• DNA methylation;
  • epigenetic;
  • non-obstructive azoospermia;
  • obstructive azoospermia;
  • testicular tissue


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

The objective of this study was to assess genome-wide DNA methylation in testicular tissue from azoospermic patients. A total of 94 azoospermic patients were recruited and classified into three groups: 29 patients presented obstructive azoospermia (OA), 26 displayed non-obstructive azoospermia (NOA) and successful retrieval of spermatozoa by testicular sperm extraction (TESE+) and 39 displayed NOA and failure to retrieve spermatozoa by TESE (TESE−). An Illumina Infinium Human Methylation27 BeadChip DNA methylation array was used to establish a testicular DNA methylation pattern for each type of azoospermic patient. The OA and NOA groups were compared in terms of the relative M-value (the log2 ratio between methylated and non-methylated probe intensities) for each CpG site. We observed significantly different DNA methylation profiles for the NOA and OA groups, with differences at over 9000 of the 27 578 CpG sites; 212 CpG sites had a relative M-value >3. The results highlighted 14 testis-specific genes. Patient clustering with respect to these 212 CpG sites corresponded closely to the clinical classification. The DNA methylation patterns showed that in the NOA group, 78 of the 212 CpG sites were hypomethylated and 134 were hypermethylated (relative to the OA group). On the basis of these DNA methylation profiles, azoospermic patients could be classified as OA or NOA by considering the 212 CpG sites with the greatest methylation differences. Furthermore, we identified genes that may provide insight into the mechanism of idiopathic NOA.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

In infertile men, the incidence of azoospermia (i.e. the absence of spermatozoa after centrifugation of total ejaculate in at least two semen analyses with a 3-month interval (WHO, 2010)) is 10–20%. The condition affects 1% of the general male population (Bhasin et al., 1994; Foresta et al., 1995). Nevertheless, medical procedures such as testicular sperm extraction (TESE) can be envisaged in azoospermic patients. If spermatozoa can be retrieved by TESE (Schoysman et al., 1993), it may be possible to offer patients access to assisted reproduction techniques based on intracytoplasmic sperm injection (ICSI) (Devroey et al., 1995).

One can define two types of azoospermia:

  1. Obstructive azoospermia (OA) results from an obstruction at some position along the genital tract (the epididymis, the vas deferens or, in very rare cases, the ejaculatory duct). In the literature, OA accounts for about 40% of all azoospermia (Jarow et al., 1989; Thonneau et al., 1991). OA is mainly because of inflammation, vasectomy or the congenital bilateral absence of the vas deferens (CBAVD), which is frequently induced by mutation of the gene for the cystic fibrosis transmembrane receptor (CFTR). After a testicular biopsy, a histological examination can establish whether spermatogenesis is quantitatively and qualitatively normal (according to the European Association of Urology guidelines) (Huyghe et al., 2008; Robin et al., 2010; Oates, 2012) and thus confirm the diagnosis. In more than 90% of cases (Nicopoullos et al., 2004), sperm retrieval is successful.
  2. Non-obstructive azoospermia (NOA) corresponds to the failure or dysfunction of spermatogenesis. Although some aetiologies are commonly observed (such as an abnormal karyotype (Klinefelter syndrome or isochromosome Yp), chromosome Yq microdeletions, a history of cryptorchidism, chemo/radiotherapy, a history of inguinoscrotal surgery or testicular tumours, etc.), most cases of NOA are idiopathic. Histopathological analysis can indicate the failure of spermatogenesis at various stages (in decreasing order of severity): Sertoli cell only (SCO) syndrome (no spermatozoa), pre- or post-meiotic germ cell maturation arrest, hypospermatogenesis (few spermatozoa) and mixed histopathological patterns. In contrast to the situation in OA, sperm retrieval is successful in only about 50% of NOA patients (Nicopoullos et al., 2004; Carpi et al., 2009). At present, there are no clinical or biological criteria that enable the accurate prediction of successful retrieval – although scoring criteria have recently been suggested (Boitrelle et al., 2011).

Hence, the failure of spermatogenesis cannot currently be explained in all cases (Huyghe et al., 2008) – even though a number of epigenetic models have recently emerged as potential aetiologies. In fact, it has been reported that sperm DNA methylation of a number of imprinted genes is abnormal or aberrant in azoospermic men (Marques et al., 2004; Kobayashi et al., 2007; Filipponi & Feil, 2009; Boissonnas et al., 2010). Methylation defects have been also observed in testicular cells (i.e. spermatozoa isolated by micromanipulation from testicular biopsies or by vasectomy reversal) from azoospermic patients (Marques et al., 2010; Minor et al., 2011). Furthermore, overall sperm DNA methylation levels are lower in infertile men than in control patients – although this finding is still subject to debate (Benchaib et al., 2003; Aoki et al., 2006).

