Investigation of the genetic etiology in male infertility with apparently balanced chromosomal structural rearrangements by genome sequencing.

Apparently balanced chromosomal structural rearrangements are known to cause male infertility and account for approximately 1% of azoospermia or severe oligospermia. However, the underlying mechanisms of pathogenesis and etiologies are still largely unknown. Herein, we investigated apparently balanced interchromosomal structural rearrangements in six cases with azoospermia/severe oligospermia to comprehensively identify and delineate cryptic structural rearrangements and the related copy number variants. In addition, high read-depth genome sequencing (GS) (30-fold) was performed to investigate point mutations causative of male infertility. Mate-pair GS (4-fold) revealed additional structural rearrangements and/or copy number changes in 5 of 6 cases and detected a total of 48 rearrangements. Overall, the breakpoints caused truncations of 30 RefSeq genes, five of which were associated with spermatogenesis. Furthermore, the breakpoints disrupted 43 topological-associated domains. Direct disruptions or potential dysregulations of genes, which play potential roles in male germ cell development, apoptosis, and spermatogenesis, were found in all cases (n = 6). In addition, high read-depth GS detected dual molecular findings in case MI6, involving a complex rearrangement and two point mutations in the gene DNAH1. Overall, our study provided the molecular characteristics of apparently balanced interchromosomal structural rearrangements in patients with male infertility. We demonstrated the complexity of chromosomal structural rearrangements, potential gene disruptions/dysregulation and single-gene mutations could be the contributing mechanisms underlie male infertility.

INTRODUCTION

Infertility is defined as the failure to achieve pregnancy after 12 months or more of regular unprotected intercourse.1 Approximately 15% of couples are affected by infertility, and among them, male factor contributes to 50%.2 Male factor infertility is often defined by abnormal semen parameters, which could be associated with other medical conditions, developmental factors, lifestyle factors, and genetics. However, the exact nature of these associations is still largely unclear.3 The most common etiology of male infertility is primary testicular failure, resulting in quantitative impairment of spermatogenesis, accounting for 75% of male factor infertility.4 However, the etiology of primary testicular failure is unknown in about 40% of cases, in which genetic factors are thought to contribute to a significant proportion of these cases.5

Genetic factors account for at least 15% of male infertility, and they have been known to contribute to all four major etiological categories, including (1) spermatogenic quantitative defects, (2) ductal obstructions or dysfunction, (3) hypothalamic–pituitary axis disturbances, and (4) spermatogenic qualitative defects.5 The genetic landscape of male infertility is highly complex and heterogeneous, with over 2000 genes involved in spermatogenesis and testicular function.6 Diagnosing the genetic causes of male infertility has clinical implications for prognosis of testicular sperm retrieval and personalizing therapy,5 improving the reproductive health and general health of the couple.

Men with azoospermia have the highest risk of being affected by genetic factors (25%).5 There are two well-known genetic causes of nonobstructive azoospermia. The most common causative genetic defects include chromosome Yq11 microdeletions of the azoospermia factor region (AZF), which can be detected in 13% of men with azoospermia.7 In patients with complete AZFb or AZFb + AZFc deletions, testicular biopsy usually reveals Sertoli cells only or diffuse early maturation arrest; hence, these patients must use donor sperm or adoption. In contrast, men with partial AZFa deletions often have hypospermatogenesis, with much better surgical sperm retrieval rates.8

Chromosomal abnormalities including sex chromosomal abnormalities (SCA) also cause male infertility. Among them, balanced chromosomal abnormalities (BCA; such as translocation and insertion), defined as exchange of chromosomal segments between two or more chromosomes without apparent gains or losses detectable by karyotyping, account for 0.5%–1.0% of patients with severe oligospermia or azoospermia. Male infertility caused by BCA may be explained by the failure of pairing of homologous elements on derivative chromosomes during meiosis I. There is also evidence that failure of pairing of homologous elements promotes association between the quadrivalent and the X-Y bivalent, which affects sperm count.9 However, conception can still be achieved in a significant proportion of male BCA carriers. In addition, semen analysis showed no significant differences in BCA carriers and males with normal karyotype.10 In contrast, GS-based investigations of BCAs have suggested that gene disruption or dysregulation caused by structural rearrangements and cryptic complexities could also contribute to male infertility.1112 In complex rearrangements, the number of chromosomes and breakpoints involved, the location of breakpoints, and their relative sizes are presumed to affect fertility.13

Although GS has been applied to investigate the molecular breakpoints of chromosomal structural rearrangements in patients with abnormal phenotypes,14151617 standard GS (with small-insert DNA) is not well suited for identifying breakpoint junctions, particularly those that are mediated by repetitive elements.12 To overcome this limitation, we developed an in-house mate-pair genome sequencing method which utilizes large-insert size DNA (3–8 kb) libraries, sequenced at a low read-depth.18 Our previous studies demonstrated that this method increased the sensitivity in the detection of BCAs and demonstrated the detection of additional cryptic and complex rearrangements in karyotypically simple chromosomal abnormalities.17 In this study, we aim to investigate BCA breakpoint junctions and the related copy number deletions/duplications by mate-pair GS and identify other genomic variants that potentially contribute to male infertility.

