Article Text
PDF
Abstract
Background Infertility affects approximately 15% of couples worldwide with male infertility being responsible for approximately 50% of cases. Although accumulating evidence demonstrates the critical role of the X chromosome in spermatogenesis during the last few decades, the expression patterns and potential impact of the X chromosome, together with X linked genes, on male infertility are less well understood.
Methods We performed X chromosome exome sequencing followed by a two-stage independent population validation in 1333 non-obstructive azoospermia cases and 1141 healthy controls to identify variant classes with high likelihood of pathogenicity. To explore the functions of these candidate genes in spermatogenesis, we first knocked down these candidate genes individually in mouse spermatogonial stem cells (SSCs) using short interfering RNA oligonucleotides and then generated candidate genes knockout mice by CRISPR-Cas9 system.
Results Four low-frequency variants were identified in four genes (BCORL1, MAP7D3, ARMCX4 and H2BFWT) associated with male infertility. Functional studies of the mouse SSCs revealed that knocking down Bcorl1 or Mtap7d3 could inhibit SSCs self-renewal and knocking down Armcx4 could repress SSCs differentiation in vitro. Using CRISPR-Cas9 system, Bcorl1 and Mtap7d3 knockout mice were generated. Excitingly, Bcorl1 knockout mice were infertile with impaired spermatogenesis. Moreover, Bcorl1 knockout mice exhibited impaired sperm motility and sperm cells displayed abnormal mitochondrial structure.
Conclusion Our data indicate that the X-linked genes are associated with male infertility and involved in regulating SSCs, which provides a new insight into the role of X-linked genes in spermatogenesis.
- male infertility
- non-obstructive azoospermia (NOA)
- X chromosome exons
- spermatogonial stem cells (SSCs)
Statistics from Altmetric.com
Introduction
Male infertility is one of the major issues of human health, the aetiology of which remains largely unknown. In contrast to the relatively small Y chromosome, the X chromosome contains approximately 5% of genomic DNA and has ~1100 annotated genes.1 Males normally have only one copy of the X chromosome, resulting in males being hemizygous for nearly 5% of the genome. Thus, de novo or rare mutations in X-linked genes cannot be compensated for and will exhibit an X-linked trait regardless of the dominant or recessive nature of the allele.
By screening the genes expressed in mice spermatogonia, Wang et al found that the X chromosome harbours many genes that were expressed specifically in male germ cells, suggesting that the X chromosome may play a prominent role in spermatogenesis.2 Given the existence of meiotic sex chromosome inactivation (MSCI) during the process of meiosis, it was proposed that mammalian X chromosomes are enriched for genes involved in early spermatogenesis.2–4 Recent studies have demonstrated that MSCI-initiated silencing persists beyond meiosis.5–7 It was reported that some of the genes identified in spermatogonia were inactivated during meiosis, but the expression levels of some of these genes increased again after meiosis was completed, even in the stage of round spermatids.8–12
Our previous genome-wide association study (GWAS) identified three susceptibility loci for non-obstructive azoospermia (NOA) in the Han Chinese population.13 While most variants identified so far confer relatively small increments of risk, a large portion of the heritability of complex traits has not been well explained by GWAS, and the causality of these loci has not been elucidated yet.14 15 As males are hemizygous for the X chromosome, unlike autosomal chromosomes, X linked variants cannot be well identified through GWAS due to limitations. Additionally, traditional heritability studies that rely on pedigree-based linkage analyses are impossible due to the fact that infertile men lack offspring. Thus, we proposed that the variants in X linked genes may make up a significant proportion of male infertility causes.
Given that the function of the X chromosome has not been fully investigated thus far, we sequenced the exons of all the genes located on the X chromosome and identified low-frequency variants in BCORL1 (BCL6 corepressor-like 1), MAP7D3 (MAP7 domain containing 3), ARMCX4 (armadillo repeat containing, X linked 4) and H2BFWT (H2B histone family member W, testis specific) associated with male infertility. In addition, knockout of Bcorl1 using CRISPR-Cas9 results in a significant reduction of mature sperm and sterility. Our findings underscore the importance of the X linked genes in spermatogenesis.
Materials and methods
Ethics statement
All participants were voluntary and completed a written informed consent before taking part in this research.
Study design and participants
We designed a three-stage case-control study that first employed X chromosome exons sequencing in individuals diagnosed with NOA and a subsequent two-stage replication with large-scale follow-up genotyping of identified candidate variants in the large cohorts of NOA-affected subjects and controls.
