Genetic and functional analysis reveals novel mutations in meiotic genes underlying non-obstructive azoospermia

Abstract

Purpose

To identify novel pathogenic mutations in meiotic genes underlying non-obstructive azoospermia (NOA) through whole-exome sequencing analysis.

Materials and methods

Whole-exome sequencing was performed on 31 NOA patients. Novel variants were identified through bioinformatic analysis and validated by Sanger sequencing. The pathogenicity of variants was assessed through functional studies including protein structural analysis, conservation analysis, and minigene splicing assays.

Results

We identified novel pathogenic mutations in four meiotic genes. Most significantly, we discovered the first human case of a homozygous nonsense mutation in MAEL (c.514C > T, p.Gln172Ter) in a patient from a consanguineous family. Additionally, we found novel compound heterozygous mutations in MSH5 (c.648-2A > G and c.1133T > C) and REC114 (c.659_706dup and c.123C > A), and a heterozygous splice-site mutation in DMRT1 (c.968-2A > G). Functional analyses revealed that these mutations disrupted critical meiotic processes through aberrant splicing, protein truncation, or structural alterations.

Conclusion

Our findings expand the mutation spectrum of meiotic genes in male infertility and provide new insights into the genetic basis of spermatogenic failure. The identification of novel mutations, particularly the first reported homozygous MAEL mutation, enhances our understanding of meiotic arrest in human spermatogenesis and provides valuable information for genetic counseling.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-025-03505-6.

Introduction

Male infertility is a multifactorial heterogeneous pathology affecting approximately 50% of infertile couples, representing a significant global health concern with profound social and psychological implications [1, 2]. Among various forms of male infertility, non-obstructive azoospermia (NOA) represents the most severe phenotype, characterized by the complete absence of spermatozoa in the ejaculate due to spermatogenic failure rather than physical obstruction [3, 4]. The condition poses a substantial challenge to reproductive medicine, as it affects approximately 1% of the male population and 10–15% of infertile men [5]. NOA patients typically exhibit diverse testicular histological patterns, including Sertoli cell-only syndrome (SCOS), maturation arrest (MA), and hypospermatogenesis [6, 7]. The developed micro-dissection testicular sperm extraction and intracytoplasmic sperm injection can help a small number of NOA patients to obtain their biological offspring. In contrast, this pathogenic genetic risk might also be transmitted to the next male generations who confront infertility.

Spermatogenesis is a highly orchestrated process involving three key phases: mitotic proliferation of spermatogonia, meiotic division of spermatocytes, and post-meiotic development of spermatids [8]. This complex biological process requires the precise temporal and spatial expression of numerous genes, making it particularly vulnerable to genetic disruptions [9]. Any errors during this process can lead to spermatogenic failure and consequently, male infertility [10]. Despite significant advances in reproductive medicine, the genetic etiology of NOA remains largely unknown, with only approximately 15–20% of cases currently explained by known genetic factors such as chromosomal abnormalities, Y chromosome microdeletions, or single gene mutations [11, 12]. This knowledge gap represents a critical scientific problem, as understanding the genetic basis of NOA is essential for developing accurate diagnostic tools, improving genetic counseling, and potentially developing targeted therapeutic strategies. Recent advances in next-generation sequencing technologies, particularly whole-exome sequencing (WES), have revolutionized our understanding of the genetic basis of NOA [13]. Several genes critical for meiotic recombination and chromosome synapsis have been implicated in NOA pathogenesis, including TEX11, MSH4, and SYCP3 [14–16]. TEX11 is an X-linked gene essential for spermatogenesis, particularly during meiosis, where it plays a critical role in maintaining the synaptonemal complex and facilitating crossover formation. TEX11 knockout in mice results in meiotic arrest and azoospermia [17]. In humans, Yatsenko et al. identified TEX11 mutations in 2.4% of 289 men with azoospermia and in 15% of 33 patients with meiotic arrest, highlighting its significance in male infertility [18]. It has been demonstrated that loss of Msh4 or Msh5 results in defects of prophase I progression in mice, with almost complete failure of homologous synapsis, and cell death prior to pachynema [19–21]. Also, whole-exome sequencing (WES) of pedigree studies reported mutations in MSH4 were associated with NOA [15]. Concerning the SYCP3 gene, which encodes a DNA-binding protein involved in the synapsis of germ cell meiosis, a null mutation in mice leads to azoospermia with meiotic arrest. In humans, the function is compromised by the mutant protein via dominant negative interference [22]. Additionally, mutations in genes such as STAG3, SYCE1, and SPO11 have been shown to cause meiotic arrest in both mouse models and human patients [23–25]. STAG3 is involved not only in DBS repair, but also in the formation of chromosomal axis and cohesion of sister chromatids after DNA replication. Very recently, a homozygous STAG3 missense variant cosegregated with the infertility phenotype in a consanguineous family including a proband with NOA [26]. These findings are consistent with Stag3 KO mice, showing an early prophase I arrest and apoptosis in both male and female germ cells [27]. SYCE1 is a member of the synaptonemal complex, which links homologous chromosomes during prophase I of meiosis. Homozygous mutations in this gene are associated with NOA [28, 29]. SPO11 is essential to initiate meiotic recombination and formation of the synaptonemal complex between homologous chromosomes [30]. To date, a SPO11 homozygous missense variant has been identified in two brothers with MA [31]. Most recently, a likely deleterious variant in RBBP7, MOV10L1, and PNLDC1 were identified as a novel cause of maturation arrest [32–34], providing new insights into the molecular mechanisms underlying spermatogenic failure. These discoveries have significantly advanced our understanding of the molecular mechanisms underlying spermatogenic failure, yet they only explain a small fraction of NOA cases. The identification of additional genetic factors is crucial for several reasons: first, it would improve our understanding of the fundamental biological processes governing spermatogenesis. Second, it would enable more comprehensive genetic screening and counseling for affected individuals, and third, it could potentially lead to the development of novel therapeutic approaches.