In contrast to genetic defects (which alter the DNA sequence itself), epigenetic defects consist of variation in chromatin condensation that alter DNA expression during the development and cell proliferation (Strahl & Allis, 2000; Jenuwein & Allis, 2001; Jaenisch & Bird, 2003). DNA methylation (the addition of a methyl group to C5 on cytosine) is the best-studied epigenetic mechanism. In epithelial cells, for example, 80% of cytosines in consecutive CpG dinucleotides are methylated (Beck & Rakyan, 2008). In contrast, cytosines in CpG islands (CGIs: regions with a high frequency of CpG sites) are mainly non-methylated. High-density CGI regions co-localize with 72% of transcriptionally active gene promoter regions (Saxonov et al., 2006). Like genetic changes, epigenetic changes can be stable and heritable.

By using an Illumina Infinium Human Methylation27 BeadChip (Illumina Inc, San Diego, CA, USA) DNA methylation array, we sought to define DNA methylation profiles in testicular tissue by comparing NOA patients with OA patients with normal spermatogenesis (included as controls). We hypothesized that these data would improve our understanding of the mechanism(s) underlying NOA.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information


Ninety-six patients were recruited through the Andrology Clinic at Poissy Saint-Germain Hospital (Poissy, France) (see Table 1). All patients presented azoospermia (i.e. the absence of spermatozoa after centrifugation of total ejaculate in two semen analyses 3 months apart (WHO, 2010)). Regardless of the aetiology, we offered all patients access to TESE. Prior to TESE, each patient underwent a comprehensive andrologic evaluation to determine the aetiology of OA or NOA; we recorded the body mass index, smoking habit, medical history, testicular volume (according to an ultrasound assessment), epididymis and vas deferens characteristics on palpation and, lastly, the endocrine profile (serum FSH, inhibin B and testosterone). If indicated, ultrasound imaging of the urogenital tract and testis was performed to rule out cancer (in patients with a history of cryptorchidism and suspected NOA, for example) or to identify the obstructed portion of the genital tract (in cases of suspected OA). The results of the testosterone assay were used to check on the innocuousness of any previous surgery and/or rule out hypogonadotrophic hypogonadism.

Table 1. Patient classification
GroupTESE resultsEtiology of azoospermiaNumber
  1. Hypo, hypospermatogenesis; SCO, sertoli cell only; CBAVD, congenital bilateral absence of vas deferens; CFTR, cystic fibrosis transmembrane receptor; TESE, testicular sperm extraction.

Obstructive azoospermiaPositiveTotal29
CBAVD with CFTR mutations14
Non-obstructive azoospermiaPositiveTotal26
Hypo and SCO11
Hypo and spermatogenesis arrest3
Spermatogenesis arrest6

For patients with normal testicular volumes and a normal endocrine profile, OA was suspected. In a second stage, epididymal markers (α-glucosidase and l-carnitine), seminal vesicle markers (fructose) and prostate markers (citrate, zinc, phosphatase acid) were assayed, to support a diagnosis of OA. Moreover, an elastase assay was used to determine whether OA was congenital or because of infection. In a third and final stage, a karyotype analysis and screening for CFTR mutations were performed (Huyghe et al., 2008; Robin et al., 2010).

We defined isolated OA (regardless of the presence or absence of CFTR mutations) as the presence of at least one of the following factors:

  1. a previous history of male accessory gland infection (including clinical symptoms of inflammation of the prostate gland, seminal vesicles, vas deferens and/or epididymis) (Marconi et al., 2009).
  2. suggestive clinical signs: CBAVD (vas deferens not obvious or palpable), normal testicular volumes (>15–16 mL), dilation of the seminal ducts upstream of the site of obstruction (vas deferens not obvious or palpable).
  3. depressed semen markers, as a guide to the obstruction site.
  4. a normal hormone profile (including FSH and inhibin B).

If OA could be excluded, NOA was suspected. Semen marker assays, genetic counselling, karyotyping and screening for Yq microdeletions (i.e. DAZ gene deletion) were performed.

The project was approved by the local institutional review board and all patients gave their written, informed consent to participation.


Testicular biopsy collection

Testicular sperm extraction was performed using the ‘window’ technique under general anaesthesia (Tournaye et al., 1997). A vertical, 3 cm incision was made through the anterior scrotal skin, dartos and tunica vaginalis at the upper pole of the testis (near the head of the epididymis) and the testis was exteriorized. This enabled assessment of the testicular volume and consistence and the potential presence of nodules and/or a dilated epididymis. A longitudinal, 1 cm incision in the tunica albuginea was then made. Using sharp scissors, several tissue samples were removed from the middle, contralateral, lower and upper poles of the testis. Each biopsy was about 5–10 mm in length (weight: about 200 mg) and contained about 25–30 tubular cross sections (which should be representative of the testis as a whole). These samples were promptly placed in flushing medium (Origo, France) pre-warmed to 32 °C. The different incisions were closed with non-absorbable suture thread.

