Copy number gain of VCX, X-linked multi-copy gene, leads to cell proliferation and apoptosis during spermatogenesis

Male factor infertility affects one-sixth of couples worldwide, and non-obstructive azoospermia (NOA) is one of the most severe forms. In recent years there has been increasing evidence to implicate the participation of X chromosome in the process of spermatogenesis. To uncover the roles of X-linked multi-copy genes in spermatogenesis, we performed systematic analysis of X-linked gene copy number variations (CNVs) and Y chromosome haplogrouping in 447 idiopathic NOA patients and 485 healthy controls. Interestingly, the frequency of individuals with abnormal level copy of Variable charge, X-linked (VCX) was significantly different between cases and controls after multiple test correction (p = 5.10 × 10−5). To discriminate the effect of gain/loss copies in these genes, we analyzed the frequency of X-linked multi-copy genes in subjects among subdivided groups. Our results demonstrated that individuals with increased copy numbers of Nuclear RNA export factor 2 (NXF2) (p = 9.21 × 10−8) and VCX (p = 1.97 × 10−4) conferred the risk of NOA. In vitro analysis demonstrated that increasing copy number of VCX could upregulate the gene expression and regulate cell proliferation and apoptosis. Our study establishes a robust association between the VCX CNVs and NOA risk.


INTRODUCTION
Male infertility, a common reproductive disorder, is estimated to affect approximately one-sixth of couples worldwide [1]. A significant proportion of male infertility is accompanied by idiopathic azoospermia, most often presenting as non-obstructive azoospermia (NOA), defined as the absence of sperm in the ejaculate without the obstruction of reproductive tract. Certain pathogenic genetic causes, especially the Yq microdeletions, in male infertility are well established so far [2][3][4]. However, www.impactjournals.com/oncotarget Chromosome Inactivation (MSCI) [8,9]. Recently, Zhang et al. noticed that some testis-specific genes could be detected throughout all stages of spermatogenesis, indicating that they escape MSCI [10]. And it was estimated that 70% of these genes were expressed in post-meiotic stages [10]. Interestingly, the majority of these genes are multi-copy X-linked genes, moreover, the more copies of a gene, the higher chance to escape from X inactivation, accounting for higher expression level of multi-copy X-linked genes than that of single-copy X-linked genes [11].
With the rapid development of genome-scanning technologies [12], the structural variations in human genome were emerged and fully uncovered [13]. One type of structural variations is copy number variation (CNV) that generally defined as a segment of DNA (1 kb or larger), presenting at a variable copy number compared to a reference genome [13]. It has been demonstrated that CNV was associated with numerous diseases [14], including Alzheimer's disease, Parkinson disease, congenital scoliosis [15][16][17]. Recently, increasing evidence implied that CNVs were crucial determinant of spermatogenesis and might contribute to male infertility [18][19][20], while little is known about the relationship between X-linked gene CNVs and NOA.
To systematically analyze X-linked gene CNVs and their effects on NOA, we performed comprehensive molecular analyses in 447 idiopathic NOA men and 485 healthy controls. Eventually, the CNVs of Variable charge, X-linked (VCX) and Nuclear RNA export factor 2 (NXF2) were found to be associated with NOA, subsequently in vitro analysis was carried out to clarify the potential role of VCX in spermatogenesis.

Characteristics and clinical parameters of the study population
Our study included 447 NOA patients and 485 fertile controls. The mean age and BMI in control group were 30.1 and 24.3, and in case group was 29.5 and 23.3, respectively (Table 1). There were no significant differences identified between the case and control groups with regard to their age, smoking and drinking status. However, significant differences were identified in BMI between cases and controls (Table 1).

Y-hg distribution between the case and control groups
To test for the potential influence of genetic backgrounds, 14 Y chromosome binary markers were used to define 14 Y-hgs in patients and normal subjects (controls). No significant difference in the Y-hg distribution was found between the healthy control group and the NOA case group (Supplementary Material, Table S1), which suggested that the genetic background, mainly Y-hgs, may not affect our results of the present association study.