The clinical significance of this research is underscored by the current limitations in NOA diagnosis and treatment [35]. The majority of patients rely on invasive procedures such as testicular sperm extraction (TESE), often with uncertain outcomes [36]. Moreover, since NOA men currently depend on surgical sperm retrieval with assisted reproductive technology to father biological children, these pathogenic genetic variants might be inherited and passed on to future generations [37]. This underscores the critical importance of comprehensive genetic screening, not only for accurate diagnosis but also for its potential prognostic value in clinical decision-making and genetic counseling [38]. Furthermore, identifying novel genetic factors is crucial for developing potential therapeutic strategies, including targeted gene therapy approaches [39].

In this study, we aimed to address these critical gaps in our understanding of NOA by performing WES analysis on 31 patients diagnosed with NOA showing either MA or SCOS phenotypes. Our research focused on identifying novel pathogenic variants, particularly in genes involved in meiotic processes, with the goal of expanding the genetic landscape of NOA and providing new insights into the molecular basis of spermatogenic failure. Here, we report several novel mutations in genes critical for spermatogenesis, the findings have the potential to significantly advance our understanding of male infertility, improve diagnostic capabilities, and inform future therapeutic strategies.

Materials and methods

Study subjects

Thirty-one unrelated patients with NOA were recruited from the Women’s Hospital, School of Medicine, Zhejiang University. The diagnosis of NOA was based on the following criteria: (1) no sperm detected in the ejaculate after centrifugation at 3000 g for 15 min on at least three occasions; (2) normal karyotype (46, XY); (3) absence of Y chromosome microdeletions; (4) no history of chemotherapy, radiotherapy, or other known causes of infertility. Testicular biopsies were performed for histological examination to confirm the diagnosis. Written informed consent was obtained from all participants. This study was approved by the Ethics Committee of Women’s Hospital, School of Medicine, Zhejiang University (Approval No. [IRB-20240397-R]).

Semen analysis

Semen samples were collected after 3–7 days of sexual abstinence and analyzed according to the World Health Organization guidelines (WHO, 5 th edition, 2010) [40]. After complete liquefaction at 37 °C for 30 min, routine semen analysis was performed. For patients with no visible sperm in initial microscopic examination, samples were centrifuged at 3000 g for 15 min and the pellet was examined for the presence of spermatozoa. To confirm the diagnosis of azoospermia, this assessment was repeated on at least three different occasions with intervals of more than 1 week.

Whole-exome sequencing and bioinformatics analysis

Genomic DNA was extracted from peripheral blood using the TIANamp Genomic DNA Kit (Tiangen Biotech, Beijing, China). WES was performed using the MGISEQ-2000 platform following standard protocols. Raw reads were assessed using FastQC and filtered using Trimmomatic. Clean reads were aligned to the human reference genome (hg19) using Burrows-Wheeler Aligner (BWA). PCR duplicates were removed using GATK’s MarkDuplicates.

To identify novel causative variants, we established a stringent filtering strategy: variants with minor allele frequency < 1% in population databases (ExAC, gnomAD, and 1000 Genomes), located in exonic or splice-site regions, functionally relevant (non-synonymous, stop-gain/loss, frameshift, or splice-site variants), and predicted to be deleterious by at least three prediction tools. Further prioritization was given to genes with high testicular expression according to The Human Protein Atlas database (https://www.proteinatlas.org/), with special attention to variants not previously reported in human azoospermia cases. The filtered variants were classified according to the American College of Medical Genetics and Genomics (ACMG) guidelines [41].

Sanger sequencing verification

Candidate pathogenic variants were validated by Sanger sequencing. PCR primers were designed using Primer3 software (primer sequences available in Supplementary Table 1). PCR products were sequenced on an ABI 3730xl DNA Analyzer (Applied Biosystems).