The biopsy samples were then divided into two portions, as previously described (Albert, 2005):

  1. One portion was used as the source of spermatozoa for ICSI. This sample was cut up with a scalpel in 500 μL of flushing medium. When the spermatozoa were retrieved (i.e. in ‘TESE + subjects’), they were cryopreserved prior to subsequent ICSI. In our centre, ICSI is never performed on the same day as TESE. When spermatozoa were not observed, the biopsy was classified as ‘TESE−’. In all cases, the seminal tubule remnants were centrifuged and stored in a cryostraw in liquid nitrogen, as described elsewhere (Ghalayini et al., 2011). These remnants were used for DNA extraction in this study.
  2. The portion part of the testicular biopsy was placed in Bouin solution for later examination, with a view to establishing the histopathological profile and confirming the aetiology of azoospermia.

Patient classification (Table 1)

Patients were selected and grouped according to their clinical characteristics, histopathological analysis, hormonal status, semen markers and TESE results. Patients with Klinefelter syndrome or Yq microdeletions were excluded. All included patients had a normal karyotype. Two samples were excluded from our analysis after data normalization and quality control, and so our final analysis included 94 patients. The latter was divided into three groups, according to the azoospermia aetiology and the TESE results: (i) OA patients (n = 29), (ii) NOA patients with spermatozoa retrieved by TESE (the TESE+ group; n = 26), and (iii) NOA patients with no spermatozoa retrieved by TESE (the TESE− group; n = 39). The mean patient ages were 35 ± 1.35 ± 1 and 34 ± 1 for the OA and NOA groups respectively.

DNA extraction

After the testicular biopsy, only the remnants of seminal tubules (stored in a cryostraw at −196 °C in liquid nitrogen) were used for DNA extraction with a Qiagen QIAmp DNA minikit (Qiagen, Inc., Valencia, CA, USA). After complete defrosting at room temperature, the testicular tissue sample was placed in a microcentrifuge tube with medium (FertiCult flushing medium or DMSO; FertiPro N.V., Beernem, Belgium) and the supernatant was carefully withdraw. Next, 180 μL of Buffer ATL and 20 μL of proteinase K were added. The sample was incubated at 56 °C for 2 h until cell lysis was complete. After the addition of 200 μL of Buffer AL, the mixture was incubated at 70 °C for 10 min. Next, 200 μL of cold ethanol (100%) was added to precipitate the DNA. The DNA was purified through a spin column with Buffer AW1 and AW2, according to the manufacturer's instructions. The DNA filtrate was recovered with 35 μL of Buffer AE and the extracted DNA was stored at −20 °C.

The Illumina Infinium methylation assay

We used the Illumina Infinium Human Methylation27 BeadChip DNA methylation array, which analyses 27 578 CpG sites (spanning 14 495 genes of all types and roughly half the human genome). All CpG sites were located in the proximal promoter regions of genes. The distance to the transcription start site varies from 0 to 1499 bp, with a mean ± SD value of 389 ± 341 bp.

Bisulfite conversion was performed with an EZ DNA Methylation Kit (Zymo Research Corp., Irvine, CA, USA) in a 96-well plate, according to the manufacturer's instructions. The methylation assay was performed according to the Illumina protocol via a manual process. Only non-methylated cytosines were converted into uracil. Bisulfite-converted sites and non-converted (i.e. methylated) sites were analysed simultaneously by hybridization to site-specific probes attached to two types of beads coupled to different fluorochromes: one for non-methylated cytosine sites and one for methylated cytosine sites. The Illumina Infinium Human Methylation27 BeadChip especially targets 14 475 reference sequences and 12 833 well-annotated genes described in the NCBI CCDS database (build36). Statistically, all CpG sites were analysed 20 times per slide. Quality control was performed to remove poor-performing CpG sites and samples, as determined by each detected p-value. This metric is defined as 1 minus the p-value computed from the background model and corresponds to the likelihood that the signal was distinguishable from negative controls.

Statistical analysis

Data normalization

In the current Infinium assay design, some of the CpG sites are measured in the red channel (when the final, extended base is A or T) and others are measured in the green channel (when the final, extended base is C or G). To remove the sources of bias usually observed in these Illumina channel data, methylated and non-methylated probes intensities were normalized. We applied the most popular quantile normalization method, in which one assumes that the intensity distribution of the pooled methylated and non-methylated probes is essentially the same in all the various samples. The first step in normalization was colour balance adjustment, followed by background level correction.

