsulated contents noninvasively from blood. To investigate this hypothesis, we determined miRNA expression in microtubule associated protein 1B (MAP1B)-enriched serum EVs derived from neurons from a large cohort of individuals with schizophrenia and nonpsychiatric comparison participants. We observed dysregulation of miRNA in schizophrenia subjects, in particular those with treatment-resistance and severe cognitive deficits. These data support the hypothesis that schizophrenia is associated with alterations in posttranscriptional regulation of synaptic gene expression and provides an example of the potential utility of tissue-specific EV analysis in brain disorders.

Fig. 1.. miRNA dysregulated in immunofractionated serum EVs from schizophrenia cases.

(A) Significant differential expression (FDR < 0.1) of neuronal miRNA was observed for schizophrenia subjects (SZ) (n = 221) compared to nonpsychiatric comparison subjects (CO) (n = 256), (B) cognitive-deficit subtype of schizophrenia (CD) (n = 111) compared to nonpsychiatric comparison subjects and (C) CD compared to cognitively spared subtype of schizophrenia (CS) (n = 110).

Fig. 2.. Differential expression of neuronal-origin miRNA in serum EVs from subjects with TRS.

(A) In the cases-only analysis, significant differential expression (FDR < 0.1) of neuronal miRNA was observed for TRS (n = 42) compared to non-TRS (n = 179). (B) In the full cohort analysis, including nonpsychiatric comparison subjects as non-TRS, significant differential expression (FDR < 0.1) of neuronal miRNA was observed for TRS (n = 42) compared to non-TRS (n = 435, 179 cases with non-TRS and 256 nonpsychiatric comparison subjects) with adjustment (adj) and (C) without adjustment (unadj) for case-comparison status (SZ and CO).

Fig. 3.. Association of gene sets from dysregulated miRNA in schizophrenia subjects.

Forest plot of MAGMA gene set association results for the three sets of predicted targets from differentially expressed miRNA between groups. Each panel displays the results constructing a model using liberal or conservative definitions of the boundaries for each gene, as well as a model with conservative boundaries additionally covaried for cortical gene expression for each gene. Conservative boundaries extend the gene 5 kb upstream and 1.5 kb downstream to capture regulatory variation, while liberal boundaries are 35 and 10 kb upstream and downstream, respectively. The MAGMA β-coefficients for the gene set term are plotted with the error bars representing the standard error of the coefficient.

Fig. 4.. Cognitive deficit–associated miRNAs are enriched with target genes associated with neural processes.

Predicted targets of miRNA dysregulated in schizophrenia cases with cognitive deficit compared to nonpsychiatric comparison subjects with binding sites for at least two miRNAs. The gene list was mapped to predefined pathways and tested (hypergeometric) for overrepresentation using Consensus Path Database. Circles (referred to as nodes) represent molecular concepts in which miRNA targets are enriched, node size represents the number of miRNA targets and node color represents the P value. Connected molecular concepts are represented by lines (referred to as edges) between nodes where the edge width represents a relative measure of shared molecules between concepts and edge color represents the quantity of miRNA targets that are shared between nodes. Edges are filtered to minimum relative overlap 0.08 to highlight the closest relationships between nodes and some edges omitted to improve readability, while filtered and all edges figure is available in supplementary materials (fig. S9). Among the enriched pathways are “cholinergic synapse” (q value = 0.024), “Nervous system development” (q value = 1.411 × 10⁻⁴), “BDNF signaling pathway” (q value = 1.695 × 10⁻⁴), “EPO signaling pathway” (q value = 0.005), and “Reelin signaling pathway (q value = 0.003).”

Fig. 5.. Differentially expressed miRNA within schizophrenia.

(A) miR-1246 was the only miRNA consistently dysregulated across all contrasts (CD versus CS, CD versus CO, TRS versus nonTRS, and SZ versus CO). miR-4521, miR-5100, and miR-7704 were the next most consistently dysregulated (CD versus CS, CD versus, CO and TRS versus non-TRS), followed by miR-203a-3p and miR-3178 (CD versus CO and TRS versus non-TRS), and last, miR-451a (CD versus CS and CD versus CO) and miR-486-5p (CD versus CO and SZ versus CO). (B) Differential expression of miR-1246 (log₂ fold change) across all comparisons. Schizophrenia with severe cognitive deficits (CD, n = 111), Schizophrenia with spared cognition (CS, n = 110), nonpsychiatric comparison subjects (CO, n = 256), TRS (n = 42), and subjects who are not treatment resistant include cases and comparison (non-TRS, n = 435).

Fig. 6.. Enrichment for neuronal-origin EVs from serum.

Schematic summary of immunofractionation of serum to enrich for neuronal origin EVs and subsequent extraction of neuronal RNA for small RNA sequencing. (A) Neurons, and other brain cell types, release EVs to the extracellular milieu both constitutively and in response to activity (refer to introduction). Released EVs, present in the interstitial fluid that surrounds neural cells, exit the brain parenchyma to reach the blood through two possible routes; the vascular pathway, guarded by the blood-brain barrier (BBB) and the cerebrospinal fluid (CSF) pathway via perivascular and perineural spaces. Given the relatively large size and negative charge of EVs, their transport to the vasculature via the BBB is uncertain, although there are reports of blood to brain transit of EVs via vesicle transcytosis (89). The CSF pathway is the more likely route of exit, although the specific mechanisms and relative contributions are unclear (90). Nevertheless, EVs have been recovered from adult human postmortem brain tissue, CSF, and peripheral biofluids (refer to discussion) and proteomics of neuron-derived plasma EVs have informed Alzheimer’s disease–related pathology (–13). (B) Whole blood was obtained from study participants during clinical interview, processed to serum, and stored at −80°C. (C) As described in methods, serum aliquots (100 μl) were incubated with anti-MAP1B–coupled magnetic beads allowing removal of the supernatant (containing unbound EVs, including those of nonneuronal origin), while leaving an enriched fraction of relatively homogeneous neuronal origin EVs for RNA extraction and small RNA sequencing.