Regular Research Paper

J Sleep Res. (2017)

Sleep duration is associated with sperm chromatin integrity among young men in Chongqing, China XIAOGANG WANG1,* , QING CHEN1,*, PENG ZOU1, TAIXIU LIU2, MIN MO1, HUAN YANG1, NIYA ZHOU1, LEI SUN1, HONGQIANG CHEN1, XI LING1, K A I G E P E N G 1 , L I N A O 1 , H U I F A N G Y A N G 2 , J I A C A O 1 and Z H I H O N G C U I 1 1

Key Laboratory of Medical Protection for Electromagnetic Radiation, Ministry of Education of China, Institute of Toxicology, College of Preventive Medicine, Third Military Medical University, Chongqing, China; 2School of Public Health, Ningxia Medical University, Yinchuan, China;

Keywords chromatin integrity, sleep duration, sperm Correspondence Jia Cao, PhD, Institute of Toxicology, College of Preventive Medicine, Third Military Medical University, Chongqing 400038, China. Tel.: +023-6875-2289; fax: +023-6875-2276; e-mail: [email protected]; and Zhihong Cui, PhD, Institute of Toxicology, College of Preventive Medicine, Third Military Medical University, Chongqing 400038, China. Tel.: +023-6875-2291; fax: +023-6875-2276; e-mail: [email protected] *These authors contributed equally to this work. Accepted in revised form 4 September 2017; received 21 March 2017

SUMMARY

This study explores whether sleep duration is associated with sperm chromatin integrity. To do so, we conducted a three-phase panel study of 796 male volunteers from colleges in Chongqing (China) from 2013 to 2015. Sleep duration was measured using a modified Munich Chronotype Questionnaire. Sperm DNA integrity was examined via Sperm Chromatin Structure Assay and Comet assay. Setting 7–7.5 h day 1 of sleep duration as a reference, either longer or shorter sleep duration was associated negatively with high DNA stainability (HDS) (P = 0.009), which reflected the immaturity of sperm chromatin. The volunteers with > 9.0 h day 1 sleep and those with ≤ 6.5 h day 1 sleep had 40.7 and 30.3% lower HDS than did volunteers with 7–7.5 h day 1 sleep. No association was found between sleep duration and DNA fragmentation index or Comet assay parameters. This study suggests that sleep duration is associated with sperm chromatin integrity. Further studies are required to validate these findings and investigate the mechanism underlying this association.

DOI: 10.1111/jsr.12615

INTRODUCTION Sleep is necessary for health and wellbeing, yet modern industrialized societies have become sleep-deprived. It is common knowledge that people are getting 1–2 h less sleep per night compared with their ancestors (Roenneberg, 2013). According to previous publications, either restricted or excessive sleep duration is associated with increased disease risk, including obesity, diabetes, hypertension, ulcerative colitis and total mortality (Ananthakrishnan et al., 2014; Cappuccio et al., 2008; Lauderdale et al., 2008; Liu et al., 2013). However, research into the relationship between sleep duration and semen quality is just beginning. Recently, studies on animals showed that sleep deprivation will lead to a decrease in the number of live sperm and sperm motility (Alvarenga et al., 2015; Choi et al., 2016). Further, an inverse U-shaped association between sleep duration and semen parameters was found in 796 young males, thus ª 2017 European Sleep Research Society

indicating that sleep duration may play a pivotal role in the regulation of male reproductive health (Chen et al., 2016). It should be noted that all the publications on this topic investigated only routine semen parameters. Although these parameters have been found to be helpful in diagnosing male infertility, their power to evaluate a man’s fertility is limited, as some people who have normal semen parameters can still be infertile (Bonde et al., 1998; Guzick et al., 2001; Van Der Steeg et al., 2011). Sperm chromatin integrity has been documented as an independent predictor of male infertility (Bungum et al., 2012). Recently, certain indirect and direct clues have suggested that sleep duration can affect chromatin integrity, i.e. sleep restriction changed the level of reactive oxygen species which were important interruptors of chromatin structure (Alvarenga et al., 2015; Lobascio et al., 2015; Mathangi et al., 2012; Novotny et al., 2013; Villafuerte et al., 2015). Moreover, a study on 26 human volunteers reported that sleep restriction disrupted the gene expresson