The discovery stage included 96 NOA cases and 96 healthy male controls recruited from the Nanjing Center of Reproductive Medicine between March 2010 and January 2013. The first replication stage (replication I) included 689 NOA cases recruited from the infertility clinic at the Affiliated Hospitals of Nanjing Medical University in the Jiangsu Province (NJMU Infertile Study) and 515 healthy male controls from the same hospitals during the same period. The second replication stage (replication II) included 548 NOA cases sampled from the Renji Hospital, Shanghai, and 530 healthy male controls from the same hospital. Some cohorts within the sample sets have been used in previously published data.16 17
All infertile subjects were genetically unrelated ethnic Han Chinese men diagnosed with idiopathic NOA and selected on the basis of comprehensive andrological examination, including semen analysis, examination of medical history, a series of physical examinations, scrotal ultrasound, hormone analysis, karyotyping and Y chromosome microdeletion screening. All controls with normal reproductive function were from the early pregnancy registry of the same hospitals, whose wives were in the first trimester of pregnancy and confirmed as having healthy babies 6–8 months later.
To identify potential confounders, a questionnaire was used to collect additional demographic information, including personal background, lifestyle factors, occupational and environmental exposures, genetic risk factors, sexual and reproduction status, etc. 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. Semen analysis for sperm concentration, motility and morphology was performed following WHO criteria.18 Subjects with NOA had no detectable sperm in the ejaculate after evaluation of the centrifuged pellet. To differentiate from obstructive azoospermia, only idiopathic azoospermia patients with small and soft testis, normal fructose and neutral alpha glucosidase in seminal plasma were included in this study. Each individual was double examined to ensure the reliability of the diagnosis, and the absence of spermatozoa from both replicate samples was used to indicate azoospermia.
X exons sequencing variant calling
Genomic DNA was isolated from peripheral blood samples by using QIAamp DNA Blood kits (Qiagen, Germany) and loaded on a 1% agarose gel for integrity check. The X exons regions were selected based on NCBI and UCSC database and the target regions were 2 MB containing 805 genes (online supplementary table S4). Customer designed Agilent SureSelect Arrays, which contained 30 463 probes and covered 98.9% of X exons, were used to capture DNA by hybridisation (Agilent Technologies, USA). The hybridisation and enrichment were carried out on an AB 2720 Thermal Cycler (Life Technologies, USA). After incubating the hybridisation mixture for 16 or 24 hours at 65°C, captured DNA was enriched by the SureSelect Target Enrichment Kit (Agilent Technologies) with 98°C for 30 s, 10 cycles of 98°C for 10 s, 60°C for 30 s, 72°C for 30 s and 72°C for 5 min. The enriched DNA was used to construct a library following the manufacturer’s standard procedure for the Truseq DNA Sample Preparation Kit (Illumina, USA). Sequencing was performed on the Illumina HiSeq 2000 to generate paired-end 100 bp reads. Each sample was sequenced to an average depth of 115×, with nearly 90% of the targeted regions covered by ≥2×, nearly 86% covered by ≥10× and nearly 76% of the targeted regions covered by ≥30×.
Supplemental material
Identifying candidate variants from the sequence data
After checking quality of sequencing reads by FastQC (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/), low-quality data were trimmed by FASTX-Toolkit (http://hannonlab.cshl.edu/fastx_toolkit/index.html). For calling of variants, good quality reads were mapped to the human reference genome (UCSC, hg19) using the Burrows-Wheeler Aligner.19 The Genome Analysis Toolkit was used for realignment, base quality score recalibration and duplicate removal.20 In brief, VarScan and SAM tools were used to identify variants in X exons data with in-house parameters.21 22 Polymorphisms referenced in dbSNP132 or the 1000 Genomes Project with a minor allele frequency over 1% was removed from further study. A list of variants was generated considering variants (missense, insertion and deletion) with MAF <1% in coding regions. Such variants were finally annotated by PolyPhen-2, SIFT and Annovar based on two databases (hg19 and 1000 Genome Project).23–25
Sanger sequencing validation
To verify the results of the next-generation sequencing described above, PCR amplification and Sanger sequencing were performed on putative variants identified by target sequencing. Target regions were amplified using the GeneAmp PCR System 9700 (Applied Biosystems, USA) and then sequenced with the 3730 DNA Analyzer (Applied Biosystems). The primers for these selected loci are presented in online supplementary table S5.
Supplemental material
Follow-up genotyping of screened genetic variants by SNPscan
Candidate variants were genotyped in additional cases and controls by a custom-by-design 48-Plex SNPscan Kit (Genesky Biotechnologies, China). This kit was developed according to patented SNP genotyping technology by Genesky Biotechnologies, which was based on double ligation and multiplex fluorescence PCR.26 In order to validate the genotyping accuracy using SNPscan Kit, 5% duplicate samples were analysed by single nucleotide extension using the Multiplex SNaPshot Kit (Applied Biosystems); the concordance rates were >99%.
