Genetic Association Between Androgen Receptor Gene CAG Repeat Length Polymorphism and Male Infertility: A Meta-Analysis

Supplemental Digital Content is available in the text


INTRODUCTION
I nfertility has become a threat to more and more couples and aroused lots of attention during recent years. About 50% infertile cases are attributed to male factor. Nevertheless, the etiology of half of male infertility cases remains obscure. 1 Many factors potentially compromising male reproductive ability have been searched, postulated, and studied.
Androgen, mainly secreted by Leydig cells in male testis, is essential for male sex differentiation, development, spermatogenesis, and sexual behavior. It is mediated by the androgen receptor (AR), a member of the steroid hormone receptor superfamily. The gene of AR is located at chromosome Xq11-12, which has 8 exons and 7 introns. 2 The protein encoded by exon 1 of the AR gene is linked to AR transcriptional activity. 3 The exon 1 of AR gene contains CAG and GGC trinucleotide polymorphic repeats which can respectively encode for polyglutamine and polyglycine. Previous studies have reported that the lengths of the 2 polymorphisms vary in different people. [4][5][6] The usual variation of the AR-CAG trinucleotide repeat length is 11 to 36,7 and the median number is 22 in Caucasians. 8 AR-CAG trinucleotide repeats were found possibly in association with male fertility in 1991 for the first time, 9 and a great many studies followed to further that investigation. But the results are divergent.
Some studies demonstrated that longer length of AR-CAG trinucleotide repeats was associated with increased risk of male infertility, which is typical of impaired spermatogenesis with different severity. Also, it was suggested that AR-CAG tracts longer than 40 repeats give rise to Kennedy disease, a fatal neuromuscular disease accompanied by reduced sperm quality. 10 Other reports, however, did not provide corresponding results. Whether AR-CAG trinucleotide repeats is linked to male infertility has not been fully revealed. We hope that an updated meta-analysis based on the data obtained from recent studies and all published literature could be helpful to the deeper exploration into this field.

Study Design
Our study was composed of searching eligible reports, extracting data, analyzing data, and conducting sensitivity analysis and testing publication bias. We classified infertile men into different groups according their ethnic races and sperm concentration in hopes of exploring more potential factors which lead to male infertility. Ethical approval was waived because this study is a meta-analysis.

Searching Strategy
To link AR-CAG trinucleotide repeat length and risk of male infertility, we searched the PubMed, China National Knowledge Infrastructure (CNKI), VIP Database for Chinese Technical Periodicals (VIP), WanFang Med Database (Wan-Fang) to retrieve all articles available before August 2015 without language restrictions. The keywords used were a combination of ''AR, CAG, male infertility.'. All reports were independently identified by 2 authors, and references cited in all original reports and review articles were examined through manual retrieval to identify other potentially eligible publications.

Inclusion Criteria and Exclusion Criteria
The included studies should be the ones that evaluates the association between AR-CAG trinucleotide repeat length and risk of male infertility, includes a case group of male patients with male infertility, which some of them were divided into azoospermia group and oligospermia group by semen analysis according to World Health Organization guidelines 11 , includes a control group of proven or presumed fertile men, and includes sufficient allele frequency data for extraction. Studies were excluded because the design of the study is not rigorous, data of study is incomplete, repetition of the published literature, and a meta-analysis or a review. The procedure of articles screening is illustrated in a flow chart ( Figure 1). Data from those reports were extracted and summarized in Table 1. Some reports not included in this meta-analysis but of reference value are also shown in Table 1 (marked by superscript a). All reports collected in Table 1 were listed according to its date of publication. 1,12-65

Data Extraction
All eligible reports were screened independently by 3 investigators (BP, RL, and YC) to extract available data according to the prespecified selection criteria. Disagreement was resolved by discussion with coauthors. The sample size, mean, and standard deviation (SD) of case and control group were extracted. In some studies, [12][13][14][15][16][17][18][19][20][21][22][23][24] azoospermic and oligospermic groups were specially classified and the data of each group were respectively extracted. Information of the range of AR-CAG repeat length, geographic location of the study, year of publication, ethnicity of participants was also noted. In several articles, 1,13,14,17,19,25 -28 we derived SD from standard error (SE): SD ¼ SE Â n, where n represents the size of case or control groups.

Data Analysis
All statistical analyses were carried out using Stata (Version 9.2, StataCorp LP, College Station, TX), and P <0.05 were considered to be significant. Adding to the study on the total infertile case group, we conducted studies on the group of azoospermia and oligospermia. When it comes to oligospermic group, we classified it into the mild oligospermic group (sperm concentration >5 Â 10 6 and <20 Â 10 6 /mL) and severe oligospermic group (sperm concentration <5 Â 10 6 /mL). In some studies, the definition of mild oligospermia and severe oligospermia was different from other reports. 12,19,20 Since the difference was not too wide to influence the study, so the data of those reports were also included in the study of mild or severe oligospermia. To explore the effect of race of study participants on results, we also conducted studies based on the race (Caucasian, Asian, Mixed races, and Unclarified race). The standardized mean difference (SMD) of the AR-CAG repeats length and its 95% confidence interval (95% CI) were used as statistic. Homogeneity of the included studies was tested. When P <0.1 or I 2 >50%, there is a high extent of heterogeneity between studies and random-effect model was used; when P >0.1 and I 2 <50%, no heterogeneity is found between studies, so fixedeffect model was used. The impact of quality of the included studies on the results was estimated by sensitivity analysis. Funnel plot studies were used to evaluate the publication bias.