1

2

X. Wang et al.

of both circadian rhythm and the oxidative stress pathway, and chromatin modification was found to be affected in whole blood (Moller-Levet et al., 2013). Nevertheless, to our knowledge, the association between sleep duration and sperm chromation integrity has not been reported to date. To address this gap, we measured sperm chromatin integrity using two methods [Sperm Chromatin Structure Assay (SCSA) and Comet assay] and analysed its association with sleep duration in a non-clinical population of 796 males in China. MATERIALS AND METHODS Study design The Male Reproductive Health in the Chongqing College Students (MARHCS) cohort was established to investigate the effect of environmental factors and socio-psycho-behavioural factors on male reproductive health. We conducted the baseline survey in June 2013 and the follow-up surveys (twice) during the same season (May–June) in 2014 and 2015. At each visit, the subjects were asked to provide semen samples, undergo physical examination and complete a composited questionnaire, including responses regarding sleep-related issues. The study was approved by the Ethics Committees of the Third Military Medical University. We also received signed informed consent from each participant. Population The male college students were recruited from the Chongqing University town in China. Volunteers with any abnormal situation in their urogenital system and volunteers with abstinence time less than 2 days or longer than 7 days were excluded. In the end, 872 volunteers participated in our survey and 796 of them were eligible from the baseline survey. A total of 656 (82.4%) and 568 (71.4%) of these subjects participated in the follow-up surveys conducted in 2014 and 2015. Questionnaire We used a modified Munich Chronotype Questionnaire (MCTQ) for the measurement of sleep duration. We calculated sleep duration according to information regarding certain time-points, including the time needed to get ready for sleep (differentiated from the time to go to bed), the duration from wake to sleep after someone went to bed and the time to wake (differentiated from the time to get out of bed). During the follow-ups, sleep disturbance was also measured using the Pittsburgh Sleep Quality Index (PSQI). A higher PSQI total score indicated worse sleep quality (Buysse et al., 1989; Liu et al., 1996). Further, information on potential confounders such as age, abstinence time, tobacco smoking, alcohol consumption and intake of coffee, coke and tea was also obtained via the questionnaire.

Sampling of semen and physical examination The physical examination was performed by an experienced urologist. Body mass index (BMI) was calculated as weight divided by height squared (kg m 2). The subjects were assigned to private rooms to collect semen samples by masturbation. SCSA Fresh semen samples were diluted to a concentration of 2 9 106 mL 1 in Tris-NaCl-ethylenediamine tetraacetic acid buffer (TNE buffer. Diluted semen samples (50 lL) were thawed at 37 °C and treated with 100 lL acid detergent (vibrated for 30 s), then stained with 300 lL acridine orange (AO; Invitrogen, Carlsbad, CA, USA) staining solution. AO was intercalated into the DNA strand in this process. Subsequently, semen samples were loaded into the flow cytometer (Beckman FC 500 MCL/MPL; Beckman Coulter, Brea, CA, USA) and 10 000 events were tested for each sample. When excited by a blue light, the AO intercalated in double-strand DNA emits green fluorescence, whereas the AO intercalated with single-strand DNA emits red fluorescence. DFI was defined as the ratio between red and total (red+green) fluorescence intensity. Total % DFI was defined as the percentage of sperm outside the main sperm population. High DNA stainability (HDS) was the percentage of sperm with high green fluorescence. All reagents, such as the TNE buffer, acid detergent and AO staining solution, were prepared according to Evenson et al. (2013). To ensure the validity of the measurements, a standard sample was tested when every 10–15 samples were measured. SCSA and Comet assay were performed within 3–6 months after storage by the same work-group according to the same protocol. Each sample underwent one freeze–thaw cycle. Fig. 1a,b shows an example of SCSA. Comet assay The semen sample was diluted with phosphate-buffered saline to a concentration of 4 9 106/mL and 40 lL diluted semen sample was mixed with 200 lL 0.6% low meltingpoint agarose gel (Sigma Aldrich Co., St Louis, MO, USA). The mixture was then added to a slide covered with 1% normal melting-point agarose gel before assay. The slide was coverslipped and kept at 4 °C for 15 min. The coverslip was removed and the slide was submersed in lysing solution at 4 °C for 1 h. The slide was then treated with enzyme solution at 37 °C for 16 h and the slide was immersed in alkaline electrophoresis solution for 20 min, leading to the unwinding of the double-strand DNA. The slide was electrophoresed at 4 °C for 8 min, 22 V (0.714 v cm 1) and 160 mA. The slide was immersed in neutralizing solution for 10 min and fixed with ethanol for 10 min. The slide was then air-dried and stained with ethidium bromide (20 mg mL 1). An inverted fluorescence microscope (ECLIPSE TE2000-S; Nikon ª 2017 European Sleep Research Society

Sleep duration and sperm chromatin integrity

3

Figure 1. Results of Sperm Chromatin Structure Assay (SCSA) and Comet assay. (a) Scatterplot, x-axis (FL4): red fluorescence with a scale from 0 to 1024; y-axis (FL1): green fluorescence with a scale from 0 to 1024. The sperm above the horizontal line have high DNA stainability (HDS), which reflects immatureness of sperm chromatin. HDS equals the percentage of sperm above this line. For example, HDS = 7.15%. (b) Histogram of DNA Fragmentation Index (DFI). Total % DFI equals the percentage of sperm outside the main sperm population. For example, total % DFI = 23.7%. (c) The measurement frame of Comet assay. It has two subframes: one for comet and the other for background (b). The frame is drawn on the screen and analysed by CASP software automatically. Tail DNA % equals the intensity ratio between tail and whole comet. Tail length (a) equals the distance between the right-most point of head and tail.