Statistical analysis
Plink and R packages were used for data analyses.27 Associations between single nucleotide variants (SNVs) and the risk of NOA were evaluated under an allele model. The Bonferroni adjustment for multiple testing was applied and the p value for a significant result was calculated as 1.25×10–2 (0.05/4). Mann-Whitney U test and one-way analysis of variance were used to compare the gene expression levels between different groups and unpaired t-test was used to compare the knockout and WT groups, p value <0.05 was considered the threshold for significance.
Spermatogonia stem cell culture
SSCs were enriched as previously described.28 Briefly, Thy1+ germ cell populations were enriched with microbead-conjugated antibodies(catalogue no. 130-049-101; Miltenyi Biotech) using magnetic activated cell sorting. Then, Thy1+ cells were seeded at a density of ~1.0×105 cells/well in a 12-well format containing mitotically inactivated SNL-STO feeder cells. SSCs were maintained in serum-free medium supplemented with 20 ng/mL of glial cell line-derived neurotrophic factor (GDNF) (R&D Systems) and 1 ng/mL of basic fibroblast growth factor 2 (BD Biosciences).
RNA interference, real-time PCR and immunoblot analysis
SSCs were removed from feeder cells and seeded in a 24-well plate at a density of 1.0×105 cells per well before transfection; 50 nM gene-specific short interfering (si)RNAs or non-targeting scrambled oligo controls (Ribobio) were transfected using lipofectamine 2000 (Invitrogen). The siRNA sequences are listed in online supplementary table S6. At 36 hours after transfection, cells were harvested to extract RNA and protein. siRNA knockdown efficiencies were validated by qRT-PCR. The PCR primers are listed in online supplementary table S7.
Immunoblotting was performed following a standard protocol described previously.29 Briefly, protein lysates were loaded onto SDS/PAGE gels and blotted onto a nitrocellulose membrane. The blot was probed with the following antibodies: anti-Gfra1 (1:500, AF560, R&D systems), anti-Bcl6b (1:500, ab87228, Abcam), anti-Plzf (1:1000, AF2944, R&D systems), anti-Stra8 (1:500, ab49602, Abcam), anti-c-Kit (1:500, ab5506, Abcam), anti-Gapdh (1:1000, AG019, Beyotime). Protein expression was visualised with an ECL plus kit (Millipore, Billerica, Massachusetts, USA).
Generation of Bcorl1 and Mtap7d3 knockout mice
CRISPR-Cas9 genome editing was used to generate Bcorl1 and Mtap7d3 Knockout mice. The sgRNAs (online supplementary table S8) were designed to target exon 7 of Mtap7d3 and Exon 3 of Bcorl1. The pUC57 expression plasmid was obtained from Addgene (Addgene 51132). pUC57-T7-sgRNA expression vectors were linearised with DraI and the Cas9 expression plasmid pST1374-NLS-Flag-linker (Addgene 44758) was linearised by AgeI. sgRNAs were purified using the MEGAclear Kit (Ambion). Cas9 mRNA was purified using the RNeasy Mini Kit (Qiagen) and transcribed using the mMACHINE T7 Ultra Kit (Ambion). Mixture of the sgRNAs (10 ng/μL each) and Cas9 mRNA (25 ng/μL) were injected into mouse zygotes to get the mutant mice. Mutant analysis of offspring was performed using PCR and sequencing. The primers for genotyping of Bcorl1 and Mtap7d3 are presented in online supplementary table S9.
Histological analysis and immunofluorescence of mouse tissues
Mouse testicular and epididymal tissues were fixed in Modified Davidson’s Fluid for 24 hours and embedded in paraffin. Sections were cut at 4.5 µm thickness and used for H&E staining. For immunofluorescence, paraffin sections were dewaxed and rehydrated. Sections were blocked in 5% bovine serum albumin for 2 hours at room temperature (RT). Then, sections were incubated with primary antibodies overnight at 4°C. After washing with PBS, secondary antibodies were added to the samples for 2 hours at RT. Antibodies used include the following: anti-PLZF 1:200 (Santa-Cruz; sc-28319), anti-SOX9 1:1000 (Millipore; AB5535), antiproliferating cell nuclear antigen (anti-PCNA) 1:200 (Proteintech; 10 205-2-AP), Alexa Fluor 555 antimouse IgG 1:500 (Life Technologies; A31570) and Alexa Fluor 488 antirabbit IgG 1:500 (Life Technologies; A21260).
qRT-PCR
Total RNAs were extracted from the testis of mice using the RNeasy Mini Kit (Qiagen) and reverse transcribed to cDNA with PrimeScript RT (Takara). qRT-PCR assays were performed using SYBR Green Master Mix (Vazyme) in Roche LightCycler 480. The qRT-PCR primers are presented in online supplementary table S10.