PATIENTS AND METHODS

Ethics approval and case recruitment

This study was approved by The Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee (The Joint CUHK-NTEC CREC; approval No. 2020.046). All participants in this study have provided written informed consent. We studied six idiopathic nonobstructive azoospermic or severe oligospermic males with BCAs recruited from the Prince of Wales Hospital, 30–32 Ngan Shing Street, Shatin, New Territories, Hong Kong, China. BCAs were previously ascertained in these subjects by karyotyping (), and chromosome Y microdeletions have been excluded. One patient carried a simple two-way reciprocal translocation, one carried double two-way translocations, two had three-way translocations, one case had a two-way chromosomal insertion (described in our previous study18), and one case had both a complex insertion and a translocation.

Mate-pair genome sequencing

Genomic DNA (gDNA) was extracted with the DNeasy Blood & Tissue Kit (category No. 69506; Qiagen, Hilden, Germany) from each sample. gDNA was then quantified with a Qubit dsDNA HS Assay Kit (Invitrogen, Carlsbad, CA, USA). After quality control (QC), a 1-μg aliquot of gDNA was sheared to 3–8 kb for mate-pair library construction.18 The libraries were pooled and sequenced to paired-end 100 bp on an MGISEQ-2000 platform (MGI Technology Co., Ltd., Shenzhen, China). A minimal of 60 million read-pairs were generated for each case, equating to an average 4 × read-depth and 60 × to 100 × physical coverage. In addition, for each sample except MI5, 400 ng of gDNA was also subjected for standard high read-depth GS. Each library was sequenced in each lane (paired-end 150 bp) on an MGISEQ-2000 platform (MGI Technology Co., Ltd.) for a minimal of 30-fold read-depth.

Variant identification and interpretation

Genomic variants were detected using our in-house bioinformatics pipelines. (1) Structural rearrangements were detected at a resolution of 10 kb. In brief, chimeric read-pairs that aligned to different chromosomes or mapped to the same chromosome with a genomic distance of >10 kb1920 were clustered and filtered against a control dataset of >2500 GS data (to remove systematic errors due to short-read alignment or variants with high frequencies as polymorphisms). (2) Copy number variants (CNVs) were detected using our reported increment-ratio-of-coverage method at a resolution of 50 kb.212223 Rare CNVs with allele frequencies of <1% in our in-house Chinese subjects (n > 2000) were selected for review. (3) Absence of heterozygosity (AOH) was detected at a resolution of 5 Mb as previously described24 to investigate any uniparental disomy and parental consanguinity.

The high read-depth GS data analysis involved the following steps: fastq data QC and alignment (BWA-MEM), detection of single-nucleotide variants (SNVs), and small insertions/deletions (InDels) by HaplotypeCaller version 3.4 from the Genome Analysis Toolkit (GATK; Broad Institute, MA, USA).25 The variants were subsequently annotated using ANNOVAR26 and InterVAR27 with in-house and public databases.

The following criteria were used for prioritization of SNVs/InDels: (1) variants reported by ClinVAR or human gene mutation database (HGMD); (2) variants with a minor allele frequency ≤5% in the ExAC (http://exac.broadinstitute.org) and gnomAD (https://gnomad.broadinstitute.org) databases; (3) variants that are located in coding regions or exon–intron junctions; (4) variants with damaging/intolerant or splicing-change effects predicted by multiple computational algorithms (Revel, SIFT, Polyphen-2, MutationTaster, Human Splicing Finder, and MaxEntScan); and (5) variants that contain or are located in OMIM disease-causing genes. For known mutations, clinical correlation will be performed (for male infertility). Further classification of novel mutations will be performed for variants (1) located in autosomal dominant or X-linked dominant genes or (2) being homozygous or compound heterozygous in autosomal recessive or X-linked recessive genes. Potential disease-causing mutations will be selected for validation and parental confirmation when available.

Genomic variants (CNVs, AOHs, SNVs, and InDels) involving genes associated with male infertility, oligospermia or azoospermia, spermatogenesis deficiency, and ciliopathy are interpreted based on the guidelines of the American College of Medical Genetics and Genomics (ACMG) using the 5-tier classification.28

Variant verification and junction annotation

Rearrangement breakpoints were resolved to the single-nucleotide level by junction-specific PCR followed by Sanger sequencing, which enabled accurate junction annotation for interpretation.29 Sanger sequencing results (fa format) were mapped to the reference genome (GRCh37) using UCSC BLAT. To analyze and assess the functional and phenotypic association of structural rearrangements, the breakpoint junctions were annotated for (1) direct disruption of genes, (2) disruption of regulatory elements, (3) disruption of topologically associating domain (TAD) defined for hESC cell line30 (3D Genome browser: http://3dgenome.fsm.northwestern.edu/index.html), and (4) genes or regulatory elements within the same TAD as the breakpoint. Furthermore, the breakpoint junction sequence features were annotated to investigate the mechanisms involved in the rearrangements.

Orthogonal validation of CNVs and AOHs was performed using the 8 × 60K Fetal DNA Chip version 2.0 (Agilent Technologies, Santa Clara, CA, USA) microarray.3132 Sanger sequencing was performed to validate SNVs and InDels.

RESULTS

We performed mate-pair GS on six male cases with azoospermia/severe oligospermia () and apparently balanced chromosomal structural rearrangements previously ascertained by G-banded chromosome analysis, who were negative for chromosome Y microdeletions. Karyotyping revealed reciprocal-balanced translocation in one case (sample ID: MI1), two independent balanced translocations in one case (MI2), three-way balanced translocations in two cases (MI3 and MI4), a two-way chromosomal insertion in one case (MI5), and a complex insertion and translocation in one case (MI6; ). Overall, mate-pair GS detected 48 structural rearrangements. Cryptic deletions were identified in two cases (). Thirty-six rearrangements were precisely mapped to single-nucleotide resolution () by Sanger sequencing, while the remaining 12 rearrangements were verified by gap-PCR but could not be resolved to nucleotide level due to the presence of repetitive elements.

