New Genetic Factors in Male Infertility
In this study we provide evidence for an association of CNV64, CNV67 and CNV69 with spermatogenic failure through (1) a case–control study on a large study population; (2) molecular characterisation of deletions; (3) a search for functional elements in the region of interest; and (4) genotype–phenotype correlation analysis.
Concerning the patient-enriched CNVs CNV64 and CNV69, two clues support their association with impaired spermatogenesis: (1) deletion carriers had an increased probability of impaired spermatogenesis compared with non-carriers (OR=1.9 and 2.2, respectively); and (2) semen quality in terms of total sperm count and total motile sperm count was significantly impaired in carriers compared with non-carriers. Type B CNV69 (the larger one) was significantly more represented in patients than in controls, suggesting that this deletion pattern may account for the potential deleterious effect of CNV69 on sperm production. This may be related to the closer position of type B deletion (7.8 kb) to an upstream insulator compared with type A, the proximal breakpoint of which maps to 2.6 kb downstream (see online supplementary figure S4). Importantly, through the breakpoint definition we developed a simple diagnostic tool for type B deletion, allowing other laboratories to further explore the role of this deletion in other ethnic groups.
We confirmed in more than 1200 subjects the recurrent (1.1%) and 'patient-specific' feature of CNV67. Unfortunately, it was impossible to obtain a fine mapping of CNV67 breakpoints because of the presence of highly repetitive sequences and the incomplete assembly of the currently available reference sequence of the human X chromosome derived from 16 different individuals. SFV analysis provided a more accurate estimate of the minimum and maximum deletion size of CNV67—for instance, when heterozygosis is observed, deletion is surely not present. Specifically, SFV analysis in the CXorf40A ampliconic region suggests that different deletion breakpoints might exist at the proximal edge of the CNV minimum size (see online supplementary figure S2), determining the involvement of the CXorf40A copy mapping inside the maximum size. Accordingly, one of our patients (carrier 10–314) displays the extent of the larger deletion that would remove part of the CXorf40A coding sequence. With regard to SFV analysis in the MAGEA9 ampliconic region, the variants are not fully informative. However, the homozygous pattern observed for most variants in our carriers allows speculation that the proximal copy of the MAGEA9 gene might be involved. It is especially evident in the case of the two brothers, given that the deletion carrier showed homozygosity whereas his brother was heterozygous for the SFVs mapping to the MAGEA9 copy. Although we were unable to formally demonstrate the removal of the proximal copy of the MAGEA9 gene on the basis of the above results, we can speculate that this gene is directly affected or the deletion affects its regulatory elements. For instance, large-scale CNVs might change the three-dimensional structure of chromatin, which is seemingly crucial for correct gene regulation.
We previously suggested that Cancer Testis gene dosage variation may play a role in CNV-related spermatogenic failure; accordingly, MAGEA9 belongs to this gene family. MAGEA9 is an ampliconic gene reported as independently acquired on the human X chromosome, since no orthologs could be detected in the murine X chromosome. Independently acquired X-linked genes are predominantly expressed in the testis with a specific expression of multi-copy genes in male germ cells. It can therefore be speculated that the loss of MAGEA9 copies would affect spermatogenesis. Moreover, according to gene ontology, MAGEA9 is mainly involved in the regulation of gene expression, DNA methylation, reproduction and spermatogenesis, reflecting its potential involvement in transcriptional and epigenetic regulatory mechanisms of gametogenesis. Finally, CNV67 may also affect the regulation of HSFX1/2, another independently acquired X-linked multi-copy gene with testis-specific expression.
Pedigree analyses of two CNV67 carriers indicated that this deletion is maternally inherited, thus not affecting female fertility. This is in accordance with the lack of expression of MAGEA9 in the ovary. The family of patient 11–041 is especially informative since the pathological semen phenotype of the carrier (11–041) versus his normozoospermic non-carrier brother is a strong indicator for a pathogenic effect of the deletion on spermatogenesis.
For the first time we provide evidence for a significant association between recurrent X-linked deletions and impaired sperm production. Strikingly, CNV67, which is specific to spermatogenic failure, resembles AZF deletions on the Y chromosome. This finding merits further investigations in order to elucidate the structural complexity of this region and to provide a feasible substrate for fine molecular characterisation and large-scale diagnostic testing.
