Validation of a microRNA target site polymorphism in H3F3B that is potentially associated with a broad schizophrenia phenotype

Abstract

Despite much progress, few genetic findings for schizophrenia have been assessed by functional validation experiments at the molecular level. We previously reported evidence for genetic linkage of broadly defined schizophrenia to chromosome 17q25 in a sample of 24 multiplex families. 2,002 SNPs under this linkage peak were analyzed for evidence of linkage disequilibrium using the posterior probability of linkage (PPL) framework. SNP rs1060120 produced the strongest evidence for association, with a PPLD|L score of 0.21. This SNP is located within the 3'UTR of the histone gene H3F3B and colocalizes with potential gene target miR-616. A custom miRNA target prediction program predicted that the binding of miR-616 to H3F3B transcripts would be altered by the allelic variants of rs1060120. We used dual luciferase assays to experimentally validate this interaction. The rs1060120 A allele significantly reduced luciferase expression, indicating a stronger interaction with miR-616 than the G allele (p = 0.000412). These results provide functional validation that this SNP could alter schizophrenia epigenetic mechanisms thereby contributing to schizophrenia-related disease risk.

Fig 1. Linkage disequilibrium between 1,544 SNPs and broad schizophrenia spectrum phenotype.

PPLD|L values for 1,544 SNPs, including five MirSNPs (Table 1), from chr17: 74,684,647 to 83,257,441 (GRCh38), were calculated using KELVIN v2.4.0 and plotted vs physical distance. The MirSNP rs1060120 in H3F3B produced a PPLD|L of 0.21, notably higher than the remaining SNPs.

Fig 2. Dual luciferase reporter assay of broad schizophrenia-associated MirSNP rs1060120.

The average normalized value for all three trials of the dual luciferase assay are plotted, with error bars representing the standard deviation for the A allele alone (A), the A Allele co-transfected with a nonspecific scramble miRNA mimic (A SCR), the A allele co-transfected with miR-616 (A 616), the G allele alone (G), the G allele co-transfected with a nonspecific scramble miRNA mimic (G SCR), and the G allele co-transfected with miR-616 (G 616). Data was analyzed with ANOVA using a term for the trial, in accordance with the randomized complete block design. There was evidence of a treatment effect (p < 1 x 10⁻¹⁵). Luciferase expression for the A construct co-transfected with miR-616 was significantly lower than for the G construct co-transfected with mir-616 (p-adjusted = 0.000412)(indicated by *). The nonspecific scramble miRNA mimic was not significantly different from the untreated constructs of either allele.