Background Azoospermia is diagnosed when sperm cells are completely absent in the ejaculate even after centrifugation. It is identified in approximately 1% of all men and in 10%–20% of infertile males. Non-obstructive azoospermia (NOA) is characterised by the absence of sperm due to either a Sertoli cell-only pattern, maturation arrest, hypospermatogenesis or mixed patterns. NOA is a severe form of male infertility, with limited treatment options and low fertility success rates. In the majority of patients, the cause for NOA is not known and mutations in only a few genes were shown to be causative.
Aim We investigated the cause of maturation arrest in five azoospermic infertile men of a large consanguineous Bedouin family.
Methods and results Using whole genome genotyping and exome sequencing we identified a 4 bp deletion frameshift mutation in TDRD9 as the causative mutation with a Lod Score of 3.42. We demonstrate that the mutation results in a frameshift as well as exon skipping. Immunofluorescent staining with anti-TDRD9 antibody directed towards the N terminus demonstrated the presence of the protein in testicular biopsies of patients with an intracellular distribution comparable to a control biopsy. The mutation does not cause female infertility.
Conclusion This is the first report of a recessive deleterious mutation in TDRD9 in humans. The clinical phenotype recapitulates that observed in the Tdrd9 knockout mice where this gene was demonstrated to participate in long interspersed element-1 retrotransposon silencing. If this function is preserved in human, our data underscore the importance of maintaining DNA stability in the human male germ line.
- germ line DNA stability
Statistics from Altmetric.com
Infertility is usually defined as a failure of conception after 12 months of unprotected intercourse1 and male infertility accounts for 30%–55% of infertile couples. Azoospermia is diagnosed when sperm is completely absent in the ejaculation even after centrifugation. It is identified in approximately 1% of all men and in 10%–20% of infertile males.2 3 Azoospermia is divided into two major groups: (1) obstructive azoospermia, in which the process of spermatogenesis is active but genital tract obstruction blocks the transport of sperm. (2) Non-obstructive azoospermia (NOA) accounting for approximately 60% of men with azoospermia, in which spermatogenesis is inactive and thus sperm cells are not generated. The pathophysiology of NOA is characterised by the absence of sperm cells due to a Sertoli cell-only histology pattern, where only Sertoli cells line the seminiferous tubules (Sertoli cell-only syndrome (SCOS), maturation arrest, where spermatogenesis is not complete and may be arrested at one or more levels of spermatogenesis; hypospermatogenesis, where complete spermatogenesis is present but only few mature sperm cells can be observed or mixed patterns.4 NOA is a severe form of male infertility, with limited treatment options and low fertility success rates.2
Genetic factors contribute to 21%–29% of NOA,3–6 approximately 28% of azoospermic men presenting with genetic alterations can be diagnosed, allowing for adequate counselling before proceeding to assisted reproductive technology.7 However, 12%–41% of NOA cases are idiopathic, most likely reflecting unknown genetic or other factors. Currently the three most relevant genetic causes of azoospermia are: mutations in the cystic fibrosis (CFTR) gene (associated with congenital bilateral absence of the vas deferens and obstructive azoospermia), chromosomal abnormalities (such as Klinefelter syndrome) and microdeletions on chromosome Y, causing NOA.8 Sixteen autosomal spermatogenic disorders (SPGFs) are presently reported in the OMIM. Only three of them represent azoospermia with complete absence of sperm in ejaculation and only two mutated autosomal genes encoding proteins with different functions were reported in the OMIM database. SPGF1 is an autosomal recessive form of spermatogenic failure associated with defects in meiosis. The gene causing SPGF1 is not yet known. SPGF14 (MIM 615842) is caused by a heterozygous frameshift mutation in the ZMYND15 gene (MIM 614312). ZMYND15 is suggested to be a histone deacetylase-dependent transcriptional repressor that controls normal temporal expression of haploid genes during spermiogenesis.9 SPGF15 (MIM 616950) is caused by a homozygous splice site mutation in the SYCE1 gene (MIM 611486).10 The SYCE1 