Discussion
In this study we provide evidence for an association of CNV64, CNV67 and CNV69 with spermatogenic failure through (1) a case–control study on a large study population; (2) molecular characterisation of deletions; (3) a search for functional elements in the region of interest; and (4) genotype–phenotype correlation analysis.
Concerning the patient-enriched CNVs CNV64 and CNV69, two clues support their association with impaired spermatogenesis: (1) deletion carriers had an increased probability of impaired spermatogenesis compared with non-carriers (OR=1.9 and 2.2, respectively); and (2) semen quality in terms of total sperm count and total motile sperm count was significantly impaired in carriers compared with non-carriers. Type B CNV69 (the larger one) was significantly more represented in patients than in controls, suggesting that this deletion pattern may account for the potential deleterious effect of CNV69 on sperm production. This may be related to the closer position of type B deletion (7.8 kb) to an upstream insulator compared with type A, the proximal breakpoint of which maps to 2.6 kb downstream (see online supplementary figure S4). Importantly, through the breakpoint definition we developed a simple diagnostic tool for type B deletion, allowing other laboratories to further explore the role of this deletion in other ethnic groups.
We confirmed in more than 1200 subjects the recurrent (1.1%) and 'patient-specific' feature of CNV67. Unfortunately, it was impossible to obtain a fine mapping of CNV67 breakpoints because of the presence of highly repetitive sequences and the incomplete assembly of the currently available reference sequence of the human X chromosome derived from 16 different individuals. SFV analysis provided a more accurate estimate of the minimum and maximum deletion size of CNV67—for instance, when heterozygosis is observed, deletion is surely not present. Specifically, SFV analysis in the CXorf40A ampliconic region suggests that different deletion breakpoints might exist at the proximal edge of the CNV minimum size (see online supplementary figure S2), determining the involvement of the CXorf40A copy mapping inside the maximum size. Accordingly, one of our patients (carrier 10–314) displays the extent of the larger deletion that would remove part of the CXorf40A coding sequence. With regard to SFV analysis in the MAGEA9 ampliconic region, the variants are not fully informative. However, the homozygous pattern observed for most variants in our carriers allows speculation that the proximal copy of the MAGEA9 gene might be involved. It is especially evident in the case of the two brothers, given that the deletion carrier showed homozygosity whereas his brother was heterozygous for the SFVs mapping to the MAGEA9 copy. Although we were unable to formally demonstrate the removal of the proximal copy of the MAGEA9 gene on the basis of the above results, we can speculate that this gene is directly affected or the deletion affects its regulatory elements. For instance, large-scale CNVs might change the three-dimensional structure of chromatin, which is seemingly crucial for correct gene regulation.
We previously suggested that Cancer Testis gene dosage variation may play a role in CNV-related spermatogenic failure; accordingly, MAGEA9 belongs to this gene family. MAGEA9 is an ampliconic gene reported as independently acquired on the human X chromosome, since no orthologs could be detected in the murine X chromosome. Independently acquired X-linked genes are predominantly expressed in the testis with a specific expression of multi-copy genes in male germ cells. It can therefore be speculated that the loss of MAGEA9 copies would affect spermatogenesis. Moreover, according to gene ontology, MAGEA9 is mainly involved in the regulation of gene expression, DNA methylation, reproduction and spermatogenesis, reflecting its potential involvement in transcriptional and epigenetic regulatory mechanisms of gametogenesis. Finally, CNV67 may also affect the regulation of HSFX1/2, another independently acquired X-linked multi-copy gene with testis-specific expression.
Pedigree analyses of two CNV67 carriers indicated that this deletion is maternally inherited, thus not affecting female fertility. This is in accordance with the lack of expression of MAGEA9 in the ovary. The family of patient 11–041 is especially informative since the pathological semen phenotype of the carrier (11–041) versus his normozoospermic non-carrier brother is a strong indicator for a pathogenic effect of the deletion on spermatogenesis.
For the first time we provide evidence for a significant association between recurrent X-linked deletions and impaired sperm production. Strikingly, CNV67, which is specific to spermatogenic failure, resembles AZF deletions on the Y chromosome. This finding merits further investigations in order to elucidate the structural complexity of this region and to provide a feasible substrate for fine molecular characterisation and large-scale diagnostic testing.