Associations of sperm telomere length with semen parameters, clinical outcomes and lifestyle factors in human normozoospermic samples

Many studies have demonstrated that lifestyle factors can affect sperm quality and fertility. Sperm telomere length (STL) has been reported as potential biomarker or sperm quality. However, no studies have investigated how lifestyle factors can affect STL and associated clinical outcomes.


| INTRODUC TI ON
Telomeres are complex ribonucleoprotein structures with repetitive DNA sequences (5'-TTAGG-3') that function as chromosomal stabilizing elements. 1,2 Repetitive cell divisions lead to telomere attrition. 3 Once telomeres reach a critical short length, genome stability and cell division are affected, which can result in an increase in DNA damage and cell apoptosis. 4 Furthermore, the rate of telomere shortening can be accelerated by the impact of lifestyle factors, environmental agents and increased activity of reactive oxygen species (ROS). 5,6,7,8 In male germ cell development, telomere length increases from spermatogonia to spermatozoa and is inversely correlated with the expression of telomerase activity. 9,10 In human sperm cells, sperm telomere length (STL) increases with age and positively correlates with male parental age at conception. 11,12 In a young student population, Ferlin et al first demonstrated a shorter STL in men having a low sperm count (threshold = 39 million), following WHO guidelines 13 compared with normozoospermic subjects and a significant and positive correlation between STL and sperm count. 14 In another study from the same research group, a positive correlation between STL and sperm count in a normozoospermic selection of the population was demonstrated. 15 Additionally, men with idiopathic infertility or having varicocoele have lower sperm counts and shorter STLs compared with fertile men who recently fathered a child. 16,17 Furthermore, Lafuente et al demonstrated that STL in infertile men was negatively correlated with sperm concentration and progressive motility. 18 An increasing number of studies have found that STL may play a role in reproduction as a molecular marker of sperm quality.
Investigating the link between STL and fertility outcomes, Yang et al showed STL was positively associated with the quality of embryos generated in clinical assisted reproductive technologies (ART).
Moreover, there was no significant association between STL and the fertilization rate. 19 Additionally, the reduction in STL was associated with an increased DNA fragmentation index. 17 DNA-damaged spermatozoa give rise to reduced embryo quality and development, but fertilization is not impaired, and pregnancy outcomes are not strongly affected. 20 In another study, the occurrence of an ongoing pregnancy after in vitro fertilization (IVF) treatment was null in a limited number of patients with abnormal relative STL. 21 Therefore, STL measurement is suggested to predict clinical outcomes. 22 Several studies have demonstrated an impact of lifestyle factors on telomere lengths, mainly in leucocytes (for reviews, see. 23 Lifestyle factors have also been stated to potentially affect sperm quality and fertility. 24 However, no studies have investigated the association of STL and lifestyle with sperm quality and ART outcomes. The aim of this study was to investigate any association between human STL with normozoospermic sperm parameters and clinical ART outcomes, and its potential link to participants' lifestyles. Ninety-four subjects' sample and data were randomly allocated for this study (including 3 that had to be discarded due to confounding anomalies or failure of oocyte culture).

| Study subjects and sample preparation
Semen samples were produced by masturbation after 2-5 days of abstinence. Semen analysis was carried out according to World Health Organization 13 guidelines. Briefly, following semen liquefaction after 30-60 min, semen volume was calculated by subtracting the weight of a pre-weighted container from the weight of the container with the semen. Sperm concentration was measured using a haemocytometer, and total count was calculated by multiplying the sperm concentration by the volume of the whole ejaculate. Sperm motility was assessed by measuring the percentages of progressive, non-progressive and immotile spermatozoa, according to the WHO guidelines.
For this study, only subjects with a normozoospermic semen sample (total sperm count ≥ 39 × 10 6 / ejaculate, concentration ≥ 15 × 10 6 /mL, progressive motility ≥ 32%, semen volume ≥ 1.5 mL, 66 participants) were included, leading to a total study population of 66 male patients for STL analysis Figure 1. After collecting the remaining of the sample from the hospital, the neat semen was centrifuged and washed to separate the spermatozoa from the seminal plasma. Sample sperm concentration was measured (see above), and the sample was pelleted and frozen at − 80°C until sperm DNA extraction.

| Sperm DNA extraction
DNA was extracted from up to 25 × 10 6 spermatozoa using the Qiagen Blood and Tissue Midi Kit (Qiagen, UK). Sperm pellets were investigations are warranted to identify the potential variation of STL overtime to clarify its significance as a potential biomarker in ART.

