SEARCH

SEARCH BY CITATION

Keywords:

  • azoospermia;
  • Sertoli cell only syndrome;
  • single nucleotide polymorphism;
  • spermatogenesis

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

Sertoli cell only (SCO) syndrome is the predominant histology for men with non-obstructive azoospermia (NOA) and is usually of unexplained aetiology. Studies in mouse models indicated that the X-linked gene glucocorticoid-induced leucine zipper (GILZ) is essential for survival and differentiation of spermatogonia, and meiosis. GILZ deficiency results in a rapid and progressive loss of germ cells with SCO tubules and sterility in adults. The role of GILZ in human fertility has not been examined. Here we show that GILZ is localized to spermatogonia and spermatocytes in the human testis in a pattern analogous to that seen in mice. To assess the potential for an association between GILZ variants and human infertility, we sequenced the entire protein-coding regions of the GILZ gene in 65 SCO and 87 fertile Australian men. We identified six genetic variants, three of which had not been reported previously. Three variants, 107018665 G>A, 107018485 C>G and 106959283 C>T, were found at a low frequency only in SCO men. Although none of the identified variants changed the protein code, sequence analysis indicated that two variants, 107018665 G>A and 107018485 C>G, would completely abolish the exonic splicing enhancer (ESE)-binding motifs for the splicing factors SF2/ASF and SC35 respectively. This result prompted an assessment of whether these two variants were associated with male infertility in a separate population of men. We used a PCR-based SNP detection approach to screen an additional 52 NOA and 153 fertile Australian men, and 86 SCO and 54 fertile American men. None of these men carried either of these two variants. The cumulative allelic frequency of these variants is less than 1% in SCO men and no association with fertility status was observed. Our study suggests that GILZ variants are not common causes of SCO and NOA in Australian or American men.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

Infertility affects one in 20 men in developed societies and for at least half, the underlying aetiology is not known (de Kretser, 1997; Boivin et al., 2007; McLachlan & O'Bryan, 2010). Approximately 60% of men present for infertility treatment with some form of spermatogenic failure with parameters ranging from a complete absence of spermatozoa in the ejaculate (azoospermia), reduced numbers of spermatozoa in the ejaculate compared with the majority of fertile men (oligospermia), a high percentage of abnormal sperm morphology (teratospermia), seriously compromised sperm motility (asthenospermia), or a combination of these defects. For men with at least a few spermatozoa in their ejaculate, artificial reproductive technologies such as in vitro fertilization (IVF) and intra-cytoplasmic sperm injection (ICSI), while not providing a cure, have provided a means to achieve paternity. For men with the most severe form of spermatogenic failure, wherein no germ cells are observed in the testis (termed Sertoli cell only, SCO) options are, however, limited to the use of donor spermatozoa or testicular sperm extraction (TESE) for ICSI, either by random biopsy or by micro-dissection to identify tubules more likely to contain active spermatogenesis (Su et al., 1999; Ramasamy et al., 2005).

Recent studies from a number of groups have identified glucocorticoid-induced leucine zipper (GILZ) as a critical regulator of male fertility in the mouse (Bruscoli et al., 2012; Romero et al., 2012; Suarez et al., 2012). GILZ is an X-linked gene and was first identified as a glucocorticoid-induced anti-inflammatory protein (Beaulieu & Morand, 2011). The absence of GILZ results in complete failure of spermatogenesis ascribable to extensive and progressive apoptosis of spermatogonia and meiosis arrest in prophase I, resulting in a SCO phenotype in adults (Bruscoli et al., 2012; Romero et al., 2012; Suarez et al., 2012). The findings reveal a critical role for GILZ in spermatogonia survival and differentiation, and meiosis in the mouse. On the basis of these findings, we hypothesized that GILZ is a candidate human infertility gene.

In this study, we assessed the localization of GILZ protein in the human testis and potential associations between GILZ genetic variants in infertile men with SCO and non-obstructive azoospermia (NOA).