In the case with the simplest rearrangement, two breakpoints were detected in a reciprocal translocation in case MI1. In the most complex case (MI6), 19 rearrangements were detected with deletions of four segments. Of the 48 rearrangements, 39 disrupted genes. Overall, a total of 30 unique RefSeq genes () and 43 unique TADs were disrupted (). Potential gene disruption, dysregulation, and/or point mutations were investigated in each case to reveal the underlying mechanisms of azoospermia/severe oligospermia.

Simple-BCA: case MI1

The 39-year-old patient suffered from severe oligospermia with a karyotype of 46,XY,t(3;19)(p21.3;q13.3). Mate-pair GS revealed a simple reciprocal balanced translocation without copy number changes at the breakpoint junctions. Mate-pair GS revised the affected bands on both chromosomes to seq[GRCh37] t(3;19)(p14.3q13.2). Validated by PCR and Sanger sequencing, the breakpoint junctions were found to directly truncate the gene ERC2 at intron 12. The expression of ERC2 has been reported to be heat-sensitive in rat spermatocytes and round spermatid.33 Expression of genes that are heat-sensitive may be crucial because spermatogenesis occurs at approximately 3°C lower than body temperature.33 Mate-pair GS revealed no additional clinically significant CNVs and AOH. In addition, high read-depth GS not only confirmed the findings from mate-pair GS but also indicated that no SNVs/InDels associated with azoospermia/oligospermia were detected. As ERC2 is likely haploinsufficient (gnomAD: pLoF=0.99), its disruption could be associated with the phenotype in the patient.

Two independent sets of BCA: case MI2

This patient has primary infertility and azoospermia. The patient has two independent balanced translocations 46,XY,t(5;9)(p13.3;p22),t(7;21)(p13;q22.1) reported by G-banded chromosome analysis. Mate-pair GS revealed two balanced translocations t(7;21)(p13;q22.11) and t(5;9)(p14.1;p23) with a 15.4-kb deletion seq[GRCh37] del(9)(p23) chr9:g.10081969_10097410del at one of the breakpoint junctions (intron 3 of gene PTPRD). Sanger sequencing enabled fine mapping of three rearrangements (). Two breakpoints had 12-bp and 15-bp sequence insertions, and one had a 3-bp microhomology (). These features suggested that the structural rearrangements were formed by DNA replication-based mechanisms.

Two reciprocal translocations of case MI2 with a karyotype of 46,XY,t(5;9)(p13.3;p22),t(7;21)(p13;q22.1). The upper panel shows the normal chromosomes 5, 7, 9, and 21, and the lower panel shows the respective derivative chromosomes. Bars in red, green, orange, and purple indicate chromosomal segments on chromosomes 5, 7, 9, and 21, respectively. Each dotted line indicates a breakpoint junction on each chromosome. A small white bar indicates a copy number loss of 15.4 kb involving the intron 3 of PTPRD in chromosome 9. Disruption of candidate genes associated with male infertility is shown above the chromosomes. PTPRD: protein tyrosine phosphatase receptor type D; chr: chromosome; der: derivative chromosome.

The genes COA1 and PTPRD were disrupted. Among them, PTPRD has been reported as a candidate gene for nonobstructive azoospermia.3435 PTPRD is also predicted to be haploinsufficient (pLoF=1; gnomAD). Although four TADs were disrupted involving a total of 41 genes, none of the genes located within the disrupted TADs was associated with azoospermia/severe oligospermia. Mate-pair GS and high read-depth GS revealed no additional SNVs/InDels, CNVs, and AOH associated with male infertility in this case. Thus, the disruption of PTPRD may potentially explain the infertility.

Three-way BCA: case MI3

The patient has severe oligospermia and karyotype results indicated a three-way balanced translocation 46,XY,t(4;11;6)(q22;q21;q16). However, mate-pair GS detected additional cryptic complexities underlying the translocations which resulted in an overall of seven rearrangements without CNVs detected at the breakpoint junctions (). Among six rearrangements fine-mapped by Sanger sequencing, the features of the breakpoint junctions included blunt ends in four rearrangements and 1 bp microhomology in the other two rearrangements, suggesting end-joining as the rearrangement formation mechanism.

Three-way translocation of case MI3 with a karyotype of 46,XY,t(4;11;6)(q22;q21;q16). The upper panel shows the normal chromosomes, and the lower panel shows the derivative chromosomes. Seven rearrangements were detected by mate-pair genome sequencing, revealing more complex rearrangements than a three-way translocation. Bars in red, green, and orange indicate chromosomal segments on chromosomes 4, 6, and 11, respectively. Segments originating from chromosome 4 were found on both der(6) and der(11). Each dotted line indicates the breakpoint on each chromosome. Bars in white represent cryptic chromosomal segments on 4q21.21q21.23, while boxes in light green indicate a chromosomal segment of 6q14.2q16.2. Segments that have an inverse genomic orientation are shown by reversed numbers. Disruption of candidate genes associated with male infertility is shown above the chromosomes. CFAP299: cilia and flagella associated protein 299; chr: chromosome; der: derivative chromosome.