Differential analysis

Du et al. (2010) have shown that the M-value (the log2 ratio of methylated and non-methylated probe intensities) is approximately homoscedastic across the entire methylation range and is more statistically valid than the commonly used β-value (where methylation levels are quantified using the ratio of intensities for methylated probes and the sum of both methylated and non-methylated probes) in a differential statistical analysis. Hence, differences in methylation between the OA group and NOA group (as M-values) were tested by computing the moderated t-statistic (Smyth, 2004) for each probe.

To adjust for multiple testing and to control the false discovery rate (FDR), we applied the Benjamini–Hochberg–Yekutieli method and a permutation test (run on 100 000 iterations). However, biological interpretation of the β value is more intuitive than that of the statistically valid M-value. We therefore performed a cluster analysis (according to Ward's method) on the β value of the CpG sites selected by the differential statistical analysis. Statistical analyses were performed using R 2.13.1 ( and Bioconductor 2.8 software (Gentleman et al., 2004).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Genome-wide DNA methylation profiling identifies methylated genes in testicular tissues

According to our quality control method, 641 CpG sites (corresponding to 293 genes) had a median intragroup detection p-value of less than 0.05 and were subsequently removed from the analysis. Two samples were excluded from our analysis after data normalization and control data.

The NOA and OA groups differed significantly in terms of their respective DNA methylation profiles at over 9000 CpG sites, with an FDR below 0.0001. The absolute difference in M-values between OA and NOA was higher than 1 for a total of 2787 CpG sites, higher than 2 for 1324 CpG sites and higher than 3 for 212 CpG sites (corresponding to 194 genes) (Table S1). For the TESE+ and TESE− NOA groups as a whole, the DNA methylation pattern revealed 78 hypomethylated CpG sites (where the differences in M-values were negative, corresponding to 68 hypomethylated genes) and 134 hypermethylated CpG sites (where the differences in M-values are positive, corresponding to 126 hypermethylated genes), when compared with the OA group. The distribution by general gene function is presented in Table 2. Forty-seven of the 194 genes had an identified or hypothetical testicular function and 39 of the 47 were hypermethylated in the NOA group. Of these 39 genes, 14 were testis-specific (Table 3).

Table 2. List of 194 differentially methylated genes involved in azoospermia according to our results
FunctionGenes numberHypomethylatedHypermethylated
  1. ORF, open reading frame.

Expression regulation541341
Cell proliferation and differentiation321616
Hypothetical function protein or ORF30228
Molecular transport19514
Signal transduction19118
Apoptosis regulation1459
Immune response1064
Post-translational modifications413
With identified or hypothetical testicular function47839
Testicular specific14014
Testicular expression20812
Gonad specific202
Supposed testicular expression505
Hypothetical function606
Table 3. Genes testicular specific
SymbolNameChromosome localisationlogFCNull mice described
TUBA4A or H2-ALPHATubulin, Alpha-4a2q353,10No
SPATA16Spermatogenesis-Associated Protein 163q26.313,62No
PDHA2Pyruvate Dehydrogenase, Alpha-24q22.33,82No
DDX4 or VASADEAD (Asp-Glu-Ala-Asp) Box Polypeptide 45q11.23,29No
DNAH8Dynein, Axonemal, Heavy Chain 86p21.23,99No
ELOVL2Elongation of Very Long Chain Fatty Acids-Like 26p24.23,17Zadravec et al. (2011)
HIST1H1THistone Gene Cluster 1, H1 Histone Family, Member T6p223,24Fantz et al. (2001)
ANKRD7Ankyrin Repeat Domain-Containing Protein 77q31.313,84No
TXNDC3 or NME8Thioredoxin Domain-Containing Protein 37p14.13,46No
PRKACGProtein Kinase, Camp-Dependent, Catalytic, Gamma9q133,12No
RNF17Ring Finger Protein 1713q12.123,14Pan et al. (2005)
TCEB3BTranscription Elongation Factor B Polypeptide 3B (Elongin A2)18q21.13,11No
MAGEC2Melanoma Antigen Family C, 2Xq27.23,01No
PAGE1P Antigen Family, Member 1Xp11.233,29No

Patient clustering

Lastly, by considering only the 212 CpG sites with an absolute M-value difference higher than 3 when comparing the OA and NOA groups, we performed a patient clustering analysis (Fig. 1). Two major branches were defined by the methylation data and correlated well with the type of azoospermia. The first branch (Br1) contained a total of 33 patients: 26 OA patients (n = 26 out of 29; 89.7%) and 7 NOA patients (n = 7 out of 65; 10.8%). Six of the seven NOA patients were TESE+. The TESE− NOA patient had testicular atrophy.


