High frequencies of Non Allelic Homologous Recombination (NAHR) events at the AZF loci and male infertility risk in Indian men

Deletions in the AZoospermia Factor (AZF) regions (spermatogenesis loci) on the human Y chromosome are reported as one of the most common causes of severe testiculopathy and spermatogenic defects leading to male infertility, yet not much data is available for Indian infertile men. Therefore, we screened for AZF region deletions in 973 infertile men consisting of 771 azoospermia, 105 oligozoospermia and 97 oligoteratozoospermia cases, along with 587 fertile normozoospermic men. The deletion screening was carried out using AZF-specific markers: STSs (Sequence Tagged Sites), SNVs (Single Nucleotide Variations), PCR-RFLP (Polymerase Chain Reaction - Restriction Fragment Length Polymorphism) analysis of STS amplicons, DNA sequencing and Southern hybridization techniques. Our study revealed deletion events in a total of 29.4% of infertile Indian men. Of these, non-allelic homologous recombination (NAHR) events accounted for 25.8%, which included 3.5% AZFb deletions, 2.3% AZFbc deletions, 6.9% complete AZFc deletions, and 13.1% partial AZFc deletions. We observed 3.2% AZFa deletions and a rare long AZFabc region deletion in 0.5% azoospermic men. This study illustrates how the ethnicity, endogamy and long-time geographical isolation of Indian populations might have played a major role in the high frequencies of deletion events.

The haplogrouping of 973 infertile and 587 fertile control men, using Y chromosome binary markers 40,41 revealed 8 distinct Y haplogroups in both cases and controls (Fig. 2). We compared the haplogroups of infertile men with or without AZF deletions and fertile control men; however, we failed to detect any specific deletion type that occurred only in a particular haplogroup background. We detected the two major haplogroups R1a-M17 and H1a-M82 with equal frequencies in both fertile and infertile men with/without deletions (Fig. 2). We also observed that the distributions of haplogroups were not different between the cases and the controls with/without deletions, suggesting that the haplogroups have no role in defining the deletion types and risk associations in infertile men.

Discussion
Yq microdeletions are well-established causative factors for quantitative decline of spermatozoa and can lead to spermatogenic failure 46,47 . In the present study, we identified very high frequencies of classical Yq microdeletions of the AZF regions in infertile men, whereas no such deletion was observed among controls. This further strengthens the idea that the classical Yq microdeletions are a cause of spermatogenic failure in the Indian idiopathic infertile men. Our study revealed a very high frequency of deletion events (a total of 29.4%) in Indian infertile men, compared to other populations 4,12,[48][49][50][51][52][53][54][55][56] . We observed 16.4% of classical Yq microdeletions, and these varied greatly in frequency among the populations, mainly due to the ethnic background, geographical   The AZFc partial deletions (or) gr/gr deletions (gr/gr + b2/b3 and b3/b4 inversion followed by gr/gr deletions) ( www.nature.com/scientificreports www.nature.com/scientificreports/ region or case-control selection criteria 4,48-56 . The AZFc region is extremely rich in repetitive sequences and is organized as amplicons, and therefore a number of possible partial AZFc deletions (gr/gr, b1/b2, b2/b3, b3/b4) are proposed to be important risk factors for spermatogenic failure 15,19,20,26,27,29,37,42,[57][58][59][60][61][62] . Interestingly, we identified AZFc partial deletions (gr/gr, b1/b3, b2/b3, b3/b4) in a total of 127 infertile men that accounted for 13.1%, with the impact of different DAZ and CDY1 copies deletions and their associations leading to spermatogenic failure and male infertility. However, a few partial AZFc deletion studies have failed to show any such association with male infertility 28,31,63,64 .

