Evidence of CNIH3 involvement in opioid dependence

Abstract

Opioid dependence, a severe addictive disorder and major societal problem, has been demonstrated to be moderately heritable. We conducted a genome-wide association study in Comorbidity and Trauma Study data comparing opioid-dependent daily injectors (N=1167) with opioid misusers who never progressed to daily injection (N=161). The strongest associations, observed for CNIH3 single-nucleotide polymorphisms (SNPs), were confirmed in two independent samples, the Yale-Penn genetic studies of opioid, cocaine and alcohol dependence and the Study of Addiction: Genetics and Environment, which both contain non-dependent opioid misusers and opioid-dependent individuals. Meta-analyses found five genome-wide significant CNIH3 SNPs. The A allele of rs10799590, the most highly associated SNP, was robustly protective (P=4.30E-9; odds ratio 0.64 (95% confidence interval 0.55-0.74)). Epigenetic annotation predicts that this SNP is functional in fetal brain. Neuroimaging data from the Duke Neurogenetics Study (N=312) provide evidence of this SNP's in vivo functionality; rs10799590 A allele carriers displayed significantly greater right amygdala habituation to threat-related facial expressions, a phenotype associated with resilience to psychopathology. Computational genetic analyses of physical dependence on morphine across 23 mouse strains yielded significant correlations for haplotypes in CNIH3 and functionally related genes. These convergent findings support CNIH3 involvement in the pathophysiology of opioid dependence, complementing prior studies implicating the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamate system.

Figure 1. Epigenetic Landscape of the Six Intronic CNIH3 SNPs

Rs1369846 and rs298733 are located within retrotransposons.^, Evidence of epigenetic functionality for rs10799590 includes the location of this SNP within a fetal brain specific enhancer.^, Fetal brain H3K4me1 data indicate that it is within a H3K4me1 peak; DNaseI hypersensitivity data suggest that it is in an open chromatin state in fetal brain. This enhancer mark on rs10799590 was specific to fetal brain, but was not in CD4+ T cells, breast luminal epithelial cells, adult liver, fetal heart, fetal lung, melanocyte, or keratinocytes. The results of motif analyses (Supplementary Table 6) predict that rs10799590 is within the binding site of transcription factor Tal1 with the G allele having significantly higher binding affinity than the A allele.

Figure 2. Amygdala Habituation

(a) Statistical parametric map illustrating mean bilateral amygdala habituation across all DNS participants (left MNI coordinates: x = −20 y = −8 z = −16, k_E = 144, t = 9.03, P < 0.001; right MNI coordinates x = 22 y = −6 z = −14, k_E = 83, t = 8.97, P < 0.001) (b) A allele carriers had greater right amygdala habituation (i.e., less persistent activation) relative to G allele homozygotes. The Y axis indicates habituation with greater values indicating a larger decrease in activation over time. See Figure S3 for a depiction of activation across blocks.