K E Y W O R D S
fertilization, lifestyle factors, male infertility, semen parameters, sperm telomere length incubated at 4°C overnight in the lysis buffer supplemented with RNase (10 mg/mL, Sigma, UK) and Proteinase K (1 mg/mL, Sigma, UK). To complete sperm lysis, a dithiothreitol solution (5 mM, Sigma, UK) was added and first incubated at 4°C for 1 h on a rotator and second placed at 37°C for 1 hour. 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 alcohol residual. DNA was quantified in TE buffer using a Synergene HT nucleic acid plate (BioTek, UK).

| Sperm telomere length measurement
The STL of the DNA samples was analysed using real-time PCR, as previously described, 25 with some adjustments. Amplification of the telomeres was performed using the primers Tel O Forward

| ART procedures
Ovulation induction was achieved using conventional down-regulation involving pituitary desensitization. Recombinant folliclestimulating hormone (FSH) was administrated by a step-down protocol. When three or more follicles reached ≥ 17 mm diameter, beta-human chorionic gonadotrophin (β-hCG) hormone was administered and oocytes were recovered 36 hour later by ultrasound-guided retrieval. In in vitro fertilization (IVF) treatment, metaphase II (MII) oocytes were inseminated with ~ 1×10 5 prepared spermatozoa/ml in IVF culture medium (G-IVF, Vitrolife) at 37˚C, 6% CO 2 . In intracytoplasmic sperm injection (ICSI) treatment, a single motile spermatozoon with apparent normal morphology F I G U R E 1 Repartition of samples. Flow chart showing exclusion criteria for repartition of samples. ART: assisted reproductive technologies, STL: sperm telomere length; WHO: World Health Organization was microinjected into each MII oocyte. Oocytes were examined 16-18 h after insemination for the presence of two pronuclei, which is indicative of successful fertilization. Thus, the fertilized oocytes were in vitro-cultured and their development was monitored. After 3 or 5 days, one or two good quality graded embryos were transferred into the uterus. After a fresh ART treatment, any spare embryos could be stored for future use, depending on their quality. Embryo implantation was assessed 12-14 days after fertilization by measuring the level of β-hCG hormone in the woman blood. Clinical pregnancy was confirmed 5-7 weeks after embryo transfer by the presence of gestational sac (foetal heart) detected by ultrasound scan. Biochemical pregnancy (early miscarriage) and late miscarriages were also recorded and were respectively identified as the absence of a foetal heart after a positive β-hCG and a pregnancy loss occurring after the detection of foetal heart. Births were reported including gestational age at delivery, birthweight and gestational age-adjusted birthweight.

| Clinical ART outcome analysis
Clinical ART data were extracted from the current integrated data management system (AcuBase) at St Mary's Hospital, Manchester.
Each clinical ART outcome was analysed combining data from both IVF and ICSI treatments. Fertilization rates were calculated using the number of embryos at the pronuclei stage divided by the total number of either microinjected or inseminated oocytes. Embryo cleavage rates were calculated as the number of cleaved embryos 2-3 days after fertilization divided by the total number of fertilized oocytes. Ectopic pregnancy was counted as a successful pregnancy.
To analyse the gestational age at birth and birthweights, these outcomes were referred to live birth events and only birthweights from singleton pregnancies were analysed due to the major biological characteristics of twin pregnancies. Gestational age-adjusted birthweights were calculated for using the Gestation-Related Optimal Weight (GROW) formula as described in Castillo et al. 26  All factors containing continuous numerical data were also analysed to their respective STL values using direct correlation. In this cohort, 64 males out 65 were White Europeans. To examine biometric factors, thresholds for age and BMI were respectively applied at 35 years (starting age for male fertility decline 27 ; and 25 kg/m 2 (to separate underweight/normal and overweight/obese). Weekly caffeine intake and alcohol consumption were calculated and reported respectively in mg of caffeine and alcohol units of alcohol following guidelines. 28,29 Physical activity was analysed using both a binary score linked to the number of exercise hours and the Godin-Shephard Leisure Score Index (LSI) threshold defined at 24. 30 The unit had a strict policy to not treat smokers,hence, no participants identified themselves as being a current smoker. Regarding the food frequency analysis (units per week), all single food items were grouped into food groups such as red meat, poultry and cruciferous vegetables. Food supercategories were created from food groups following similar biological classification (ie vegetables include cruciferous, leafy green and other vegetables, presented with a '1' exponent in Table 6. Two patients had minor missing information about consumption frequencies in the questionnaire and were excluded from the STL and food frequency analysis Figure 1 , which was carried out on 63 males of the cohort in Table 6.