gene encodes a protein that is a member of the synaptonemal complex, a tripartite structure that physically links homologous chromosomes during prophase I.11 Meiotic arrest and azoospermia were additionally reported to be caused by mutations in the X-linked TEX11 gene,12 13 accounting for 1%13 to 2.4%12 of azoospermia in men and 15% in men diagnosed with azoospermia with meiotic arrest.12 TEX11 regulates homologous chromosome synapsis and double-strand DNA break repair and thus, it is critical for synaptonemal complex formation and the chiasma in chromosomal crossover. The importance of proteins that participate in meiosis has also been demonstrated by the identification of mutations in SYCP3 that encodes a component of the synaptonemal complex.14 Heterozygous mutations in SOHLH1, a gene encoding a testis specific transcription factor essential for spermatogenesis cause a Sertoli cell-only pattern.15 A mutation in the NR5A1 gene encoding the steroidogenic factor 1 was found in one Pakistani patient with azoospermia with complete bilateral meiotic arrest and normal values of follicular-stimulating hormone (FSH), luteinising hormone (LH) and testosterone.16 In a patient with the spermatogenesis process mainly blocked at the spermatocyte stage, a dominant negative mutation was identified on HSF2 encoding the heat shock transcription factor 2.17 Finally, mutations in PRM1 and PRM2 that encode for the protamines that condense DNA in the sperm cause severe oligozoospermia.18
In the present study, we demonstrate that NOA can be caused by a mutation in TDRD9 that is essential for silencing long interspersed element (LINE)-1 retrotransposon in the mouse male germ line.19 Furthermore, we have shown that this mutation does not affect female fertility. The importance of maintaining germ line DNA stability through the repression of autonomous DNA elements has not been demonstrated previously in human male fertility.
Patients and methods
We have recruited a Bedouin extended family with a high proportion of consanguineous marriages. The study was approved by the Soroka Medical Center institutional review board, and all participants have signed a written informed consent prior to participation. Among the five patients with the TDRD9 mutation, one individual was diagnosed as having cryptozoospermia (six immotile sperm with small head found after centrifugation in one of the semen samples) and four patients had azoospermia (no sperm found at any of the semen analyses). In addition, two out of the five patients were born with cryptorchidism and have had orchiopexy (testicles retained in the inguinal canal and surgically replaced to the scrotum in childhood). Two other patients also had type 1 diabetes mellitus. All of them have had endocrine studies including FSH, LH, prolactin and testosterone—all the results in the normal range. No hypergonadotropic or hypogonadotropic hormone levels have been detected. Physical examination of the patients and ultrasound of the scrotal area has demonstrated normal size testicles and no significant varicocele. One patient had a large hydrocele which was treated surgically. The surgical technique of testicular sperm extraction (TESE) in patients with azoospermia is described elsewhere.20 Briefly, after stabilisation of the testicle, a small incision in the testicles’ mid-portion was performed, cutting through the skin, tunica vaginalis and tunica albuginea. A piece of testicular tissue was cut with small scissors and placed in a Petri dish containing ‘Follicle Flush Medium’ (Sydney IVF Fertilization Medium, Cook Medical). The testicular tissue was fragmented, minced and examined fresh under the microscope for the presence of spermatozoa in a wet preparation. Once spermatozoa were found, the surgical procedure was terminated. If no spermatozoa were observed, several specimens of tissue were extracted from different sites of the testicle as well as from the contralateral testicle and the procedure terminated. In addition, all patients except one (IV-2) had a karyotype and a chromosome Y-deletions investigated. The karyotypes were all 46xy, while two out of five patients had an inversion on chromosome 9. There were no Y chromosome microdeletions detected (table 1).
Testicular tissues staining
Testicular biopsies were fixed in 4% paraformaldehyde (Sigma) and paraffin embedded. Sections of 5 µm were placed on superfrost plus slides (Thermo, Braunschweig, Germany) for H&E staining.