Genes CEP162 and CFAP299 were disrupted by the breakpoint junctions. Among them, the role of CFAP299 in spermatogenesis has been implicated in animal studies. Messenger RNA of CFAP299 is predominantly expressed in mouse testis at increased levels from 2–8-week-old testes in the cytoplasm of spermatogonia and primary spermatocytes.36 Silencing of CFAP299 resulted in an increase in the number of apoptotic cells and arrested cells at the G2/M phase.36 As such, CFAP299 plays a potential role in spermatogenesis in regulating cell apoptosis.36 In addition, five TADs were disrupted, involving 16 genes. However, none of the genes were known to be associated with spermatogenesis. Mate-pair GS and high read-depth GS confirmed that there were no additional clinically significant SNVs/InDels, CNVs, and AOH associated with azoospermia/oligospermia detected in this case. Therefore, disruption of CFAP299 could be a candidate cause of the infertility.

Three-way BCA: case MI4

The azoospermic patient carried a three-way reciprocal translocation: 46,XY,t(8;12;10)(q24.1;p13;q22), as shown in . Mate-pair GS revealed additional cryptic complexities, with a total of seven rearrangements. There were no CNVs involved at the breakpoint junctions. All breakpoints were precisely mapped with Sanger sequencing. The sequence features showed blunt-ends in four rearrangements and 1 bp microhomology in the other three rearrangements. The formation of these rearrangements was likely mediated by end-joining mechanism.

Three-way translocation of case MI4 with a karyotype of 46,XY,t(8;12;10)(q24.1;p13;q22). (a) The upper panel shows the normal chromosomes, and the lower panel shows the derivative chromosomes. Seven breakpoint junctions were detected by mate-pair genome sequencing. Bars in red, green and orange indicate chromosomal segments in chromosomes 8, 10 and 12, respectively. Each dotted line indicates the breakpoint locus on each chromosome. Bars in white represent chromosomal segments on 12p12.3. Characters in a reverse direction indicate a reverse orientation of the chromosomal segments in derivative chromosomes. Disruption of candidate genes associated with male infertility is shown above the chromosomes. (b) Sequence resolved breakpoint junction of a cryptic rearrangement der(12) reveals blunt end as the sequence characteristics. (c) The 3D genome map (http://3dgenome.fsm.northwestern.edu/) visualizing the 3D interaction and topologically associating domains. The black triangle shows a topologically associating domain disrupted by a breakpoint (shown by the vertical black line). Genes encompassed under each topologically associating domain are listed in the lower panel. EPS8: epidermal growth factor receptor pathway substrate 8; DHSs: DNase I hypersensitive sites; chr: chromosome; der: derivative chromosome.

Three genes including EPS8, SOX5, and CTNNA3 were disrupted. EPS8 is known to be sensitive to haploinsufficiency (gnomAD: pLoF=0.97) and has been implicated to play a role in the induction of spermatid release from seminiferous epithelium during spermiation.37 Loss of EPS8 at the apical Sertoli cell–spermatid interface induced premature spermatid release and infertility.37 In addition, seven TADs were disrupted by the breakpoint junctions, which contained a total of 49 genes. Among them, PDE3A is a candidate gene associated with hypogonadism and Kallmann syndrome.38 Translocation breakpoints near the PDE3A gene have been reported in translocation carriers with Kallmann syndrome, potentially due to positional effects.39 In addition, female mice with PDE3A homozygous deletions were infertile, and ovulated oocytes were arrested in the germinal stage and could not be fertilized.40 Mate-pair GS and high read-depth GS found no additional SNVs/InDels, CNVs, and AOH associated with male infertility in this case. The disruption of EPS8 and the potential dysregulation of PDE3A are the possible explanations for infertility.

Two-way insertion: case MI5

This patient suffered from severe oligoasthenospermia and has been described in our previous study.18 An apparently balanced chromosomal insertion was identified by karyotyping analysis: 46,XY,ins(6;2)(q23;p13p22). Mate-pair GS detected nine rearrangements and indicated that the acceptor chromosome was likely involved in a chromothripsis event. Among them, six rearrangements were fine mapped by Sanger sequencing, including 1–2 bp microhomology (n = 3), small insertion (1 bp and 9 bp; n = 2), and blunt end (n = 1).

Seven genes were directly truncated by the breakpoints; however, none of which have been implicated in azoospermia/severe oligospermia. Eight TADs involving a total of 65 genes were disrupted. Although no direct evidence has shown associations of SLC17A5 (directly disrupted by the breakpoint junction) with male infertility, the gene EEF1A1 whose super enhancer was highly correlated to SLC17A5 was haploinsufficient (gnomAD: pLoF=0.98) and essential for spermatogenesis.41 Therefore, male infertility might be explained by the dysregulation of EEF1A1 expression through disruption of SLC17A5, leading to spermatogenesis failure.18 In addition, mutations in MCM9 are known to result in ovarian dysgenesis. MCM9-knockout male mice were sterile, resulted by arrest of spermatocytes in meiotic prophase I, while MCM9-knockout female mice were also sterile as their ovaries only contained arrested primary follicles and frequently developed tumors.42 There was insufficient DNA to pursue high read-depth GS in this case. Dysregulation of gene EEF1A1 and/or MCM9 by the structural rearrangement might explain the infertility in this case.

