Association of TUSC1 and DPF3 gene polymorphisms with male infertility

Abstract

Purpose

Recently, genome-wide association studies of a Hutterite population in the USA revealed that five single nucleotide polymorphisms (SNPs) with a significant association with sperm quality and/or function in ethnically diverse men from Chicago were significantly correlated with family size. Of these, three SNPs (rs7867029, rs7174015, and rs12870438) were found to be significantly associated with the risk of azoospermia and/or oligozoospermia in a Japanese population. In this study, we investigated whether the rs10966811 (located in an intergenic region between the TUSC1 and IZUMO3 genes) and rs10129954 (located in the DPF3 gene) SNPs, previously related to family size, are associated with male infertility. In addition, we performed association analysis between rs12348 in TUSC1 and rs2772579 in IZUMO3 and male infertility.

Methods

We genotyped 145 patients with infertility (including 83 patients with azoospermia and 62 with oligozoospermia) and 713 fertile controls by PCR-RFLP technique for polymorphism. Because rs10966811 has no restriction sites, the SNP rs12376894 with strong linkage disequilibrium was selected as an alternative to rs10966811.

Results

There was a statistically significant association between rs12376894 proxy SNP of rs10966811 and oligozoospermia. Also, a statistically significant association between rs10129954 and azoospermia, and oligozoospermia was observed. When we assessed the relationship between rs12348 in TUSC1 and rs2772579 in IZUMO3 and male infertility traits, we found that rs12348 in TUSC1 was significantly associated with azoospermia and oligozoospermia, but rs2772579 in IZUMO3 was not associated with male infertility.

Conclusion

We found that the polymorphisms in TUSC1 and DPF3 displayed strong associations with male infertility.

Introduction

The main cause of male infertility is spermatogenic failure, of which there are three general classifications: azoospermia, oligozoospermia, and asthenozoospermia. It is thought that lifestyle factors such as cigarette smoking and alcohol abuse, as well as environmental factors, such as exposure to certain chemicals, are potential causes of spermatogenic failure [1]. In addition, genetic factors, such as gene mutations and chromosomal aberrations, can also be related to male infertility. Previously, three genome-wide association studies (GWASs) revealed that several single nucleotide polymorphisms (SNPs) were associated with male infertility including azoospermia and/or oligozoospermia [2–4]. Another GWAS, including 269 members of a Hutterite population in the USA, showed that of 41 SNPs associated with family size or birth rate, nine were also associated with reduced sperm quality and/or function in ethnically diverse men from Chicago [5]. Recently, we conducted a case-control association study to assess whether four SNPs, associated with semen parameters in men from Chicago with minor allele frequencies > 0.05 in the HapMap-JPT population, are associated with infertility (including azoospermia and oligozoospermia) in Japanese males. The SNPs assessed and their locations were as follows: rs7867029 (downstream of phosphoserine aminotransferase 1 [PSAT1], rs7174015 (in the ubiquitin specific peptidase 8 [USP8] gene, rs12870438 (in the epithelial stromal interaction protein 1 [EPSTI1] gene, and rs724078 (upstream of MAS1 oncogene-like protein [MAS1L] and downstream of ubiquitin D [UBD]). We found that rs7867029, rs7174015, and rs12870438 were significantly associated with the risk of azoospermia and/or oligozoospermia [6]. These SNPs were previously related to family size in the Hutterite GWAS. Conversely, rs724078, which was not associated with male infertility, was associated with birth rate in the previous GWAS. From this data, we predicted that other SNPs, in which significant association with family size was indicated in the previous Hutterite GWAS, may also be related to male infertility in the Japanese population. Of nine SNPs associated with reduced sperm quality and/or function in the men from Chicago, five SNPs were associated with family size in the previous GWAS [5]. These included rs10966811 (downstream of the tumor suppressor candidate 1 [TUSC1] gene) and rs10129954 (located in the D4, zinc and double plant homeodomain fingers, family 3 [DPF3] gene), as well as three SNPs (rs7867029, rs7174015, and rs12870438), which were analyzed previously [6]. In this study, we assessed whether rs10966811 and rs10129954 are associated with male infertility including azoospermia and oligozoospermia.

