The marker of alkyl DNA base damage, N7-methylguanine, is associated with semen quality in men

Sperm DNA contains a range of DNA base damage that can arise, in part, from exposure to methylating agents. However, the effects are not fully characterized and so the aim of this study was to investigate associations between semen quality and the levels of N7-methyldeoxyguanosine (N7-MedG), a marker of exposure to methylating agents, and other markers of DNA damage and DNA methylation. Sperm samples were collected from 105 men attending an assisted reproduction clinic as part of a couple undergoing treatment for infertility and semen quality assessed manually according to WHO guidelines. Semen levels of N7-MedG, quantified by immunoslotblot, were significantly higher in men with sperm concentration < 15 × 106/ml (p ≤ 0.01), semen volume < 1.5 ml (p ≤ 0.05) and also in men with any aspect of semen quality below WHO reference levels (p ≤ 0.001). Measures of neutral Comet DNA damage were correlated with semen quality in a univariate analysis but not after adjustment for N7-MedG levels. Sperm concentration was negatively associated with % methylation at the gene for DAZL but no other marker of global or gene-specific DNA methylation. Results support the hypothesis that the known toxic and DNA damaging properties of alkylating agent exposure may have direct deleterious consequences on semen quality.

Descriptive analysis of semen quality and markers of DNA damage and DNA methylation. The semen quality of the studied population is shown in Table 1 together with a descriptive analysis of markers of DNA damage and DNA methylation. There were 28 men whose sperm concentration was less than 15 × 10 6 /ml, 18 men with a % progressively motility of < 32, and 11 men whose semen volume was less than 1.5 ml. Overall there were 44 men with a below reference level of either sperm concentration, % progressive motility or semen volume. % of sperm with low DNA damage (% LDD sperm) was inversely correlated (r s = − 0.31, p < 0.001 n = 105) with % of sperm with high DNA damage (% HDD sperm) and with median % tail DNA (r s = -0.79, p < 0.001 n = 105) whereas % HDD sperm was positively correlated with median % tail DNA (r s = 0.67, p < 0.001 n = 105). N7-MedG in sperm DNA was negatively associated with % LDD sperm (r s = -0.22, p < 0.05 n = 101) and positively associated with median % tail DNA (r s = 0.20, p < 0.05 n = 101): it was not correlated with % HDD sperm (r s = 0.12, n = 101). % DAZL methylation was positively associated with % HDD sperm (r s = 0.37; p < 0.05) and with median % tail DNA (r s = 0.39; p < 0.05). There was no other association between global methylation and gene-specific methylation or markers of DNA damage.
% immotile sperm was negatively associated with % LDD sperm (r s = -0.31, p < 0.01), but positively associated with % HDD sperm (r s = 0.41, p < 0.001) and median % tail DNA (r s = 0.41, p < 0.001). There was no other correlation between semen quality and global methylation or % methylation of H19CTCF6, SNRPN, or MEG3 ( Table 2). The multiple regression model with %LDD sperm, % HDD sperm and median % tail DNA produced  (Table 3). Semen volume was not associated with any of the measured biomarkers.
