Low-frequency Genetic Variants Are Associated With Non-obstructive Azoospermia In Chinese Men

Infertility, a growing severe health problem, is affecting approximately 10-15% of the couples in the world. Studies have shown that half of all fertility problems are due to male reproductive defect including azoospermia. Azoospermia is known as the major cause of male infertility in humans and can be categorized into two major categories: obstructive versus non-obstructive. Among those, non-obstructive azoospermia(NOA) is characterized as no or little sperm in semen as a result of congenital dysfunction in spermatogenesis, and occurs in over 1% of all adult men. Therefore, it is very important to identify high-risk individuals and adopt preventive measures or treatments, which may reduce the rate of infertility effectively.The etiology of NOA is complex and multifactorial, possibly being a result of interaction between environmental and genetic factors. As reported, well known genetic causalities of NOA involve chromosomal alterations, Y chromosome microdeletions, autosomal chromosome mutations and autosomal polymorphism. Among them, the effect of polymorphic loci is relatively weak. However, it is really prevalent in the crowd.As a powerful tool in identifying the significant susceptible single nucleotide polymorphisms(SNPs) associated with complex diseases based on high throughput genotyping platform and testing millions of the whole genome SNPs in a large sample size of cases and controls, followed by several independent replications, genome-wide association study(GWAS) has been applied into the exploring of genetic etiology of NOA. Previously, our multistage GWAS has identified several risk loci at 1p13.3, 1q42.13, 1p36.32, 6p21.32, 6p12.2, 10q25.3, and 12p12.1, for NOA in Han Chinese men, implying a vital role of genetic factors in the development of NOA. However, what we have found may be just the tip of the iceberg in the etiology of NOA. Most common genetic variants with minor allele frequency(MAF) above 0.05 identified by GWAS conferred relatively small increments in risk(1.1-1.5 fold) and 89.9% of them were in the non-coding region, explained only a small proportion of heritability. Recently, low-frequency(MAF below 5%) variants have been hypothesized to convey larger relative risks than common variants and contribute meaningfully to the understanding of inherited susceptibility. Therefore, it calls for exome-wide association study(EXWAS) to systematically identify low-frequency susceptible loci or genes associated with the risk of NOA.This EXWAS comprises three phases. The 962 cases of the discovery phase were recruited from the Nanjing Center of Reproductive Medicine between April 2005 to January 2012 and the 1,348 healthy male controls were randomly selected from more than 30,000 participants in acommunity-based screening program for non-infectious diseases in Jiangsu Province. Each of the male controls had fathered at least one healthy child without assisted reproductive technologies. 714 cases and 1,398 controls from Jiangsu Province and Shanghai were included in Replication I; and Replication II with 1,454 cases and 3,850 controls were from Wuhan and Shenyang.In the discovery phase, the 962 cases and 1,348 controls were genotyped with Illumina Infinium® Human Exome Beadchip. 247,870 variants in the chip were flow into quality control. A total of 939 cases and 1,335 controls with 40,936 polymorphic variants were included after the quality control of the study variables and subjects. We calculated the P values, odds ratio(OR) and 95% confidence intervals(CI) in the the additive model by logistic regression. Variants met the following criteria were selected in the Replication I: 1) Variants showing a significant difference of P≤10-4 between cases and controls; 2) Variants with the lowest P value when there were in high linkage disequilibrium(LD)(r2>0.8). Variants showing association significance less than 0.05 in Replication I were selected to further replicate in Replication II.We further conducted gene-based analysis through both a simple burden test with MAF<0.01 and a sequence kernel associationoptimal test(SKAT-O) with MAF<0.05 in the discovery phase. A total of 6,919 genes in burden test and 7,979 genes in SKAT-O test were included.Eventually, we found strong evidence of three NOA susceptibility loci. Among them, two rare genetic variants rs2298090 in HIST1H1 E at 6p22.2 and rs200847762 in FKBPL at 6p21.33 conferred to a decreased risk of NOA(rs2298090 A>G: OR=0.29, 95%CI=0.21-0.40, P=6.93×10-14; rs200847762 G>A: OR=0.10, 95%CI=0.05-0.22, P =1.16×10-8), while low-frequency genetic variant rs11754464 G>A in MSH5 at 6p21.33 was associated with an increased risk of NOA(OR=1.78, 95%CI=1.43-2.22, P = 1.03×10-6). Conditional analysis between the 3 susceptibility loci indicated that they were independent signals of NOA susceptibility on 6p21.33 and 6p22.2. For gene-based analysis, we found significant association between NOA and the gene CFTR(P=5.77×10-7). However, the lead variant rs113857788 was not validated in Phase I. We further excluded rs113857788 and found that the association between NOA and CFTR was abolished(P=5.05×10-3).In conclusion, our study is the first one to utilize the exome array with NOA and serves as an extension to our previous examination of common variants on NOA risk. It highlights the role of low-frequency genetic variants in spermatogenesis, addresses part of the missing heritability in NOA risk and provides novel insight into the etiology of NOA. Moreover, our newly identified susceptibility markers might be applied as the genetic information reference for screening, diagnosis and treatment of NOA.