Figure 1. Hierarchical clustering of methylation data from testicular samples analysed in obstructive azoospermic (OA) patients (control, in yellow) and non-obstructive azoospermic (NOA) patients with sperm retrieval by testicular biopsy (TESE+, in light blue) or no sperm retrieval (TESE−, in dark blue). Columns represent samples and rows represent CpG sites. A heat map showing relative methylation differences at the 212 most significantly different CpG sites presented in the clustering dendrogram (red (>1.7) indicates greater methylation, green (<-2) indicates less methylation and black indicates no significant differences in the DNA methylation level when comparing OA controls and NOA patients). The methylation profiles of 212 CpG sites from human testicular biopsies from 94 individuals were clustered using Ward's algorithm to compute the distance between clusters. Two major branches were defined by our methylation data and were correlated with the type of azoospermia. Branch 1 contained 26 of the 29 OA patients and branch 2 containing 58 of the pooled 65 NOA patients. Branch 2 had two sub-branches.

Download figure to PowerPoint

The second branch (Br2) contained 61 patients in total: 3 OA patients (10.3%) and 58 NOA patients (89.2%). Branch 2 could be sub-divided into two further branches (denoted as a and b). Branch 2a did not have a particular profile and the β value (close to 0 for all genes) was quite inhomogeneous. The branch contained three OA patients, nine TESE+ NOA patients and six TESE− NOA patients and did not clearly identify an azoospermic category. For all six TESE− patients included in Br2a, azoospermia was because of meiotic arrest – an aetiology that was never observed in Br2b. Br2b contained 32 of the 39 TESE− patients and 11 TESE+ patients.

For 97% (n = 32 of 33) of the patients in Br1, spermatozoa were always retrieved by TESE. Spermatozoa were retrieved for 67% (n = 12 out of 18) of the patients in Br2a and for only 26% (n = 11 of 43) of the patients in Br2b. When considering only NOA patients, the retrieval rates were 85.7% (6 of 7), 60% (9 of 15) and 26% (11 of 43) in branches 1, 2a and 2b respectively.

Genome-wide DNA methylation profiling in SCO vs. pachytene maturation arrest

Using the same methodology, we compared cases of SCO with cases of pachytene maturation arrest. The two groups differed significantly in terms of the DNA methylation profile, with differences at 3206 CpG sites (yielding an FDR below 0.0001).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information

Here, we report on the first study of testicular DNA methylation patterns (using an Illumina Infinium Human Methylation27 BeadChip DNA methylation array) in the remnants of biopsy samples from 94 azoospermic patients.

The methylation profile for over 9000 of the 27 578 screened CpG sites differed significantly when comparing OA and NOA patients, with an FDR below 0.0001. Considering the high number of genes, this result confirms the presence of epigenetic pattern differences between NOA and OA patients. However, as we used post-TESE seminal tubule remnants for this study, the methylation profile variations might also be because of variations in the proportion of each type of testicular cell. Although this limitation might conceivably modify the results for small M-value differences (i.e. values below 1), this should not be the case for the M-values higher than 2 (i.e. with ratio differences of up to 5) observed for 1324 loci. These M-values cannot be explained by variations in the proportion of cells and strongly suggest the presence of an epigenetic modification in NOA patients. Given this apparent difference between OA and NOA patients, we performed a cluster analysis by referring only to loci with an M-value difference greater than 3 (i.e. with ratio differences of up to 8). This novel clustering by epigenetic profile generated a classification of azoospermic patients that closely matched the clinical subgroup distribution. The dendrogram had three branches. The Br1 and Br2b methylation patterns were ‘mirror images’; Br1 included 26 of the 29 OA patients and Brb2 contained 32 of the 39 TESE− NOA patients. Of the seven TESE− NOA patients not included in Br2b, six presented with meiotic arrest; this observation strengthens the hypothesis whereby the latter's epigenetic profile is quite similar to control profiles, and differs with SCO patients. This difference was confirmed by examining the methylation profiles in SCO (n = 20) and pachytene maturation arrest (n = 6); 3206 of the 27 578 screened CpG sites differed significantly when comparing these (small) groups. The aetiology of NOA, for meiotic arrest is thus probably because of a gene defect and not an epigenetic defect. In the last of the seven cases, azoospermia was because of testicular atrophy and no hypothesis could be generated. When considering only TESE− NOA patients lacking germinal cells, all but one were included in the Br2b branches associated with the opposite methylation profile to that of OA patients. The TESE+ NOA patients were spread throughout the dendrogram. The TESE+ patients with pure hypospermatogenesis were generally found in Br1 (i.e. with OA patients). The TESE+ NOA patients with partial depletion of germinal cells and many SCO tubules were preferentially located in Br2b (i.e. with the TESE− patients). Lastly, the 18-patient branch Br2a was quite heterogeneous, with three OA patients, six TESE− NOA patients with meiotic arrest and nine TESE+ NOA patients. On the basis of our methylation data, we suggest that a non-obstructive component contributed to the aetiology of the three OA patients.