Complex insertion and translocation: case MI6

This patient had a complex karyotype: 46,XY,t(4;20)(q28;q12),der(20)ins(20;3)(q12;q28q13.3)inv(3)?(q13.3q25.3) involving three chromosomes. Mate-pair GS detected multiple complex rearrangements with 19 rearrangements (). Four copy number losses were found on chromosome 3, and the largest was a 1.15-Mb deletion involving the gene CFAP91. Mutations in CFAP91 are known to cause spermatogenic failure 51 (OMIM#619177) in an autosomal recessive manner (OMIM*609910). However, no additional variants were detected in CFAP91. Among the 19 rearrangements, 12 were successfully fine mapped by Sanger sequencing. The results revealed blunt ends at five junctions, 1–2 bp microhomology at six junctions, and 1 bp insertion at one junction. The breakpoint junction features along with frequent rearrangements suggested that this complex rearrangement was likely formed by DNA replicative mechanisms ().

Complex insertion and translocation of case MI6 detected by mate-pair genome sequencing. Bars in red, green, and orange indicate chromosomal segments in chromosomes 3, 4, and 20, respectively. Each dotted line indicates the breakpoint locus on each chromosome. Bars in white present cryptic chromosomal segment in 3q13.32q28 indicating there are 22 segments after chromosome scatting. Characters in a reverse direction indicate a reverse orientation of the chromosomal segments in derivative chromosomes. Disruption of candidate genes associated with male infertility is shown above the chromosomes. Copy number losses of four chromosome segments originating on chromosome 3 are indicated in a dotted frame. TP63: tumor protein p63; chr: chromosome; der: derivative chromosome.

Overall, 14 genes were directly truncated by the rearrangements including TP63. TP63 is sensitive to haploinsufficiency (gnomAD: pLoF=1) and plays a role in mediating male germ cell development, apoptosis, and regulation of spermatogenesis.4344 In addition, 15 TADs, involving a total of 128 genes, were disrupted. Among them, three genes were implicated in male infertility, including GMNC, SOX2, and CCDC39. GMNC (also known as Gemc1 in mice) plays a role in later stages of spermatogenesis.45 Gemc1−/− mice had a significantly reduced number of round spermatids and elongating spermatids.45 SOX2 is also a candidate gene for hypogonadotropic hypogonadism and resultant testicular failure.46 At last, mutations in CCDC39 are known to cause primary ciliary dyskinesia and Kartagener syndrome in an autosomal recessive manner, of which the common features include situs inversus, infertility, and oligoasthenospermia.47

There were no additional SNVs/InDels in CFAP91, which was involved in a heterozygous deletion as described above. However, high read-depth GS detected two heterozygous SNVs in the gene DNAH1. Mutations in DNAH1 cause spermatogenic failure through multiple morphological abnormalities of the sperm flagella and primary ciliary dyskinesia in an autosomal recessive manner.4849 However, the inheritance pattern of these two SNVs (in cis- or trans-) could not be determined as parental samples were not available. Nonetheless, male infertility in this case may be contributed by dual diagnoses of a complex rearrangement and point mutations.

DISCUSSION

In this study, we investigated the genetic etiology of six cases with male infertility and apparently balanced interchromosomal structural rearrangements by mate-pair GS in combination with high read-depth GS. Overall, our study revealed that disruption or potential dysregulation of candidate genes implicated in male germ cell development, apoptosis, and spermatogenesis could be the underlying mechanisms of male infertility in these cases.

Balanced chromosomal structural rearrangements are thought to cause male infertility through failure of pairing of homologous elements.9 Recently, studies have shown male infertility in patients with BCAs may be caused by gene disruption or dysregulation by breakpoints or other genomic variants, for example, SNVs and InDels.1112 While standard high read-depth GS cannot detect chromosomal rearrangements in low complexity or highly repetitive regions of the genome,5051 our large-insert size DNA mate-pair GS approach had increased sensitivity in detecting structural rearrangements.18 We identified the composition of each derivative chromosome in each case and revealed additional rearrangements cryptic to conventional G-banded chromosome analysis. We showed direct disruption or potential dysregulation of candidate genes associated with male infertility. In addition, high read-depth GS not only reported a dual molecular findings in case MI6, including a complex rearrangement and two point mutations in an autosomal recessive gene DNAH1, but also excluded known causative variants and noncoding regions (such as deep intronic region).

Our study on interchromosomal structural rearrangements revealed the complexities of BCAs that are largely underappreciated by G-banded chromosome analysis. Except MI1, additional cryptic rearrangements and/or copy number changes were detected in the remaining cases. The simplest was a cryptic deletion (15.4 kb) in case MI2 with two independent reciprocal translocations. In comparison, the rearrangements were most complex in MI6, which had 19 rearrangements and four deletions. The proportion of cases (5/6, 83.3%) with additional cryptic complex rearrangement identified by mate-pair GS over karyotyping was significantly higher than that reported in our previous study of males carrying BCAs in recurrent miscarriage couples (5/33, 15.2%, Fisher's exact test: P = 0.0023), who were able to conceive.16 It indicated that the complexities of structural rearrangements may be correlated with the phenotypic presentation (azoospermia or severe oligospermia).

Interestingly, among these cases, direct gene disruptions and/or potential dysregulation due to TAD disruptions are the potential underlying disease-causing mechanisms. In case MI2, the cryptic deletion (15.4 kb) was located in the intron of gene PTPRD. In case MI6, there were four copy number losses including gene CFAP91; a gene known to cause spermatogenic failure 51 (OMIM#619177) in an autosomal recessive manner (OMIM*609910). We excluded the possibility of compound heterozygous disease-causing alleles in autosomal recessive genes by identifying any causative hemizygous point mutations using high read-depth GS.