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

Patient and control samples

All subjects gave their informed consent. The study was approved by the Human Research and Ethics Committees of Southern Health (Monash Medical Centre, Monash University), the Royal Women's Hospital (Melbourne, Victoria, Australia), Concorde Hospital (Sydney, NSW, Australia), and Weill Cornell Medical College (NY, USA). Genomic DNA samples from infertile men were obtained from the Monash Male Infertility Repository (Lynch et al., 2005; Jamsai et al., 2008, 2011; O'Bryan et al., 2012) and the Weill Cornell Male Infertility DNA Repository. Sequencing was initially done on genomic DNA from 65 Australian men with bilateral SCO as verified by testicular biopsy and assessed by a qualified pathologist, and 87 Australian men of proven fertility with a sperm concentration >40 × 106/mL (average concentration of 105 ± 45 × 106/mL). Following the detection of several potentially pathogenic SNPs, two additional groups of men were assessed, i.e. 52 Australian men with NOA as defined by the absence of spermatozoa in the centrifuged pellet from a semen sample, elevated serum FSH levels (>8 IU/L, mean ± SD = 28.9 ± 14.6) and low testicular volumes (<10 mL in each testis) (O'Bryan et al., 2012), and 153 Australian men of proven fertility (sperm concentration >40 × 106/mL, average concentration of 110 ± 60 × 106/mL); and 86 American men with SCO as defined by testis biopsy and 54 American fertile control men. Selection and exclusion criteria for these patients were as previously described (O'Bryan et al., 2012). In brief, patients with karyotype and other chromosomal abnormalities including Y chromosome deletions, congenital abnormalities, past cryptorchidism, orchitis, cancer and other systemic illnesses were excluded.

GILZ mutation screen

Primers were designed to amplify 441–500 bp fragments that cover all exon-intron boundaries of transcripts ENST00000315660 and ENST00000372397 (Supplementary Table 1). Primers were used for the amplification and sequencing of the PCR products in both sense and antisense directions. With this design, the protein-coding regions of all predicted isoforms of the human GILZ gene (ENSG00000157514) were 100% covered and sequenced. Sequencing was conducted at the Australian Genome Research Facility using the AB3730xl sequencing platform (Applied Biosystems, Mulgrave, Vic., Australia) as previously described (O'Bryan et al., 2012). Mutation detection was performed using the novoSNP software (VIB, Gent, Belgium) (Weckx et al., 2005) as previously described (O'Bryan et al., 2012).

Additional screen for the 107018665 G>A and 107018485 C>G variants

Two variants, i.e. 107018665 G>A and 107018485 C>G were further screened in 86 SCO American men, 52 NOA Australian men and 153 fertile Australian and 54 American fertile men using the Amplifuor SNP detection system (S7909; Chemicon, Temecula, CA, USA) as previously described (O'Bryan et al., 2012). Primers used are shown in Supplementary Table 2.

Defining GILZ localization in the human adult testis

Immunostaining was performed to define GILZ localization in the human adult testis. Testis biopsy was obtained with consent from a man undergoing vasectomy reversal and was processed and fixed, as previously described (Jamsai et al., 2008, 2011; O'Bryan et al., 2012). A mouse monoclonal GILZ antibody (sc-101194; Santa Cruz Biotechnology, Santa Cruz, CA, USA) was used and immunostaining was performed as previously described (O'Bryan et al., 2012).

Statistical analysis

Analysis of allele frequency was performed using a 2 × 2 contingency table analysis (http://www.graphpad.com/quickcalcs/contingency1.cfm). p values were determined using two-tailed Fisher's exact test. A value of p < 0.05 was defined as being statistically significant.

Sequence analysis and splice site prediction

Mapping of the identified variant was performed using VectorNTI 11 (Invitrogen, Mulgrave, Vic., Australia). Splicing motif analysis was assessed using HSF 2.4 (http://www.umd.be/HSF/) (Desmet et al., 2009) and NetGene 2.0 (http://www.cbs.dtu.dk/services/NetGene2/) computer modelling tools.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