Corporation, Tokyo, Japan) was used to examine the slide at 9200 magnification. All reagents, such as lysing and enzyme solutions, were prepared according to Han et al. (2011). Tail length and the percentage of tail DNA were evaluated using CASP software (http://casplab.com/, version 1.2.2). To ensure the validity of the measurements, a standard sample was tested within each batch of samples. Fig. 1c shows an example of the Comet assay. Statistical analysis As sleep duration was in inverse U-shaped association with sperm chromatin integrity, two methods were used to investigate the association. (1) Sleep duration was categorized into an ordinal variable with an 0.5-h increment, e.g. 6.5–7, 7–7.5, 7.5–8 h day 1, etc. The data were separated into two parts, i.e. sleep duration ≤ 7.5 and sleep duration > 7 h day 1. An association analysis was performed respectively. (2) The sleep duration was transformed into a distance from 7 to 7.5 h day 1 (see Supporting information, Fig. S1), and the association between the distance and the chromatin integrity was explored. The 7–7.5-h day 1 sleep duration was used as the reference in the above-mentioned analyses, because it was reported to be the turning-point in the inverse U-shaped association between sleep duration and other semen parameters, and was similar to the turning-point in the association between sleep duration and many other phenotypes (Ananthakrishnan et al., 2014; Heckman et al., 2017). As the sperm chromatin integrity parameters were of skewed distribution, a Jonckheere–Terpstra test was performed as a univariate analysis. The association was analysed using multivariate linear regression, and potential confounders were chosen according to previous publications on sleep and semen quality (Chen et al., 2016; Jensen et al., 2013). Sleep duration may be associated with lifestyle factors ª 2017 European Sleep Research Society

such as tobacco smoking, alcohol consumption and the intake of tea, cola and coffee. Some covariates such as age, BMI and abstinence time were also adjusted for. Sleep disturbance (as PSQI scores) was adjusted for in the followup data, as it was suggested to be associated with sleep duration and semen quality (Jensen et al., 2013). To improve the skewed distribution, the chromatin integrity parameters were transformed into a logarithmic scale before analysis and then back-transformed as the percentage change, which indicated the relative difference of chromatin integrity parameters for the subjects with different sleep duration. As we had the repeated-measurement data from three surveys and an additional potential confounder was measured in later surveys, we first analysed the association of sleep duration and chromatin integrity in the baseline data and then in the follow-up data. Further, concerning the association between total sperm number and sleep duration, the association between sleep duration and sperm count with normal/abnormal chromatin integrity was also analysed separately. Finally, the data of three surveys were integrated using a mixed model, which was capable of dealing with the correlation within the subject. The mixed-model analysis was run using SAS (SAS Institute, version 9.1); other statistical analyses were run using SPSS (IBM SPSS statistics, version 18.0). A P-value < 0.05 was considered significant. RESULTS Demographic characteristics and sperm chromatin integrity The subjects were all young, with a median age of 20 years. The median BMI in our study was 20.9 kg m 2. Approximately 2.6% of the subjects were obese in this population

4

X. Wang et al.

(criteria for the Chinese population: BMI ≥ 28). Approximately half the subjects did not consume tobacco, alcohol and tea. The median of the entire sleep duration was 7.8 h (25th and 75th percentiles: 7.3 and 8.3 h, respectively) (Table 1). No significant difference in sperm chromatin integrity or sleep duration was found between the follow-up and lost subjects (data not shown). Association between sperm chromatin integrity and sleep duration In the baseline data, we found an inverse U-shaped association between HDS and sleep duration. The peak of HDS was found in the 7–7.5-h day 1 sleep group (Table 2): in subjects with sleep duration ≤ 7.5 h day 1, the sleep duration was associated positively with HDS (P = 0.008); in subjects with sleep duration > 7.0 h day 1, the sleep

Table 1 Demographic characteristics, sleep duration and sperm chromatin integrity of the subjects Characteristics

No. of subjects

Value*

Age, years Abstinence time, days Body mass index, kg m 2 Tobacco smoking Never Ever Current Alcohol consumption Never Ever Current Tea intake Never Ever Current Cola intake, bottles weeks 0 6 Coffee intake, cups weeks 0 6 Sleep duration, h day 1 SCSA parameters HDS, % Total % DFI, % Comet assay parameters Tail DNA% Tail length, lm

796 796 795

20 (20, 21) 4 (3, 6) 20.9 (19.6, 22.7)