Conservation analysis

Multiple sequence alignment was performed using MEGA 7.0 software (https://www.megasoftware.net) to analyze the evolutionary conservation of affected amino acids. Protein sequences of orthologs from different species were obtained from the NCBI protein database, including Homo sapiens (Human), Pan troglodytes (chimpanzee), Macaca mulatta (rhesus monkey), Mus musculus (mouse), Rattus norvegicus (rat), and Loxodonta africana (elephant). The sequence alignment was conducted using the ClustalW algorithm implemented in MEGA 7.0 with default parameters. Conservation scores were calculated based on the alignment results, and the evolutionary relationships were visualized using the maximum likelihood method. Protein domains and functional motifs were annotated based on the UniProt database (https://www.uniprot.org/).

Minigene array

To evaluate the effect of splice-site variants, we constructed minigene plasmids using two different vectors. For MSH5 (c.648-2 A > G), wild-type and mutant sequences were amplified using three rounds of PCR with specific primers (Supplementary Table 2). The amplified fragments were cloned into pcMINI and pcDNA3.1 vectors using KpnI and XhoI restriction sites. The pcMINI constructs contained the sequence spanning from exon 7 to intron 8 (991 bp), while the pcDNA3.1 constructs included the region from exon 7 to exon 9 (1,155 bp).

The recombinant plasmids (pcMINI-MSH5-wt, pcMINI-MSH5-mut, pcDNA3.1-MSH5-wt, and pcDNA3.1-MSH5-mut) were transiently transfected into HEK-293T and HeLa cells according to the manufacturer’s protocol. After 36 h of culture, cells were harvested for RNA extraction and subsequent RT-PCR analysis.

To evaluate the effect of splice-site variants, we constructed minigene plasmids using two different vectors. For DMRT1 (c.968-2 A > G), wild-type and mutant sequences were amplified using three rounds of PCR with specific primers (Supplementary Table 3). The amplified fragments were cloned into pcMINI and pcDNA3.1 vectors using KpnI and NotI restriction sites. The pcMINI constructs contained the sequence spanning from intron 4 to exon 5 (683 bp), while the pcDNA3.1 constructs included the region from exon 4 to exon 5 (1268 bp).

Subsequently, the recombinant plasmids (pcMINI-DMRT1-wt, pcMINI-DMRT1-mut, pcDNA3.1-DMRT1-wt, and pcDNA3.1-DMRT1-mut) were transiently transfected into HEK-293 T and HeLa cells following the manufacturer’s instructions. The transfected cells were cultured for 36 h and then collected for further RT-PCR analysis.

Protein structure analysis

Protein secondary structures were predicted using the SOPMA method (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html). The AlphaFold3 server (https://golgi.sandbox.google.com/) was used to predict the tertiary structure of wild-type and mutant genes. The structural models of the wild-type and mutant genes were analyzed and visualized with PyMOL2.

Results

Clinical characteristics of NOA patients

Whole-exome sequencing was performed on 31 unrelated patients diagnosed with NOA, comprising 20 patients (64.5%) with SCOS and 11 patients (35.5%) with MA. All patients showed normal karyotype (46, XY) and absence of Y chromosome microdeletions. Clinical examination revealed reduced testicular volume (9.3 ± 3.8 mL). Endocrine evaluation demonstrated that serum FSH (18.1 ± 8.6 IU/L; reference range, 2.3–9.5 IU/L) and LH (10.2 ± 4.6 IU/L; reference range, 2.2–8.4 IU/L) levels were markedly elevated, while testosterone concentrations (13.0 ± 3.8 nmol/L; reference range, 6.6–35.0 nmol/L) were within the normal range.

Identification of novel variants by whole-exome sequencing

The average sequencing depth was greater than 100× with more than 99% of target regions covered by at least 20 reads. Through systematic whole-exome sequencing analysis of 31 NOA patients, we identified potentially pathogenic variants in four meiosis-related genes (MSH5, REC114, MAEL, and DMRT1). These variants were validated by Sanger sequencing and assessed for pathogenicity according to ACMG guidelines.

Compound heterozygous MSH5 mutations disrupt meiosis during spermatogenesis.

Through whole-exome sequencing, we identified compound heterozygous mutations in MSH5 (NM_172166.4) in patient M2401: a splice-site variant (c.648-2 A > G) and a missense variant (c.1133 T > C, p.Ile378Thr). Sanger sequencing confirmed autosomal recessive inheritance, with the splice mutation inherited from the mother and the missense mutation from the father (Fig. 1A). Both variants were absent from population databases (1000 Genomes, ExAC_EAS, and gnomAD_EAS), and no other pathogenic variants related to male infertility were detected.

Fig. 1.