GILZ localizes to spermatogonia and spermatocytes in the human testis

In the mouse, GILZ has been shown to localize to spermatogonia and primary spermatocytes (Bruscoli et al., 2012; Romero et al., 2012). The localization of GILZ has not been determined in the human testis. Thus, we performed immunostaining on an adult human testis biopsy to define the site of GILZ production during spermatogenesis. We found that GILZ localized primarily to the nucleus and cytoplasm of spermatogonia and spermatocytes up until the mid-pachytene stage (Fig. 1A–C). GILZ was also observed within the nuclei of peritubular cells and at low levels in round spermatids (Fig. 1A and C). The positive staining was absent when the GILZ antibody was omitted (Fig. 1D).

image

Figure 1. Glucocorticoid-induced leucine zipper (GILZ) localizes to spermatogonia and early spermatocytes in the human testis. Immunostaining of GILZ in the adult human testis reveals GILZ-positive staining (red signal) in spermatogonia and spermatocytes. (B) DAPI (marker for DNA) staining in the same fields as image A. (C) A high magnification portion of image A. SG, spermatogonia; SP, spermatocyte; R, round spermatid; eSp, elongated spermatid; PT, peritubular cells. (D) GILZ-positive staining was eliminated when the primary antibody was omitted. Scale bars: 25 μm.

Download figure to PowerPoint

Six genetic variants identified in the GILZ gene

The analysis of the Gilz knockout mouse models confirmed the critical role GILZ plays in male fertility (Bruscoli et al., 2012; Romero et al., 2012; Suarez et al., 2012). As such, we hypothesized that GILZ is a candidate human infertility gene. To define a potential association between genetic variants in the GILZ gene with human infertility, we sequenced the human GILZ gene in groups of fertile men and infertile patients with phenotypic defects analogous to the Gilz knockout mice, i.e. men who produced no spermatozoa in their ejaculates (NOA) and showed a complete loss of germ cells (SCO) on histological evaluation. In the initial mutation screen, we sequenced the entire protein-coding region and the exon-intron junctions of the GILZ gene in 65 SCO and 87 fertile Australian men. Based on the Ensembl database (release 69), the human GILZ gene is predicted to give rise to multiple transcripts that differ in promoter usage (thus different in 5′UTR) and 3′UTR. We designed our sequencing to cover transcripts ENST00000315660, which encodes a 200 amino acid protein, and ENST00000372397, which encodes a 134 amino acid protein. With this design, primers covered 100% of protein-coding regions of all predicted isoforms.

We identified six genetic variants including three variants located in 5′UTR (107019084 T>C, 107019075 C>T, 107018665 G>A), one variant (107018485 C>G) located in the protein-coding region of exon 2, which did not change protein sequence (synonymous) of transcript ENST00000315660, and two variants (106959283 C>T and 106956638 C>T) located in the overlapped region of both transcripts, i.e. intron 2 of transcript ENST00000315660/intron 1 of transcript ENST00000372397 and exon 4 (3′UTR) of transcript ENST00000315660/exon 3 (3′UTR) of transcript ENST00000315660 (Table 1 and Fig. 2A). Three variants, i.e. 106959283 C>T, 107019075 C>T and 107019084 T>C had not been previously reported (Table 1). All six variants were found at low frequency. Three variants were only found in SCO men. These were 107018665 G>A, 107018485 C>G and 106959283 C>T (Table 1).

Table 1. GILZ genetic variants in the Australian SCO and fertile men
SNP positionaSNP locationSNP typeFrequency of SNP
SCO men (= 65)Fertile men (= 87)
WTMUTWTMUT
  1. a

    Reference sequence: ENSG00000157514 (Ensembl release 69, October 2012): Chromosome X: 106,956,451-107,020,572. As an X-link gene, WT (wild-type) refers to the allele that is identical to the reference sequence, e.g. 107019084 T>C, WT = T. MUT (mutant) refers to the variant allele, e.g. 107019084 T>C, MUT = C.

  2. b

    Previously reported SNPs.

  3. c

    SNPs selected for follow-up study.