593 30 171

74.7% 3.8% 21.5%

409 10 374

51.6% 1.3% 47.2%

512 123 159

64.5% 15.5% 20.0%

273 404 100 17

34.4% 50.9% 12.6% 2.1%

1

1

duration was associated negatively with HDS (P = 0.008). For the other chromatin integrity parameters, no significant differences were found for those with different sleep duration. After adjustment for potential confounders using multivariate linear regression, the association between sleep duration and HDS remained significant (Fig. 2). Compared to those with 7–7.5 h day 1 sleep duration, HDS declined by 40.7% [95% confidence interval (CI): 58.9%, 14.5%] and 30.3% (95% CI: 53.9, 5.0) in those subjects with sleep duration > 9 h day 1 and ≤ 6.5 h day 1 (Fig. 2). To represent comprehensively the association between sleep duration and HDS, we transformed the sleep duration into the distance from 7.0 to 7.5 h day 1 sleep. After adjusting for potential confounders, a negative association was observed between sleep duration distance and HDS: each hour distance from 7 to 7.5 day 1 sleep duration was associated with a 17.6% decrease of HDS (Table 3, 95% CI: 4.9%, 28.7%, P = 0.008). We applied the same analysis to the follow-up data from the 2014 survey, and again observed a significant association between HDS and sleep duration distance (P = 0.049). Additional adjustment for sleep disturbance (PSQI score) did not influence the results substantially (Table 3, P = 0.036). Further, we integrated the data of three surveys using a multilevel model; the result was similar with the separate results of each survey: each hour of sleep duration distance was associated with an 8.2% decrease of HDS (95% CI: 2.2%, 13.8%, P = 0.009, Table 3). Many people were woken artificially and thus had a sleep duration they did not prefer. We investigated further whether the interruption of sleep duration biased the association between sleep duration and the chromatin integrity of sperm. At baseline, the subjects were asked whether they were awakened by a clock on work days or free days. A total of 496 subjects replied that they used clocks either on work days or free days, and 294 replied that they never used them. HDS, total % DFI and the Comet assay parameters were not significantly different between the two groups (Supporting information, Table S2). When clock use was adjusted for, the associations between sleep duration and HDS remained significant in the baseline data (P = 0.008) and the follow-up data (P = 0.036). In the mixed-model analysis, each hour of sleep duration distance was associated with a 9.3% (95% CI: 0.9%, 17.1%, P = 0.032) decrease of HDS in the clock users and a 9.8% (95% CI: 0.9%, 17.9%, P = 0.038) decrease of HDS in the non-clock users.

605 154 21 14 738

76.2% 19.4% 2.6% 1.8% 7.8 (7.3, 8.3)

761 753

3.9 (2.3, 9.1) 11.1 (6.9, 19.5)

Association between sleep duration and sperm counts with normal/abnormal chromatin integrity

766 766

17 (13.9, 20.6) 29.4 (25.2, 33.8)

HDS was an index that represented the proportion of sperm with abnormal chromatin, and it was reported previously that sleep duration was associated with total sperm numbers, so we analysed further whether the sperm count with abnormal chromatin (Sa) and the count of sperm with normal chromatin (Sn) were associated with sleep duration. Sa was calculated as the total sperm number multiplied by HDS; Sn was

DFI, DNA fragmentation index; HDS: high DNA stainability; MARHC: Male Reproductive Health in Chongqing College Students; SCSA: Sperm Chromatin Structure Assay. *Data represented as ‘median (25th, 75th percentiles) or percentage.

ª 2017 European Sleep Research Society

Sleep duration and sperm chromatin integrity

5

Table 2 Sperm chromatin integrity parameters of the subjects with different sleep duration: univariate analysis in the baseline survey Sleep duration, h day P-trend* ≤ 6.5 6.5–7 7–7.5 (Ref.) 7.5–8 8–8.5 8.5–9 >9 P-trend†

1

No. of subjects

HDS, %

40 81 136 177 143 78 52

0.008 3.3 (2.4, 3.2 (2.0, 4.8 (2.5, 4.2 (2.4, 4.1 (2.5, 3.9 (2.1, 3.0 (1.9, 0.008

6.1) 7.7) 11.9) 10.1) 8.4) 10.3) 5.9)

Total %DFI, %

Tail DNA, %

Tail length, lm

0.409 11.8 (7.2, 10.4 (6.9, 10.7 (6.9, 11.3 (7.3, 12.4 (6.9, 11.2 (7.1, 9.7 (5.2, 0.681

0.41 17.6 (14.1, 17.0 (13.9, 16.2 (13.5, 17.9 (14.9, 16.7 (13.5, 17.3 (14.3, 16.4 (14.0, 0.691

0.332 27.8 (25.7, 29.8 (25.1, 27.7 (23.9, 30.6 (25.7, 29.5 (25.7, 28.9 (25.0, 29.8 (25.4, 0.864

20.3) 20.7) 18.1) 19.6) 21.2) 20.3) 15.0)

19.9) 21.2) 20.4) 21.3) 20.6) 19.7) 20.2)

Data represented as median (25th, 75th percentiles). *P-values were calculated by Jonckheere–Terpstra test, assuming that sleep duration no more than 7–7.5 h day with the parameters of chromatin integrity. † P-values were calculated by Jonckheere–Terpstra test, assuming that sleep duration no less than 7–7.5 h day with the parameters of chromatin integrity.