Identification of the novel compound heterozygous MSH5 variants in a man with NOA. A Family pedigrees analysis of the family with the MSH5 variant. The black fill represents the patient, and the red frame indicates mutated positions in the Sanger sequencing results. B Testicular histopathology of M2401. P, pachytene spermatocyte; Scale bar: 10 µm. C Multiple sequence alignment of the MSH5 protein against different species. Sequence alignment of amino acids 216–228 of MSH5 from diverse species indicates that the mutated residues in MSH5 are highly conserved. The red arrow denotes the position of the variant (p.Ile378Thr), and the dotted lines indicate the positions of MSH5 variants in MSH5 protein. D The predicted effects on the secondary structure by SOPMA database

The patient was a 25-year-old male presenting with primary infertility after 3 years of regular unprotected intercourse, with normal libido and erectile function. Physical examination revealed bilateral small testes (mean testicular volume 7.2 mL). Laboratory tests showed markedly elevated serum FSH (29.2 IU/L) and LH (13.0 IU/L) levels, while total testosterone concentration (20.9 nmol/L) was within normal range (Table 1). Histopathological examination revealed complete meiotic arrest at the pachytene stage, with regular seminiferous tubules, intact basement membrane, and normal Sertoli cells. The most advanced germ cells were pachytene spermatocytes, with no post-meiotic cells observed (Fig. 1B).

Table 1.

Clinical characteristics of NOA patients

Patient ID	Age (years)	Testosterone (nmol/L)	FSH (IU/L)	LH (IU/L)	Testicular volume (mL)	Testicular histology
M2401	25	20.9	29.2	13.0	7.2	MA
M2402	26	7.7	5.7	7.3	12.0	MA
M2403	36	14.4	7.5	9.9	10.2	MA
M2404	30	9.3	18.6	13.4	8.8	MA

FSH follicle-stimulating hormone, LH luteinizing hormone, AZF azoospermia factor, MA meiotic arrest. Hormone values outside the normal range are marked in bold

The MSH5 missense variant p.Ile378Thr was predicted to be damaging by multiple in silico tools, including SIFT, PolyPhen-2, and PROVEAN, with a CADD score of 25.1. The MSH5 splice variant c.648-2 A > G showed high probability of splice-site disruption by multiple tools: dbscSNV (0.99), MaxEntScan (0.64), and SpliceAI (0.98), with a CADD score of 35 (Table 2). The MSH5 p.Ile378Thr mutation is located in the evolutionarily conserved DNA-binding domain of MSH5, and multiple sequence alignment confirmed high conservation of this residue across species (Fig. 1C). Secondary structure analysis revealed significant alterations in protein conformation, with changes in α-helix content (from 51.8 to 50.96%), extended strand (from 13.79 to 13.55%), β-turn (from 3.84 to 4.08%), and random coil (from 30.58 to 31.41%) compositions (Fig. 1D). To validate the effect of c.648-2 A > G variant, we performed minigene assays using pcDNA3.1 constructs containing MSH5 exons 7–9. RT-PCR analysis of transfected cells revealed two aberrant transcripts (Fig. 2B–E): one with a 20-bp deletion at exon 8′s 5′ end, producing a truncated 226-amino-acid protein (p.Thr217Serfs*11), and another showing complete exon 8 skipping, resulting in an altered 822-amino-acid protein (p.Leu216_Ser228 delinsPhe).

Table 2.

Comprehensive molecular and bioinformatic characterization of gene variants in NOA patients

Individual		M2401	M2402	M2403	M2404
Gene		MSH5	REC114	MAEL	DMRT1
Transcript		NM_172166.4	NM_001042367.2	NM_032858.3	NM_021951.3
Inheritance pattern		AR	AR	AR	AD
Hom/Het		Het/Het	Het/Het	Hom	Het
cDNA mutation		c.648-2 A > G /c.1133 T > C	c.123 C > A /c.659_706 dup	c.514 C > T	c.968-2 A > G
Mutation type		Splice site/missense	Nonsense/inframe insertion	Nonsense	Splice site
Protein alteration		-/p.Ile378Thr	p.Cys41Ter/p.Leu220_Gly235dup	p.Gln172Ter	-
Allele Frequency	1 KGP	0/0	0/0	0	0
	EXAC_EAS	0/0	0/0	0	0
	gnomAD_EAS	0/0	0/0	0	0
In silico bioinformatics prediction	SIFT	-/D	-/-	-	-
	PolyPhen-2	-/D	-/-	-	-
	PROVEAN	-/D	-/-	-	-
	MutationTaster	-/-	A/-	A	D
	CADD ≥ 10	35/25.1	24.7/-	38	35
	dbscSNV	0.99/-	-/-	-	0.99
	MaxEntScan	0.64/-	-/-	-	7.95
	SpliceAI	0.98/-	-/-	-	0.91

AR autosomal recessive, AD autosomal dominant, 1 KGP 1000 Genomes Project, ExAc_EAS the data of East Asian in Exome Aggregation Consortium, gnomAD_EAS the data of East Asian in the Genome Aggregation Database, D Damaging, A disease causing automatic

Fig. 2.

