Genetic investigation in non-obstructive azoospermia: from the X chromosome to the whole exome

ABSTRACT The severest form of male factor infertility is non-obstructive azoospermia (NOA), which occurs in approximately 1% of all men in reproductive age. It is common knowledge that Klinefelter Syndrome (47, XXY) and Y-chromosome microdeletions are direct causes of NOA, but in the majority of patients the etiology of this spermatogenic alteration is still unknown. The global aim of the present thesis was to enhance our understanding on genetic factors involved in non-obstructive azoospermia. The first part of the thesis focuses on the search of X-linked “AZF-like” regions. The Y-linked AZF deletions, which arise through Non-Allelic Homologous Recombination (NAHR), are the first example in andrology of functionally relevant Copy Number Variations (CNVs) causing spermatogenic failure. In analogy to the Y chromosome, the X chromosome is enriched in genes involved in spermatogenesis and its hemizygous state in males implies a direct effect of a damaging deletion making it a promising target for the discovery of new genetic factors leading to male infertility. To this purpose, we performed a multi-step bioinformatic analysis starting from all X-linked CNVs reported in UCSC Genome Browser in order to select X-linked recurrent CNVs: i) flanked by segmental duplications (SDs) and thus possibly generated by the NAHR ii) containing genes that are probably under negative selection i.e. with an inverted ratio of deletions/duplications. Following the above analysis we identified 12 X-linked CNVs (candidate “AZF-like” regions) of which 10 CNVs contained genes with a predicted role during spermatogenesis. Screening for deletions was performed in 82 idiopathic NOA patients with different testis phenotypes from pure Sertoli Cell Only Syndrome (SCOS) to partial spermatid arrest. The analysis revealed a single deletion in a patient affected by pure spermatocytic arrest removing part of the members of the Opsin gene family and possibly affecting the expression of a testis specific gene, TEX28. qPCR analysis revealed that the Opsin gene family is not expressed in germ cells and the analysis of the carrier’ testis biopsy did not reveal any impairment of TEX28 expression. Therefore, no cause-effect relationship between deletion and the testis phenotype can be established. We hypothesize that the lack of deletions in our NOA cohort may be partially due to the strictly selected testicular phenotype. Hence, we cannot exclude deletions in these regions may cause a less severe impairment of spermatogenesis. On the other hand, for the regions containing ubiquitously expressed genes, the removal of one or more of these genes may cause a more complex phenotype. Our is the first study that, through a multi-step bioinformatic analysis, provides information about potential X-linked “AZF-like” regions and represents a starting point for future large scale investigations involving patients with crypto-or oligozoospermia. The second part of the thesis focuses on the sequencing of >160.000 coding exons in NOA patients and proven normozoospermic fertile controls. We performed a Whole Exome Sequencig (WES) in a set of 18 men affected by SCOS, Spermatogenic Arrest (SGA) and normozoospermic fertile controls. We studied patients with consanguineous parents and sporadic azoospermic cases. We have identified more than 22,000 variants/subject in the exons and splice sites. Concerning patients with consanguineous parents we adopted the recessive model by selecting rare (MAF≤0.01), predicted as pathogenic, homozygous variants in genes with a putative role during early spermatogenic stages. This analytic approach allowed the identification of 3 candidate genes for male infertility: FANCA, ADAD2 and MRO. The most relevant finding concerns the patient who carried the mutation p.Arg880Gln in the FANCA gene (a functionally damaging mutation) since it is the first time that Fanconi Anemia (FA) is diagnosed following an exome analysis for idiopathic NOA. Interestingly enough, the patient’s brother, also affected by NOA, was homozygous carrier of the same mutation. Although the two brothers were not aware about having Fanconi anemia, the discovery of this genetic anomaly prompted us to perform the chromosome breakage test, through which a mosaic FA was diagnosed in both subjects. Therefore, besides diagnosing the cause of NOA, we made an important incidental finding of Fanconi Anemia (chromosome instability/cancer-prone condition), providing benefit to the siblings’ future health by allowing preventive measures. For patients with unrelated parents we applied four models: i) search for hemyzigous rare X-linked pathogenic mutations (MAF≤0.01); ii) oligogenic inheritance of low-frequency/rare mutations in genes with a putative role during early spermatogenic stages; iii) synergistic effect of genes containing low-frequency/rare mutations belonging to the same biological pathway; iv) combined effect of validated genetic risk factors for NOA (common SNPs). Finally, we also performed a high resolution X-chromosome array-CGH in sporadic patients in order to complete WES data. The first model allowed us to indentify RBBP7 as a novel X-linked candidate gene for early spermatogenic stages. So far RBBP7 has been only proposed as a key regulator during oocyte meiosis, but the expression analysis performed in our laboratory in different testis biopsies showed that the encoded protein is also overexpressed in the spermatogonial cells. Concerning the X-chromosome specific array-CGH we could not identify any relevant X-linked CNV. The second model (oligogenic inheritance) allowed the identification of three patients with single heterozygous variants and three controls with multiple heterozygous mutations. Since no patients presented more than one mutation we exclude the possibility that the azoospermic phenotype is due to digenic/oligogenic inheritance. The fact that more than one mutation in these genes has been found in three normozoospermic men suggests that it is an unlikely model for NOA. Regarding the third model (Synergistic effect of multiple low frequency mutations), the enrichment analysis in NOA patients allowed the identification of an overrepresentation of genes belonging to 19 KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways. After filtering out the pathways enriched in the control group, we could define enrichment in the “regulation of actin cytoskeleton” pathway as a candidate for impaired spermatogenesis. One patient presented multiple mutations in genes forming part of this pathway suggesting a potential pathogenic mechanism for the NOA phenotype. Concerning the disease enrichment analysis we identified an overrepresentation of genes associated with neoplasms, urogenital neoplasms and Fanconi anemia/syndrome in the patient group and not in the control group. Finally, regarding the combined effect of validated genetic risk factors (common SNPs) reported in previous GWAS we did not observe differences between patients and controls. The work presented in this thesis provides further advancement in the understanding of the genomic basis of idiopathic NOA. On one hand, our bioinformatic analysis identified 12 AZF-like regions along the X-chromosome that are candidates for further large scale screening in less severe forms of male infertility. Our WES experiments proved that this approach is able to identify novel candidate genes and to provide a genetic diagnosis in patients with consanguineous parents (FANCA mutation). We provided a clear example on how WES might lead to important incidental findings and thus to diagnose a chromosome instability/cancer-prone condition with implication on general health and disease prevention. Concerning the sporadic cases, WES allowed the identification of a novel X-linked candidate gene for impaired spermatogenesis indicating that the X-chromosome remains a highly interesting target.