Identification of deleterious variants in patients with male infertility due to idiopathic non-obstructive azoospermia

A total of 55 patients with NOA and infertility were recruited from the First Affiliated Hospital of Anhui Medical University. The couples did not achieve a clinical pregnancy during more than 12 months of unprotected intercourse. Idiopathic NOA was diagnosed by careful medical history, physical examination, laboratory examination, imaging, and testicular biopsy. All patients excluded other risk factors of infertility/NOA, including chromosomal abnormalities, Y chromosome microdeletion, cryptorchidism, radiotherapy and chemotherapy, viral or bacterial orchitis, epididymitis, epididymo-orchitis, undescended testis and hypogonadism as well as sexually transmitted infections [6, 7]. We collected whole-peripheral blood samples from patients and performed WES as described in our previous study [8]. This research was approved by the Ethics Committee of the First Affiliated Hospital of Anhui Medical University, and all patients agreed to participate and signed the informed consent form.

Based on the World Health Organization guidelines (5th edition), semen sample was collected and examined. Semen samples were collected following 3–7 days of abstinence and evaluated after liquefaction for 30 min at 37 °C. Centrifugation was performed at 3000 g to detect sperm by microscopic examination of semen when no sperm were found by routine test [9]. To confirm the clinical diagnosis, we detected sperm at three different times with an interval of more than 2 weeks.

The whole-exome was sequenced using the Illumina HiSeq platform. The original FASTQ data were mapped to the human genome using BWA software, and SAMtools and GATK were used to call genetic variants. We annotated variants into the allele frequency database (1000G, EXAC03_EAS, gnomAD_exome_EAS), deleterious prediction tools (SIFT, PolyPhen-2, Mutation Taster, and CADD), and Human Gene Mutation Database (HGMD) using ANNOVAR [10] and dbNSFP [11]. Common variants with allele frequency of > 0.05 were excluded. We focused on loss-of-function (including splicing (≤ 2 bp), stopgain, stoploss, and frameshift indels) and deleterious missense variants. The variant that was predicted to be deleterious by more than three of four software, including SIFT, PolyPhen-2, Mutation Taster, and CADD (score > 20), were defined as a deleterious variant. Moreover, variants defined as deleterious in HGMD were included in the analysis. Sanger sequencing was performed to validate the inheritance pattern of the putative pathogenic variants.

We performed H&E staining to detect the detailed histopathology of the testis in NOA patients. Testicular tissue was fixed with Bouin solution for more than 24 h. Next, gradient alcohol (75%, 85%, 90%, 95%, and 100%) was used to dehydrate the fixed testicular tissue, embedded in paraffin, and sectioned into 3-μm-thick sections. Then, H&E staining was performed to stain the nucleus and cytoplasm.

NOA16 was diagnosed with infertility for 2 years with non-contraceptive intercourse. We found no sperm in recurrent semen analysis of NOA16, and physical examination showed typical development of secondary sexual characteristics, normal epididymis, vas deferens, and testicular size. Hormone tests revealed an abnormally increased level of follicle-stimulating hormone (FSH) (Table 1). We performed a testicular biopsy and subsequent H&E staining of NOA16 and found a testicular pathological phenotype of SCOS (Fig. 1). NOA16 was carrying homozygous missense variants in FANCA (NM_000135: c.3263C > T; p.S1088F) with allele frequency < 0.05, in three databases (Table 1). In addition, three of the four tools (SIFT, PolyPhen-2, MutationTaster, CADD) predicted that this missense mutation was deleterious. Sanger sequencing confirmed that the parents were heterozygous carriers. Additionally, another patient, NOA50, exhibited potential compound heterozygous FANCA. One heterozygous variant was consistent with that in NOA16, and the other variant (NM_000135: c.1729C > G; p.P577A) was extremely rare (0, 0, 0) in the three allele frequency databases (Table 1). Four tools generally predicted that this variant was deleterious. The sex hormones of NOA50 also showed an abnormally increased level of FSH (Table 1). Unfortunately, we were unable to perform a testicular biopsy and validate these two variants based on Sanger sequencing due to the disagreement of NOA50.

