Disrupted Body Clock Genes Linked to Breast Cancer in Night Shift Workers

Breast cancer (BC) represents the most frequently diagnosed cancer among women worldwide. According to the Global Cancer Observatory, in 2022, 2.3 million women were diagnosed with BC, and 670,000 deaths have been counted globally [1]. In Europe, for the year 2024, it has been predicted that there will be the highest absolute number of deaths (89,300) in women [2]. Intriguingly, compared to less developed countries, in more developed ones there is a higher survival rate attributed to earlier diagnosis, better preventive measures, and improved access to treatment [3]. The aetiology of BC is multifaceted as it involves reproductive factors, genetic predisposition, family history, and lifestyle choices. Hormone-related factors such as early menarche, late menopause, delayed pregnancies, and lower childbirth numbers are linked to BC development. Also, women with a family history of BC or specific mutations such as BRCA1 and BRCA2 are at higher risk. In addition, multiple lifestyle factors, including low physical activity, alcohol consumption, smoking, and a high-fat diet, are known to elevate the risk of developing this tumour [4]. Likewise, several occupational exposures have been associated with an elevated risk of BC [5].

Notably, BC aetiology extends beyond this broad range of non-modifiable and modifiable risk factors, appearing to be tightly linked with circadian clock alterations as a significant contributor [6]. Circadian rhythms are pivotal for life, having the purpose of synchronizing our homeostatic processes with the rhythmic alternation of daytime and nighttime [7]. At the cellular level, they are governed by a conserved group of clock genes with established self-regulatory feedback loops, although the signal transduction pathways remain yet to be fully elucidated [8]. Clock gene alterations have been shown to significantly contribute to BC initiation, progression, and invasion, either by modifying neoplastic cell homeostasis or the tumour microenvironment [9,10]. Within the mammary epithelium, it has been proven that clock gene impairments (e.g., point mutations, alteration of transcription or translation, post-translational modifications) may lead to cell cycle deregulation, which can be correlated with tumour progression and more aggressive BC [11].

Several studies indicate a 30–50% increased BC risk in countries with high light at night (LAN) exposure compared to those with lower exposure [12]. Altered clock gene expression in BC may be caused by disrupted circadian rhythm of melatonin synthesis due to a permanent exposure to artificial LAN, which particularly relates to night shift work (NSW) [5,13,14]. In fact, it has been observed that women who are persistently exposed to NSW and, consequently, are highly exposed to LAN, such as nurses, have a higher frequency of developing BC. On these bases, since 2019, NSW has been considered probably carcinogenic to humans by the International Agency for Research on Cancer (IARC) [15]. One of the most extensive American studies, the Nurses’ Health Studies I and II (NHS I and II), assessed the risk of BC linked with NSW, confirming a heightened BC risk among workers engaged in prolonged night shifts [16,17]. This observation has been corroborated by a number of multicentric studies, confirming the existence of a positive association between NSW and BC risk, particularly in high-intensity and/or long-duration patterns [18,19].

From a mechanistical point of view, decreased nocturnal melatonin production may increase the pro-cancer effects of oestrogens, overall contributing to BC development [20]. Loss of expression of certain circadian clock genes has been found to be associated with BC progression in human studies [21]. Another hypothesized mechanism linking NSW to BC could be through the expression of some micro-RNAs (miRNAs) showing oscillations during the day [22]. Despite significant advances being made in the field of chronobiology, the molecular basis of malignant transformation in women chronically exposed to NSW is still debated. Hence, the objective of this study is to identify specific clock genes that could serve as promising biomarkers for BC while also exploring their association with varying susceptibility in women who work shifts at night. To achieve this goal, as summarized in Figure 1, a differential gene expression analysis of the clock gene candidate pool was performed using BC samples from the Gene Expression Omnibus (GEO) Dataset of NHS I and II.

These samples included nurses who developed BC and were exposed to NSW. The differential expression of clock genes was validated in multiple gene expression datasets from whole blood samples of healthy night shift workers, including female nurses, and BC tissue samples from healthy women with diverse susceptibilities to BC development. Furthermore, this analysis comprised the cell-specific expression patterns of the selected clock genes in both non-cancerous and cancerous breast tissues, along with their functional pathways and upstream epigenetic regulation.

Nurse Health Studies (NHSs) are among the largest prospective investigations concerning the risk factors associated with major chronic diseases in nurses working night shift. NSW was studied as a factor increasing BC risk [16,17,18,25]. The NHS gene expression dataset, which we investigated in our study, was previously employed to define BC molecular subtypes, their association with body mass index, adiposity, and inflammatory dietary patterns [26,27,28]. It includes a total of 1577 breast samples, including 623 paired cancerous and non-cancerous ones. We considered only paired samples to analyse the differential expression of the core clock genes (Table S1), known to orchestrate biological rhythms at the cellular level and, when deregulated, to contribute to BC development and progression [21,29].