Materials and methods

Subjects and clinical trait measurements

In total, 145 consecutive patients who presented with infertility at the Department of Urology, St. Mariana University Hospital, Kanagawa Prefecture, Japan, were recruited from 2000 to 2011. Of these patients, 83 (aged 33.0 ± 5.6 years; mean ± SD) were diagnosed as having azoospermia, and 62 (aged 34.8 ± 6.2 years; mean ± SD) were diagnosed as having oligozoospermia. Some of the individuals in this study and diagnostic criteria have been described in previous reports [6–8]. Briefly, semen analysis was performed in accordance with the fourth edition of World Health Organization (WHO) Laboratory Manual for the Examination of Human Semen [9]. According to the criteria described in the fourth edition of WHO guidelines (1999), patients with azoospermia were diagnosed on the basis of semen analysis (absence of sperm in ejaculate), serum hormone levels, and the results of physical examinations. Oligozoospermia was defined as a sperm concentration of less than 20 × 10⁶/mL. We excluded patients with any other known cause of infertility (i.e., obstructive azoospermia, varicocele, cryptorchidism, hypogonadotropic hypogonadism, karyotype abnormalities, or complete deletion of AZF a, b, or c). Deletions in AZF a, b, and c were analyzed according to European Academy of Andrology and the European Molecular Genetics Quality Network best practice guidelines [10].

The 713 fertile Japanese men (aged 31.1 ± 4.8 years; mean ± SD) used as the control group were the partners of pregnant women who attended obstetric clinics in four cities in Japan (Sapporo, Kanazawa, Osaka, and Fukuoka). Some of the fertile individuals in this study and the eligibility criteria for participants have been described in previous reports [11–15]. Briefly, the eligibility criteria for the male participants were as follows: age 20–45 years at the time of invitation by the hospital, and that both the participant and his mother were born in Japan and currently living there as well. In addition, the pregnancy of their partner had to have been the result of conception by sexual intercourse and not through fertility treatment.

SNP selection and genotyping

For the association analysis, the SNP rs12376894 was selected as an alternative to rs10966811, because rs10966811 has no restriction sites, and rs10966811 and rs12376894 are in strong linkage disequilibrium (LD) (r ² = 0.91, |D′| = 1), according to data from JPT in 1000 genomes phase 1 release version 3 dataset [16]. rs12348 and rs2772579, located in TUSC1 and IZUMO family member 3 (IZUMO3), respectively, were selected as tag SNPs using SNPinfo Web Server (http://snpinfo.niehs.nih.gov/) [17], owing to their strong LD (r ² > 0.8), minor allele frequency > 0.05, and a minimum of two SNPs tagged each.

Genomic DNA was extracted from the peripheral blood samples of subjects using a QIAamp DNA blood kit (Qiagen; Tokyo, Japan) as previously described [6–8, 12–15]. The rs12376894, rs10129954, rs12348, and rs2772579 SNPs were detected by restriction fragment length polymorphism PCR using the following primer sets: rs12376894, 5′-CGGTGTAGAAAATGCGACCAA-3′ (forward) and 5′-ATCACCCCATCGTCAGTCAT-3′ (reverse); rs10129954, 5′-TTGGCTCTGAACCCAAGCAA-3′ (forward) and 5′-TGATTGGTCTCCCAGCCTCT-3′ (reverse); rs12348, 5′-TCGCCGCAACCTTTCAGATA-3′ (forward) and 5′-CCGCGGATACCTTGGACTAC-3′ (reverse); and rs2772579, 5′-GCGGGAAAAGTGGAGTCAGA-3′ (forward) and 5′-AGATTTCCCAGCAAGGAGCC-3′ (reverse). DNA from each subject was amplified using EmeraldAmp MAX PCR Master Mix (TaKaRa Bio Inc., Otsu, Japan) under the following PCR cycling parameters: initial denaturation at 94 °C for 3 min; 30 cycles of 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min; and final extension for 3 min at 72 °C. The resulting PCR products were then digested at 37 °C for 3 h using the following restriction enzymes: EarI (rs12376894), HpyCH4IV (rs10129954), PvuII (rs12348), and Hpy166II (rs2772579); all of which were sourced from New England Biolabs Japan Inc., Tokyo, Japan. The digested products were separated by electrophoresis on 2.5% agarose gel. The following fragment sizes were used for allele identification on gels: for rs12376894, 594 bp (A-allele) and 287 + 307 bp (G-allele); for rs10129954, 738 bp (T-allele) and 365 + 373 bp (C-allele); for rs12348, 633 bp (T-allele) and 269 + 364 bp (C-allele); and for rs2772579, 250 bp (G-allele) and 74 + 176 bp (C-allele).

Statistical analysis

The fertile cohort was used as the control, and odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated using logistic regression analysis. All statistical analyses were performed using R version 3.1.2 (The R Project for Statistical Computing [http://www.r-project.org]), and statistical significance was declared at P value < 0.05. Hardy−Weinberg equilibrium (HWE) was assessed in both patients with infertility and fertile control samples with a threshold of 0.05 by using Pearson chi-square test for genotypes. LD was calculated using PLINK version 1.07 [18] with JPT in 1000 genomes phase 1 release version 3 dataset [16] or genotyped fertile control data in this study.