NOA51 was found to carry a homozygous frameshift variants in SYCE1 (NM_001143763, c.689_690del; p.F230fs) with extremely rare allele frequencies (0, 0, 0.0001) in three databases (Table 1). The levels of hormones in NOA51 were within the normal range (Table 1). We performed a testicular biopsy with the agreement of NOA51 and found the phenotype of MA (Fig. 2). Sanger sequencing confirmed the homozygous and heterozygous carriers in NOA51 and his biological father.

We found two NOA patients carrying heterozygous variants in DMRT1 and related amino acids in the functional domain. NOA22 carries NM_021951:c.425C > T; p.A142V, which was extremely rare (0, 0, 0) in the three allele frequency databases (Table 1). Three of the four tools generally predicted that this variant was deleterious. Sanger sequencing was performed to validate the heterozygous state. The sex hormones of NOA22 also showed abnormally increased FSH value (Table 1). Based on the agreement of NOA22, we performed a testicular biopsy and found only Sertoli cells (Fig. 3). The other patient, NOA25 carrying NM_021951: c.340G > A; p.V114M, also exhibited extremely rare (0, 0, 0) in the three allele frequency databases (Table 1). Four general tools predicted that this variant would be deleterious. Sanger sequencing was performed to validate the heterozygous state. Abnormal hormones, including increased FSH levels, were found in NOA25 (Table 1). Based on the agreement of NOA25, we performed a testicular biopsy and found that NOA25 exhibited the pathological phenotype of SCOS (Fig. 3). In addition, after an exhaustive understanding of microsurgical testicular sperm extraction, NOA25 consented to undergo the procedure. This patient succeeded in achieving pregnancy using microsurgical testicular sperm extraction and intracytoplasmic sperm injection. Unfortunately, we were unable to validate these variants based on Sanger sequencing because of the disagreement of the offering DNA sample of parents.

We also found that NOA42 carries a missense variant in the dominant PLK4 gene (NM_001190799, c.2785A > G; p.M929V) (Table 1). All allele frequencies in the three databases were zero. Four tools generally predicted that this variant was deleterious. Sanger sequencing validated the heterozygous state of c.2785A > G. An abnormal hormone of the increased FSH value was found in NOA42 (Table 1). We performed testicular biopsy with the agreement of NOA22, and only found Sertoli cells (Fig. 4). However, we were unable to validate these variants based on Sanger sequencing because of the disagreement for DNA sample collection of parents.

TEX11 is a well-known X-linked NOA pathogenic gene. This study identified two novel putatively pathogenic hemizygous variants in two patients with relatively different phenotypes (Table 1). NOA39 carries a missense variant (NM_031276:c.466A > G; p.M156V) with an allele frequency of < 0.01, which is inherited from his mother. The HGMD database defines this variant as being deleterious. In NOA39, the hormone levels were as follows: FSH 8.44 IU/L, LH 6.33 IU/L, T 10.75 nmol/L, estradiol 97 pmol/L, and prolactin 8.92 ng/ml. NOA39 showed hypospermatogenesis based on a testicular biopsy test (Fig. 5). The other patient, NOA49, carried a frameshift deletion NM_031276:c.559_560del, p.M187fs with zero allele frequency. The frameshift variant was a functional loss, resulting in severe structural and functional impairment of the TEX11 protein. In NOA49, the hormone levels were as follows: FSH 4.02 IU/L, LH 5.33 IU/L, T 13.34 nmol/L, estradiol 132 pmol/L, and prolactin 10.24 ng/ml. The phenotype of NOA49 was different from that of NOA39 and exhibited MA (Fig. 5). Sanger sequencing validated these two variants in patients. However, we cannot validate the inheritance pattern because of the disagreement of the offering DNA sample of parents.