The splicing pattern and outcomes of the MSH5 mutation were demonstrated via a minigene assay. A Minigene construction strategy. B Sequencing results of minigene construction, with the wild type at the top and the mutated type at the bottom. C Agarose gel electrophoresis showing the sizes of the wild-type and mutated bands. D Minigene shear diagram. E Sequencing results for the splicing bands. The red arrow indicates the mutation position

Based on these functional analyses and according to ACMG guidelines, the MSH5 c.648-2 A > G variant was classified as pathogenic (PVS1, PM2_Supporting, and PP4), while the MSH5 p.Ile378Thr variant was considered likely pathogenic (PM1, PM2_Supporting, PP3_Moderate, and PP4). These findings suggest that both mutations severely impair MSH5 function during meiosis, leading to spermatogenic arrest at the pachytene stage.

Novel compound heterozygous mutations in REC114 were identified in the MA patient

Whole-exome sequencing identified compound heterozygous variants in REC114 (NM_001042367.2) in patient M2402: an insertion variant (c.659_706 dup) resulting in an in-frame duplication of 16 amino acids (p.Leu220_Gly235 dup), and a nonsense variant (c.123 C > A) creating a premature stop codon (p.Cys41*). Sanger sequencing confirmed the c.659_706 dup variant was inherited from the mother, while paternal DNA was unavailable for testing (Fig. 3A). Both variants were novel and absent from all population databases (1000 Genomes, ExAC_EAS, and gnomAD_EAS).

Fig. 3.

Identification of compound heterozygous variants in the REC114 gene in an infertile patient (M2402) leading to NOA. A Pedigree of the family affected by the REC114 mutation. The arrow indicates the male azoospermic proband. B Testicular histopathology of M2402. P, pachytene spermatocyte; Scale bar: 10 µm. C Multiple sequence alignments of REC114 protein from different species. Arrows (red) and dotted lines show (red) the positions of novel REC114 variants identified in the present study. D Secondary structure information from SOPMA server. E Schematics of the REC114 3D protein structure

The patient was a 26-year-old male who presented with a 4-year history of primary infertility. Physical examination showed normal testicular size bilaterally (mean testicular volume 12.0 mL). Hormonal evaluation revealed normal FSH (5.7 IU/L) and LH (7.3 IU/L) and testosterone levels (7.7 nmol/L) (Table 1). Testicular biopsy revealed maturation arrest at the pachytene stage of meiosis I (Fig. 3B).

The REC114 p.Cys41* variant occurs in exon 1 and is predicted to result in a severely truncated protein lacking all functional domains. The duplicated region (p.Leu220_Gly235 dup) lies within a highly conserved domain essential for REC114’s interaction with MEI4, as demonstrated by multiple sequence alignment across species (Fig. 3C). Protein structure prediction using SOPMA showed that both variants significantly alter the secondary structure composition. For REC114 p.Cys41*, the truncation eliminates most structural elements, while REC114 p.Leu220_Gly235 dup causes changes in α-helix (from 28.57 to 34.04%), extended strand (from 13.91 to 11.7%), and β-turn (from 5.64 to 5.32%) content (Fig. 3D). Three-dimensional modeling using AlphaFold3 predicted that these structural changes would disrupt the protein’s tertiary structure, particularly affecting the MEI4-binding interface. Comparative analysis between wild-type and mutant structures using PyMOL revealed that changes in spatial conformation occurred and might disrupt critical hydrogen bond networks, suggesting impaired protein stability and protein–protein interactions (Fig. 3E).

Based on comprehensive functional analyses and following ACMG guidelines, the REC114 p.Cys41* variant meets criteria PVS1 (null variant), PM2 (absent from population databases), and PP4 (phenotype highly specific for disease), and was classified as pathogenic, while p.Leu220_Gly235 dup satisfies PM2, PM1 (critical functional domain), PM4 and PP3, it was classified as likely pathogenic. The identification of these novel REC114 mutations expands our understanding of genetic factors causing meiotic arrest in human spermatogenesis.

First report of homozygous MAEL mutation causing maturation arrest of spermatogenesis in human

Whole-exome sequencing identified a novel homozygous nonsense mutation in MAEL (NM_032858.3: c.514 C > T, p.Gln172 Ter) in patient M2403 from a consanguineous family. Sanger sequencing confirmed autosomal recessive inheritance, with both parents being heterozygous carriers of the variant (Fig. 4A). This variant was absent from population databases (1000 Genomes, ExAC_EAS, and gnomAD_EAS).

Fig. 4.

A homozygous mutation (c.514 C > T) of MAEL in the patient from a consanguineous pedigree. A Family pedigree analysis of the family with variant in MAEL. Proband is indicated by black arrow. B Testicular histopathology of M2403. RS, round spermatid; ES, elongated spermatid; Scale bar: 10 µm. C Orthologous alignment shows that mutated amino acid is highly conserved, which underlines the likely pathogenic impact of the variant. D Schematic diagram of the domains of the MAEL protein, and the dotted lines indicate the positions of variants in MAEL protein. E Secondary structures of MAEL predicted by the SOPMA. F The predicted part three-dimensional structure of mutated MAEL residues by SWISS-MODEL software. WT, wild type; MUT, mutant

The patient was a 36-year-old male with a 5-year history of primary infertility despite regular unprotected intercourse. Physical examination revealed bilateral small testes (mean testicular volume 10.2 mL). Hormonal analysis showed normal FSH (7.5 IU/L) but elevated LH (9.9 IU/L) levels, with testosterone concentration (14.4 nmol/L) within normal range (Table 1). Testicular biopsy showed maturation arrest of spermatogenesis. In the seminiferous tubules, no mature spermatozoa were found, although a few round/elongated spermatids were observed (Fig. 4B).