Immunofluorescence staining of testicular biopsies
Testicular sections of 5 µm were placed on superfrost plus slides for immunofluorescence staining of TDRD9 protein as previously described.21 After washing, non-specific adhesions sites in the tissues were blocked by 5% normal donkey serum (Biological Industries) for 30 min at room temperature. After removing the blocking buffer, the first antibody of rabbit polyclonal to TDRD9 (Abcam, Cambridge, UK; ab118427) and/or polyclonal goat anti-acrosin (c-46284, Santa Cruz Biotechnology, Santa Cruz, California, USA) and/or MAGE-A (6C1) (sc-20034, Santa Cruz Biotechnology), which is a mouse monoclonal antibody, were added. After overnight incubation at 4°C, the slides were washed and the relevant secondary antibodies were added for 40 min at room temperature: donkey anti-rabbit IgG (Cy3), Jackson Immuno Research (West Grove, PA, USA) donkey anti-goat IgG (DyLight 488), donkey antirabbit IgG (DyLight 488) (Bethyl Laboratories, Montgomery, Texas, USA). After washing, the slides were dried and 4′,6-diamidine-2′-phenylindole dihydrochloride (DAPI) (Santa Cruz Biotechnology) was added to the tissues and the slides were stuck with cover slides. The negative control was incubated with normal rabbit IgG (figure 4A), Jackson Immuno Research, instead of the first antibody for TDRD9. Slides were examined for staining using Nikon eclipse 50 i microscope (Tokyo, Japan).
DNA was extracted from all individuals of generations IV and V and III 1–4 (figure 2A). Lymphoblastoid cell line was established for IV-1. Genotyping was performed on all parents of the affected patients except III-1 and all the affected except IV-1, using Affymetrix GeneChip Human Mapping SNP_5 arrays. Dedicated software (KinSNP)22 was used to automatically search the microarray results for homozygous regions consistent with linkage. Exome sequencing was performed on the DNA of one affected make (IV-2) as detailed in Reish et al.23
Verification of the mutation
PCR amplification of exon 5 of the TDRD9 gene (NM_153046) was performed using the forward: CAACAGAGGACACCAGGCTA and reverse: GCACAAAGACAGGGTTCACA (annealing temperature 62°C) primers. Direct sequencing of the PCR products was performed after cleaning the products using ExoSap (ThermoScientific) on an ABI PRISM 3100 DNA Analyzer with the BigDye Terminator v.1.1 cycle sequencing kit (Applied Biosystems, Carlsbad, California, USA) according to the manufacturer’s protocol. The PCR for screening of the population was done using the forward: TTTATATGACAACTGGAGTCC and reverse: GATATGTGTGAATTCCATCA (annealing temperature 54°C) primers. The products were separated on a 12% polyacrylamide gel.
Analysis of splicing
RNA was extracted from lymphoblastoid cells of patient IV-1 and from lymphoblastoid cells of a male fertile individual control, using the Direct Zol RNA miniprep (Zotal). Five micrograms of RNA were used to prepare cDNA by the SuperScript Kit (Invitrogen). PCR was performed on the cDNA using primers in exon 4 forward: GACCCTGGGAGGTGTGGT and exon 7 reverse: TGATGAATGTGCTCCAAATCA that contain the mutation on exon 5. The PCR products were separated on 2% of agarose gel, extracted from the gel using Zymoclean Gel DNA Recovery Kit (Zymogene) and sequenced as above.
We have recruited a Bedouin extended family (figure 2A) with a high proportion of consanguineous marriages. All patients, except one who was defined as cryptozoospermia (IV-5), were diagnosed with azoospermia primarily due to consecutive semen analyses with no sperm cells detected even after centrifugation. Serum gonadotropin values were found to be normal (table 1) distinguishing these patients from the most common forms of non-obstructive azoospermia. Three out of five patients have undergone testicular sperm extraction (TESE) and the testicular tissue samples were searched for sperm cells by embryologists at the in vitro fertilisation (IVF) laboratory. In addition, a sample of testicular tissue was evaluated in the pathology laboratory. The IVF laboratory search has observed a few immotile sperm cells with small head in the ejaculation of the patient defined with cryptozoospermia (IV-5). After TESE, another four sperm cells were extracted. Ten oocytes were injected with sperm cells found in the ejaculation and in the TESE, but no fertilisation was observed. In the other two TESE patients, no mature sperm was detected in the IVF laboratory, and the histological samples have shown spermatogenic arrest (table 1).