X-linked multi-copy gene copy number variations and NOA
Overall, seven X-linked gene copy numbers were determined in 447 NOA patients and 485 healthy controls using the AccuCopy method. The distributions of copy number of seven genes in case and control groups were shown in Table 2. We found that the frequency of individuals with abnormal copy number of NXF2 (OR, 2.46, 95% CI 1.15−5.25, p = 1.66 × 10 −2 ) and VCX (OR, 2.53, 95% CI 1.60−4.02, p = 5.10 × 10 −5 ) in NOA group was significantly higher than that in control group, while the frequency of FAM47 (OR, 0.27, 95% CI 0.07−0.95, p = 2.89 × 10 −2 ) was significantly lower in NOA groups ( Table 2). However, only the association between VCX and NOA risk retained after Bonferroni correction ( Table 2).
To discriminate the effects of copy number gain or loss in these genes on NOA, we subdivided the subjects into three subgroups: the common level copy group, the less than common level group and the more than common level group. The detailed distributions were shown in Table 3. Our results demonstrated that 10 out of 485 (~2%) was found with decreased NXF2 copy number in the control group, while no one was found in the NOA group. On the contrary, 22 out of 447 (~5%) were found with increased NXF2 copy number in NOA group, while no one was found in the control group. Namely, decreased NXF2 copy number showed protective against NOA (p = 2.20 × 10 −3 ), while increased copy number conferred the risk of NOA (p = 9.21 × 10 −8 ).

VCX expression in germ cells and seminal plasma
To explore the transfection efficiency of VCX in 293 and GC cell lines, we measured the VCX expression before and after transfection. As it showed (Supplementary Material, Figure S1A-S1C), the VCX expression was significantly increased after transfected. Besides, to investigate whether the VCX copy number gains lead to mRNA overexpression, the mRNA level of VCX in seminal plasma was detected, and we found that the mRNA expression of VCX was increased in NOA (Supplementary Figure S1D).

Effects of VCX on cell proliferation, cell apoptosis and cell cycle
By searching Database of Genomic Variants (DGV), we found that most frequent structure variations in NXF2 were copy number loss, and didn't match what we observed that gaining copy was a risk in NOA. Additionally, its frequency was much lower (27/873, 3%) than VCX (2/9, 22%). Thus, in the functional study, we only included VCX.

DISCUSSION
X chromosome was derived from autosomal chromosome and under natural selection, and its gene content changed largely, some of which were multi-copy genes [21]. The roles of X-linked CNVs in cognitive disorders [22,23], primary ovarian insufficiency [24] and premature ovarian failure [25] have been investigated extensively, while few studies have reported the associations between X-linked CNVs and male infertility. Recently, accumulating evidence revealed that infertile men had significantly higher duplication or deletion load than normal controls [26][27][28].    In this study, we hypothesize that the copy number variations of X-linked multi-copy genes confer the risk of male infertility. We explored the role of X-linked CNVs in 447 individuals with NOA and 485 healthy controls by applying AccuCopy assay. Amazingly, we found that the gaining more copies of VCX and NXF2 were associated with NOA.
NXF2, encoded by multigene families on X chromosome, is a germ cell-specific gene [29]. It has been demonstrated that NXF2 play functions in male meiosis,  (E, F, G) Cell counts in S-phase of cell cycle were markedly reduced in VCX group. Each data point represented the mean ± SE from three separate experiments in which treatments were performed in triplicate. *P < 0.05. www.impactjournals.com/oncotarget spermatogonial proliferation as well as maintenance of spermatogonial stem cells [30,31]. And the deletion of NXF2-NXF3 genomic region on the X chromosome in male germline causes male infertility in mice [32]. Contrary to previous studies, our results demonstrated that the copy number increase of NXF2 conferred the risk of NOA. Functional studies, including both the gain and loss of this gene, are needed to explore the potential mechanisms in the future.
VCX localizes on cytogenetic band Xp22.3 and contains 4 copies [33]. From DGV, we found most structure variations in Xp23 are gaining the gene copies (including VCX). Thus, we overexpressed VCX in vitro for the further investigation. As one member of novel cancer/testis (CT) antigen [34], VCX exclusively expresses in testis and is most likely restricted to male germ cells [35], moreover, its expression is regulated epigenetically [36]. It has been demonstrated that VCX was overexpressed in a subset of non-small cell lung cancer (NSCLC) cell lines and tumor tissues [37,38], suggesting that it has oncogenic role [34]. To explore the role of VCX in spermatogenesis, functional analysis was performed in 293T, GC-1 and GC-2 cell lines, and our results demonstrated that VCX promoted cell apoptosis and inhibited cell growth, contributing to spermatogenesis impairment [39], which was probably by mediating mitochondria-dependent apoptosis pathway [40] or p53-Bax pathway [41]. Moreover, VCX delayed cell-progression in G1 to S transition, resulting in cell division disorder and spermatogenic failure.
In conclusion, our findings revealed that copy number gain of VCX attributed to NOA by an underlying mechanism that induced cell apoptosis, inhibited cell proliferation and disrupted cell cycle progression, thus VCX may be a potential biomarker for the susceptibility to NOA.

