Night Shift Work, MTNR1B rs10830963 Polymorphism, and Prostate Cancer Risk: Findings from a Prospective, Population-Based Study

Prostate cancer is the second most frequently diagnosed cancer and the fifth leading cause of death in males worldwide (1). Numerous studies demonstrate that aging (2), ethnicity (3), and family history of prostate cancer (4) are the well-established risk factors of prostate cancer. Other modifiable lifestyle factors, such as alcohol consumption (5) and cigarette smoking (6) were also associated with prostate cancer.

Night shift work, which wreaks havoc on the circadian rhythm, has been classified as a probable carcinogen by the International Agency for Research on Cancer (7). However, to date, the association between night shift work and prostate cancer has not been clearly established. A recent meta-analysis (8) concluded that an association between rotating/night shift work and prostate cancer cannot be confirmed with the available current data. Specifically, out of nine prospective cohort studies, only one shows a significant association, and others failed to find a significant association. Evidence has shown that genetic and environmental factors both contribute to the genesis of prostate cancer (9), which raises the concern that the inconsistent findings from previous studies may be partly due to potential interaction between night shift work and genetic variation in genes involved in the circadian rhythm.

Previous studies show that the melatonin receptor 1B gene (MTNR1B) rs10830963 influences the dynamics of melatonin secretion (10, 11). Melatonin, which plays an important role in regulating circadian rhythm, could inhibit the growth and progression of prostate cancer through the promotion of androgen receptor exclusion, modulation of prostate cancer metabolism, inhibition of angiogenesis, regulation of neuroendocrine differentiation, and induction of apoptosis (12). Furthermore, the bright light exposure during night shift work also dysregulates the production of melatonin (13). Besides, it is well known that rs10830963G increases the risk of type 2 diabetes mellitus (T2DM; ref. 14). Of interest, numerous studies indicate that there is an inverse association between T2DM and prostate cancer (15–17).

In light of these findings, we hypothesized that MTNR1B rs10830963 interacted with night shift work to modify the risk of prostate cancer. In the current study, we aimed to examine the potential interaction between night shift work and rs10830963 on the incidence and mortality of prostate cancer among male employed participants from the UK Biobank cohort.

The UK Biobank is a population-based cohort that recruited approximately 500,000 participants in the United Kingdom between 2006 and 2010 (18). Participants visited one of 22 assessment centers throughout the United Kingdom and underwent detailed baseline assessment including characterization of sociodemographics, lifestyles, medical history if they consented to participate. In the current study, inclusion criteria were as follows: (i) male; (ii) being in paid employment or self-employed; (iii) having available genotyping data, and (iv) no diagnosis of prostate cancer at baseline (Fig. 1). UK Biobank received ethical approval from the National Health Service (NHS) National Research Ethics Service (16/NW/0274). All participants gave written informed consent before enrolment in the study.

Information about night shift work analyzed in the current study was self-reported at baseline.

Night shift work was obtained by asking “Does your work involve night shifts?” with responses of (i) never/rarely, (ii) sometimes, (iii) usually, (iv) always, (v) do not know, or (vi) prefer not to answer. Night shift work is defined as a work schedule that involves working through the normal sleeping hours, for instance, working through the hours from 12 am to 6 am. The information about night shift work was collected from participants who indicated they were in paid employment or self-employed and excepting those who indicated their work never or rarely involved a shift pattern, as defined by their answers to “Does your work involve shift work?”. The night shift was participants reported taking night shift work. It was defined as “night shift work” with any response of “sometimes,” “usually,” and “always.” In this variable (ID 3426), “do not know” and “prefer not to answer,” were treated as missing data.

Participants were genotyped on the Affymetrix UK Biobank Lung Exome Evaluation Axiom array or the Applied Biosystems UK Biobank Axiom array. MTNR1B rs10830963 SNP (chromosome 16) was amongst the directly genotyped SNPs of the UK Biobank. Quality control and imputation were conducted centrally by Haplotype Reference Consortium and UK10K haplotype resource. Testing for Hardy–Weinberg equilibrium revealed that the SNP did not deviate from the expected genotype proportion.

The primary outcome is the incidence of prostate cancer, and the secondary outcome is the mortality from prostate cancer. Information on prostate cancer diagnosis was obtained from the cancer register in UK Biobank. UK Biobank obtains data on cancer diagnoses through cancer registries. Cancer registries acquire information on cancer diagnosis from a variety of sources including hospitals, cancer centers and treatment centers, hospices and nursing homes, private hospitals, cancer screening programmers, other cancer registers, general practices, death certificates, Hospital Episode Statistics and Cancer Waiting Time data. The date and cause of death were obtained from death registries of the NHS Information Centre for participants from England and Wales and the NHS Central Register Scotland for participants from Scotland. At the time of analysis, cancer registries were available with the last linkage date of December 14, 2016 and mortality data were available up to March 23, 2021. Therefore, for incident prostate cancer, we used this date (December 14, 2016) as the end of follow-up unless death or admission occurred first. Concerning mortality from prostate cancer, we censored follow-up at this date (March 23, 2021) or the date of death, whichever came first.