(d) The b3/b4 inversion followed by gr/gr deletion or vise versa
In two meta-analyses, one consisting of seven studies reported significant association of gr/gr deletions with less motile sperm with low sperm count 65 , and another comprising of 18 case-control studies also established a strong relationship between gr/gr deletion and male infertility 66 . A few independent studies have also reported that the gr/gr deletion was more common among infertile men with azoo/oligozoospermia than in men with normozoospermia, suggesting that the deletion might be a significant risk factor for spermatogenic failure 30,32,58,67,68 . However, others failed to show any phenotypic impact of gr/gr deletions on spermatogenic failure 57,69,70 . Therefore, it is extremely important to study the frequencies of AZFc partial deletions and their association with www.nature.com/scientificreports www.nature.com/scientificreports/ fertility in Indian idiopathic infertile men. Our study showed that gr/gr deletions are more frequent among men with oligozoospermia (11.4%) than azoospermia (4.6%) and than in oligoteratozoospermia (2.1%); as expected, the prevalence is very low in controls (1.53%) ( Table 1), suggesting that these partial deletions might also be a significant risk factor for spermatogenic failure in Indian idiopathic infertile men. Therefore, ethnic-specific differences in gr/gr deletion frequencies and their association with infertility are evident.
The previous studies have suggested that the gr/gr deletion frequency in the patient group was higher in the Asians (~10%) compared to the Europeans (~4.5%) 71 . Nevertheless, it is yet to be clarified whether the partial deletions (gr/gr, b1/b2, b2/b3, b3/b4) and their association is because of the lack of DAZ copies or due to other intervening genes that are also deleted. Some studies have suggested that the putative deletions of the BPY2 and CDY1 genes were associated differentially with the distinct DAZ gene copy deletions that affect sperm pathology leading to infertility 25,72 . A few studies have also reported that the gr/gr deletion was neutral because of unknown compensatory mechanisms that had rescued the deleterious gr/gr deletion effect 25,42,60 .
Previous studies have reported that the gr/gr deletion was fixed in haplogroups D2b and Q1 in the Japanese and Chinese populations, respectively 15,37,42 . In the Northern Eurasian population, the b2/b3 partial deletion was fixed with haplogroup N 16 . However, some other studies have proposed that the b2/b3 deletion is different in different haplogroups 58,69,74 . Haplogrouping of 973 infertile and 587 normozoospermic fertile men with 24 Y chromosome binary markers 40,41 in the present study revealed 8 major haplogroups, of which H1a-M82 and R1a-M17 were the two major haplogroups among both infertile and fertile men (Fig. 2). In India, the overall frequencies of haplogroups H1a-M82 and R1a-M17 were reported to be 40% and 17%, respectively 77 . Our results are also consistent with the general trend of Indian populations, where H1a-M82 is the most frequent haplogroup. These 8 haplogroups are common throughout India and are present among all the four major linguistic families 78 . Haplogroup H1a-M82 is an autochthonous haplogroup, whereas R1a-M17 is shared with the West Eurasian populations. Our previous studies have revealed that the people of Indian subcontinent are unique in their origin and differ significantly from the rest of the world in terms of their genetic affinities and disease www.nature.com/scientificreports www.nature.com/scientificreports/ susceptibility 79,80 . Therefore, heterogeneity in terms of the haplogroups observed among Indians and the Chinese are not surprising 80 .
Though our study included a substantial total sample size, subsamples such as oligozoospermic and oligoteratozoospermic patients remain small, which will have limited some statistical inferences. Even after the analysis of the Y chromosome partial deletions, the etiology remains unknown in a large proportion of the infertile men. Further analyses of genes, not only of the Y chromosome, but also of the X chromosome and the autosomes are required to understand the genetic causes of male infertility in a greater percentage of the idiopathic infertile cases.

Conclusions
Our study revealed a very high frequency of AZF deletion events in Indian infertile men (29.4%) compared to other populations. We observed 16.4% of AZF region deletions exclusively in infertile men, consisting of AZFa (3.2%), AZFb (3.5%), AZFc (6.9%), AZFbc (2.3%) and AZFabc (0.5%). However, these deletion frequencies differ greatly in different populations, mainly due to the ethnic background/case-control selection criteria. We also identified partial AZFc deletions (gr/gr, b1/b3, b2/b3, b3/b4) in 127 infertile men (13.1%). Therefore, ethnic-specific differences in the AZFc partial deletion frequencies and their association with infertility are also evident. We found that gr/gr deletions are more frequent among oligozoospermic (11.4%), than azoospermic (4.6%) or oligoteratozoospermic (2.1%) patients, and that as expected the prevalence is very low in controls (1.53%), suggesting that these partial deletions might be a significant risk factor for spermatogenic failure (low sperm counts) in Indian idiopathic infertile men. Some studies have suggested that AZFc partial deletions are fixed in specific haplogroups. However, in the present study, we found that the distribution of haplogroups was not different between cases and controls with/without deletions, and that all deletions were rare in controls, suggesting that haplogroup has no role in determining risk associations in Indian infertile men. Thus, in our study we found some AZF deletion events that explained the infertility in these idiopathic infertile men. Indian populations are unique in their origin and have been practicing endogamy for the last two thousand years, and therefore it is important to add a study of the frequencies of AZF deletions on the Y chromosome and their association with fertility in Indian idiopathic infertile men to similar studies from other parts of the world.