Of the 194 genes with an M-value difference greater than 3, 54 regulated protein expression, 32 were involved in cell proliferation or differentiation and 14 were involved in apoptosis. A total of thirty genes were classified as open reading frames or as encoding hypothetical proteins. Other genes (n = 64) were involved in molecular transport, the immune response, signal transduction, post-translational modifications or metabolism.

Forty-seven of these genes had an identified or hypothetical testicular function and 14 appeared to have a specific action in testicular tissue. Murine knockout models had been created for three of these genes and male animals were fertile in one case (HIST1H1T (Fantz et al., 2001)) and infertile in the two other cases (ELOVL2 (Zadravec et al., 2011) and RNF17 (Pan et al., 2005)).

The array-based DNA methylation assay used in this study has also been applied to compare sperm DNA methylation in (i) men with abnormal protamine P1/P2 ratios, (ii) men having undergone IVF with poor embryo quality, and (iii) fertile men with normal semen parameters (Aston et al., 2011). A ‘control’ methylation pattern was observed in 25 of the 28 patients. Only the remaining three patients displayed an altered methylation pattern across a large number of CpG sites. In contrast, our present results showed that over 9000 genes differed when comparing OA and NOA patients. This disparity is probably because of the nature of the testicular tissue analysed and the various aetiologies of azoospermia. In our study, a mixture of several types of cell was analysed. Altered spermatogenesis was based on clinical and biological examinations and the retrieval of spermatozoa by TESE, rather than on in vitro fertilization outcomes or the P1/P2 ratio.

In the second study (Pacheco et al., 2011) comparing infertile patients with fertile patients, 9189 CpG sites with altered methylation were associated with low-motility sperm samples. The great majority of these sites (80%) were hypomethylated (vs. just 40% in this study). However, we only considered loci with a FDR below 0.0001, rather than the value of 0.05 previously reported. Furthermore, the fact that spermatozoa were mainly transcriptionally inactive may explain the value of 80% (Li et al., 2005).

In fact, recent research has demonstrated the existence of epigenetic abnormalities in the spermatozoa of infertile men. Abnormal DNA methylation of imprinted genes has been linked to spermatogenesis failure in patients with moderate oligozoospermia (Kobayashi et al., 2007; Marques et al., 2008; Boissonnas et al., 2010). However, abnormal DNA methylation at imprinted genes in testicular spermatozoa from patients with azoospermia did not appear to be associated with NOA or OA (Marques et al., 2010).

Analysis of DNA methylation of imprinted genes in the spermatozoa retrieved from male patients with NOA and OA might enable us to identify the potential consequences of abnormal DNA methylation, such as spermatogenesis failure in NOA male patients or obstruction in OA male patients. In a recent study (Minor et al., 2011), spermatozoa isolated from testicular tissue from patients with azoospermia in general and OA in particular had significantly lower DNA methylation levels at the H19 differentially methylated regions DMR than fertile men did.

Moreover, exposure to endocrine disruptors has been linked to abnormal sperm DNA methylation at imprinted and non-imprinted genes (Anway et al., 2005; Stouder & Paoloni-Giacobino, 2010). A number of parameters restricted to the testicular environment may be responsible for abnormal methylation in male patients with obstruction.

Fine needle aspiration cytology (FNAC, a simple, minimally invasive procedure) of the testicles has been suggested as a front-line investigative technique in patients with azoospermia, to determine the condition's aetiology and define the prognosis for TESE. In view of our results, one could design a new prognostic tool by combining FNAC with an analysis of the DNA methylation status of a few of the 195 genes identified in this study. If a negative FNAC corroborates the DNA methylation analysis, TESE could be considered (considering that the sperm retrieval rates were 85.7, 60 and 26% in NOA branches 1, 2a and 2b respectively). Patient counselling would be improved and couples could be invited to consider a strategy other than TESE (such as the use of donor spermatozoa).

Given that variations in the methylation profile could conceivably be because of variations in the proportion of each type of testicular cell in the sample, laser microdissection or germ cell fractionation prior to DNA isolation might also constitute a way of confirming our present data. Both approaches have their own limitations. Microdissection of more than 1000 cells (the number currently required for the methylation analysis) would be prohibitively time consuming. However, one can expect that fewer cells will be required in the future. Germ cells fractionation seems impossible, a large amount of material is necessary and biopsying this type of testicular fragment would run the risk of inducing hypogonadism.