In silico analysis using multiple prediction tools suggested the pathogenic nature of this variant. The nonsense mutation was predicted to be disease-causing by MutationTaster (prob = 0.999), with a CADD score of 38, well above the pathogenic threshold of 20 (Table 2).

The MAEL nonsense mutation (p.Gln172Ter) is located in the highly conserved MAEL domain, with multiple sequence alignment demonstrating 100% conservation of the glutamine residue across diverse species (Fig. 4C). This mutation introduces a premature termination codon, resulting in a truncated protein lacking the essential MAEL domain and C-terminal intrinsic disordered region (IDR) (Fig. 4D). Protein structure analysis using SOPMA revealed substantial changes in secondary structure elements, particularly affecting α-helix and β-sheet compositions (Fig. 4E). Three-dimensional modeling by AlphaFold3 predicted complete disruption of the protein’s spatial conformation, with the truncated protein lacking critical regions required for its endonuclease activity and biomolecular condensate formation (Fig. 4F).

Based on the functional analyses and considering previous studies showing that Mael^−/− mice exhibit meiotic arrest and male infertility [42], this variant was classified as pathogenic according to ACMG guidelines (PVS1, PM2 and PM3). This represents the first report of a MAEL homozygous mutation causing human male infertility, providing direct evidence for its essential role in human spermatogenesis.

DMRT1 splice-site mutation impairs spermatogenic progression

Through whole-exome sequencing, we identified a heterozygous splice-site mutation in DMRT1 (NM_021951.3: c.968-2 A > G) in patient M2404. The variant was located at the highly conserved intron 4/exon 5 junction and was absent from major population databases (1000 Genomes, ExAC_EAS, and gnomAD_EAS). Parental DNA was unavailable for segregation analysis (Fig. 5A).

Fig. 5.

Identification of the novel heterozygous DMRT1 variant in a man with NOA. A The family affected by the variant in DMRT1. The red frame indicates mutated positions in the Sanger sequencing results. B Testicular histopathology of M2404. P, pachytene spermatocyte; Scale bar: 10 µm. C Sequence alignment of amino acids 323–329 of DMRT1 from diverse species indicates that the mutated residues in DMRT1 are highly conserved. D The strategy for minigene construction and the accompanying schematic diagram are provided. E The results of sequencing the minigene construction are presented, with the wild type displayed at the top and the mutated variant positioned below. F Agarose gel electrophoresis illustrates the sizes of both the wild-type and mutated bands. G Minigene shear diagram. H Sequencing results for the splicing bands are shown, with the mutation site indicated by a red arrow

The patient was a 30-year-old male with a 3-year history of primary infertility despite regular unprotected intercourse. Physical examination revealed bilateral small testes (mean testicular volume 8.8 mL). Endocrine evaluation showed markedly elevated serum FSH (18.6 IU/L) and LH (13.4 IU/L) levels, while testosterone concentration (9.3 nmol/L) remained within normal range (Table 1). Histopathological examination revealed complete arrest of spermatogenesis at the primary spermatocyte stage, characterized by normal seminiferous tubule architecture with preserved spermatogonia and Sertoli cells, but absence of post-meiotic cells (Fig. 5B).

The DMRT1 c.968-2 A > G variant affects the canonical 3′ acceptor splice-site “AG” sequence, and in silico analysis predicted high probability of splice-site disruption (Table 2). Three independent splicing prediction tools showed consistent results: SpliceAI, RDDCsc, and FF, indicating approximately 90% probability of altered splicing (Supplementary Figure S1). Multiple sequence alignment demonstrated that this splice region is highly conserved across species (Fig. 5C).

To experimentally validate the splicing defect, we performed minigene assays using pcDNA3.1 constructs containing DMRT1 exon 4 (145 bp), partial intron 4 (946 bp), and exon 5 (155 bp). The wild-type and mutant constructs were confirmed by Sanger sequencing (Fig. 5D, E) and transfected into HeLa and HEK-293T cells. RT-PCR and sequence analysis revealed two aberrant transcripts (Fig. 5F, H): one showing a 16-bp deletion at the 5′ end of exon 5, resulting in a frameshift and premature termination (p.Val323 Alafs*59, 380 amino acids), and another with a 19-bp deletion, producing a different truncated protein (p.Phe324Lysfs*57, 379 amino acids).

Based on these functional analyses and according to ACMG guidelines, the splice-site mutation was classified as pathogenic (PVS1, PM2_Supporting, and PP4). The discovery of this new mutation in DMRT1 enhances our comprehension of the genetic elements responsible for meiotic arrest during human spermatogenesis.