Computer-assisted sperm analyser
For sperm motility and counts analysis, sperms were extracted from cauda epididymis and incubated in sperm medium (GENMED) for 5 min at 37°C, and analysed using Hamilton Thorne’s-TOX IVOS system. Criteria for diagnosis of normal and abnormal spermatozoa were performed under the WHO guidelines.
Transmission electron microscopy
Sperm were fixed with 2.5% glutaraldehyde overnight at 4°C. The sperm were then embedded in Epon 812, lead citrate and uranyl acetate were used to stain the ultrathin slides. The images were taken using transmission electron microscopy (TEM) (Tecnai G2 Spirit Bio TWIN, FEI).
TUNEL assay
TUNEL assay was performed with the TUNEL BrightRed Apoptosis Detection Kit (Vazyme) according to the manufacturer’s protocol. Images were obtained with LSM800 laser scanning confocal microscope (ZEISS).
Western blot analysis
Proteins were extracted with 8 M urea lysis buffer containing 1 mM Phenylmethanesulfonyl fluoride (PMSF). The protein samples were separated in a 7.5% SDS-PAGE gel and transferred to Polyvinylidene Fluoride (PVDF) membranes. The PVDF membranes were blocked in 5% non-fat milk and primary antibodies were incubated overnight at 4°C. The PVDF membranes were incubated at RT for 2 hours with secondary antibodies after washing three times with Tris Buffered saline Tween (TBST) buffer. The signals were detected via the SuperSignal West Femto Chemiluminescent Substrate (Thermo Scientific). Antibodies used include the following: antiheat shock protein 70 (anti-HSP70) 1:1000 (Santa Cruz, sc-32239), anti-CR6-interacting factor 1 (anti-CRIF1) 1:1000 (Santa Cruz, sc-374122).
Results
X chromosome exons sequencing
Chromosome X exon capture sequencing was performed on 96 pairs of NOA and control DNA samples recruited from Nanjing Medical University affiliated hospitals. On average, each sample was sequenced to an average depth of 115×, with nearly 90% of the targeted regions covered by ≥2× (online supplementary figure S1). Sequencing data were deposited in NCBI SRA with accession number SRP072055. A total of 541 nonsense, missense or essential splice-site SNVs that were in protein-coding regions and had a minor allele frequency (MAF) of <0.01 (1%) from the variant profiles were identified across all NOA case and control subjects. In total, 453 SNVs were identified only in samples from NOA cases, 65 SNVs were identified only in controls and 23 SNVs were identified in both NOA cases and controls (online supplementary table S1). Thus, the estimated mutation frequency for NOA is higher compared with controls (online supplementary figure S2A). For the nucleotide substitutions, the mutation spectrum showed a predominance of C>T/G>A transitions (42.5%), followed by A>G/T>C transitions (21.0%) and C>G/G>C transversions (13.2%) in study subjects (online supplementary figure S2B). Most of these changes are non-synonymous SNVs, which have the potential to affect gene function.
Because of the exploratory nature of the sequencing screening, the following additional criteria were applied to identify SNVs for the next stage of validation: (a) non-synonymous SNVs or splicing; (b) predicted to be functional by SIFT and PolyPhen-2 and (c) only occurred in cases with MAF <0.01. In total, 104 SNVs, corresponding to 77 genes, were identified in the screening stage that had low MAF in the 1000 genome database were included in the validation stage (online supplementary table S2). Additionally, to verify the results from next-generation sequencing, 24 mutations (nearly 25%) were randomly selected and validated using Sanger sequencing.
Independent validation of candidate variants
Next, a two-stage validation study was conducted for 104 promising variants. Cohort I consisted of 689 NOA cases and 515 healthy male controls, the detailed distributions of these 104 selected variants are shown in online supplementary table S2. In total, 12 variants showed a significantly different distribution between the NOA group and the healthy control group (p<0.05). Among these, seven variants showed consistent associations with those identified in the discovery cohort (online supplementary table S2). For validation cohort II, an additional 548 NOA cases and 530 healthy male controls were genotyped to verify significant associations of these seven variants. Finally, ARMCX4 chrX: 100745021C>T, BCORL1 chrX: 129149363 T>G, H2BFWT chr X: rs201332803 A>G and MAP7D3 chrX: 135310906 T>C showed significant associations in the same direction as those observed in the discovery cohort and in validation cohort I (table 1). The remaining SNVs (AFF2 chrX rs142559324A>G, BCORL1 chrX rs144988023 C>T and TEX13A chrX 104464194 C>T) did not show significant associations with NOA (online supplementary table S3). These four genetic variants are reported to be with allele frequencies <0.01 according to the 1000 Genomes Project and ExAC Browser. Further evaluation revealed that two of the four SNVs were predicted to be potentially deleterious by PolyPhen-2, MutationTaster and SIFT (table 1).