Results

The genotype and allele frequencies of rs12376894 and rs10129954 SNPs among 713 fertile controls and 83 patients with azoospermia, and 62 with oligozoospermia are summarized in Table 1. Two SNPs were in HWE in the both patients with infertility and fertile controls.

Table 1.

Genotype and allele frequencies of rs12376894 and rs10129954 SNPs in the subjects

							Control		Case
							(n = 713)		Azoospermia (n = 83)		Oligozoospermia (n = 62)
SNP	Chr	Position (NCBI build 37)	Closest Genes	Location	Dist.(kb)	Allelea	Genotypesb	AFc	Genotypesb	AFc	Genotypesb	AFc
rs12376894	9	25,268,867	TUSC1	dwnst.	445.4	A/G	152/364/197	0.47	13/45/25	0.43	22/25/15	0.56
			IZUMO3	upst.	687.8
rs10129954	14	73,150,701	DPF3	intron		C/T	1/89/623	0.06	2/16/65	0.12	1/17/44	0.15

Chr chromosome, dwnst. downstream, upst. upstream, Dist. distance, TUSC1 tumor suppressor candidate 1, IZUMO3 IZUMO family member 3, DPF3, D4, zinc and double plant homeodomain fingers, family 3

^aAllele indicates control subject minor/major allele

^bGenotypes indicate control subject minor homozygote/heterozygote/major homozygote

^cAF indicates the allele frequency of the control subject minor alleles

To verify whether rs12376894 and rs10129954 SNPs are associated with male infertility traits, we carried out a case-control study using a logistic regression analysis under three comparative genetic models (additive, recessive, or dominant). The results of the logistic regression analysis are summarized in Table 2. There was a statistically significant association between rs12376894 and oligozoospermia in the recessive model (OR = 2.03, 95% CI = 1.17–3.52, P = 0.012). rs10129954 displayed significant association with azoospermia in all three models: log-additive (OR = 2.05, 95% CI = 1.21–3.46, P = 7.4 × 10⁻³), recessive (OR = 17.60, 95% CI = 1.58–196.00, P = 0.020), and dominant (OR = 1.92, 95% CI = 1.09–3.38, P = 0.025); and with oligozoospermia in two models: log-additive (OR = 2.83, 95% CI = 1.61–4.97, P = 2.9 × 10⁻⁴), and dominant (OR = 2.83, 95% CI = 1.57–5.11, P = 5.6 × 10⁻⁴). rs10129954 is located in an intronic region of DPF3. Therefore, DPF3 protein may affect spermatogenesis, although there is no evidence. rs12376894 is located in an intergenic region between TUSC1 and IZUMO3. Then, we speculate that TUSC1 or IZUMO3 may be related to male infertility. We performed a further association analysis between TUSC1 and IZUMO3 polymorphisms and male infertility; rs12348 and rs2772579 were selected as tag SNPs respectively, according to the pairwise LD measurement. rs12348 is located in a 3′ untranslated region (UTR) of TUSC1, and rs2772579 is located in a 5′ UTR of IZUMO3. The genotype and allele frequencies of rs12348 and rs2772579 among 713 fertile controls, as well as 83 patients with azoospermia, and 62 with oligozoospermia are summarized in Table 3. These SNPs were in HWE in both patients with infertility and fertile control group. Using a logistic regression analysis, a statistically significant association between rs12348 and azoospermia was observed in the recessive model (OR = 1.82, 95% CI = 1.07–3.10, P = 0.028), and oligozoospermia in two models: log-additive (OR = 1.87, 95% CI = 1.29–2.71, P = 8.8 × 10⁻⁴), and recessive (OR = 3.16, 95% CI = 1.82–5.50, P = 4.5 × 10⁻⁵) (Table 4). However, no association was found between rs2772579 and male infertility traits in three genetic models.

Table 2.