33.1) 34.0) 34.3) 34.7) 33.2) 32.9) 32.4)

1

were in linear association

1

were in linear association

Table 3 Association between HDS of sperm and sleep duration distance from 7–7.5 h day 1 HDS, % change per h day 1 sleep duration distance

Data †

Baseline data 1st follow-up data (adjusted for PSQI)‡ Integrated data of the three surveys§

Figure 2. Association between high DNA stainability (HDS) of sperm and sleep duration: multivariate analysis in the baseline survey. The association between HDS and sleep duration was analysed by linear regression, adjusted for age, body mass index, abstinence time, tobacco smoking, alcohol drinking, intake of tea, coffee and cola. To improve the skewed distribution of HDS, it was transformed into logarithmic scale before analyses and then back-transformed into percent change. Taking the HDS of the subjects with 7–7.5 h day 1 sleep as the reference (solid triangle), the results indicated the percent change in HDS (solid circle) for the subjects with a certain sleep duration. The error bars indicated the 95% confidential interval. P-values were also given, assuming that sleep duration below 7–7.5 h day 1 and sleep duration above 7–7.5 h day 1 were, respectively, in linear association with HDS.

calculated as the total sperm number minus Sa. We found that both Sa and Sn were associated with sleep duration in an inverse U-shaped pattern (Supporting information, Table S1). On average, each hour of sleep duration distance was associated with an 18.7% (95% CI: 10.4%, 26.2%, P < 0.0001) decrease of Sa and a 9.6% (95% CI: 3.2%, 15.6%, P = 0.004) decrease of Sn. According to the method described by Zeka et al. (2006), the decrease in Sa was ª 2017 European Sleep Research Society

95% CI

P-value

17.6 6.5

28.7, 12.1,

4.9 0.5

0.008* 0.036*

8.2

13.8,

2.2

0.009*

This table represented the associations between sleep duration distance and high DNA stainability (HDS) of sperm. The distance indicated the difference (in absolute value) between a sleep duration and 7–7.5 h day 1 sleep, as illustrated in Supporting information, Fig. S1. The analysis was performed first in the baseline survey, and then in the first follow-up survey for additional adjustment for Pittsburgh Sleep Quality Index (PSQI), and last in the integrated data of all three surveys. To improve the skewed distribution of HDS, it was transformed into logarithmic scale before analyses and then back-transformed into percentage change. The results indicated the percentage change in HDS with each hour difference from 7–7.5 h day 1 sleep duration. This percentage was a relative ratio compared to the HDS of the subjects with 7–7.5 h day 1 sleep. The 95% confidence interval (CI) and P-value were also given. *P < 0.05. † Analysed by linear regression, adjusted for age, body mass index, abstinence time, tobacco smoking, alcohol drinking, intake of tea, coffee and cola. ‡ Analysed by linear regression, adjusted for PSQI besides age, body mass index, abstinence time, tobacco smoking, alcohol consumption, intake of tea, coffee and cola. § Analysed by mixed model, adjusted for age, body mass index, abstinence time, tobacco smoking, alcohol drinking, intake of tea, coffee and cola.

larger than the decrease in Sn (95% CI: 3.1, 21.1, P = 0.07). This was consistent with the result that sleep duration distance was associated negatively with HDS.

6

X. Wang et al.

DISCUSSION Based on a population of 796 male college students, for the first time we found an inverse U-shaped association between sleep duration and HDS. The phenomenon was observed repeatedly in the baseline and follow-up surveys. Our results reinforced the hypothesis that sleep behaviour is associated with semen quality, emphasizing further the necessity to investigate comprehensively the effect of sleep behaviour on male reproductive health. Higher HDS means there is a lack of normal exchange from histones to protamines, a process important for chromatin condensation (Evenson, 2013). Studies that focus on the clinical utility of SCSA parameters are growing. Some studies have reported that higher HDS was correlated with lower progressive motility, normal morphology rate of sperm, poorer fertilization rates and pregnancy with assisted reproductive techniques (Nijs et al., 2009; Novotny et al., 2013; Payne et al., 2005; Virro et al., 2004; Zini et al., 2009). It is reported that the HDS > 15% indicates an adverse clinical consequence (Virro et al., 2004). However, the results from different studies have not always been consistent. Buck Louis et al. (2014) did not find an association between HDS and time to pregnancy. It is not completely clear what HDS represents regarding reproductive function, but it can be regarded as a potential marker of sperm chromatin integrity when applied to assisted reproduction. The association between sleep duration and HDS in our study indicates that there could be an alteration in chromatin integrity of sperm when a man’s sleep duration changes. The HDS was relatively low in the young and healthy subjects in the present study, but it varied by one-third (3.0 versus 4.8%), depending on sleep duration. It is sensible, therefore, to anticipate that this effect could lead to a more substantial change in the sperm of more susceptible individuals, such as elderly or subfertile males. Unlike HDS, total % DFI and the parameters of the Comet assay indicate DNA strand breakage. In the present study, however, no association was found between sleep duration and total % DFI or the Comet assay parameters. The negative results for both assays were mutually supportive, thereby suggesting that unsuitable sleep duration was not associated with DNA breakage. It is accepted widely that the association of sleep duration and hazardous phenotypes is usually U-shaped (Ananthakrishnan et al., 2014; Liu et al., 2013). Regarding male reproductive health, a previous paper reported an inverse U-shaped association between sleep duration and total sperm numbers. However, in the same population, the present study showed that either longer or less sleep duration was associated with lower HDS. The simultaneous associations of sleep duration with HDS and total sperm count are not attributable to the correlation between these two semen parameters (P = 0.732 using Spearman’s correlation). These two parameters are thought to represent different aspects of fertility. HDS indicates the proportion of incompletely differentiated sperm in the epididymis. For 7–7.5-h day 1 sleep

