Over 90% of schizophrenia-associated genetic variations are in non-protein-coding regions with poorly understood functions. These regions include non-coding RNAs (ncRNAs) and regulatory sites that dynamically control transcription and synthesis of a large but defined set of proteins to coordinate cell-specific functional activity and behaviour. Single nucleotide variations (SNVs) in small ncRNAs, such as microRNAs (miRNAs), are good candidates to help understand the regulatory genetic code where SNVs may change the binding affinity of miRNA:target interactions affecting cell signalling pathways and function relevant to disease risk.
Aims:
(1) To develop a systems-based analysis to predict loss-of-function non-coding variations in miRNA genes and their regulatory sites;
(2) To validate miRNA:target binding affinity loss or gain caused by non-coding SNVs using a luciferase reporter system;
(3) To functionally validate the role of specific miRNA variations using time-course live cell imaging.
Methods:
miRNA SNVs identified in schizophrenia-associated miR-137/2682 loci were selected for computational inference of disrupted miRNA-target networks. Changes in miRNA-target binding affinity were tested by cloning putative regulatory sites into a luciferase reporter system. Luciferase activity was measured and compared between mimic miRNA variant and reference conditions. The functional role of miRNA variation was assessed by liposomal transfection of mimic miRNA variant and reference into SH-SY5Y neuronal cells followed by a time-course live imaging of single cells (Phase Focus platform).
Results:
We identified a loss-of-function SNV in the miRNA-2682 gene which is embedded in the miRNA-137 locus in complete linkage disequilibrium with one of the strongest regions associated with schizophrenia. Our study proposes a novel integrative approach to predict and experimentally validate a mechanistic hypothesis for the causal effect of SNVs based on binding affinity changes between miRNA and target transcripts with downstream effects in cellular function and behaviour. miRNA-2682 target network significantly overlaps with several genes known to be implicated in schizophrenia. We showed experimental evidence based on time-course live cell imaging that miRNA variations in miRNA-2682 mature sequence can disrupt gene networks controlling cell cycle, growth, motility and proliferation. These fundamental cellular functions have been previously found to be affected in schizophrenia patient-derived olfactory neuronal stem cells when compared to healthy controls.
Conclusion:
Our work shows, for the first time, how a single nucleotide variation in miRNA sequences can affect the direct binding and regulation of several target genes to promote critical changes in fundamental cellular functions recapitulating key schizophrenia endophenotypes at the molecular and cellular levels.