Sensitivity Analysis and Evaluation of Publication Bias
Sensitivity analysis was conducted by wiping out 1 report each time to find whether the result of the study would be changed. Results remained the same in analysis on overall infertile men, Asian race, mixed race, unclarified race, azoospermic group, and mild oligospermic group. In the analysis on Caucasian race, the result was changed by wiping out 1 of the reports. 43 The study showed no association between AR-CAG repeat length and male infertility. In severe oligospermic group, the result showed that increased AR-CAG repeat length could be a risk for male infertility after wiping out 2 of the reports. 16,19 These results may be caused by different inclusion standard and number of cases. No obvious publication bias was found based on the shape of funnel plot studies ( Figure S3, http://links.lww.com/MD/A750).

DISCUSSION
Spermatogenesis disorder has a complex pathogenesis. Androgen functions through combining with AR to stimulate  or inhibit the expression of relevant gene. The number of glutamine encoded by (CAG)n is essential to the structure and function of AR molecule and its cofactors. The normal range of CAG repeat length is considered between 16 and 29. 66 Reports on association between male infertility and gene polymorphism of androgen receptor mostly focused on Caucasian race and Asian race. Our study suggests that Asian men are more easily affected by the abnormality of AR-CAG polymorphism. Our study shows that the overall infertile population is typical of an increased AR-CAG repeat length. When it comes to racial factors, Asian, Caucasian, and mixed races all show the same result. In recent years, many studies classified infertile patients according to sperm concentration, such as azoospermia and oligospermia. We conducted analysis on those  [22][23][24] did not provide data for the overall infertile patients so they were not included in the analysis for the overall analysis. But they were included in analysis on subgroups of azoospermia, severe oligospermia, and mild oligospermia. One report 13 included studies on 2 different ethnic groups, so we consider it as 2 separate studies. § When P <0.01 or I 2 >50%, there is a high extent of heterogeneity between studies and random-effect model was used; when P >0.01 and I 2 <50%, no heterogeneity is found between studies, so fixed-effect model was used.
jj Sperm concentration <5 Â 10 6 /mL. ô Sperm concentration <20 Â 10 6 /mL and >5 Â 10 6 /mL. groups to explore the possible relationship between the severity of defect spermatogenesis and AR-CAG repeat length. Azoospermia was found to be associated with increased AR-CAG repeat length, but this result was not found in either severe oligospermia or mild oligospermia. That indicates oligospermia could be a result of many more complex factors and increased AR-CAG repeat length could result in severe spermatogenesis disorder. The study of Liu et al 24 included in the meta-analysis is different from others. Liu et al 24 using small sample size found that the AR-CAG repeat length was 30.8 AE 1.1 in group of azoospermia, which was significantly higher than that of fertile control. The study with a relatively small sample size might lack of adequate power to draw a fair conclusion. Additionally, the gene polymorphism might differ among different geography areas and race. Therefore, future studies with larger sample size in this area are needed to verify the association between AR-CAG repeat length and male infertility. Though the result of the study was different from others, it also met the strict criteria of study and thus were included in our meta-analysis.
(GGN)n polymorphism has also been analyzed but its function is still unknown. Nevertheless, many more studies revealed that there was no association between AR-GGN repeat length and male infertility. 18,59,67 When the joint of CAG and GGC was taken into consideration, 2 haplotypes (CAG ¼ 21/ GGC ¼ 18, CAG !21/GGC !18) could make the risk of male infertility increase. 18 Whether (GGN)n polymorphism could influence the function of androgen receptor remains to be fully studied.
Meta-analysis is a quantitative systematic review and its result could be influenced by publication bias, database bias, inclusion criteria bias, and language bias. We followed the strict criteria to eliminate the ineligible reports to ensure the quality of included studies. Funnel plot studies were conducted. We searched both PubMed database and Chinese database such as CHKI, VIP, WanFang to get a more comprehensive set of data.
AR-CAG repeat polymorphism has been studied for its influence on decreased sexual function, which could lead to infertility. Increased CAG repeat length was found to compromise sexual function. 68 Nevertheless, no significant association was found between erectile dysfunction and CAG repeat length. 69 Increased CAG repeat was also found to be associated with depression and men with CAG repeat length !23 more frequently encountered decreased potency. 70 Recent studies show that hypogonadal patients with shorter CAG repeat length had a more significant improvement after testosterone replacement therapy. 71,72 Those reports indicated that failed sexual intercourse resulting from impaired sexual function could be a reason why increased CAG repeat length could lead to male infertility. Additionally, our meta-analysis suggests that AR-CAG repeat polymorphism has a relationship with male infertility, but the exact molecular mechanisms of how the CAG repeat polymorphism affects male infertility are unknown. We suppose that the secondary structure of AR mRNA sequence or the transcription factors binding site may change by the longer CAG repeat length. Our analysis suggests that further exploration of the molecular mechanism of AR-CAG repeat polymorphism and risk male infertility is demanded.
The present study had some limitations that require consideration. First, some studies with small sample size may not have enough statistical power to explore the real association. Second, our results were based on unadjusted estimates and a more precise analysis should be conducted if individual data were available, which would allow for adjustment by other covariants such as age, body mass index, smoking status, drinking status, environment factors, and lifestyle. Third, in the subgroup analyses, the number of mild oligospermia was relatively small, not having enough statistical power to investigate the association of the polymorphism with male infertility susceptibility.
In conclusion, we again confirmed the association between increased AR-CAG repeat length and defected spermatogenesis. However, its association with the severity of the disease is to be fully studied. The studies on other possible factors such as race and lifestyle are also needed. AR-CAG polymorphism could be an effective way of evaluating the risk of infertility. The studies also paved the way for the gene therapy for the male infertility. Further studies are needed to clarify the association between length of CAG repeats and male infertility as well as its mechanism.