Discussion

High-throughput sequencing has revolutionized our understanding of the genetic basis of male infertility, particularly in cases of NOA, a condition that remains poorly understood at the molecular level [43]. In this study, we performed whole-exome sequencing on 31 patients diagnosed with NOA and identified novel pathogenic or likely pathogenic variants in four meiotic genes (MSH5, REC114, MAEL, and DMRT1). Patients with MSH5, REC114, and DMRT1 mutation exhibited maturation arrest at meiosis I, highlighting the critical roles of these genes in meiotic progression. These findings significantly advance our understanding of the genetic architecture underlying spermatogenic failure and provide critical insights into the molecular mechanisms of male infertility.

Most notably, we identified that mutations in the MAEL gene cause spermatogenic arrest with an autosomal recessive inheritance pattern. In a consanguineous family, we discovered a homozygous nonsense mutation in MAEL (p.Gln172 Ter), which had not been previously reported. This represents the second report of MAEL variants associated with human azoospermia, following the recent discovery by Stallmeyer et al. [44]. Whereas compound heterozygous MAEL variants (c.799 C > T/c.908 + 1G > C) have been previously reported in this case of NOA, our findings represent the first reported instance of a homozygous MAEL mutation in humans. Consistent with previous findings [44], the M2403 patient in our study exhibited spermatogenic arrest phenotypes. MAEL is specifically expressed in the testis and localizes to both the sex body in spermatocytes and the chromatoid body in round spermatids [45, 46]. Recent studies have demonstrated that MAEL orchestrates piRNA biogenesis through liquid–liquid phase separation, a process essential for transposon repression and germ cell development [47, 48]. The mutation we identified disrupts both the MAEL domain and the intrinsically disordered region (IDR), likely impairing its interaction with PIWI proteins and compromising transposon repression. In addition, both our study and that of Stallmeyer et al. [44] observed the presence of sperm cells in testicular pathology results, which is consistent with the phenotype of Mael¹²⁹-null mutant mouse testes, further validating the role of MAEL in human fertility. This discovery underscores the importance of MAEL in meiotic progression and provides a potential target for future therapeutic interventions.

We identified compound heterozygous variants in MSH5 (c.648-2 A > G and c.1133 T > C) in patient M2401 with MA. MSH5 is highly expressed in testis and forms heterodimers with MSH4 to repair DNA double-strand breaks during meiotic recombination [49]. Studies in mouse models have demonstrated that Msh5^−/− mice exhibit meiotic arrest at the zygotene stage, characterized by defective chromosome synapsis and impaired DNA double-strand break (DSB) repair, resulting in sterility [21]. While several studies have reported MSH5 variants associated with NOA in humans [50–52], our functional analysis revealed that the novel splice-site variant c.648-2 A > G disrupts normal RNA splicing, as demonstrated through minigene assays, while the missense variant c.1133 T > C (p.Ile378 Thr) affects a highly conserved residue within the DNA-binding domain. Structural analysis and functional assays suggest that these mutations may disrupt the protein’s structure, potentially compromising MSH4-MSH5 heterodimer formation. Given that the MSH4-MSH5 heterodimer is crucial for DSB repair through homologous recombination [49], our findings suggest that these variants likely impair MSH4-MSH5 protein interactions and subsequent DNA-binding capacity, ultimately affecting meiotic progression. Although the definitive proof of pathogenicity for missense variants can only be established by functional tests, segregation studies in very large families, and/or identifying unrelated patients or families with the same mutations, reporting possible causative variants is an effective strategy to improve consistency in the interpretation of molecular findings in MSH5. This part of research provides additional evidence supporting the critical role of MSH5 in human male fertility and highlights its potential as a diagnostic marker for NOA.

Additionally, compound heterozygous variants in REC114 (c.123 C > A and c.659_706 dup) were identified in patient M2402 with MA. This is the second instance of reported REC114 variants linked to human azoospermia, subsequent to the recent findings by Xu et al. [53]. Consistent with earlier findings, the M2402 patient from our research demonstrated characteristics associated with spermatogenic arrest. Studies in mice have shown that REC114, highly expressed in the testis, is required for the formation of programmed DNA double-strand breaks and subsequent cell division of germ cells. In its absence, spermatogenesis was altered in adult male mice, as indicated by the presence of major defects in testis tubule development, specifically, the tubule diameter was smaller and tubules lacked haploid cells (spermatids and spermatozoa) [54]. Recent investigations have suggested that REC114 forms a complex with CCDC36 and MEI4 in mice, which plays a crucial role in the formation of double-strand breaks during homologous chromosome pairing and recombination [55]. REC114 consists of 266 amino acid residues. In its N terminus, it possesses a Pleckstrin homology (PH) domain, which specifically interacts with the N terminus of the coiled-coil-based tetramers formed by CCDC36. The C-terminal part of REC114 directly interacts with the N terminus of MEI4 [56]. Therefore, we speculate that the loss-of-function variant (123 C > A) resulted in truncated REC114 protein at the functionally critical domain and impaired interaction with MEI4 and CCDC36. The other variant, a 48-bp duplication in exon 6 (c.659_706 dup), does not cause a reading frame shift during translation, but it disrupts a highly conserved consensus sequence in MEI4-binding region. Given that the REC114/MEI4/CCDC36 complex is essential for DSB formation, our findings indicate that these variants may lead to a loss of complex function and abnormal DSB formation during chromosome synapsis, which, in turn, affecting the progression of meiosis. Therefore, the compound heterozygous variants we found were probably causative for infertility in patient M2402, and this finding reinforces the importance of REC114 in meiotic recombination and provides further evidence of its role in NOA pathogenesis.