Histology of testicular biopsies
Histological section of testicular biopsies of patients IV-3, and V-2 confirmed incomplete maturation arrest in testicular biopsies (figure 1). Some seminiferous tubules in both patients showed SCOS histology. Immunofluorescence staining for spermatogonial cells (using specific antibodies for MAGE-A and VASA—specific markers for spermatogonial cells) and meiotic cells (using specific antibodies for acrosin—a specific marker for meiotic/postmeiotic cells) confirmed the presence of these cells and the histology of both patients as incomplete maturation arrest (figure 1B,C).
Identification of the TDRD9 mutation and its effect on splicing and protein
Assuming homozygosity by descent of a recessive mutation as the likely cause of the disorder, we genotyped four of the patients and five of the parents as presented in figure 2A. We identified five shared homozygous regions larger than 2 cM, encompassing a total of 13.8 Mbp on the autosomal chromosomes. (X-linked inheritance can be excluded since the mutation is transmitted through two fertile brothers). Exome sequencing was performed on patient IV-2. Only one homozygous variant with allele frequencies of less than 1% in the public databases (ExAc browser, 1000 Genomes and dbSNP) was identified in the shared homozygous autosomal chromosomal regions. This variant is on chromosome 14: 104433122, del TAGT (GRCh37/hg19), in coding exon 5 of the Tudor domain containing protein 9 (TDRD9) gene, c.720_723 del TAGT causing p.Ser241Pro fs Ter 4 close to the N-terminus of the 1382-amino acid long protein. We validated this result by Sanger sequencing and verified the segregation of the variation in all available family members. All the patients with azoospermia were homozygous for the mutation, while the rest of the healthy individuals had the normal or heterozygous alleles (figure 2A,B). The calculated Lod score for the segregation using the superlink programme (http://cbl-hap.cs.technion.ac.il/superlink-snp/), assuming recessive inheritance with 99% penetrance and 0.0016 for the frequency of the mutation, was 3.42. The mutation does not affect the fertility of females since female IV-9 who is homozygous for the mutation is a mother of seven children and did not have any fertility medical intervention.
The variation was not present in our collection of Bedouin exomes of 202 individuals. To increase the number of Bedouin controls for the verification of the prevalence of the mutation, the DNA of 31 additional fertile Bedouin males was PCR amplified and the variation was visualised by separation of the PCR products on 12% polyacrylamide gels (figure 2C), none of these males had the variation. This variation was also not reported in a database of 77 Bedouins,24 thus its prevalence in the Bedouin population is less than 1/620.
The variation causes p.Ser241Pro fs Ter 4. The deletion is within the DEXDc domain and is predicted to result in the loss of the next 1141 highly conserved amino acids of the full-length protein (figure 3C). The frameshift mutation in exon 5, which is an inner exon of the gene, may lead to either nonsense-mediated decay of the mRNA25 or to rescue of translation by exon skipping as we have previously demonstrated for another gene.26 Thus, we verified the effect of the gene variant at the mRNA level. RNA was extracted from lymphoblastoid cells of patient IV-1 and from comparable cells of a control individual and cDNA was prepared. PCR was performed on the cDNA using primers in exons 4 and 7, flanking exon 5 that contains the mutation. The RT-PCR products of the patient-derived cells demonstrate two fragments in contrast to the one expected fragment of normal splicing shown in control cells (figure 3A). Sanger sequencing of the PCR products demonstrated that the large fragment of the patient (386 bp) contains exon 5 with the deleted TAGT and the smaller fragment of 305 bp that appears only in patient cells reflects skipping of exon 6 (figure 3B). Conventional splicing leading to the large fragment would produce the truncated protein p.Ser241Pro fs Ter 4, and the mutation-induced splicing with the skipping of exon 6 does not rescue the reading frame of translation and results in a protein missing all the known functional domains.