this proportion increases, while the normal sperm count also reaches its peak. This phenomenon is somewhat unexpected, but in other sleep-related studies there is already evidence that the same duration may have distinct effects on different aspects of function (Zohar et al., 2005), revealing a complex association between sleep duration and male reproductive health that should be interpreted with caution. Whether sleep duration modifies the success pregnancy rate and the development of offspring deserves further investigation in future. Currently, the mechanism of the sleep–semen association is not clear. The role of reproductive hormones may be excluded, because they were not associated with sleep duration in this population. This result was also confirmed by Jensen et al. (2013). The circadian clock may be an important factor that induces the sleep–semen association. The circadian clock exists in both the brain and peripheral tissues, including the testis (Archer et al., 2014). It has been found to be necessary for production of mature spermatozoa and fertility in animal models (Alvarez et al., 2008; Beaver et al., 2002; Tobback et al., 2012). More importantly, the circadian genes could catalyze the chromatin modifications (Doi et al., 2006). As an important regulator of circadian genes, improper sleep behaviour is likely to disrupt the circadian genes, resulting in adverse outcomes in the male reproductive system (Boden et al., 2013; Gamble et al., 2013). Scrotal heating could also be involved. Higher temperature in the scrotal area has been reported to induce an increase of HDS (Ahmad et al., 2012; Rao et al., 2016). The scrotal temperature fluctuates with diurnal variation and becomes higher at night (Lerchl et al., 1993). Hence, sleep behaviour may affect the scrotal temperature by regulating the circadian rhythm or the warm circumstances in bed. These hypotheses need in-depth validation in future studies. This study has several strengths. First, we applied MCTQ to measure the subjects’ sleep duration. The traditional methods of sleep duration assessment use direct questions requiring the subjects to recall the length of sleep. It is easy to induce a misunderstanding of the correct concept of sleep duration, which can lead to systemic error (Lauderdale et al., 2008). MCTQ avoided this shortcoming by recording the time-point of sleep behaviours that were defined and differentiated clearly from the incorrect concepts that could lead to misunderstanding. It has also been shown to be an accurate €hnle, 2006). Secondly, we tool to assess sleep duration (Ku used both SCSA and Comet assays to detect sperm chromatin integrity. The Comet assay is a classical method used to detect DNA breakage. SCSA is a developed method used to detect both DNA damage and chromatin maturation. It can analyse thousands of cells without objective analysis (Tamburrino et al., 2012). It was standardized by Evenson, and its result is reproducible and robust (Evenson, 2013). Thirdly, we used the data from baseline survey and follow-up surveys to investigate the sleep–HDS association separately and found similar results. This could help us to minimize the possibility of chance error. Further, the subjects in this study ª 2017 European Sleep Research Society

Sleep duration and sperm chromatin integrity had similar demographic characteristics and lived in a similar environment, hence the bias caused by certain important potential confounders, such as age, could be excluded. The three surveys were also completed in the same season, thus avoiding any seasonal influence on the semen parameters. There were several limitations to this study. First, the present study was a cross-sectional study. The possibility of reverse causation could not be ruled out, although it seems unlikely that semen quality would affect sleep duration. Alteration in sleep duration may be an indicator of stress or other diseases. However, the subjects were young college students and should have no idea of their personal fertility. It was unlikely that they had suffered psychological stress from infertility. Further, only four and five individuals, respectively, took sleeping pills in the follow-up surveys of 2014 and 2015, and those doses were extremely low. Thus, sleep-related medication was not likely to be a bias for our results. Studies with higher strength for causal inference, such as randomized controlled trials, would be necessary to validate such causality in the future. Secondly, no correction for multiple comparisons was performed. In fact, we investigated sleep duration using only four parameters of chromatin integrity. The inverse U-shaped association with HDS (P < 0.01) would remain significant even if Bonferroni’s correction was performed (threshold of significance = 0.05/4 = 0.0125). Thirdly, some potential confounders, such as socioeconomic status, were not included. Further studies are thus warranted to validate our findings. Fourthly, the subjects in the present study were young and healthy. Those individuals with reproductive diseases and other severe diseases were excluded which may, of course, limit the generalizability of our results to other males with different health statuses. In conclusion, we found an inverse U-shaped association between sleep duration and sperm HDS. Subjects with sleep duration between 7 and 7.5 h day 1 had the highest proportion of chromatin-abnormal sperm. This finding, in combination with the inverse U-shaped association between sleep duration and total sperm number, indicates usefully the complexity of the association between sleep duration and male reproductive health. Further studies are needed to validate our findings and investigate the mechanism underlying this association. ACKNOWLEDGEMENTS We thank the college students who participated in our study. We also acknowledge all the fieldworkers in the MARHCS study team. This work was supported by the National Natural Science Funding of China (grant no. 81402660) and the National Key Research and Development Program of China [2017YFC1002001]. CONFLICT OF INTEREST All the authors approved the final manuscript and declare no potential conflicts of interest. ª 2017 European Sleep Research Society

7

AUTHOR CONTRIBUTIONS XW and QC contributed to statistical analyses, interpretation of the data and drafted the paper. The study was conceived and designed by ZC and JC. Data acquisition was conducted by XW, QC, PZ, TL, MM, HY, NZ, LS, HC, XL, KP, LA and HY. All co-authors interpreted the data and participated in finalizing the manuscript. All co-authors approved the final version of the paper.