In patient M2404, who presented MA, we identified potentially damaging variants in DMRT1 gene (c.968-2 A > G). DMRT1 is strongly expressed in undifferentiated and less in differentiating spermatogonia. Furthermore, DMRT1 represses the meiotic pathway of male germ cells and promotes germ cells to undergo mitosis. By coordinating spermatogonial development and mitotic amplification with meiosis, DMRT1 allows abundant, continuous production of sperm [57]. Studies in mouse models have demonstrated that Dmrt1^−/− mice exhibit severe testicular hypoplasia, disorganized seminiferous tubules, and nondifferentiated Sertoli cells. In addition, germ cells were missing [58]. In humans, DMRT1 plays a key role in testis differentiation and is expressed mainly in the testis [59]. Recently, heterozygous missense mutations in DMRT1 have been identified in patients with spermatogenic disorders, and patients carrying DMRT1 mutations exhibit abnormal testicular pathological phenotypes, which are heterogeneous and characterized by SCOS, MA of germ cells, such as the stagnation of spermatogonia and spermatocytes [60, 61]. Consistent with previous findings, the M2404 patient in our study exhibited spermatogenic arrest phenotypes. Although multiple studies have indicated that variants of DMRT1 are linked to NOA in humans [60–62], our functional analysis showed that the newly identified splice-site variant c.968-2 A > G interferes with normal RNA splicing, as evidenced by minigene assays. Given that the DMRT1 loss-of-function mutations are a risk factor and potential genetic cause of human spermatogenic failure [57, 61], our results indicate that this variant could lead to the disruption of normal DMRT1 function, resulting in impaired spermatogenesis, which may be the underlying cause of NOA in patient. Considering the moderate evidence that DMRT1 is associated with male infertility [63], it would be worthwhile to perform in vitro functional studies to confirm these variants’ pathogenicity. Our findings contribute to the growing body of evidence linking DMRT1 mutations to NOA and highlight its role in maintaining spermatogonial homeostasis.

In conclusion, our comprehensive genetic analysis has identified novel pathogenic variants in four meiosis-related genes (MSH5, REC114, MAEL, and DMRT1) in patients with NOA. The discovery of these variants, particularly the first reported homozygous nonsense mutation in MAEL, substantially enriches our understanding of the genetic architecture underlying spermatogenic failure. Through detailed bioinformatic analyses and functional validation, we have demonstrated that these variants likely disrupt critical processes during spermatogenesis, including meiotic recombination, double-strand break formation, and chromosomal synapsis.

While our findings provide compelling evidence for the pathogenic role of these variants, we acknowledge certain limitations of our study. First, the relatively small cohort size may not fully represent the mutation spectrum in the general NOA population. Second, the lack of complete family segregation data for some variants necessitates further validation in larger pedigrees. Additionally, more comprehensive functional studies would strengthen our understanding of the precise molecular mechanisms through which these variants affect spermatogenesis.

Nevertheless, the consistent phenotype-genotype correlation observed in our patients, combined with previous reports, strongly supports the causative relationship between these genes and NOA. The recurrent involvement of these meiotic genes in independent cohorts highlights their fundamental importance in human spermatogenesis. Based on these findings, we strongly recommend the inclusion of MSH5, REC114, MAEL, and DMRT1 in diagnostic gene panels for NOA.

In summary, our study significantly advances the understanding of the genetic basis of NOA by identifying novel pathogenic variants in four meiosis-related genes. These findings not only enrich our knowledge of the molecular mechanisms underlying spermatogenic failure but also provide valuable insights for clinical diagnosis, genetic counseling, and potential therapeutic development. Future studies with larger cohorts and comprehensive functional analyses will further validate these findings and enhance their translational impact.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

Conception and design: Ying Liu, Jingping Li, Fan Jin, Jia Fei; patients care and evaluation: Jingping Li, Lejun Li, Yuanyuan Zhou; experiments and data analysis: Ying Liu, Jingping Li; manuscript writing: all authors; final approval of the manuscript: all authors.

Data availability

The datasets generated and analyzed during this study will not be publicly accessible in order to safeguard the privacy of individual patients. However, it may be made available to the corresponding author upon reasonable request.

Declarations

Ethical approval

IRB-20240397-R

Conflict of interest

The authors declare no competing interests.