We have verified whether the TDRD9 protein is produced in patients’ testes using a commercial TDRD9 antibody raised to a peptide within the 50–150 amino acids at the amino terminal of Human TDRD9. Indeed, in accordance to the presence of transcription of TDRD9 detected in lymphoblastoid cells, we have detected the TDRD9 protein in testes biopsies of patients V-2 and IV-3 in spermatogonial cells (mainly in the cytoplasm) and in spermatocysts/round spermatids (mainly in the nucleus) (figure 4A). The intracellular distribution was comparable to the distribution in a biopsy with complete spermatogenesis (control testis; CT) and thus is not affected by the mutation. To ascertain that the TDRD9 is expressed in spermatogonial cells and in spermatocytes/round spermatids we have costained a normal control testis with TDRD9 and MAGE-A that shows spermatogonial cells (figure 4B) and acrosin that shows meiotic/postmeiotic cells (figure 4C), both showed colocalisation.
About a half of infertility cases are related to a male factor and azoospermia in particular occurs in a 5% of infertile men. The aetiology of NOA is mostly idiopathic and only a minority of patients is found to be carriers of a defective karyotype or a Y-chromosome deletion. Therefore, a continuous search aimed to discover the underlying aetiologies that induce azoospermia is of paramount importance.
In this study, we demonstrate for the first time the association between NOA and a mutation in a TDRD gene. The frameshift mutation in exon 5 of the TDRD9 gene: c.720-_723 del TAGT is predicted to result in a truncated TDRD9 protein: Ser241Pro fs Ter 4 missing all the known functional domains. The mutation segregates as expected in the family members, with a Lod score of 3.42 and was not detected in 620 chromosomes of Bedouin controls.
Tdrd9 was shown to be highly expressed in germ cells in male and female mice, and by immunohistology it was detected in mitotic spermatogonia, meiotic spermatocytes (predominantly at the pachytene stage) and haploid spermatids in the testis.19 Although Tdrd9 is also expressed in female germ cells, the fertility of the females is not affected, similarly to the mouse, where it was demonstrated that there were no defects in meiotic progression and or cellular structure of oocytes of females without Tdrd9.19 Our immunofluorescence staining for TDRD9 showed results similar to those found in the mouse testis; TDRD9 was found in spermatogonial cells, spermatocytes and round spermatids in testicular biopsies of the patients and control. This may suggest a similar role for TDRD9 in human and mouse testes. The normal function of TDRD9 was demonstrated to be essential for silencing of Line-1 retrotransposon in the male mouse germ line.19 Line-1 is the most abundant class of autonomous retrotransposons. One mechanism to control them is by RNA interference and its related systems, which use small non-coding RNAs that guide the effector complex, including argonaute proteins, to degrade and/or suppress target mRNAs. Germ cells are equipped with specific members of the argonaute subfamily, the piwi proteins, which interact with piwi-interacting small RNAs (piRNAs). The piRNAs are complementary to either sense or antisense transposon transcripts. Sense strand piRNAs are preferentially associated with piwi like RNA-mediated gene silencing 2 (MILI), whereas antisense piRNAs interact with piwi like RNA-mediated gene silencing 4 (MIWI2), two mouse piwi proteins (reviewed in Ref. 27). It was suggested that the Tudor domains of TDRD proteins interact with the piwi proteins MILI and MIWI2, and their concerted action is essential for retrotransposon control and DNA methylation in the germ line.27 TDRD9 is essential for the function of MIWI2, its absence causes male mice infertility accompanied by L1 derepression, accumulation of L1 viral particles and massive damage of genomic DNA in germ cells.19 TDRD9 protein has in addition to the Tudor domain several domains whose function has not been directly studied, from the amino terminus to the carboxy terminus: DEXDc, HELICc, HA2 and the Tudor. The DEXDc domain is a member of the DEAD-like helicase, a diverse super family of helicases that use ATP hydrolysis to break DNA and RNA.28 HELICc is a 102 amino acids domain, of the helicase superfamily c-terminal domain, found in proteins belonging to the helicase superfamilies 1 and 2. HA2 is a Helicase associated domain of about 101 amino acids that is found in a diverse set of RNA helicases, its function is unknown; however, it seems likely to be involved in nucleic acid binding. The presence of these domains in TDRD9 suggests that it may have additional role/s directly on the piRNA or their interaction with L1 in addition to the action on MIWI2. The frameshift mutation in our patients is in the DEXDc which is the most amino terminal, and thus all other domains will be absent in the mutated protein.