REFERENCES Ahmad, G., Moinard, N., Esquerre-Lamare, C., Mieusset, R. and Bujan, L. Mild induced testicular and epididymal hyperthermia alters sperm chromatin integrity in men. Fertil. Steril., 2012, 97: 546–553. Alvarenga, T. A., Hirotsu, C., Mazaro-Costa, R., Tufik, S. and Andersen, M. L. Impairment of male reproductive function after sleep deprivation. Fertil. Steril., 2015, 103: 1355–1362 e1. Alvarez, J. D., Hansen, A., Ord, T. et al. The circadian clock protein BMAL1 is necessary for fertility and proper testosterone production in mice. J. Biol. Rhythms, 2008, 23: 26–36. Ananthakrishnan, A. N., Khalili, H., Konijeti, G. G. et al. Sleep duration affects risk for ulcerative colitis: a prospective cohort study. Clin. Gastroenterol. Hepatol., 2014, 12: 1879–1886. Archer, S. N., Laing, E. E., Moller-Levet, C. S. et al. Mistimed sleep disrupts circadian regulation of the human transcriptome. Proc. Natl Acad. Sci. USA, 2014, 111: E682–E6891. Beaver, L. M., Gvakharia, B. O., Vollintine, T. S., Hege, D. M., Stanewsky, R. and Giebultowicz, J. M. Loss of circadian clock function decreases reproductive fitness in males of Drosophila melanogaster. Proc. Natl Acad. Sci. USA, 2002, 99: 2134–2139. Boden, M. J., Varcoe, T. J. and Kennaway, D. J. Circadian regulation of reproduction: from gamete to offspring. Prog. Biophys. Mol. Biol., 2013, 113: 387–397. Bonde, J. P., Ernst, E., Jensen, T. K. et al. Relation between semen quality and fertility: a population-based study of 430 first-pregnancy planners. Lancet, 1998, 352: 1172–1177. Buck Louis, G. M., Sundaram, R., Schisterman, E. F. et al. Semen quality and time to pregnancy: the Longitudinal Investigation of Fertility and the Environment Study. Fertil. Steril., 2014, 101: 453–462. Bungum, M., Bungum, L., Lynch, K. F., Wedlund, L., Humaidan, P. and Giwercman, A. Spermatozoa DNA damage measured by sperm chromatin structure assay (SCSA) and birth characteristics in children conceived by IVF and ICSI. Int. J. Androl., 2012, 35: 485–490. Buysse, D. J., Reynolds, C. F. III, Monk, T. H., Berman, S. R. and Kupfer, D. J. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res., 1989, 28: 193–213. Cappuccio, F. P., Taggart, F. M., Kandala, N. et al. Meta-analysis of short sleep duration and obesity in children and adults. Sleep, 2008, 31: 619. Chen, Q., Yang, H., Zhou, N. et al. Inverse U-shaped association between sleep duration and semen quality: longitudinal observational study (MARHCS) in Chongqing, China. Sleep, 2016, 39: 79– 86. Choi, J. H., Lee, S. H., Bae, J. H. et al. Effect of sleep deprivation on the male reproductive system in rats. J. Korean Med. Sci., 2016, 31: 1624–1630. Doi, M., Hirayama, J. and Sassone-Corsi, P. Circadian regulator CLOCK is a histone acetyltransferase. Cell, 2006, 125: 497–508. Evenson, D. P. Sperm chromatin structure assay (SCSA(R). Methods Mol. Biol., 2013, 927: 147–164.