Study population and sample collection
Study subjects were volunteers from the affiliated hospitals of Nanjing Medical University between 2010 and 2014 (NMU Infertile Study) and described previously [42,43]. All infertile male subjects were Han Chinese men determined to have idiopathic NOA and selected on the basis of comprehensive andrological testing, including examination of medical history, physical examination, semen analysis, scrotal ultrasound, hormone analysis, karyotyping and Y chromosome microdeletion screening. Those with a history of cryptorchidism, vascular trauma, orchitis, obstruction of the vas deferens, vasectomy, abnormalities in chromosome number or microdeletions of the azoospermia factor region on the Y chromosome were excluded from the study. Subjects with NOA had no detectable sperm in the ejaculate after evaluation of the centrifuged pellet. To ensure the reliability of the diagnosis, each individual was examined twice, and the absence of spermatozoa from both replicate samples was taken to indicate azoospermia. The controls had fathered one or more healthy child and were frequency matched to the cases on the basis of age and area of residence. The protocol and consent form were approved by the Institutional Review Boards of Nanjing Medical University. All activities involving human subjects were done under full compliance with government policies and the Helsinki Declaration. Totally, 447 patients with nonobstructive azoospermia and 485 matched healthy male controls were recruited in this study. Written informed consents were obtained from all of them. After completing a questionnaire with detailed information, each subject donated 5 ml of blood for genomic DNA extraction.  [44,45] were used to define 14 Y-hgs following the nomenclature recommended by the Y Chromosome Consortium [46] and its update described by Sengupta et al [47].
The SNaPshot (Applied Biosystems, Foster City, CA) minisequencing reaction assay was used for polymorphism genotyping [48] The detailed experimental procedures have been well described in our previous study [49].

CNV genotyping using the AccuCopy method
The AccuCopy technique, a CNV genotyping method based on multiplex fluorescence competitive amplification, was recently developed by Genesky Biotechnologies (Shanghai, China), and well described by Du et al [50]. Limited by the highly homologous sequences and the detect threshold of AccuCopy, only seven gene families were recruited in our study. In order to ensure the reliability of our results, we chose three target genomic segments within the CNV region for the CT45 and FAM47 genes, and two for the NXF2, CSAG, CTAG, H2AFB1, and VCX. According to the reaction condition and stability, these target segments were divided into two panels. The reference genome sequences were obtained from the UCSC Genome Browser (http://genome.ucsc. edu/ genome assembly hg19). Additionally, three reference segments used for normalization were screened and chosen at three loci of 2P, 16P and 20q. The forward (F) and reverse (R) primers of these segments and the size of PCR product amplified from human genomic DNA are provided in Supplementary Table S2, respectively. AccuCopy assay was modified as one multiplex competitive PCR amplification followed by one labeling extension, in which one of PCR primers for each fragment was synthesized with addition of a universal sequence, i.e. 5′ACACGACCGGTAACGCTTAGA3′ at 5′ end so that the PCR products can be FAM-labeled in a subsequent extension reaction using a 5′ FAM-modified primer i.e. FAMEF: 5′ [FAM] ACACGACCGGTAACGCTTAGA3'. All primers mentioned above were synthesized at Sangon Biotech (Shanghai, China). The competitive DNAs for the three reference and sixteen target segments were designed and synthesized in double strand by Genesky Biotechnologies and provided in a 200 × mixture. The sequences of synthesized competitive DNAs were almost the same as their reference homologies in the human reference genome except 2 bp deletion introduced. The detailed experimental procedures were described previously with minor modifications.

Statistical analysis
Statistical analyses were carried out using Stata 10.0 statistical software package (Stata Corp, LP). Pearson's chi-squared test was used to demonstrate differences of categorical variables such as drinking and smoking status between cases and controls. Student's t-test was used to show differences in continuous variables such as age and body mass index (BMI). Associations between CNVs and the risk of NOA were evaluated by computing odds ratios (ORs) and their 95% confidence intervals (CIs) from logistic regression under different models. The Bonferroni adjustment for multiple testing was applied. The P-value for a truly significant result was calculated as 7.14 × 10 −3 (0.05/7).

Functional analysis of VCX CNV
By searching the DGV, we found that there are seven structure variations in VCX and most of them were copy number gains (Supplementary Material, Figure S2). Human VCX cDNA sequence (NM 013452.2) was synthesis by Realgene biotechnology company and inserted in expression vector pcDNA3.1 (+). The recombinant plasmid was confirmed by endonucleases digesting and DNA sequencing.