| Statistical analysis
Data are represented as mean ± the standard deviation (SD) or median with the interquartile range (IQR) using GraphPad Prism (La Jolla, CA) and IBM SPSS (Chicago, IL) software. Data normality was assessed using the Shapiro-Wilk normality test. Depending on normality results, either parametric Welch's t test or non-parametric Wilcoxon-Mann-Whitney tests were performed to determine differences between two study categories for biometric, health and education/job data (expect for education level, Bonferroni's one-way ANOVA; Table 5. ANCOVA tests were performed to test differences while justifying for male age and BMI between study categories for ART outcomes Table 4 and lifestyle/dietary factors Table 5. Linear regression analyses were performed using a general linear model univariate analysis to determine correlations between STL and study parameters while adjusting for age, BMI and sporadically any other appropriate factors (sexual abstinence for sperm parameters in Table 3 and number of oocytes or embryos for ART outcomes in Table 4. P-values < .05 were considered statistically significant. P-values < .08 were reported as non-significant trends.

| General semen characteristics and clinical outcomes
Male biometrics and basic semen characteristics are shown in Table 1 and are reported as the mean ± standard deviation (SD) or as the median (interquartile range, IQR) and minimal-maximal values.
All normozoospermic semen parameters are based on the WHO lower reference values, with values greater than the lower reference values established by 13 guidelines. 13 ART clinical outcomes are presented in Table 2

| Sperm telomere length and normozoospermic semen parameters
No direct associations were observed between STL and semen parameters including total sperm count, concentration, semen volume, progressive motility, grade A motility, percentage of immotile spermatozoa and sexual abstinence in a normozoospermic population Table 3.

| Sperm telomere length and clinical ART outcomes
STL, in vitro fertilization procedures and clinical outcomes are presented in Table 4. When comparing STL in the ART procedures, no significant differences were observed between IVF and ICSI, and as such for the further analyses of STL with clinical ART outcomes, we used the combined data from both treatments. This revealed that STL positively correlated with the rate of fertilization Figure 2 , r = .32, P = .0097, adjusted for BMI: r = .36, P = .004). In a limited number of cases where all embryos failed to develop, these embryos tended to have been fertilized from lower STL spermatozoa. Nevertheless, for the couples who had cleaved embryos, there were no significant associations between STL and embryo cleavage rate. In cumulating outcomes from fresh and frozen cycles, we observed a non-significant trend where STL was higher in couples who had successfully implanted embryos compared to couples with failed implantation (P = .051).
However, STL was not found significantly different or associated with post-implantation events including clinical pregnancy, successful live birth event and gestational age-adjusted birthweight. Table 5 shows the correlation between STL and biometric factors.

| Sperm telomere length, biometrics, education and general health
No significant difference in STL was identified between thresholddefined groups for age or BMI. The STL of the normozoospermic men demonstrated no correlation with age (r = −.10, P = .4528) and BMI (r = .06, P = .6575). Additionally, no associations were found with the degree of education or between those who had worked in the last 24 h prior to ART treatment. Approximately 12% of participants had suffered from flu and fever, but no difference in STL was found (P = .8180). Participants suffering from an illness caused or made worse by work tended to have a lower STL (P = .0575).

| Sperm telomere length, lifestyle and dietary factors
When investigating common lifestyle factors such as smoking, drinking caffeine and exercise, which are known to influence leucocyte telomere length (for reviews, see 31, we found no differences in STL for each lifestyle factor Table 5  (non-smokers vs ex-smokers: P = .4117) or being exposed to cigarette smoke (P = .9426).

Dietary patterns including meat consumption in the last 24 h
were also investigated together with STL Table 6. All male participants had a meat-inclusive diet. The difference in STL was not significantly linked to either dietary pattern (meat only vs meat + fish) or meat consumption. Overall food consumption frequencies of the last three months prior ART treatment were analysed for food categories such as meat, dairy, fruit, vegetables and more Table 6.
None of the food categories and single food items (data not shown) were significantly associated with STL. The consumption of processed meat tended to be inversely correlated with STL (r = −.24, P = .068),however, the correlation coefficient remains low.

| D ISCUSS I ON
Exploring the role of STL in semen quality, male fertility and reproductive clinical outcomes is a growing area of interest in reproductive medicine. In this study, we identified a positive correlation between the STL and the fertilization rate in normozoospermic samples and a very close trend towards higher STL in successful embryo implantation rates. However, we found no association between STL and sperm parameters. Furthermore, occupational health, diet and lifestyle factors were not correlated with STL. Altogether, our data suggest that STL might have an important role mechanistically for fertilization rate regardless of sperm parameters and lifestyle factors.
Recent studies demonstrated that STL decreased in ICSI patients with previous low fertilization rate (n = 10) 32 and positively correlated with embryo quality, 19 which is in agreement with data presented in the present study. In assisted conception practice, it is accepted that implantation and pregnancy outcomes are improved by the transfer of good quality embryos compared with low-quality ones. 33,34 However, in a recent study, Cariati and colleagues noted that no ongoing pregnancies were observed from patients having an atypical STL following IVF. 21 In another study, it was demonstrated that spermatozoon with shorter STL was associated with lower natural pregnancy rates compared with couples who did achieve pregnancy naturally with longer STL. 18 Herein, a trend was observed in which

Semen parameters (n = 65) GLM univariate analysis, r (P-values)
Total sperm count (×10 6  Note: A general linear model (GLM) univariate analysis was used to identify correlations between sperm telomere length and sperm parameters while adjusting for age, body mass index and sexual abstinence duration. n, number of semen samples; r, correlation coefficient.