The limitations of our study include (1) limited sample size as there were six cases enrolled in this study, partly due to the rarity of apparently balanced chromosomal structural rearrangements and negative AZF region deletions in patients with severe oligospermia or azoospermia (0.5%–1.0%); (2) no parental samples available to study the inheritance mode of the detected rearrangements variants; for the two point mutations identified in DNAH1, long-read sequencing may resolve the haplotype spanning the two variants that are approximately 33 kb apart; however, no additional sample was available for this experiment; and (3) no additional samples were available for transcriptome analysis to confirm the potential consequences of the genomic variants. Nonetheless, our study provided a comprehensive composition of genomic structural rearrangements in each case and showed that the disruption or potential dysregulation of candidate genes could be the underlying causes of male infertility.

Overall, our study provided molecular characteristics of apparently balanced interchromosomal structural rearrangements in patients with male infertility. We showed the complexity of chromosomal structural rearrangements involved in male infertility and revealed that disruption or potential dysregulation of genes potentially implicated in male gametes development, apoptosis, and spermatogenesis could be a cause for infertility.

AUTHOR CONTRIBUTIONS

MHKC, YL, KWC, XK, and ZD designed the study. JPWC collected the samples and followed up the clinical outcomes. MHKC, YL, MS, YKK, and ZD performed the analysis and data interpretation. MHKC, YL, PD and XZ conducted the validation. MHKC, YL, KWC, and ZD wrote the manuscript. All authors read and approved the final manuscript.

COMPETING INTERESTS

All authors declare no competing interests.

Table 1

Case summary of the azoospermic/oligospermic cohort with apparently balanced structural rearrangements identified

Case ID	Clinical indication	FSH (mIU ml−1)	LH (mIU ml−1)	T (ng ml−1)	Age (year)	Karyotype	Rearrangements identified by mate-pair GS (n)	Number of cryptic copy number variants involved in the rearrangements (n)
MI1	Severe oligospermia	4.9	5.8	2.2	39	46,XY,t(3;19)(p21.3;q13.3)	2	0
MI2	Primary infertility, and azoospermia	NA	NA	NA	25	46,XY,t(5;9)(p13.3;p22),t(7;21)(p13;q22.1)	4	1
MI3	Severe oligospermia	5.7	5.5	5.9	24	46,XY,t(4;11;6)(q22;q21;q16)	7	0
MI4	Azoospermia	5.0	7.0	4.4	26	46,XY,t(8;12;10)(q24.1;p13;q22)	7	0
MI5	Severe oligospermia, asthenozoospermia, and teratospermia	4.4	5.5	5.9	28	46,XY,ins(6;2)(q23;p13p22)	9	0
MI6	Azoospermia	NA	NA	NA	NA	46,XY,t(4;20)(q28;q12),der(20)ins(20;3)(q12;q28q13.3)inv(3)?(q13.3q25.3)	19	4

FSH: follicle-stimulating hormone; LH: luteinizing hormone; T: testosterone; NA: not available; GS: genome sequencing

Supplementary Table 1

RefSeq genes encompassed in the copy number variation regions

Case ID	Chromosome	Start	End	Length	Band	Genes
MI2	9	10,081,969	10,097,410	15,441	9p23	PTPRD
MI6	3	117,792,055	118,443,877	651,823	3q13.32	EU250752
MI6	3	118,759,395	119911093	1,151,699	3q13.32q13.33	IGSF11, TEX55, UPK1B, B4GALT4, ALT4-AS1, ARHGAP31, AP31-AS1, TMEM39A, POGLUT1, TIMMDC1, CD80, ADPRH, PLA1A, OPDC2, POPDC2, COX17, FAP91, CFAP91, NR1I2, GSK3B, MIR6529, GPR156, PR156
MI6	3	189,298,977	189,332,375	33,399	3q28	TP63
MI6	3	189,675,304	190,261,962	586,659	3q28	P3H2, P3H2-AS1, LDN16, CLDN1, CLDN16, TMEM207, L1RAP, IL1RAP

Supplementary Table 2

Precise breakpoints junction features delineated by Sanger sequencing or breakpoints detected by mate-pair genome sequencing