Identified low-frequency variants associated with NOA risk
Next, a meta-analysis of the genotype data combining both validation cohorts were performed. All four variants reached the significance threshold (p<1.25×10−2) for NOA susceptibility (table 1). To further extend our analyses, the GTEx database was searched to see whether these four variants were quantitative trait loci (eQTL) variants. We did not find significant eQTLs in the available datasets.30
Gene expression analysis of three candidate genes in SSC
Spermatogenesis is a stem cell-dependent process, supported by self-renewal and differentiation of SSCs.31 An imbalance between self-renewal and differentiation of SSCs could induce male infertility, including NOA. We performed homology analysis for ARMCX4, BCORL1, H2BFWT and MAP7D3 and found mouse homologues for three genes except for H2BFWT. Next, we evaluated the expression levels of these three genes in mouse SSCs based on our previous RNA sequencing data. By withdrawing and replacing GDNF, which served as an irreplaceable factor for SSC self-renewal,32 we found that the mRNA levels of Bcorl1 and Mtap7d3 were changed significantly after GDNF was withdrawn, and that the mRNA level of Mtap7d3 increased significantly after GDNF was replaced (online supplementary figure S3).
Additionally, in the presence of retinoic acid (RA), an essential factor for SSC differentiation,33 to the cultured SSCs, resulted in a significant increase in the mRNA level of Armcx4 (online supplementary figure S3). Results were confirmed using qRT-PCR (online supplementary figure S3). Based on these findings, Bcorl1, Mtap7d3 and Armcx4 were selected for further functional characterisation in mouse SSCs.
Bcorl1, Mtap7d3 or Armcx4 knockdown affected the maintenance of SSC-enriched germ cells in vitro
Cultures of SSC-enriched germ cells were derived from 6-dpp LacZ-expressing Rosa donor mice (figure 1A), and the presence of SSCs was confirmed by functional transplantation assays (figure 1B). To explore whether these candidate genes (Bcorl1, Mtap7d3 and Armcx4) have important roles in the maintenance of SSC-enriched germ cells in vitro, we first knocked down these three genes individually in SSC-enriched germ cells using siRNA oligonucleotides. BLAST analysis revealed that these siRNA oligonucleotides were not homologous to any known mammalian genes other than candidate genes. The knockdown efficiency was nearly 50% (figure 1C).
Proliferation, self-renewal and differentiation analysis of SSCs in vitro after knocking down Bcorl1, Mtap7d3 or Armcx4. (A) Spermatogonial stem cells (SSC) clumps established from 6-dpp LacZ-expressing Rosa donor mice were incubated with X-Gal. (B) Colonies of donor-derived spermatogenesis were detectable in recipient testes 2 months post-transplantation. Each blue colony derives from a single SSC. (C) Relative expression levels of Bcorl1, Mtap7d3 and Armcx4 in SSC-enriched germ cells after transfection with their short interfering (si)RNAs, respectively. (D) SSC-enriched germ cells number at day 3, 5 and 7 after transfection with Bcorl1, Mtap7d3 and Armcx4 siRNA. (E) Relative mRNA levels of SSC self-renewal (Gfra1, Bcl6b and Plzf) and differentiation (Stra8, c-Kit) markers after knocking down Bcorl1, Mtap7d3 or Armcx4. (F–G) Immunoblot analysis of SSC self-renewal and differentiation markers after knocking down Bcorl1, Mtap7d3 or Armcx4. Error bars are mean±SEM. *P<0.05, **p<0.01, ***p<0.001.
Next, their effects on the maintenance of SSC-enriched germ cells was evaluated and it was found that the total number of germ cells decreased progressively at day 3, 5 and 7 after silencing of Mtap7d3. After silencing Bcorl1, the total number of germ cells was significantly decreased at day 7 compared with control (figure 1D). Unlike Mtap7d3 and Bcorl1, silencing of Armcx4 increased the total germ cell numbers (figure 1D). These data suggested that knocking down Bcorl1, Mtap7d3 and Armcx4 could affect the maintenance of SSC-enriched germ cells in vitro.