107019084 T>CExon 1 (5′ UTR)Non-coding650861
107019075 C>TExon 1 (5′ UTR)Non-coding650861
107018665 G>Ab,c (TMP_ESP_X_107018665)Exon 2 (5′ UTR)Non-coding641870
107018485 C>Gb,c (rs138575903)Exon 2Synonymous (leucine 55)641870
106959283 C>TIntron 2Non-coding641870
106956638 C>Tb (rs55822441)Exon 4 (3′ UTR)Non-coding650843
image

Figure 2. Location of six glucocorticoid-induced leucine zipper (GILZ) genetic variants identified in SCO and NOA Australian men. (A) Schematic of two human GILZ transcripts ENST00000315660 and ENST00000372397. Unfilled boxes represent untranslated region (UTR), filled boxes represent protein-coding exons, arrows indicate variant positions in relation to the gene, previously reported variants are indicated by * and variants selected for following up study are indicated by #. (B) Bioinformatics prediction of the potential splicing effect resulted by variant 107018665 G>A. (C) Bioinformatics prediction of the potential splicing effect resulted by variant 107018485 C>G. Wild-type and variant alleles are shown as in a reverse stand (open reading frame).

Download figure to PowerPoint

Glucocorticoid-induced leucine zipper is an X-linked gene, thus all the identified variants are in homozygous state. To assess the potential effects of all identified variants on the expression of GILZ gene, we performed splicing prediction using the computer modelling tools HSF 2.4 and NetGene 2.0. First, we used the HSF tool to determine the effects on splicing factor-binding sites. These included (i) the analysis of exonic splicing enhancer (ESE) for the major splicing factors SF2/ASF, SC35, SRp40 and SRp55, (ii) defining the defect on hnRNP motifs, (iii) defining the silencer motifs (ESS, EIEs, IIEs) and (iv) defining potential new branch points and cryptic donor and acceptor splice sites. Our analysis revealed that two variants found in only SCO men, i.e. 107018665 G>A and 107018485 C>G were predicted to effect the splicing factor-binding site. Neither branch points nor cryptic donor/acceptor splice sites were identified in these six variants. Consistent with the above results, the analysis using NetGene 2.0 tool, which determines the presence of branch points and cryptic donor/acceptor splice sites of all six variants, showed that none of these variants would create branch points or cryptic donor/acceptor splice sites.

The substitution of G by A in the 107018665 G>A variant would completely abolish the ESE binding motif for the splicing factors SF2/ASF (Table 2 and Fig. 2B) and the substitution of C by G in the 107018485 C>G variant would completely abolish the ESE binding motif for the splicing factor SC35 (Table 2 and Fig. 2C). SF/ASF2 and SC35 are the major spliceosomal proteins that facilitate pre-mRNA splicing (Zuo & Manley, 1993; Tacke & Manley, 1995; Liu et al., 2000). A change in the sequence of a splicing factor-binding motif within pre-mRNAs would be expected to have a major impact on splicing.

Table 2. The effects of variants 107018665 G>A and 107018485 C>G on binding sites for the splicing factors SF2/ASF and SC35
VariantLinked SR proteinReference motif (value 0–100)Threshold valuesEffect of the variant on the SR protein motif
  1. Reference motif refers to wild-type sequence. Reference motif values refer to the degree of sequence conservation of the reference (wild-type) sequence to common ESE motifs. High reference (wild-type) motif values above the threshold indicate that the wild-type sequence is closely related to the most highly conserved ESE sites. Site broken (-100) indicates that the presence of the variant completely abolishes the splicing factor-binding site.

107018665 G>ASF2/ASFCTCCGAG (72.69)70.51Site broken (-100)
107018485 C>GSC35GGTCTTTG (81.25)75.05Site broken (-100)

These results prompted us to assess whether the 107018665 G>A and 107018485 C>G variants are associated with male infertility. We screened for the presence of these two variants using a PCR-based SNP detection method in an additional 52 NOA and 153 fertile Australian men, and 86 SCO and 54 fertile American men. We found that none of these men carried either of these two variants (Table 3).

Table 3. Frequency of the 107018665 G>A and 107018485 C>G variants
Subject group/genotypeSNP 107018665 G>ASNP 107018485 C>G
WT (G)MUT (A)WT (C)MUT (G)
  1. Values within parentheses are expressed in percentage. Variant frequency of the Australian men derived from direct DNA sequencing data shown in Table 1 (n = 65 SCO and 87 fertile men) and SNP-based PCR detection method (= 52 NOA and 153 fertile men).