The associations derived from different comparative genetic models (additive, recessive, and dominant) between two SNPs (rs12376894 and rs10129954) and azoospermia, and oligozoospermia

SNP	Model	Azoospermia		Oligozoospermia
		OR (95% CI)	P value	OR (95% CI)	P value
rs12376894	Log-additive (risk allele, A)a	0.84 (0.61–1.17)	0.31	1.43 (0.99–2.07)	0.60
	Recessive (AA vs. AG + GG)	0.69 (0.37–1.27)	0.23	2.03 (1.17–3.52)	0.012
	Dominant (AA + AG vs. GG)	0.89 (0.54–1.46)	0.63	1.20 (0.65–2.19)	0.56
rs10129954	Log-additive (risk allele, C)a	2.05 (1.21–3.46)	7.4 × 10 −3	2.83 (1.61–4.97)	2.9 × 10 −4
	Recessive (CC vs. CT + TT)	17.60 (1.58–196.00)	0.020	11.70 (0.72–189.00)	0.084
	Dominant (CC + CT vs. TT)	1.92 (1.09–3.38)	0.025	2.83 (1.57–5.11)	5.6× 10 −4

Data are shown as odds ratio (OR), 95% confidence interval (CI), and P value

Bold numbers indicate P value < 0.05

^aLog-additive, additive model in log-odds scale

Table 3.

Allele and genotype frequencies of rs12348 in TUSC1 and rs2772579 in IZUMO3 in the study group

					Control		Case
					(n = 713)		Azoospermia (n = 83)		Oligozoospermia (n = 62)
SNP	Position (NCBI build 37)	Genes	Function	Allelea	Genotypesb	AFc	Genotypesb	AFc	Genotypesb	AFc
rs12348	25,677,217	TUSC1	3′-UTR	T/C	112/345/256	0.40	21/37/25	0.48	23/23/16	0.56
rs2772579	24,545,695	IZUMO3	5′-UTR	G/C	28/250/435	0.21	1/25/57	0.16	3/21/38	0.22

TUSC1 tumor suppressor candidate 1, IZUMO3 IZUMO family member 3

^aAllele indicates control subject minor/major allele

^bGenotypes indicate control subject minor homozygote/heterozygote/major homozygote

^cAF indicates the frequency of the control subject minor alleles

Table 4.

The associations derived from different comparative genetic models (additive, recessive, and dominant) between the SNPs, rs12348 (in TUSC1) and rs2772579 (in IZUMO3), and azoospermia and oligozoospermia

SNP	Model	Azoospermia		Oligozoospermia
		OR (95% CI)	P value	OR (95% CI)	P value
rs12348	Log-additive (risk allele, T)a	1.37 (0.99–1.89)	0.058	1.87 (1.29–2.71)	8.8 × 10 −4
	Recessive (TT vs. TC + CC)	1.82 (1.07–3.10)	0.028	3.16 (1.82–5.50)	4.5 × 10 −5
	Dominant (TT + TC vs. CC)	1.30 (0.79–2.13)	0.30	1.61 (0.89–2.90)	0.11
rs2772579	Log-additive (risk allele, G)a	0.70 (0.45–1.09)	0.11	1.02 (0.65–1.60)	0.93
	Recessive (GG vs. GC + CC)	0.30 (0.04–2.22)	0.24	1.24 (0.37–4.21)	0.73
	Dominant (GG + GC vs. CC)	0.71 (0.44–1.16)	0.18	0.99 (0.58–1.68)	0.97

Data are shown as odds ratio (OR), 95% confidence interval (CI), and P value

Bold numbers indicate P value < 0.05

^aLog-additive, additive model in log-odds scale

To understand the relationship between rs12376894 and rs12348, we performed a conditional logistic regression analysis for azoospermia, and oligozoospermia adjusted for each SNP. The strength of association between two SNPs and azoospermia or oligozoospermia was not reduced (Table 5). Additionally, we calculated the pairwise LD between rs12376894 and rs12348 using genotyped fertile control data in this study pairwise r ² values were low (r ² = 0.005). Therefore, it is suggested that rs12376894 and rs12348 are independently related to azoospermia and/or oligozoospermia.

Table 5.

Conditional logistic regression analysis for azoospermia, and oligozoospermia under three genetic models (additive, recessive, and dominant)

SNP	Model	Azoospermia		Oligozoospermia
		OR (95% CI)	P value	OR (95% CI)	P value
rs12376894	Log-additive (risk allele, T)a	0.86 (0.62–1.19)	0.36	1.47 (1.01–2.14)	0.042
	Recessive (TT vs. TC + CC)	0.69 (0.37–1.28)	0.23	2.10 (1.20–3.68)	9.2 × 10 −3
	Dominant (TT + TC vs. CC)	0.90 (0.55–1.47)	0.67	1.23 (0.67–2.25)	0.50
rs12348	Log-additive (risk allele, T)a	1.36 (0.98–1.87)	0.065	1.91 (1.32–2.77)	6.4 × 10 −4
	Recessive (TT vs. TC + CC)	1.82 (1.06–3.10)	0.029	3.24 (1.85–5.66)	3.6 × 10 −5
	Dominant (TT + TC vs. CC)	1.29 (0.79–2.12)	0.31	1.63 (0.90–2.94)	0.11