Uncontrolled retrotransposon activity results in genomic instability and germ cell death. This was demonstrated in mice deleted for genes involved in transposon silencing, for example, the putative DExD-box helicase MOV10-like-1 Moloney leukaemic virus 10-like 1 (MOV10L1) that interacts with piwi proteins,29 MILI,30 Dnmt3L that might have a function in the de novo methylation of dispersed repeated sequences.31
Tdrd proteins that are expressed in germ cells are divided into two groups. One group, consisting of TDRD1, TDRKH, TDRD9 and TDRD12, functions in piRNA biogenesis and retrotransposon silencing, while the other group including RNF17/TDRD4 and TDRD5-7 are required for spermiogenesis. These Tdrd proteins play distinct roles during male germ cell development. Disruption of STK31/TDRD8 does not affect fertility of either sex in mice. TDRD5 (also called meiosis arrest female 1, MARF1) is essential for oogenesis and the development of healthy offspring.32 TDRD proteins were demonstrated to be essential for male mice fertility including Tdrd1,33 Tdrd12,34 Tdrkh35 and Tdrd5.36 Only recently was a TDRD1 polymorphic variant in the 5’-UTR of exon 1 reported to associate with spermatogenic failure susceptibility in a dominant model in the Han Chines.37 In this study, a genotyping analyses for five SNPs in four piRNA pathway genes was carried out in a case–control study including 342 cases and 493 controls. Two SNPs on the TDRD9 gene in 3’UTR in exon 35 were tested as well but not found to confer susceptibility. Another study was carried in cryptorchid boys with impaired mini-puberty that develop infertility despite timely and successful surgical treatment. A genome-wide RNA expression analysis of 18 cryptorchid and 4 control testes found impaired expression of the genes important for transposome silencing: DDX4, MAEL, MOV10L1, PIWIL2, PIWIL4 and TDRD9 genes in the group of cryptorchid boys at high risk of infertility.38 In relation to this report, it may be of significance that two of our patients also had cryptorchidism. Finally, mutations in TDRD7 in human cause both paediatric cataract and glaucoma and in mice in addition to cataract and glaucoma also male infertility.39
Our study is the first to demonstrate the importance of the gene TDRD9 known to be involved in silencing of Line-1 retrotransposon in the male germ line for enabling fertility. This is also the first report of the participation of a member of the TDRD protein family in male fertility in humans. If the function of TDRD9 reported in mice is identical in human, the importance of maintaining germ cell DNA stability joins the other processes important for male fertility to proceed the spermatogenic maturation process that were identified through genes mutated in humans: meiosis (TEX11, SYCE1, SYCP3) and transcription factors (ZMYND15, SOHLH1, HSF2). This information is also important for the expected results of TESE and further consultation of the family.
We are grateful to the affected persons and their families, whose cooperation made this study possible.
Contributors MA performed the genetic anayses.
IH-V was responsible of the semen and testicular tissue analyses, definition of the sperm defect and in vitro fertilisations.
AH participated in recruiting the patients and clinical data.
AZ participated in recruiting the patients and clinical data.
MA-A performed the immunostaining analyses on the patients ‘testicular biopsies’.
VD performed the pathohistological analyses.
VCS participated in designing the genetic study.
EL coordinated the clinical research and participated in recruiting the patients and clinical data.
MH was responsible for the histology and immunostaining analyses on the patients ‘testicular biopsies’.
RP designed the genetics analyses and coordinated the research.
Funding This study was partially supported by The Natural Science Foundation of China (NSFC) - Israel Science Foundation (ISF) (NSFC-ISF) (1183/14), (MH and EL). This work was partly supported by an internal grant from Ben-Gurion University of the Negev, Faculty of Health Sciences.
Competing interests None declared.
Ethics approval Soroka Medical Center Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The manuscript presents a mutation in a gene not previously associated with human disease that causes male infertility. There is no additional unpublished data from the study that is available.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.