We considered several factors to be potential confounders: age (Field ID 21022), ethnic background (Field ID 21000), education (highest qualification they owned; Field ID 6138), location derived from assessment center (Field ID 54), Townsend deprivation index reflecting area-level socioeconomic deprivation (Field ID 189), smoking status (Field ID 20116), alcohol intake frequency (Field ID 1558), living with wife or partner (Field ID 6141), obesity derived from BMI (Field ID 21001), family history of prostate cancer derived from illness of father (Field ID 20107, Code 13) and illnesses of siblings (Field ID 20111, Code 13), ever had prostate specific antigen tests (Field ID 2365), insomnia (Field ID 1200), and sleep duration (Field ID 1160), and chronotype (Field ID 1180). Detailed information is provided in Supplementary Table S1.

Data were presented as the mean ± SD for continuous variables and percentages for categorical variables. Cox regression analysis was used to calculate HRs and 95% confidence intervals (CI) for associations of night shift and MTNR1B rs10830963 with prostate cancer. We created a multiplicative interaction term between the night shift work and MTNR1B rs10830963 to test the interaction on the incidence and mortality of prostate cancer with included main effects. Participants who were diagnosed with prostate cancer before the date of baseline assessment were excluded from the current study. The primary analysis (model 1) was the model adjusted for age (continuous), ethnicity (White/other), education (college or university degree/other), location (England/Wales/Scotland), and Townsend deprivation index (continuous). Model 2 was additionally adjusted for lifestyles including smoking status (never/previous/current), alcohol intake frequency (never or special occasions only/once or twice a week or one to three times a month/three or more times a week), and living with wife or partner or not (yes/no). Further adjustment (model 3) was made for obesity (yes/no), family history of prostate cancer (yes/no), and ever had prostate specific antigen tests (yes/no). In the final model (model 4), we further controlled for sleep duration (<6 hours/6–8 hours/>8 hours), insomnia symptoms (never/rarely/sometimes/usually), and chronotype (morningness/eveningness).

To maximize the sample size, participants with missing variables were excluded in each model and the number of cases and censored participants were shown in the tables. On the basis of the final model, adjusted cumulative incidence curves were then generated to show the risk of prostate cancer according to night shift work and MTNR1B rs10830963 genotypes (19). The proportional hazards assumption was examined by creating a product term of follow-up time and night shift work, and we found no significant deviation from the assumption. The sample size was sufficient enough for statistical power in Cox regression models for the incidence but not the mortality of prostate cancer (20). However, mortality from prostate cancer is also a meaningful outcome for prostate disease, and therefore the results of mortality were presented in Supplementary Tables. Because of no death from prostate cancer among the GG carriers who took night shift work, we combined the GC carriers and GG carriers and presented the results for associations of night shift and MTNR1B rs10830963 with mortality from prostate cancer with four groups. Besides, we conducted the sensitivity analyses restricted to participants who were unrelated individuals of white British ancestry. To test the potential variations in different subgroups, we repeated the analyses on the incidence of prostate cancer stratified by chronotype. Sensitivity analyses and subgroup analyses on the mortality from prostate cancer were not further conducted because of limited events. All statistical analyses were performed using SPSS 26.0 and statistical significance was prespecified at a P < 0.05.

This research was conducted using the UK Biobank Resource under application number 58082.

General characteristics of the male employed participants, separated by night shift work and rs10830963 polymorphism, are shown in Table 1. From a total sample of 133,416 participants, the mean (SD) age at the initial recruitment was 53.13 (7.34) years, and genotype distribution of rs10830963 was 52.8% for CC, 39.4% for CG, and 7.8% for GG. A total of 15,298 (11.5%) male workers reported taking night shift works. The mean follow-up period is 7.76 years for the incidence of prostate cancer and 12.17 years for the mortality from prostate cancer, and the maximum follow-up period was 10.74 years and 15.01 years, respectively. During the follow-up period, 2646 (1.98%) participants were diagnosed with prostate cancer and 245 participants died because of prostate cancer. Missing data for each variable were shown in Table 1. Those who took night shift work were younger, and there was a greater proportion of Wales and Scotland population, non-White ethnicity, lower education, greater degree of material deprivation, not living with wife or partners, smokers but consumed alcohol less frequently, obesity, sleep for abnormal durations, insomnia symptoms, and eveningness type. In addition, a smaller proportion of night shift workers reported a family history of prostate cancer and ever had prostate specific antigen tests than those without night shift work.