Materials and Methods ethical statement and samples of infertile and fertile men. The Institutional Ethical Committees
(IECs) of the participating institutes approved the study. The experiments were carried in accordance with the relevant guidelines and regulations approved for research on human samples. All the experimental protocols were approved by the IEC of the Centre for Cellular and Molecular Biology (CCMB). Before blood sample collection, the subjects underwent detailed medical and physical examinations. Informed written consents were obtained from all of 973 infertile and 587 fertile control men. The blood samples of 973 infertile men, consisting of 771 azoospermia (complete absence of sperm), 105 oligozoospermia (low sperm count) and 97 oligoteratozoospermia (low sperm count with abnormal shape and size) patients, were collected from the Genetic Clinic, Institute of Reproductive Medicine, Kolkata, India. The blood samples of the remaining 40 oligoteratozoospermic men were collected from the Infertility Institute and Research Centre, Hyderabad, India. In both the hospitals, a team of doctors (urologists and andrologists) performed detailed clinical investigations, which included semen analyses, and recorded complete case histories. In the hospitals, the blood samples were subjected to karyotyping and endocrinological assays, such as for follicle-stimulating hormone (FSH), luteinizing hormone (LH), testosterone (T), prolactin (PRL) and thyroid-stimulating hormone (TSH). Patients included in the study did not exhibit any obstructions, pelvic injury or major illness, karyotype abnormalities or endocrinological defects.
The 587 fertile normozoospermic control men with matched ethnic backgrounds had normal semen parameters (>20 × 10 6 sperm/ml semen fluid with normal motility and morphology), according to the World Health Organization guidelines 81,82 , and normal levels of inhibin B, testosterone T, LH and FSH. They volunteered themselves to be included in this study as controls. About 5.0 ml of blood were collected from 537 and 50 infertile men from 2 hospitals: a) The Genetic Clinic, Institute of Reproductive Medicine, Kolkata, India and b) The Infertility Institute and Research Centre, Hyderabad, India, respectively. In addition, all the controls had fathered at least one child, each with proven paternity by STR-based DNA fingerprinting (Profiler Plus; Applied Biosystem, Foster City, USA), and were enrolled in the study after obtaining informed written consents. DNA was isolated from all the blood samples of infertile and fertile control men using the method published elsewhere 83 . polymerase Chain Reaction (pCR). Primer sequences of the STS, and SNV markers were obtained from (www.ncbi.nlm.nih.gov/entrez/) and synthesized using an ABI394 oligo-synthesizer (Perkin Elmer, Foster City, California, USA). The Polymerase Chain Reaction (PCR) was performed in 0.2 ml thin-walled tubes using 50 ng of DNA, 10 pM of the STS primers mentioned above, 100 μM dNTPs, 10X PCR buffer containing 1.5 mM MgCl 2 , and 2 units of AmpliTaq Gold (Perkin Elmer). Amplification was carried out in a MJ Research Thermal Cycler (Waltham, MA 02451, USA) using the amplification conditions: 94 °C for 5 minutes, 35 cycles at 94 °C for 45 seconds, 60 °C for 45 seconds and 72 °C for 1 minute, followed by the final extension at 72 °C for 5 minutes. The PCR products were size fractionated using 2% agarose gel electrophoresis and detected by staining with ethidium bromide.
Identification of Y-chromosomal haplogroups. All the infertile (973) and fertile control (587) men were haplogrouped using 24 Y chromosome binary markers 40,41 . PCR was carried out for all 24 binary markers, the amplified products were then directly sequenced using Sanger sequencing, and the haplogroups were assigned based on the sequence. statistical analysis. The types of deletion combinations observed in our study were tabulated (Table 1), and comparisons were made between each category versus control using biostatistical tools available online (http:// faculty.vassar.edu/lowry/VassarStats.html). To confirm the results, statistical tests were repeated at least twice. P values less than 0.05 were considered as statistically significant changes.  TaqI-digested DNA from an infertile man with a DAZ3 deletion confirmed by hybridized with 49f probe. Although DNA of both fertile and infertile men were run in the same gel, transferred, and hybridized, they were not in the adjacent lanes. Therefore, we have cropped the images and placed the relevant adjacent for better comparison. We have provided the unedited autoradiogram as a Supplementary File.