F I G U R E 2
Positive correlation between STL and fertilization rate sperm samples with higher STLs tend to have an increased embryo implantation rate. However, in the present study, we observed live birth rates occurred with some of the lowest and highest STL values in our cohort. However, there was no STL association with birthweight or gestational age. Therefore, it could be hypothesized that a higher STL may be associated with increased fertilization rates and subsequent increased embryo development/ quality, leading to improved implantation rates and pregnancy outcomes. Decreased STL was repeatedly observed in patients with oligozoospermia compared with normozoospermic samples, indicating that shorter STL may be associated with impaired spermatogenesis through segregation errors during meiosis. 14,16,18,21 Studies investigating the association between STL and sperm quality have shown a positive correlation with total sperm count and STL. 14,21 In other studies using normozoospermic samples, STL was found to be positively linked to sperm progressive motility/ vitality and negatively linked to sperm DNA fragmentation 15, but no association was detected with male age, sexual abstinence, total sperm count, concentration and morphology. However, in another study, there was no association noted between STL and sperm parameters. 16 In the present study, we found no correlation between sperm motility, count or concentration with STL. The current assessment of sperm quality as using the WHO guidelines 13 does not take molecular andrology pathways (such as STL and sperm DNA integrity) into account. 35 Therefore, the quality of spermatozoa and its fertilization capacity may be dependent on molecular integrity in addition to the conventional sperm parameters.
In humans, lifestyle and environmental factors are known to influence telomere length in leucocytes. 31,36,37 This telomere attrition is associated with smoking cigarettes, low physical activity and stress exposure 38-40, though a recent large-scale study with repeated measures did not support any effect of alcohol consumption on leucocyte telomere length. 41 Importantly, however, these parameters have never been investigated in relation to STL. The results presented here suggest that neither occupational nor lifestyle factors associated with STL.
Similar to the other lifestyle factors, dietary patterns and food frequencies have been associated negatively and positively to leucocyte telomere lengths. Shorter telomeres have been associated with the level of consumed total fat and saturated fatty acid 42 whereas, consumption of seeds, nuts, legumes and coffee correlated with increase telomerases. Further recent studies support that a diet with more plant-based food items and reduced processed meat might be beneficial for increased telomeres and longer life, according to research on the Mediterranean diet. 43,44 Interestingly, the effect of diet interventions has been shown to have only minor to no effect on telomeres. 45,46 In our cohort, STL was not associated with any food categories (such as meats, vegetables and fruits) or specific food item consumption. Processed meat consumption had a non-significant, but mild negative TA B L E 4 Correlation between clinical outcomes and STL. (A) ANCOVA tests were performed to test differences in STL between groups while controlling for the number of embryos used, male age and BMI. (B) Direct correlations between STL and continuous values of ART outcomes were assessed using Pearson's correlation tests Age is a well-investigated parameter for telomere lengths, and advanced age is associated with a decrease in leucocytes 54 and increase in spermatozoa. 55 The ageing protection mechanism of telomere length in leucocytes is well defined in the literature, whereas the mechanism of STL elongation remains currently unclear. This elongation phenomenon is inherited from the father to offspring independently to the offspring sex 56 and seems to be conserved across species. 57,58 Interestingly, this paternal inheritance is also evident in leucocytes. 59 Older parental age at conception has been shown to have a negative effect on both offspring viability in mice 60 and the overall fitness of offspring in humans. 61,62 As a confounding factor, ageing is linked to fertility issues, such as the decline in male hormones and sperm quality. 63 66 Therefore, given the low number of possible leucocyte or RC relative to sperm cell numbers in our study samples, we are confident that the STL results are representative of the spermatozoa.
This study demonstrates a positive association between STL and the fertilization rate in a normozoospermic population.
However, we did not find any association between the sperm parameters and the individual biometric and lifestyle factors. The data presented in this present study set the scene for larger longitudinal studies to explore the dynamics of STL in humans in relation to lifestyle influences and to clarify its significance as a potential biomarker in ART, and the inheritance effect of a longer STL on ART offspring's fertility.

ACK N OWLED G EM ENTS
The authors would like to thank the staff at St. Mary's Hospital for their assistance with patient recruitment and data collection.

CO N FLI C T O F I NTE R E S T
None declared. Note: General linear model univariate analyses were performed to identify associations between STL and frequencies of food categories while adjusting for male age and male body mass index. Supercategories are notified by a '1' in exponent; n, number of participants; r, correlation coefficient.