In conclusion, our analysis of testicular gene methylation patterns was able to distinguish between OA and NOA and enabled the classification of azoospermic patients according to the TESE results. If combined with FNAC, analysis of testicular gene methylation patterns could be developed as a new prognostic tool for TESE. A methylation analysis of the 195 selected genes may thus be able to predict the failure or success of TESE. This topic merits further investigation. The DNA methylation patterns in TESE− NOA patients should be confirmed with an mRNA expression analysis and correlated with immunohistochemical data on protein expression and localization.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information
  • Albert M. (2005) Modalités techniques de récupération et de congélation des spermatozoïdes. Andrologie 15, 223226.
  • Anway MD, Cupp AS, Uzumcu M & Skinner MK. (2005) Epigenetic transgenerational actions of endocrine disruptors and male fertility. Science 308, 14661469.
  • Aoki VW, Emery BR & Carrell DT. (2006) Global sperm deoxyribonucleic acid methylation is unaffected in protamine-deficient infertile males. Fertil Steril 86, 15411543.
  • Aston KI, Punj V, Liu L & Carrell DT. (2011) Genome-wide sperm deoxyribonucleic acid methylation is altered in some men with abnormal chromatin packaging or poor in vitro fertilization embryogenesis. Fertil Steril 97, 285292.
  • Beck S & Rakyan VK. (2008) The methylome: approaches for global DNA methylation profiling. Trends Genet 24, 231237.
  • Benchaib M, Ajina M, Lornage J, Niveleau A, Durand P & Guerin JF. (2003) Quantitation by image analysis of global DNA methylation in human spermatozoa and its prognostic value in in vitro fertilization: a preliminary study. Fertil Steril 80, 947953.
  • Bhasin S, de Kretser DM & Baker HW. (1994) Clinical review 64: pathophysiology and natural history of male infertility. J Clin Endocrinol Metab 79, 15251529.
  • Boissonnas CC, Abdalaoui HE, Haelewyn V, Fauque P, Dupont JM, Gut I et al. (2010) Specific epigenetic alterations of IGF2-H19 locus in spermatozoa from infertile men. Eur J Hum Genet 18, 7380.
  • Boitrelle F, Robin G, Marcelli F, Albert M, Leroy-Martin B, Dewailly D et al. (2011) A predictive score for testicular sperm extraction quality and surgical ICSI outcome in non-obstructive azoospermia: a retrospective study. Hum Reprod 26, 32153221.
  • Carpi A, Sabanegh E & Mechanick J. (2009) Controversies in the management of nonobstructive azoospermia. Fertil Steril 91, 963970.
  • Devroey P, Liu J, Nagy Z, Goossens A, Tournaye H, Camus M et al. (1995) Pregnancies after testicular sperm extraction and intracytoplasmic sperm injection in non-obstructive azoospermia. Hum Reprod 10, 14571460.
  • Du P, Zhang X, Huang CC, Jafari N, Kibbe WA, Hou L et al. (2010) Comparison of Beta-value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinformatics 11, 587.
  • Fantz DA, Hatfield WR, Horvath G, Kistler MK & Kistler WS. (2001) Mice with a targeted disruption of the H1t gene are fertile and undergo normal changes in structural chromosomal proteins during spermiogenesis. Biol Reprod 64, 425431.
  • Filipponi D & Feil R. (2009) Perturbation of genomic imprinting in oligozoospermia. Epigenetics 4, 2730.
  • Foresta C, Ferlin A, Bettella A, Rossato M & Varotto A. (1995) Diagnostic and clinical features in azoospermia. Clin Endocrinol (Oxf) 43, 537543.
  • Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S et al. (2004) Bioconductor: open software development for computational biology and bioinformatics. Genome Biol 5, R80.
  • Ghalayini IF, Al-Ghazo MA, Hani OB, Al-Azab R, Bani-Hani I, Zayed F et al. (2011) Clinical comparison of conventional testicular sperm extraction and microdissection techniques for non-obstructive azoospermia. J Clin Med Res 3, 124131.
  • Huyghe E, Izard V, Rigot JM, Pariente JL & Tostain J. (2008) Optimal evaluation of the infertile male. 2007 French urological association guidelines. Prog Urol 18, 95101.
  • Jaenisch R & Bird A. (2003) Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat Genet 33(Suppl.), 245254.