8

X. Wang et al.

Gamble, K. L., Resuehr, D. and Johnson, C. H. Shift work and circadian dysregulation of reproduction. Front. Endocrinol., 2013, 4: 92. Guzick, D. S., Overstreet, J. W., Factor-Litvak, P. et al. Sperm morphology, motility, and concentration in fertile and infertile men. N. Engl. J. Med., 2001, 345: 1388–1393. Han, X., Zhou, N., Cui, Z. et al. Association between urinary polycyclic aromatic hydrocarbon metabolites and sperm DNA damage: a population study in Chongqing, China. Environ. Health Perspect., 2011, 119: 652–657. Heckman, C. J., Kloss, J. D., Feskanich, D., Culnan, E. and Schernhammer, E. S. Associations among rotating night shift work, sleep and skin cancer in Nurses’ Health Study II participants. Occup. Environ. Med., 2017, 74: 169–175. Jensen, T. K., Andersson, A. M., Skakkebaek, N. E. et al. Association of sleep disturbances with reduced semen quality: a crosssectional study among 953 healthy young Danish men. Am. J. Epidemiol., 2013, 177: 1027–1037. €hnle, T. Quantitative Analysis of Human Chronotypes. DissertaKu € t fu € r Biologie, LMU Mu € nchen, Germany, 2006. tion, Fakulta Lauderdale, D. S., Knutson, K. L., Yan, L. L., Liu, K. and Rathouz, P. J. Self-reported and measured sleep duration: how similar are they? Epidemiology, 2008, 19: 838–845. Lerchl, A., Keck, C., Spiteri-Grech, J. and Nieschlag, E. Diurnal variations in scrotal temperature of normal men and patients with varicocele before and after treatment. Int. J. Androl., 1993, 16: 195–200. Liu, X., Tang, M. and Hu, L. Reliability and validity of the Pittsburgh sleep quality index. Chin. J. Psychiatry, 1996, 29: 103–107. Liu, Y., Wheaton, A. G., Chapman, D. P. and Croft, J. B. Sleep duration and chronic diseases among U.S. adults age 45 years and older: evidence from the 2010 Behavioral Risk Factor Surveillance System. Sleep, 2013, 36: 1421–1427. Lobascio, A. M., De Felici, M., Anibaldi, M., Greco, P., Minasi, M. G. and Greco, E. Involvement of seminal leukocytes, reactive oxygen species, and sperm mitochondrial membrane potential in the DNA damage of the human spermatozoa. Andrology, 2015, 3: 265–270. Mathangi, D. C., Shyamala, R. and Subhashini, A. S. Effect of REM sleep deprivation on the antioxidant status in the brain of Wistar rats. Ann. Neurosci., 2012, 19: 161–164. Moller-Levet, C. S., Archer, S. N., Bucca, G. et al. Effects of insufficient sleep on circadian rhythmicity and expression amplitude of the human blood transcriptome. Proc. Natl Acad. Sci. USA, 2013, 110: E1132–E1141. Nijs, M., Creemers, E., Cox, A. et al. Chromomycin A3 staining, sperm chromatin structure assay and hyaluronic acid binding assay as predictors for assisted reproductive outcome. Reprod. Biomed. Online, 2009, 19: 671–684. Novotny, J., Aziz, N., Rybar, R. et al. Relationship between reactive oxygen species production in human semen and sperm DNA damage assessed by Sperm Chromatin Structure Assay. Biomed Pap Med Fac Univ Palacky, Olomouc, Czechoslovakia, 2013, 157: 383–386.

Payne, J. F., Raburn, D. J., Couchman, G. M., Price, T. M., Jamison, M. G. and Walmer, D. K. Redefining the relationship between sperm deoxyribonucleic acid fragmentation as measured by the sperm chromatin structure assay and outcomes of assisted reproductive techniques. Fertil. Steril., 2005, 84: 356–364. Rao, M., Xia, W., Yang, J. et al. Transient scrotal hyperthermia affects human sperm DNA integrity, sperm apoptosis, and sperm protein expression. Andrology, 2016, 4: 1054–1063. Roenneberg, T. Chronobiology: the human sleep project. Nature, 2013, 498: 427–428. Tamburrino, L., Marchiani, S., Montoya, M. et al. Mechanisms and clinical correlates of sperm DNA damage. Asian J. Androl., 2012, 14: 24–31. Tobback, J., Boerjan, B., Vandersmissen, H. P. and Huybrechts, R. Male reproduction is affected by RNA interference of period and timeless in the desert locust Schistocerca gregaria. Insect Biochem. Mol. Biol., 2012, 42: 109–115. Van Der Steeg, J. W., Steures, P., Eijkemans, M. J. et al. Role of semen analysis in subfertile couples. Fertil. Steril., 2011, 95: 1013– 1019. Villafuerte, G., Miguel-Puga, A., Rodriguez, E. M., Machado, S., Manjarrez, E. and Arias-Carrion, O. Sleep deprivation and oxidative stress in animal models: a systematic review. Oxidat. Med. Cell. Longevity, 2015, 2015: 234952. Virro, M. R., Larson-Cook, K. L. and Evenson, D. P. Sperm chromatin structure assay (SCSA) parameters are related to fertilization, blastocyst development, and ongoing pregnancy in in vitro fertilization and intracytoplasmic sperm injection cycles. Fertil. Steril., 2004, 81: 1289–1295. Zeka, A., Zanobetti, A. and Schwartz, J. Individual-level modifiers of the effects of particulate matter on daily mortality. Am. J. Epidemiol., 2006, 163: 849–859. Zini, A., Phillips, S., Courchesne, A. et al. Sperm head morphology is related to high deoxyribonucleic acid stainability assessed by sperm chromatin structure assay. Fertil. Steril., 2009, 91: 2495– 2500. Zohar, D., Tzischinsky, O., Epstein, R. and Lavie, P. The effects of sleep loss on medical residents’ emotional reactions to work events: a cognitive-energy model. Sleep, 2005, 28: 47–54.

SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article: Figure S1. Computation of sleep duration “distance”. Table S1. Association between sleep duration “distance” and the count of sperms with abnormal/normal chromatin integrity Table S2. Parameters on sperm chromatin integrity of the subjects

ª 2017 European Sleep Research Society