Australian SCO + NOA (n = 65 + 52 = 117)116 (99.15)1 (0.85)116 (99.15)1 (0.85)
Australian fertile men (n = 87 + 153 = 240)240 (100)0 (0)240 (100)0 (0)
American SCO (= 86)86 (100)0 (0)86 (100)0 (0)
American fertile men (n = 54)54 (100)0 (0)54 (100)0 (0)

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

The complexity of fertility regulation is highlighted by the identification of over 400 genes that are critical to male and/or female fertility in the mouse (Matzuk & Lamb, 2008) and predictions that several thousand more genes are likely to be involved. Despite a continuing success in defining novel fertility genes using mouse models, only a modest number of these genes have been defined as critical for fertility in humans (Aston & Carrell, 2012). The major obstacles in the translation of mechanistic data derived from the mouse into a clinically relevant outcome for humans include an extremely high degree of genetic heterogeneity in most human populations and the limited availability of clinically well-defined infertile samples and samples from proven fertile control men.

In this study, we have used two such high quality repositories to examine the potential association between defects in the GLIZ gene and human infertility. We have confirmed GLIZ localizes to spermatogonia and spermatocytes in the human testis in a pattern that is analogous to that seen in the mouse. Sequencing of the protein-coding region of GILZ in Australian SCO and fertile men identified a total of six genetic variants, of which three variants were novel. Moreover, three variants, i.e. 107018665 G>A, 107018485 C>G and 106959283 C>T, were found in low frequency only in SCO men. Although none of the identified variants resulted in a change in the protein code, sequence analysis and splicing predictions indicated that variants 107018665 G>A and 107018485 C>G would completely abolish ESE-binding motifs for the splicing factors SF2/ASF and SC35, respectively, and are thus likely to be highly pathogenic. These results prompted us to further assess if variants 107018665 G>A and 107018485 C>G are a common cause of SCO. We subsequently used a PCR-based SNP detection approach to screen additional 52 NOA and 153 fertile Australian men and 86 SCO and 54 fertile American men. None of these men carried either of the two variants. As such the overall allelic frequency of these variants is less than 1% in SCO men and no association with infertility status was observed (p = 0.33). Our study suggests that GILZ variants are not a common cause of SCO and NOA in Australian or American men.

These results do not, however, preclude the possibility that 107018665 G>A and 107018485 C>G are more frequent in other less genetically diverse population or that less genetically disruptive SNPs may lead to hypospermatogenesis in some patients. Assuming that human GILZ is biologically similar to mouse GILZ, we would, however, still anticipate that even hypomorphic mutations would lead to a progressive decline in testicular germ cell content, ascribable to a decline in the stem cell population, and ultimately lead to SCO. Knowledge of hypomorphic alleles would, however, allow for the screening of women for carrier status and/or the collection of spermatozoa or testicular tissue from young men who carry such mutations. Gametes could be stored until such time when children are desired.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

This work was partially supported by grants from Monash IVF (M.K.O'B., D.J., and R.I.M.L.) and the Theresa and Frederick Dow Wallace Fund of New York Community Trust (P.N.S.). M.K.O'B. and R.I.M.L. are the recipients of Fellowships from the National Health and Medical Research Council (NHMRC). We thank Jo Merriner, Stephanie Smith and Anna Mielnik for technical assistance.

Disclosure

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

D.J., A.G., P.J.S., P.N.S. E.M. and M.K.O'B. have nothing to declare. R.I.M.L. has a consultancy arrangement with, and equity interest in, Monash IVF.

Authors contributions

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

D.J. and M.K.O'B. designed the study. A.G. conducted the experiments. D.J. M.K.O'B., P.J.S., P.N.S. R.I.M.L and E.M. analysed the data. D.J. M.K.O'B. and E.M. wrote the manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Disclosure
  9. Authors contributions
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
andr76-sup-0001-TableS1-S2.docxWord document114K

Table S1. Primers used for PCR amplification and sequencing of the human GILZ gene.

Table S2. Primers used for screening of the 107018665 G>A and 107018485 C>G variants.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.