Associations between semen quality and DNA damage and DNA methylation in men with semen quality above and below WHO referent values. To assess whether DNA damage and DNA methylation levels were associated with clinically more relevant factors, we examined whether levels were different in men whose semen quality was either below or above WHO reference values (Table 4). Specifically, N7-MedG levels, % HDD sperm and median % Tail DNA were higher in those men whose sperm concentration was < 15 × 10 6 ml. In contrast, % LDD sperm was lower and % HDD sperm and median %tail DNA higher, in men with < 32% progressively motile sperm. Furthermore, N7-MedG levels were higher in those men with a semen volume was less than 1.5 ml. Overall, in any men whose sperm concentration or % progressively motile sperm or semen volume was below the reference category, N7-MedG levels, % HDD sperm and median %tail DNA were significantly higher and % LDD sperm significantly lower compared to men whose semen quality was above the reference value. In this analysis there was no association between any DNA methylation marker and semen quality above or below the WHO reference value. In a logistic regression analysis (Table 5), N7-MedG levels were significantly associated with low sperm concentration (OR 6.46, 95%CI 1.51-27.7) even after adjustment (OR 6.42, 95%CI 1.47-28.0) for other markers of DNA damage, % HDD sperm and median % tail DNA. Whilst % HDD sperm and % median tail DNA were significantly associated with having low progressive motility in a univariate analysis (Table 5), after adjustment

Discussion
In this study of sperm DNA damage and DNA methylation of men attending for ART as part of a couple, the main finding was that N7-MedG levels were significantly associated with adverse semen quality even after adjustment for other markers of DNA damage as measured by the Comet assay. Measures of DNA damage assessed by the Comet assay were correlated with % progressive motility and % immotile sperm and differed in men whose semen quality was below WHO reference levels in a univariate analysis but not after adjustment for N7-MedG levels. Sperm concentration was negatively associated with % methylation of DAZL but no other marker of DNA methylation was associated with adverse semen quality in this study. N7-MedG is a marker of exposure to exogenous and endogenous DNA methylating agents and has been detected in a wide range of human tissues 13 . The precise contribution of exogenous (e.g. tobacco specific nitrosamines) and endogenous (e.g. s-adenosylmethionine and other agents formed in situ through N-nitrosation of amines) methylating agents to N7-MedG levels is unknown. Smoking is an exclusion criterion for treatment in our NHS centre and none of the men in the study were self-reported current smokers making it very unlikely that smoking was associated with N7-MedG levels.
There is also evidence that obesity is associated with DNA damage levels in that higher levels of DNA fragmentation (ie cells that did not have a detectable nucleus but only a Comet tail as detected by the alkaline Comet assay) were reported in obese men as compare to controls 20,21 . However, we found no correlation between BMI and DNA damage levels (data not shown). However, the finding in this study that N7-MedG levels were significantly associated with adverse semen quality is consistent with our previous report that N7-MedG levels were higher in men diagnosed with male factor infertility 5 and implicates DNA base damage, and not just DNA strand breaks, as a contributing cause of male infertility. The mechanistic basis of this association is unclear but it might be related to the known chemical instability of the N7-MedG adduct leading to DNA strand breaks 16 and it is worthy of note that N7-MedG levels were negatively associated with %LDD sperm and positively associated with www.nature.com/scientificreports/ median % tail DNA. In this context N7-MedG instability might represent a previously uncharacterised exposure route to increased DNA strand breaks in human sperm. Alternatively, the association between N7-MedG and adverse semen quality may be due to toxic DNA protein cross-links 17 or due to the presence of other more toxic adducts such as O 6 -methylguanine or N3-methyladenine that will be formed at the same time as N7-MedG 22 .