Data are shown as odds ratio (OR), 95% confidence interval (CI), and P value with adjusted with each SNPs

Bold numbers indicate P value < 0.05

^aLog-additive, additive model in log-odds scale

Discussion

In this study, we identified that the SNPs rs12376894 and rs10129954, which have been associated with family size in a genome-wide association study (GWAS) of a Hutterite population in the USA [5], were significantly correlated with azoospermia and/or oligozoospermia in a Japanese population. rs12376894 is located 445.4 kb downstream of TUSC1 and 687.8 kb upstream of IZUMO3. Therefore, we speculate that TUSC1 or IZUMO3 may be contribute to spermatogenesis, and assessed whether TUSC1 and IZUMO3 polymorphisms are related to male infertility. We found that rs12348 in TUSC1 was associated with azoospermia and oligozoospermia; an especially strong association was identified in the recessive model (TT vs. TT + CC). However, rs2772579 in IZUMO3 was not associated with male infertility. TUSC1 has been suggested as a candidate tumor suppressor gene in several cancers [19–23], and it is reported that variation in TUSC1 and the surrounding region is associated with risk of cancers [24, 25]. It is also reported that TUSC1 variants are related to tanning ability [24], supernumerary teeth occurrence [26], amongst other traits. However, no evidence that TUSC1 contributes to spermatogenesis has yet been published. TUSC1 is an intronless gene, and major transcripts TUSC1-L (2.0 kb long) and TUSC1-S (1.5 kb long) have been detected in a wide range of human adult tissues [19]. Interestingly, two transcripts were most strongly expressed in human testes and another 1.1-kb transcript has been detected only in this organ [19]. Similarly, in mouse tissues, TUSC1-S was most strongly expressed in the testis [19]. Therefore, TUSC1 may have an important function in the testis. On the other hand, IZUMO3 was also strongly expressed in the testis. However, IZUMO3 is essential for the function of the gamete fusion rather than spermatogenesis [27].

rs12376894, which displayed significant association with oligozoospermia, is located 445.4 kb downstream of TUSC1. According to LD and conditional logistic regression analysis adjusted for each SNP, rs12376894 and rs12348 are individually related to male infertility. It is known that SNPs within noncoding regions (intronic, intergenic, etc.) may influence expression regulation of genes both local to and distant from the locus [28]. Though significant expression quantitative trait loci (eQTL) of rs12376894 were not found in the GTEx Portal database [29], this SNP may contribute to the regulation of TUSC1 expression.

Liu et al. reported that rs10129954 is strongly associated with idiopathic male infertility (asthenozoospermia, oligozoospermia, and oligoasthenozoospermia) in Chinese Han people [30]. In the current study, we found a similar relationship in a Japanese population. rs10129954 is located in an intronic region of DPF3. DPF3 protein is a member of the BRG1-associated factor (BAF) chromatin remodeling complex (SWI/SNF-like complex), with binding acetylated and methylated H3 and H4 lysine residues, and regulates the transcription during heart and muscle development [31]. It is known that the SWI/SNF chromatin remodeling complex has an important role throughout spermatogenesis [32, 33], especially during male meiosis [34]. DPF3 is strongly expressed in the human testis as well as in the ovaries, heart, and brain [35]. It is suggested that DPF3 may contribute to spermatogenesis via chromatin remodeling. However, it is unclear whether the intronic variant rs10129954 influences DPF3.

In conclusion, we found that polymorphisms in TUSC1 and DPF3 displayed strong associations with male infertility. The effect of these polymorphisms on spermatogenesis remains unknown. To study this, functional analyses using genetically modified animals are required. In addition, in this study, the sample size of case subject was not very large. Therefore, further studies with larger sample size are also required to achieve the sufficient statistical power.

Funding information

This study was supported in part by the Ministry of Health and Welfare of Japan (1013201) (to T.I.), Grant-in-Aids for Scientific Research (C) (26462461) (to Y.S.) from the Japan Society for the Promotion of Science, the European Union (BMH4-CT96-0314) (to T.I.), the Suzuki Urinary Foundation (to Y.S.), and a research program for the development of intelligent Tokushima artificial Exome (iTEX) from Tokushima University.

Compliance with ethical standards

This study was approved by the ethics committees of the University of Tokushima and St. Marianna Medical University. All participants provided written informed consent.

Conflict of interest

The authors declare that there are no conflict of interest.