The expression level of SSC self-renewal and differentiation markers changes after Bcorl1, Mtap7d3 or Armcx4 knockdown
We examined the effects of silencing Bcorl1, Mtap7d3 or Armcx4 on SSC self-renewal and differentiation by examining the expression of SSC self-renewal (Gfra1, Bcl6b and Plzf) and differentiation (Stra8 and c-Kit) genes using qRT-PCR and immunoblots. As shown in figure 1E, the Gfra1 mRNA level was decreased in SSC-enriched germ cells after a reduction of Mtap7d3, while the Stra8 was decreased after a reduction of Armcx4. Immunoblot analysis revealed that the protein levels of GFRA1 was downregulated in the Bcorl1 and Mtap7d3 knockdown cells, whereas STRA8 was downregulated in the Armcx4 knockdown cells compared with control (figure 1F–G). In addition, the downregulation of Mtap7d3 resulted in decreased levels of BCL6B (figure 1F–G). These findings suggest that silencing of Mtap7d3 or Bcorl1 may inhibit SSC self-renewal while silencing of Armcx4 may inhibit SSCs differentiation.
Male Bcorl1 knockout mice are infertile
Given that silencing of Bcorl1 and Mtap7d3 affected SSC function in vitro and both genes are highly expressed in mouse testis (online supplementary figure S4A and S5A), we generated Bcorl1 and Mtap7d3 knockout mice via CRISPR-Cas9 technology to study their roles in spermatogenesis in vivo. Two guide RNAs (gRNAs) were designed to target exon 7 of Mtap7d3 and exon 3 of Bcorl1, respectively. We injected both gRNAs and Cas9 mRNA into C57BL/6J zygotes, and obtained the first-generation Bcorl1 mutant mice with frameshift mutation (22 bp deletion) (figure 2A) and Mtap7d3 mutant mice with frameshift mutation (1 bp deletion) (online supplementary figure S5B), which was confirmed by sequencing and qRT-PCR (online supplementary figure S4B and S5C).
Bcorl1-deficient mice display impaired spermatogenesis and sterility. (A) Schematic diagram of sgRNAs targeting at Bcorl1 locus and sequences of modified Bcorl1 loci detected in founders. PAM sequences are highlighted in red. sgRNAs targeting sequences are in green. The mutations in blue. (B) All Bcorl1-/Y male mice were infertile. (C) Testis/body weight ratio of wild-type (WT) and Bcorl1-/Y mice. (D) Computer-assisted sperm analyser analysis of sperm quality in WT and Bcorl1-/Y mice. Sperm counts (E), motility (F) and progressivity (G) were decreased in Bcorl1-/Y mice. (H) H&E of testes from WT and Bcorl1-/Y mice. Scale bar, 100 and 50 µm. (I) H&E staining of epididymides. Scale bar, 100 µm. (J) H&E staining of sperms. Scale bar, 50 µm. Error bars are mean±SEM. *P<0.05, **p<0.01, ****P<0.0001.
Heterozygous female mice were intercrossed with wild-type (WT) males to obtain male Bcorl1 and Mtap7d3 knockout mice. Strikingly, all male Bcorl1 −/Y mice were sterile (figure 2B), while Mtap7d3−/Y males exhibited normal fertility (online supplementary figure S5D). Bcorl1-/Y males exhibited significantly smaller testes, ~70% of WT by weight (figure 2C). Computer-assisted sperm analyser was used to examine sperm quality, which showed great reductions in both sperm counts and motility in Bcorl1 males (figure 2D–G, online supplementary movies S1 and S2). In comparison to WT mice, no significant differences were found in both testes weight and sperm quality in Mtap7d3-/Y males (online supplementary figure S5E-G).
Supplementary video
Supplementary video
H&E staining showed no obvious histopathological differences in germ cell types and cellularity between Bcorl1-/Y and WT male testes (figure 2H), and spermatozoa morphology appeared normal in Bcorl1-/Y mice (figure 2J). However, H&E stained epididymis sections presented drastically reduced spermatozoa in male Bcorl1-/Y mice(figure 2I).
In order to uncover which step of spermatogenesis was affected, immunofluorescence staining of PLZF and SOX9 were performed. No significant differences of PLZF and SOX9 positive cells was observed between WT and Bcorl1-/Y mice (online supplementary figure S6A-C). Next, the PCNA, a key factor in the proliferation of mouse spermatogonia, was detected by immunofluorescence staining.34 The results showed that the number of PCNA-positive spermatogonia cells was significantly reduced in Bcorl1-/Y males (figure 3A,B). Interestingly, counting spermatocytes and spermatids at different stages revealed that the numbers of spermatocytes, round spermatids and elongating spermatids were significantly reduced in Bcorl1-/Y males (online supplementary figure S7A,B). There was no difference in germ cell apoptosis in testes between Bcorl1-/Y and WT mice (online supplementary figure S8).
Effect of Bcorl1 deficiency on spermatogonia proliferation. (A) Immunofluorescence staining for proliferating cell nuclear antigen (PCNA) in wild-type (WT) and Bcorl1-/Y mice. (B) PCNA-positive cells per tube. Error bars are mean±SEM. *P<0.05.