The latter 2 adducts have as far as we can ascertain not yet been reported in human sperm DNA but their presence in sperm DNA may have more deleterious consequences as O 6 -methylguanine is a known pro-mutagenic adduct 19 and N3-methyladenine can induce chromosome aberrations 23 . The post-fertilisation effects of these adducts would depend not only on its levels in individual sperm but also the repair capacity of the oocyte. The Comet assay measures a variety of different types of DNA damage (e.g. single and double strand breaks and apurinic sites) whether induced for example, by oxidative stress, chemical modifications of DNA or DNA repair processes 24,25 . The precise contribution of these different mechanisms to DNA damage in any one sample is unknown. The neutral Comet assay has been reported to measure principally double-strand breaks as Alu1 induced DNA fragmentation was higher with the neutral Comet assay than the alkaline comet assay 26 . This, though, is controversial with other authors reporting that the neutral comet assay measures single strand nicks when the DNA remains attached to the nuclear matrix 27 . As the extent of strand breaks increases, very large stretches of DNA can become detached from the matrix and move out of the nucleus and into the gel 28 . We refer to these events as DNA damage, recognising that the neutral Comet assay measures a mix of single and double strand breaks. The relationship between DNA damage detected by the neutral Comet assay and semen quality is currently unclear. The DNA fragmentation index in sperm, as assessed by the neutral Comet assay, has been negatively associated with sperm concentration and morphology in one study 1 , and with sperm viability, motility and morphology but not sperm concentration 29 . Similarly, a weak correlation was observed between tail % intensity and % motility, % normal morphology but not sperm concentration 30 . Conversely, no associations between semen quality and % tail DNA were detected in 80 healthy male volunteers 31 . It is unclear whether these differences result from differences in analysis protocols, descriptors of DNA damage or populations studied, though in one study associations between indices of semen quality varied with how the DNA damage was characterised 32 . One additional interpretation is that the differences in associations may result from differences in other types of DNA damage e.g. N7-MedG present within human sperm.
Previous studies have reported associations between environmental exposures and/or DNA damage and sperm DNA methylation in specific genes, suggesting, at least, that DNA damage and DNA methylation share a common exposure or cause and potentially that the effects of such exposure are mediated, in part, by their epigenetic effects. Associations between reactive oxygen species and H19-IgF2 methylation have been described 33 and also methylation levels of IGF2 and KCNQ1 but not H19, MEG3, PEG3, SNRPN were associated with DNA fragmentation index 34 . We also examined but found no associations between markers of DNA damage and global DNA methylation and methylation of genes that have been previously associated with male infertility 35 . This suggests that the effects of alkylating agent exposure may not be mediated through changes in DNA methylation, but rather via the direct effects of DNA damage. The exact source/origin of DNA damage detected by the Comet assay still remains unknown but there is little evidence in the present study of these sources also impacting on the methylation of sperm. If such associations do exist, they will likely depend upon the initial exposures causing the DNA damage. DAZL methylation was negatively associated with sperm concentration, consistent with previous studies 36 . Higher DAZL methylation was found in asthenozoospermia and oligozoospermia men 37 . In contrast, no associations were detected with global methylation or methylation of other genes (H19, MEG3, NR3C1, SNRPN) that have previously been associated with semen quality 35 ; this may be due to population or sample size differences.
Like previous studies examining associations between DNA damage and semen quality, the main weakness of this study is that it is cross-sectional in nature and so unable to determine causative associations between the measured biomarkers and semen quality. In addition, there is the possibility that the levels of N7-MedG reported here may reflect levels occurring in contaminating leucocytes and round cells (for example). However, as these samples were used for clinical purposes, numbers of other cells were routinely assessed and samples identified as containing high levels of leucocytes and round cells were omitted from the study. Furthermore, seminal round cells have been reported to be mostly immature post-meiotic germ cells 38 and leucocytes account for maximum 5% of total number of round cells in the seminal fluid 39 . The alternative possibility that low levels of leucocytes and round cells influenced the results is considered unlikely as their N7-MedG levels would have had to have been much higher than observed here , at levels only previously observed in patients receiving methylating agent chemotherapy 13 . We are thus confident that the results herein are representative of sperm N7-MedG levels.
However, our results are consistent with current knowledge regarding the toxicological consequences of alkylating agent exposure in other tissues. This study and other studies indicate that sperm DNA contains a range of different types of DNA damage that might be associated with semen quality and have more deleterious consequences post-fertilisation.

Materials and methods
Study population and ethical approval. 105  www.nature.com/scientificreports/ WHO guidelines 40 in the hospital by biomedical and clinical scientific staff, and according to the UK NEQAS Reproductive Sciences scheme 41 . Samples were to be used for clinical treatment and those with any significant round cell/white blood cell count or samples with leukospermia were omitted from the study. After collecting the remaining sample from the hospital, the neat semen was centrifuged to separate the sperm from the seminal plasma and an aliquot of the neat sperm pellet was directly used to determine the level of DNA damage by using the neutral Comet assay.
