Allelic differences between Europeans and Chinese for CREB1 SNPs and their implications in gene expression regulation, hippocampal structure and function, and bipolar disorder susceptibility

Abstract

Bipolar disorder (BD) is a polygenic disorder that shares substantial genetic risk factors with major depressive disorder (MDD). Genetic analyses have reported numerous BD susceptibility genes, while some variants, such as single-nucleotide polymorphisms (SNPs) in CACNA1C have been successfully replicated, many others have not and subsequently their effects on the intermediate phenotypes cannot be verified. Here, we studied the MDD-related gene CREB1 in a set of independent BD sample groups of European ancestry (a total of 64,888 subjects) and identified multiple SNPs significantly associated with BD (the most significant being SNP rs6785[A], P=6.32 × 10(-5), odds ratio (OR)=1.090). Risk SNPs were then subjected to further analyses in healthy Europeans for intermediate phenotypes of BD, including hippocampal volume, hippocampal function and cognitive performance. Our results showed that the risk SNPs were significantly associated with hippocampal volume and hippocampal function, with the risk alleles showing a decreased hippocampal volume and diminished activation of the left hippocampus, adding further evidence for their involvement in BD susceptibility. We also found the risk SNPs were strongly associated with CREB1 expression in lymphoblastoid cells (P<0.005) and the prefrontal cortex (P<1.0 × 10(-6)). Remarkably, population genetic analysis indicated that CREB1 displayed striking differences in allele frequencies between continental populations, and the risk alleles were completely absent in East Asian populations. We demonstrated that the regional prevalence of the CREB1 risk alleles in Europeans is likely caused by genetic hitchhiking due to natural selection acting on a nearby gene. Our results suggest that differential population histories due to natural selection on regional populations may lead to genetic heterogeneity of susceptibility to complex diseases, such as BD, and explain inconsistencies in detecting the genetic markers of these diseases among different ethnic populations.

Figure 1

Genomic and linkage disequilibrium (LD) structure of CREB1 as well as association results with bipolar disorder. The results in the discovery sample were shown in red circle, and the results in the replication sample were shown in blue square, and the results in the combined sample were shown in green triangle. The LD color scheme was defined using the R-squared and the LD value of the paired SNPs was calculated using the r² algorithm. The r² value for each color was: black (r²>0.90), dark gray (0.2 < r² < 0.6), gray (0.01 ≤ r² < 0.2), white (r² ≤ 0.01).

Figure 2

Effect of the risk single-nucleotide polymorphism (SNP) rs2709370 on hippocampal function. During memory recall, carriers of the risk allele (C) of rs2709370 exhibit significantly decreased allele-dosage-dependent hippocampal activation in the left hippocampus (Z = 3.97, P<0.01, family wise error corrected for multiple testing across region of interest). Each red dot represents size of effect in one subject and reflects hippocampal activation. Number of subjects in each group: CC = 14, AC = 99, AA = 166.

Figure 3

Effects of risk single-nucleotide polymorphisms (SNPs) on CREB1 mRNA expression. (a) Results in the lymphoblastoid cell lines of 75 healthy European subjects. (b) Results in the prefrontal cortex of healthy Caucasian and African-American subjects (N = 261).

Figure 4

Global distribution of CREB1 single-nucleotide polymorphisms (SNPs) in world populations. (a) Global distribution of rs6785 (risk SNP) allele frequencies in world populations. The derived allele [A] is the risk allele. (b) Global distribution of rs2254137 (non-risk SNP) allele frequencies in world populations.