Case ID	Breakpoint	Junction	Confirmation method	Feature
MI1	1	Chr19:(+)(42340154):ttctctt: chr3:(−)(55995592)	Sanger	Insertion, 7bp
MI1	2	Chr19:(−)(42340158)::chr3:(+)(55995594)	Sanger	Blunt end
MI2	1	Chr21:(+)(34271591): attaataaatgtccg: chr7:(+)(43760345)	Sanger	Insertion, 15bp
MI2	2	Chr7:(+)(43760342):acatttattaat: chr21:(+)(34271591)	Sanger	Insertion, 12bp
MI2	3	Chr9:(+)(100820{48~50})::chr5:(+)(2767695{5~7})	Sanger	MH, 3bp
MI2	4	Chr5:(+)(27676953)::chr9(+)(10,097,410)	gap-pcr	NA
MI3	1	Chr4:(+)(8224610{9})::chr11:(−)(8240970{1})	Sanger	MH, 1bp
MI3	2	Chr4:(+)(82204576)::chr11:(+)(82409700)	Sanger	Blunt end
MI3	3	Chr4:(−)(8594093{2})::chr4:(+)(8220457{4})	Sanger	MH, 1bp
MI3	4	Chr6:(+)(93377684)::chr4:(−)(85940930)	Sanger	Blunt end
MI3	5	Chr4(+)(81610923)::chr6:(−)(93377690)	gap-pcr	NA
MI3	6	Chr6:(−)(84894850)::chr4:(+)(81611154)	Sanger	Blunt end
MI3	7	Chr6:(+)(84894849)::chr4:(+)(82247031)	Sanger	Blunt end
MI4	1	Chr10:(+)(1141977{7}):chr10:(−)(6804828{8})	Sanger	MH, 1bp
MI4	2	Chr12:(+)(1516307{9}):chr10:(+)(1141977{7})	Sanger	MH, 1bp
MI4	3	Chr12:(−)(15166059):chr12:(+)(15810632)	Sanger	Blunt end
MI4	4	Chr12:(+)(15810630)::chr12:(+)(19890288)	Sanger	Blunt end
MI4	5	Chr12:(+)(19890255)::chr12:(+)(19978665)	Sanger	Blunt end
MI4	6	Chr12:(+)(2371854{8})::chr8:(+)(12056580{0})	Sanger	MH, 1bp
MI4	7	Chr8:(+)(120565798)::chr10:(+)(68048323)	Sanger	Blunt end
MI5	1	Chr6(+)(74138566)::chr6(+)(74346180)	gap-pcr	NA
MI5	2	Chr6:(+): 114469971::chr2:(−):58708053	Sanger	Blunt end
MI5	3	Chr2:(−): 5816279{1}::chr6:(+): 11446997{0}	Sanger	MH, 1bp
MI5	4	Chr6:(+):119963985:: aatctttta::chr2:(+):58708055	Sanger	Insertion, 9bp
MI5	5	Chr2:(+):8401992{2-3}::chr2:(+):3816018{1-2}	Sanger	MH, 2bp
MI5	6	Chr2(+)(58162618)::chr6(+)(119964044)	Gap-pcr	NA
MI5	7	Chr6(+)(162909640)::a::chr6(−)(74343934)	Sanger	Insertion ,1bp
MI5	8	Chr6(−)(7413867{0-1})::chr6(+)(16290964{2-3})	Sanger	MH, 2bp
MI5	9	Chr2:(+)(38160180*)::chr2(+)(84019924)	gap-pcr	NA
MI6	1	Chr20:(+)(40948748)::chr3:(−)(189676111)	gap-pcr	NA
MI6	2	Chr3:(+)(183349269)::chr20:(+)(40948749)	gap-pcr	NA
MI6	3	Chr3:(−)(18334986{8-9})::chr3:(−)(11874283{5-6})	Sanger	MH, 2bp
MI6	4	Chr3:(+)(117789537)::chr3:(+)(119912673)	Sanger	Blunt end
MI6	5	Chr3:(−)(125756425)::chr3:(+)(190261730)	Sanger	Blunt end
MI6	6	Chr13:(+)(19091001{5-6})::chr3:(−)(12575642{3-4})	Sanger	MH, 2bp
MI6	7	Chr3:(+)(19044791{0})::chr3:(+)(19091001{5})	Sanger	MH, 1bp
MI6	8	Chr3:(+)(18933161{8})::chr3:(+)(19044790{9})	Sanger	MH, 1bp
MI6	9	Chr3:(+)(189107560)::chr3:(+)(189105082)	gap-pcr	NA
MI6	10	Chr3:(+)(186592260)::chr3:(−)(186024159)	gap-pcr	NA
MI6	11	Chr3:(−)(182591739)::chr3:(+)(182591232)	gap-pcr	NA
MI6	12	Chr3:(+)(186676625):t: chr3:(−)(182591177)	Sanger	Insertion ,1bp
MI6	13	Chr3:(−) 122637474::chr3:(+)(120152667)	Sanger	Blunt end
MI6	14	Chr3:(+)(120152668)::chr3:(+)(190980450)	Sanger	Blunt end
MI6	15	Chr3:(+)(189295694)::chr3:(−)(190980449)	Sanger	Blunt end
MI6	16	Chr4:(+)(10139960{2~3})::chr3:(−)(11873198{6~7})	Sanger	MH, 2bp
MI6	17	Chr3:(−)(11873198{6-7})::chr4:(+)(10139960{2-3})	Sanger	MH, 2bp
MI6	18	Chr3:(−)(118443877)::chr3:(+)(189107419)	gap-pcr	NA
MI6	19	Chr3:(−)(190910012)::chr3:(−)(190447866)	gap-pcr	NA

MH: microhomology; NA: not available

Supplementary Table 3

RefSeq genes directly truncated by one or more breakpoints. Genes associated with male infertility are bolded

Genes
ATP11B	IGSF11	EPS8
CEP162	IL1RAP	ERC2
CFAP299	LOC105377975	FSTL1
CGAS	OSTN	GPR156
COA1	P3H2	HDAC2-AS2
CTNNA3	PRKN	TBC1D2
DGKG	PTPRD	TP63
EMCN	PTPRT	VRK2
RMDN2	SLC17A5	SOX5
SEMA5B	SLC41A3	ST6GAL1

Supplementary Table 4

RefSeq genes located within topologically associated domains disrupted by one or more breakpoints