Extraction of DNA from sperm. Briefly, genomic DNA was extracted from up to 4 × 10 7 sperm cells using the Qiagen Blood and cell culture DNA Midi Kit except that the proteinase K and ribonuclease A digestion step was carried out at 4 °C overnight. The following day, 0.5 ml of 5 mM DTT solution was added to each tube, mixed and incubated again in the cold room on a rotor mixer for 1 h, followed by incubation for another 1 h at 37˚C in water bath. Following the manufacturer's instructions, the samples were added into gravity flow columns. Eluted from the columns, DNA was isopropanol and ethanol precipitated and rapidly air dried to remove residual alcohol. The concentration of extracted sperm DNA in 1X TE buffer was measured first by Nano-drop and then by a picogreen assay 5 .
N7-MedG slotblot analysis of sperm DNA. Immunoslotblot analysis of N7-MedG in sperm DNA has been described previously 5,12 . Methylated DNA standards were prepared by reacting CT-DNA with N-nitroso-N-methylurea. Briefly, DNA samples were sonicated (Sonicator, Bandeline electronic sonopuls, UW 2070, Probe: SH70G) for 10 s at 50% power to obtain DNA fragments of size between 400 and 1500 bp and these were incubated in alkaline solution to induce imidazole ring opening. After neutralisation, DNA samples were heat denatured and then 1 µg of DNA immobilised on a nitrocellulose membrane (Whatman: Portran, BA79, pore size 0.1 µm). Membranes were incubated first with a rabbit polyclonal antibody raised against 2,6-diamino-4-hydroxy-5-N-methylformamidopyrimidine (me-Fapy) conjugated to keyhole limpet hemocyanin which specifically detected N7-MedG in mice treated with N-Nitroso-N-methylurea 42 and then a polyclonal goat antirabbit-horseradish peroxidase conjugate. Membranes were incubated with ECL reagent (Amersham Hyperfilm) and exposed to an X-ray film (Fuji) which was subsequently developed and scanned using a HP Scanjet 485. The intensity of the bands for the samples and the standards was quantified by using ImageJ program and a standard curve was generated and subsequently the levels of N7-MedG in the samples were calculated. To confirm DNA binding the membrane was stained with a propidium iodide solution and scanned with a Typhoon scanner (GE Healthcare) at 610 emission and 560 absorption wavelengths and the intensity of DNA bands in the obtained image quantified using ImageJ.
Single-cell gel electrophoresis (Comet) assay. DNA damage was assessed by a neutral comet assay according to previous studies but with minor modifications 1,43 . Sperm (~ 2 × 10 4 ) were suspended in LMP agarose and applied to a microscope slide pre-coated with 1% NMP agarose. After gel setting, the slides were immersed firstly in freshly made precooled lysis buffer (2.5 M NaCl, 100 mM EDTA, 10 mM Tris-base, 10% DMSO and 1% Triton X-100, pH 10) at 4 °C, then in fresh lysis buffer containing 10 mM DTT and finally fresh lysis buffer containing proteinase K (100 μg/ml) to lyse the cells. The slides were then equilibrated for 20 min in cold electrophoresis buffer (300 mM sodium acetate 100 mM Tris, pH9) then electrophoresis was carried out at 13 V, 120-125 mA for 1 h. After neutralisation in cold neutralization buffer (0.4 M Tris-base pH 7.5), slides were stained with SYBR gold (Invitrogen, UK) used according to the manufacturer's instructions and were then examined using an Olympus fluorescence microscope (OLYMPUS, U-RFL-T; Filter: FITC; Camera: Q IMAG-ING, RETIGA-SRV, FAST 1394; Software: QCapture Pro 6.0) and around 35 images containing in total around 400 sperm (Comet) were obtained. Then, % tail DNA was determined by using TriTeK CometScoreTM Freeware version 1.5 (TriTeK Corp, US) in 400 sperm.