Bcorl1 deficiency induces abnormal mitochondrial sheath
Mitochondria are known to produce energy for sustaining sperm motility.35 Since impaired sperm motility was observed in Bcorl1-/Y mice, we hypothesised that mitochondrial function was compromised in Bcorl1 knockout mice. The ultrastructure of sperm cells from Bcorl1 knockout and WT mice was studied by TEM. The central pair microtubules, outer dense fibres and peripheral microtubule doublets did not appear to be affected (figure 4A). However, the mitochondrial sheath structures of sperm from Bcorl1-/Y mice were abnormal with the presence of large vacuoles (figure 4A). Western blot analysis confirmed reduced levels of the mitochondrial markers CRIF1 and HSP70 in Bcorl1-/Y mice (figure 4B) which further indicates deficient mitochondrial function in Bcorl1-/Y mice.
Bcorl1 is required for sperm mitochondria. (A) Transmission electronmicroscopy (TEM) displayed that the structures of mitochondrial sheath were abnormal and presented vacuoles in Bcorl1-/Y mice (red arrow). (B) Western blot analysis of CR6-interacting factor 1 (CRIF1) and heat shock protein 70 (HSP70) from wild-type (WT) and Bcorl1-/Y mice testes.
Collectively, these data suggest that our findings strongly suggest that Bcorl1 plays a critical role in male infertility.
Discussion
The aetiology of NOA is still largely unknown, and studies have suggested that rare or de novo mutations may be involved. Because X chromosome is enriched for spermatogenesis genes and, it is hemizygous in males, mutations in X linked genes could display direct effects on sperm development and be difficult to detect in a human population.36 To date, a number of mutations in X linked genes have been identified to be with spermatogenesis failure and male infertility,37 38 but no systematic study has been conducted to evaluate the X-linked mutations attributed in NOA.
Given that the X linked genes are predominantly expressed in both premeiotic and postmeiotic spermatogenic cells, one might anticipate that loss-of-function mutations in X linked genes would perturb male gametogenesis.2 12 In the present study, X chromosome exons sequencing was carried out in a case-control cohort and validated these findings in two independent cohorts. Four single nucleotide variants on X chromosome, corresponding to the genes BCORL1, MAP7D3, ARMCX4 and H2BFWT, were observed at a higher rate in individuals with NOA.
MAP7D3 belongs to the MAP7 family, which regulates microtubule assembly and stability via its interaction with tubulin and microtubules and has been reported to promote breast cancer growth and metastasis.39 40 Since GDNF has been regarded as the primary regulator of SSC self-renewal,41 the significant change of Mtap7d3 expression levels resulting from withdrawal and replacement of the growth factor suggests Mtap7d3 as an important GDNF pathway regulator of SSC self-renewal. Knockdown of Mtap7d3 downregulated the levels of GFRA1 and BCL6B, two GDNF-dependent regulators of SSC self-renewal, further confirming its important role in the GDNF pathway. However, no significant changes in the testis weight or sperm number in the Mtap7d3-/Y male mice was found. The only observed phenotype in Mtap7d3-/Y male mice was a slight decrease in sperm motility. Given that in vitro, the total number of SSC-enriched germ cells decreased progressively after silencing of Mtap7d3, we hypothesised that Mtap7d3-/Y may affect the function of SSC over time thus leading to male infertility with increasing age. However, there was no observed difference in fertility between Mtap7d3-/Y and WT mice over time (online supplementary figure S9). This might be due to the redundancy of paralogs in Map7 gene family or compensation by other genes during early spermatogenesis.
ARMCX4 belongs to the armadillo repeat-containing family of proteins, which interact with other proteins in diverse cellular processes.42 For example, it has been reported to regulate mitochondrial trafficking through interaction with Miro and Trak2 in neurons.43 However, very little is known about the function of this gene in the reproductive system. In our study, we found that low level of Armcx4 inhibited SSC differentiation activity through downregulating Stra8, a differentiation marker of SSCs and, in turn, promoted proliferation of the SSCs-enriched germ cells by inducing the increase of the total number of SSCs. Considering the low expression level of Armcx4 in adult mouse testis,44 we did not validate Armcx4’s role in spermatogenesis in vivo.
Through the use of both in vivo and in vitro models, our study demonstrates that BCORL1 contributes to spermatogenesis. BCORL1 is a transcriptional corepressor that is found tethered to promoter regions by DNA-binding proteins. This protein can interact with several class II histone deacetylases to repress transcription.45 BCORL1 is highly expressed in human testis. However, previous studies indicated that mutations in this gene were associated with various diseases including chronic myeloproliferative neoplasms, adult acute myelogenous leukaemia and breast cancer.46–48 It had not been previously linked to male infertility. This study demonstrates, for the first time, that loss of Bcorl1 causes spermatogenesis failure. But the phenotypes were different between humans (azoospermia) and mice (reduced sperm count and motility). To explore these phenotypic differences, RGV database (https://rgv.genouest.org/) was used to analyse the expression of BCORL1 in spermatogenesis between human and mice. And we found the different expression patterns of BCORL1 during human and mouse spermatogenesis (online supplementary figure S10). In human, BCORL1 is mainly expressed in type A pale spermatogonia, leptotene and zygotene spermatocytes. However, Bcorl1 is mainly expressed in primitive type A spermatogonia, round spermatids and elongated spermatids in mice. This result can well explain the phenotypic differences between human and mice.