Supplementary Table S2 shows that there were no significant associations of night shift work (HR: 1.07; 95% CI: 0.92–1.25) and rs10830963 polymorphism (GC vs. CC; HR: 0.94; 95% CI: 0.86–1.03; GG vs. CC; HR: 0.89; 95% CI: 0.75–1.05) with prostate cancer, but a significant interaction was found between night shift work exposure and rs10830963 polymorphism on the incidence of prostate cancer (P_interaction = 0.009). Table 2 shows the results of Cox regression analysis of rs10830963 genotypes and night shift work on prostate cancer among male employed participants. Among non–night shift workers, allele G was not associated with prostate cancer compared with CC carriers. Among night shift workers, compared with those with CC genotype, GC carriers had a significantly lower risk of prostate cancer and it retained significance upon adjustment for multiple confounders (HR: 0.69; 95% CI: 0.51–0.93). Similar associations were more evident for GG carriers, which exhibited a significantly decreased risk of prostate cancer (HR: 0.33; 95% CI: 0.15–0.70) and this association remained strong in the final model. The cumulative incidence curves for cases among male employed participants are shown in Fig. 2.

Supplementary Table S3 further shows the results from the perspective of treating night shift work as exposure. Among CC carriers, night shift workers exhibited a significantly increased risk of prostate cancer (HR: 1.29; 95% CI: 1.06–1.55) compared with non–night shift workers. On the contrary, among GG carriers, night shift workers showed a lower risk than non-shift workers (HR: 0.45; 95% CI: 0.20–1.01). Sensitivity analyses were further conducted restricted to those unrelated individuals of White British ancestry (Supplementary Table S4) and subgroups analyses on the associations of night shift work with the incidence of prostate cancer were also conducted stratified by chronotype (Supplementary Table S5). The findings were largely consistent with those in the main analyses.

Supplementary Table S6 shows that there were no significant associations of night shift work (HR: 1.03; 95% CI: 0.63–1.68) and rs10830963 polymorphism (GC/GG vs. CC; HR: 0.93; 95% CI: 0.71–1.22) with mortality from prostate cancer, and no significant interaction was found (P_interaction = 0.51). Supplementary Table S7 shows the results of Cox regression analysis of rs10830963 genotypes and night shift work on mortality from prostate cancer among male employed participants. Among non–night shift workers, allele G was not associated with prostate cancer compared with CC carriers. Among night shift workers, compared with those with CC genotype, GC carriers had a modestly lower risk of prostate cancer, although not statistically significant (HR: 0.68; 95% CI: 0.26–1.76). Supplementary Table S8 further shows the results from the perspective of treating night shift work as exposure. Among CC carriers, there was a slightly higher risk among night shift workers (HR: 1.19; 95% CI: 0.63–2.23) compared with non–night shift workers, but it was not statistically significant. Among G carriers, there was a trend for a decreased risk among night shift workers (HR: 0.85; 95% CI: 0.39–1.84).

To our knowledge, this is the first large cohort study to explore the interaction between night shift work and the MTNR1B rs10830963 on the risk of prostate cancer. In this study, we found a significant interaction between shift work exposure and rs10830963 polymorphism on the incidence of prostate cancer. At a first glance, night shift work was not associated with prostate cancer in the main effect analyses. However, when stratifying rs10830963 polymorphism, we observed that rs10830963G was associated with a decreased risk of prostate cancer among night shift workers in a dose–response fashion. From the other perspective, when treating night shift work as an exposure, night shift work was associated with an increased risk of prostate cancer only among CC carriers, but not allele G carriers, which further supports that allele G may play a protective effect against night shift work on the prostate cancer. The sensitivity analyses further confirmed our findings. In terms of mortality from prostate cancer, similar trends were also observed although they were not statistically significant.

This interaction may explain the lack of association between night shift work and prostate cancer risk observed in some studies (21–23). Barul and colleagues (21), based on a large population-based case–control study on prostate cancer in Canada, reported the absence of associations between night shift work and prostate cancer. In line with those studies, the main effect analysis in our study also did not find any significant association between night shift work and prostate cancer. However, when taking the rs10830963 polymorphism into consideration, diverse associations between night shift work and prostate cancer risk were observed among different genotypes carriers. Similar to our findings, a case-control study conducted by Liu and colleagues also found some polymorphisms such as rs6983561 and rs16901966, interacting with environmental factors, contribute to the risk of prostate cancer (24). Our finding from a large prospective study adds to a growing body of evidence that prostate cancer is a consequence of individual genetic and environmental components.