  • Jarow JP, Espeland MA & Lipshultz LI. (1989) Evaluation of the azoospermic patient. J Urol 142, 6265.
  • Jenuwein T & Allis CD. (2001) Translating the histone code. Science 293, 10741080.
  • Kobayashi H, Sato A, Otsu E, Hiura H, Tomatsu C, Utsunomiya T et al. (2007) Aberrant DNA methylation of imprinted loci in sperm from oligospermic patients. Hum Mol Genet 16, 25422551.
  • Li T, Vu TH, Ulaner GA, Littman E, Ling JQ, Chen HL et al. (2005) IVF results in de novo DNA methylation and histone methylation at an Igf2-H19 imprinting epigenetic switch. Mol Hum Reprod 11, 631640.
  • Marconi M, Pilatz A, Wagenlehner F, Diemer T & Weidner W. (2009) Impact of infection on the secretory capacity of the male accessory glands. Int Braz J Urol 35, 299308; discussion 308-9.
  • Marques CJ, Carvalho F, Sousa M & Barros A. (2004) Genomic imprinting in disruptive spermatogenesis. Lancet 363, 17001702.
  • Marques CJ, Costa P, Vaz B, Carvalho F, Fernandes S, Barros A et al. (2008) Abnormal methylation of imprinted genes in human sperm is associated with oligozoospermia. Mol Hum Reprod 14, 6774.
  • Marques CJ, Francisco T, Sousa S, Carvalho F, Barros A & Sousa M. (2010) Methylation defects of imprinted genes in human testicular spermatozoa. Fertil Steril 94, 585594.
  • Minor A, Chow V & Ma S. (2011) Aberrant DNA methylation at imprinted genes in testicular sperm retrieved from men with obstructive azoospermia and undergoing vasectomy reversal. Reproduction 141, 749757.
  • Nicopoullos JD, Ramsay JW, Almeida PA & Gilling-Smith C. (2004) Assisted reproduction in the azoospermic couple. BJOG 111, 11901203.
  • Oates R. (2012) Evaluation of the azoospermic male. Asian J Androl 14, 8287.
  • Pacheco SE, Houseman EA, Christensen BC, Marsit CJ, Kelsey KT, Sigman M et al. (2011) Integrative DNA methylation and gene expression analyses identify DNA packaging and epigenetic regulatory genes associated with low motility sperm. PLoS One 6, e20280.
  • Pan J, Goodheart M, Chuma S, Nakatsuji N, Page DC & Wang PJ. (2005) RNF17, a component of the mammalian germ cell nuage, is essential for spermiogenesis. Development 132, 40294039.
  • Robin G, Boitrelle F, Leroy X, Peers MC, Marcelli F, Rigot JM et al. (2010) Assessment of azoospermia and histological evaluation of spermatogenesis. Ann Pathol 30, 182195.
  • Saxonov S, Berg P & Brutlag DL. (2006) A genome-wide analysis of CpG dinucleotides in the human genome distinguishes two distinct classes of promoters. Proc Natl Acad Sci USA 103, 14121417.
  • Schoysman R, Segal L, Van Der Zwalmen P, Nijs M, Bertin G, Cittadini E et al. (1993) Assisted fertilization with epididymal spermatozoa. Acta Eur Fertil 24, 712.
  • Smyth GK. (2004) Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol 3, Article3.
  • Stouder C & Paoloni-Giacobino A. (2010) Transgenerational effects of the endocrine disruptor vinclozolin on the methylation pattern of imprinted genes in the mouse sperm. Reproduction 139, 373379.
  • Strahl BD & Allis CD. (2000) The language of covalent histone modifications. Nature 403, 4145.
  • Thonneau P, Marchand S, Tallec A, Ferial ML, Ducot B, Lansac J et al. (1991) Incidence and main causes of infertility in a resident population (1,850,000) of three French regions (1988-1989). Hum Reprod 6, 811816.
  • Tournaye H, Camus M, Vandervorst M, Nagy Z, Joris H, Van Steirteghem A et al. (1997) Surgical sperm retrieval for intracytoplasmic sperm injection. Int J Androl 20(Suppl. 3), 6973.
  • WHO (2010). WHO Laboratory Manual for the Examination and Processing of Human Semen, 5th edn. World Heath Organisation, Cambridge University Press, Cambridge, UK.
  • Zadravec D, Tvrdik P, Guillou H, Haslam R, Kobayashi T, Napier JA et al. (2011) ELOVL2 controls the level of n-6 28:5 and 30:5 fatty acids in testis, a prerequisite for male fertility and sperm maturation in mice. J Lipid Res 52, 245255.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. References
  8. Supporting Information
andr117-sup-0001-TableS1.xlsxapplication/msexcel26KTable S1. CpG sites genes with a M-value higher than 3.

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.