In our study, we also found that reduced expression of Bcorl1 could inhibit the proliferation of SSC-enriched germ cells and downregulate GFRA1. Spermatogenesis depends on the normal proliferation of spermatogonia.49 We observed that knockout of Bcorl1 impaired spermatogonia proliferation. In addition, we unexpectedly found the abnormal brain development in Bcorl1-/Y mouse (data not shown). And previous studies also found that mutations of BCORL1 were associated with neurodevelopmental disorder and intellectual disability.50 51 In the further study, we will focus on the mechanisms underlying the phenotype.
BCORL1 was reported to be a core component of polycomb repression complex 1 (PRC1), which was recruited by KDM2B to bind to DNA. Without BCORL1, the PRC1 complex cannot form and interact with KDM2B to regulate gene expression.52 Other studies have demonstrated that inactivation of PRC1 impairs the expression of Sall4, which codes for a transcription factor essential for spermatogenic differentiation and proliferation,53 and that during spermatogenesis, loss of function of PRC1 resulted in the gradual loss of spermatogonia stem cells and severe proliferation and differentiation defects. These phenotypes mirror the phenotypes observed in this study, suggesting further research is needed to understand how Bcorl1 is involved in SSC differentiation.
Lastly, abnormal mitochondrial sheath structure was observed in sperm cells of Bcorl1-/Y mice. The normal structural and functional mitochondria sheath plays an important role in energy maintenance of sperm movement.54 55 CR6-interacting factor 1 is an essential mitoribosome factor involved in biosynthesis of mitochondrial oxidative phosphorylation (OXPHOS) subunits and plays important role in the insertion of OXPHOS polypeptides into the mitochondrial membrane in mammals.56 57 HSP70 is essential for the homeostasis and function of mitochondria.58 The reduced expression levels of CRIF1 and HSP70 in Bcorl1-/Y mice suggest that Bcorl1 plays an important role in the maintenance of mitochondrial sheath integrity by regulating CRIF1 and HSP70. Together these data suggest that Bcrol1 plays multiple rolls in spermatogenesis, resulting in a complex molecular phenotype of infertility.
In summary, this study has demonstrated that mutations in X linked genes likely contribute to male infertility. Using a multisystem approach, we identified three X linked genes in spermatogenesis. BCORL1, MAP7D3 and ARMCX4 all showed significant association with NOA. In addition, these findings provide a possible explanation for how X linked genes affect spermatogenesis, opening new avenues to identify novel genes for male infertility. These findings also provide new perspective from the view of X linked inheritance to better understand the aetiology of male infertility and to develop strategies for targeted therapies.
Acknowledgments
The authors would like to thank all the study participants, research staff and students who took part in this work. The authors would also like to thank Dr John D. Roberts (NIEHS) for critically reading the manuscript.
References
Footnotes
CL, YZ and YQ are joint first authors.
CL, YZ and YQ contributed equally.
Contributors Yankai Xia and Xinru Wang directed the study, obtained financial support and were responsible for study design. Chuncheng Lu performed overall project management along with Yan Zhang and Yufeng Qin. and drafted the initial manuscript. Qiaoqiao Xu, Yiqiang Cui, Jian-hua Mao and Xuejiang Guo performed statistical analysis. Min Wang and Jintao Zhang were responsible for sample processing and managed the genotyping data. Yunfei Zhu and Xin Zhang were responsible for subject recruitment and sample preparation. Xin Wu, Ran Zhou and Xiang Wei were responsible for functional analysis in SSCs. Mingxi Liu, Bo Hang, Zhibin Hu, Hongbing Shen, Antoine M Snijders and Zuomin Zhou conceived of the study, and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript.
Funding This work was supported by the National Key R&D Program of China (2019YFC1005100), the National Natural Science Foundation of China (81471500, 81630085 and 81671461) and the Priority Academic Programme for the Development of Jiangsu Higher Education Institutions (Public Health and Preventive Medicine).
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval This study was approved by the institutional review board of Nanjing Medical University, China (FWA00001501), and conducted according to the Declaration of Helsinki.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available in a public, open access repository. Sequencing data were deposited in NCBI SRA with accession number SRP072055.