Prior to the analysis of the participant samples, we validated our neutral comet assay by incubating sperm with up to 30 µM H 2 O 2 and this clearly showed that the assay was able to detect increases in DNA damage induced by a DNA damaging agent. We found a dose-dependent decrease in the % of sperm with few strand breaks (r 2 = 0.077: < 2% at the highest H 2 O 2 concentration ) and a dose-dependent increase in median levels of % tail DNA (r 2 = 0.91) and % sperm with high levels of DNA damage (r 2 = 0.90: 20% at the highest H 2 O 2 concentration). A pilot reproducibility study also showed little variance from day to day.
Global DNA methylation. Genomic DNA was heat-denatured and digested to nucleosides by nuclease P1 (Sigma, UK) and then alkaline phosphatase (Sigma, UK). 5-methylcytosine levels were then quantified using the Global DNA Methylation ELISA Kit (Cambridge Biosciences, UK), following the manufacturer's instructions.
Gene specific methylation. DNA samples underwent bisulphite conversion using the EpiTect Fast DNA Bisulphite Kit (Qiagen, UK), according to the manufacturer's instructions. Pyrosequencing target regions were amplified using the PyroMark PCR Kit (Qiagen, UK) as follows: PCR reactions were performed in a total volume of 25 μl consisting of 12.5 μl 2 × PCR master mix, 2.5 μl 10 × CoralLoad, 0.75 μl (3 μM) of each forward and reverse primer, 2 μl (~ 10 ng) bisulfite-converted DNA and 6.5 μλ H 2 O. Primer sequences covering the target CpG regions were either purchased as pre-designed CpG assays from Qiagen UK (Supplementary data  Table 1), or custom oligonucleotides were designed using the PyroMark Assay Design software (Qiagen, UK) and purchased through Life Technologies (UK) (Supplementary Table 2). Gradient PCR was initially carried out to determine the optimal annealing temperature for amplification ranging from 52 to 60 °C and cycling conditions were carried out as follows: 95  www.nature.com/scientificreports/ products were frozen at -20 °C until pyrosequencing. PCR products were immobilised to Streptavidin Sepharose High Performance beads (GE Healthcare Life Sciences, Uppsala, Sweden), denatured and the pyrosequencing primer (Supplementary Table) was annealed to the template. Pyrosequencing was performed on a PyroMark 24 Pyrosequencing system (Qiagen). Results were analysed on the PyroMark Q24 software to determine the % methylation for each CpG site.

Statistical analyses.
Semen quality was assessed both as a continuous variable and after dividing the population into two based upon WHO reference values for semen quality 44 (i.e. sperm concentration > 15 × 10 6 /ml, % progressive motility > 32% and a semen volume > 1.5 ml). We established the Comet thresholds blind to the semen quality status. The distribution of % tail DNA in the sperm samples was not normal with the log % tail DNA showed an almost trimodal distribution in 75% of the samples, ~ 14% of the samples showed biomodal distributions while the remaining 11% samples showed multimodal distribution. It was possible to distinguish two subgroups of sperm; sperm with low levels of DNA damage (% tail DNA < 0.005%, LDD sperm) and sperm with high levels of DNA damage (% tail DNA ≥ 7.5%, HDD sperm). Correlation (Spearmann rank analysis) and multiple regression analyses (including those markers whose associations with semen quality had a p ≤ 0.1) were conducted to examine the relationship between semen quality and markers of DNA damage and DNA methylation (mean methylation of all CpG sites analysed). This regression analysis is of limited clinical value and so biomarker levels were then compared in men whose semen quality were above or below WHO referent levels in both a univariate analysis and by logistic regression after adjustment for other DNA damage markers with P < 0.1.