Melatonin receptor 1B is one of two transmembrane receptors for melatonin and the SNP rs10830963 is in an intron unique to MTNR1B. Highly controlled laboratory studies demonstrate that the rs10830963G is associated with later melatonin offset and longer duration of elevated melatonin levels with the more obvious effect in GG than GC (11). The melatonin profile among evening chronotypes is close to those allele G carriers (10, 25). Similar to the current findings, it has been shown that night shifts increased the risk for prostate cancer among participants with early but not late chronotype (26, 27). Indeed, night shifts cause misalignment between the endogenous circadian timing system and the external light/dark cycle (28), and the biological plausibility for increased risk of cancer with night shift work is possibly related to the suppression of nocturnal melatonin secretion (29). Those allele G carriers or evening types may experience less misalignment caused by night shifts and thus they seem to have better tolerance for night shifts. However, the specific mechanism underlying the complex relationships should be followed up with experimental studies.

Besides, the protective effect of rs10830963G on prostate cancer among night shift workers may help to explain the inverse association between T2DM and prostate cancer (15–17). Consistent with the epidemiologic data, Meyer and colleagues identified that some T2DM related SNPs including SLC2A2 rs5400 Thr110 allele, TCF7L2 rs7903146 T allele, and UCP2 rs660339 Val55 allele are associated with a reduced risk of prostate cancer (30). On the other hand, results from eight case–control groups, including one West African and one Chinese, demonstrate that the risk allele for prostate cancer TCF2 rs4430796 A allele, confers protection against T2DM (31). The findings in our study provide some new clues to this genetic associations between T2DM and prostate cancer.

Our study has several strengths. First, the large-scale prospective study design provides sufficient statistical power to determine the incident prostate cancer by stratifying different genetic variances. Second, information on detailed medical history, lifestyle information, and demographic information were collected in a uniform manner in the UK Biobank. Therefore, we were also able to take potential confounders into consideration. Third, our primary outcome data, the diagnosis of prostate cancer and its date, were collected through the linkage with national registries, which acquire information on cancer diagnoses from a variety of sources. Therefore, it can reduce the risk of outcome misclassification and selective dropout in our study.

There are several potential limitations in our study when interpreting the findings. First, the information on the duration of night shift work conditions and whether they took permanent night shifts or rotating night shifts is not available in the UK Biobank. Second, this is a relatively young cohort (mean age at the initial recruitment was 53.13 years), and the follow-up time for incidence was limited to a maximum of 10 years, which might not be long enough to detect more cases during the incidence peak of prostate cancer (ages 65 or above; ref. 32). However, it allows us to examine those interactions on early-onset prostate cancer. Third, information on prostate cancer stage and grade is currently unavailable in the UK Biobank, as a result of which we could not determine whether the observed associations vary by these prostate cancer characteristics. Fourth, although the baseline data of covariates were complete, we could not have access to time-varying information to enable updating of the variation in covariates during follow-ups because of the limited UK Biobank participants with follow-up questionnaire assessments. Fifth, our study has an inadequate sample size for the mortality from prostate cancer to attain reliable power for our statistical analysis and thus further larger studies are warranted to confirm these findings. Sixth, only one well-known melatonin receptor polymorphism was examined in our study, and other polymorphisms could be investigated in the future as new evidence accumulates over time.

Our study provides insights into how MTNR1B rs10830963 modifies the associations of night shift work with prostate cancer. Compared with CC, carrying allele G may reduce the risk of prostate cancer when exposed to night shift work. Therefore, genetic screening may be suggested for those males entering into night shift work. In addition, the use of melatonin may be taken into consideration to minimize the risks among this population when night shifts are inevitably involved. Our findings have important implications for male workers and policymakers.

J.Y.W. Chan reports personal fees from Eisai. Co. Ltd outside the submitted work. Y.K. Wing reports personal fees from Eisai Co., Ltd and Lundbeck HK Limited outside the submitted work. No disclosures were reported by the other authors.

L. Yang: Conceptualization, resources, data curation, formal analysis, visualization, methodology, writing–original draft. J. Chen: Conceptualization, resources, validation, methodology. H. Feng: Data curation, writing–review and editing. S. Ai: Data curation, writing–review and editing. Y. Liu: Writing–review and editing. X. Chen: Writing–review and editing. B. Lei: Writing–review and editing. J.W.Y. Chan: Writing–review and editing. S.W.H. Chau: Writing–review and editing. L.A. Tse: Writing–review and editing. A.W.-Y. Ho: Writing–review and editing. C.S. Ho: Writing–review and editing. Y.K. Wing: Resources, supervision, funding acquisition, writing–review and editing. J. Zhang: Conceptualization, resources, supervision, investigation, project administration, writing–review and editing.

We would like to thank all the participants for their valuable contributions in the UK Biobank. The data used in the analysis are available to other researchers upon request to the UK Biobank (https://www.ukbiobank.ac.uk/).

This work was supported by the To support research activities in SAU.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.