Genes
ADCY5	COA1	FUNDC2P2	LINC00945	LOC101929130	MIR7151	PTPRT	SOX2
ADIPOQ	CORO2A	FXR1	LINC01122	LOC101929353	MIR8066	PURPL	SOX2-OT
ADIPOQ-AS1	CRYGS	GABBR2	LINC01206	LOC102723582	MIR944	PYDC2	SOX5
ALDH1L1	CTNNA2	GALM	LINC01216	LOC102724152	MRAP2	QKI	SRSF7
ALDH1L1-AS1	CTNNA3	GART	LINC01480	LOC102724604	MRPL32	RABL3	SST
ALG1L	CYB5R4	GMNC	LINC01489	LOC105374060	MRPS24	RERG	ST6GAL1
ANKS6	CYP1B1	GPR156	LINC01515	LOC105374312	MTO1	RERG-AS1	STK17A
ANXA2P3	CYP1B1-AS1	GSK3B	LINC01548	LOC105377879	MYLK-AS1	RFC4	STRAP
ARHGDIB	DDIT4L	GTF2E1	LINC01690	LOC105377975	NANS	RIPPLY2	STXBP5L
ART4	DDX43	H2AJ	LINC01793	LOC105378098	NDUFB4	RMDN2	SUCLG1
ATL2	DERA	H2AZ1-DT	LINC01795	LOC107984208	NR1I2	RMDN2-AS1	SYNJ1
ATP11B	DGKG	H4-16	LINC01809	LOC285762	NT5DC1	RNU6-1	TAF2
B3GNT8	DHFRP3	HACD2	LINC01815	LRRC58	OLIG1	RNU6-2	TBC1D2
BANK1	DKFZp451B082	HDAC2	LINC01883	LRRTM1	OLIG2	RNU6-7	TBCCD1
BCKDHA	DMAC2	HDAC2-AS2	LINC01994	LRRTM3	OOEP	RNU6-8	TMEM207
BCL11A	DMRTC2	HECW1	LINC01995	LUARIS	OSTN	RNU6-9	TMEM50B
BCL6	DNAJC12	HECW1-IT1	LINC02020	LYPD4	OSTN-AS1	ROPN1B	TMEM91
BLVRA	DNAJC19	HGD	LINC02024	MAATS1	P3H2	RPL39L	TP63
BMP3	DNAJC28	HNRNPLL	LINC02035	MAN1A1	PACRG	RTP1	TPI1P3
C12orf60	DONSON	HS3ST5	LINC02041	MAP3K7	PACRG-AS1	RTP2	TPRG1
C21orf62	DPPA5	HSPBAP1	LINC02043	MAP6D1	PACRG-AS2	RTP4	TRIM14
C21orf62-AS1	DTX3L	IFNAR1	LINC02049	MASP1	PACRG-AS3	SEC22A	UBE2D4
CAHM	ECHDC3	IFNAR2	LINC02052	MCM9	PARL	SEMA5B	URGCP
CASC6	EEF1A1	IFNGR2	LINC02103	MGP	PARP14	SFTA1P	URGCP-MRPS24
CCDC39	EMCN	IGSF11	LINC02398	MIR1255A	PARP15	SLC17A5	USP6NL
CCDC50	ENPP2	IGSF11-AS1	LINC02468	MIR198	PARP9	SLC39A8	UTS2B
CCDC66	EPS8	IL10RB	LINC02534	MIR3938	PAXBP1	SLC41A3	VRK2
CDH9	ERC2	IL10RB-DT	LINC02613	MIR3943	PAXBP1-AS1	SLC49A4	WBP11
CEACAM21	ERC2-IT1	IL1RAP	LINC02614	MIR4282	PCAT19	SLCO1B1	WDFY3-AS2
CEACAM3	ERICH4	KCNQ5	LINC02670	MIR4300	PDE3A	SLCO1B3	WDR5B
CEACAM4	ERP27	KCNQ5-AS1	LINC02671	MIR4300HG	PDE6H	SLCO1B3-SLCO1B7	WNT5A
CEACAM5	ETNK1	KCNQ5-IT1	LINCR-0002	MIR4432	PDIA5	SLCO1B7	YEATS2
CEACAM6	EXOSC5	KHDC1	LOC100131635	MIR4432HG	POLQ	SLCO1C1	YEATS2-AS1
CEACAM7	FAM162A	KHDC1L	LOC101926968	MIR4447	POLR2J4	SMCO3
CELF2	FAM184A	KHDC3L	LOC101927138	MIR4643	PPP3CA	SNAP91
CELF2-AS1	FAM86JP	KLHL24	LOC101927159	MIR548B	PRKG2	SNAR-I
CELF2-AS2	FANCL	KLHL6	LOC101928107	MIR548I1	PRKN	SNORA101A
CELF2-DT	FGF12	KLHL6-AS1	LOC101928441	MIR5682	PSMA2	SNORA63D
CEP162	FGF12-AS1	KPNA1	LOC101928882	MIR6501	PTPRD	SNORA63E
CFAP299	FLJ20021	LINC00710	LOC101928942	MIR6529	PTPRD-AS2	SNORD141A
CGAS	FRK	LINC00888	LOC101928961	MIR6854	PTPRO	SNORD141B
CGAS	FSTL1	LINC00901	LOC101929106	MIR7110	PTPRO	SON

Supplementary Table 5

Breakpoint junction sequence feature summary

Breakpoint junction feature	Number	Proportion (%)
Microhomologya	15	41.7
Small insertionb	6	16.7
Blunt end	15	41.7
Total	36

aMicrohomology ranged from 1–3 bp; bSmall insertion ranged from 1–15 bp