Among the 29 clock genes analysed, we found that 12 were significantly deregulated in BC tissues compared with non-transformed paired ones, consistently in both NHS I and NHS II subgroups. In particular, while BHLHE40, CIART, CLOCK, PDPK1, and TIMELESS were significantly over-expressed, HLF, NFIL3, NPAS3, PER1, PER3, SIM1 and TEF were significantly down-expressed. Among these genes, NPAS3 and PER3 were highly downregulated in higher-grade BC compared with lower-grade BC, while PDPK1 and TIMELESS were significantly upregulated (Figure 3). In line with our findings, in a Polish study, it was observed that CLOCK and TIMELESS were over-expressed while PER1 and PER3 were down-expressed in BC compared to normal adjacent samples [11]. Additionally, two studies independently proved that a higher expression of TIMELESS and CLOCK and a lower expression of PER3 were associated with a higher BC grade, a lower metastasis-free survival rate, and an overall worse prognosis [21,30]. Finally, very recent research confirmed that in Lum A BC, compared to paired noncancerous tissue, PER1, PER2, HLF, TEF, and NFIL3 were downregulated, whereas CLOCK, CIART, and BHLHE40 were upregulated [31].

Intriguingly, while other groups found high PER2, PER3, and CRY2 expression and low TIMELESS expression were significantly associated with either longer metastasis-free survival or longer overall survival, in the NHS datasets, we could not find any correlation between clock genes’ expression levels and recurrence-free survival (Supplementary Figure S4). This difference might be explained by the variance in survival outcomes considered within the diverse studies [21,30].

Subsequently, in order to verify rhythmic clock gene expression in healthy workers, we analysed two different gene expression datasets from PBMC samples of day shift versus night shift workers. One of the datasets included nurses undertaking different rotation shifts. Notably, for none of the studies were we able to register any change in time-dependent fluctuations in clock gene expression (Supplementary Figure S5). This result could be explained by tissue-specific physiological and metabolic gene functions [32]. Our finding is in line with other studies, where PER1 or PER2 loses its normal day shift rhythmicity following NSW, while PER3 was significantly rhythmic in both shift conditions [33,34]. Also, Hattammaru et al. carried out a transcriptome analysis from beard-hair follicle cells and highlighted that, despite the observed rhythmic variation in PER3 in night shift workers, there was substantial variability between individuals [35].

In summary, the disruptive effects on clock gene rhythmicity seem to strictly depend on the nature of the tissue examined [36]. In light of that, a differential expression analysis was performed in single-cell RNA-seq datasets of cancerous and non-cancerous breast microenvironments, highlighting the selective expression of the clock genes within specific cellular components. For instance, CLOCK and PER3 were preferentially expressed in tumour and stomal cells, while PER1 was selectively expressed in immune cells and stromal cells but not in cancer cells. Finally, TEF was selectively expressed in CAFs (Figure 4B).

Given the above-described role of clock genes in the PBMCs of healthy workers, we opted to validate the predictive value of the same genes in two datasets of breast tissues from healthy women with diverse susceptibilities to develop BC. In one dataset with samples from women with different BC risks, PER1, PER3, and TEF downregulation was associated with higher risk, further confirming what was observed in the NHS datasets. Among these genes, both PER1 and TEF revealed predictive value, as demonstrated by their significant AUC performances. In the second dataset, we found greater CLOCK upregulation in breast tissue from healthy women that developed BC years later compared those that did not, with very high and significant AUC performance. Collectively, these results suggest that PER1, TEF, and CLOCK hold considerable promise in the early detection and risk stratification of BC, potentially guiding preventative strategies and personalized interventions for women at increased risk [37].

To further investigate the role of the identified clock genes, we performed a functional analysis, which evidenced that they regulate many processes, not only related to circadian rhythms and chronobiology but also to nuclear activity, transcription, and metabolism (). Also, with the exclusion of NPAS3 and PDPK1, proteins coded by these genes are functionally linked, as proved by the STRING analysis. Overall, these findings underscore the multifaceted influence of the identified clock genes on several cellular functions and metabolic pathways, reinforcing their role as valuable biomarkers. Supplementary Table S2

Regarding epigenetic regulation, the miRNet analysis highlighted a total of 26 miRNAs that might inhibit the expression of the 12 clock genes. Among them, 10 are specifically expressed in normal breast tissue, while only 4 are in BC. Interestingly, mir-10a, mir-21, mir-107, and mir-34 were previously found to be differentially methylated in night shift workers compared to day workers; hence, they might be suggested as putative circadian miRNAs [38]. This finding suggests an explorable role for these miRNAs as biomarkers for assessing BC risk in women chronically exposed to NSW. Moreover, it provides a new avenue for implementing novel research about the epigenetic mechanisms linking NSW to increased BC risk.

In conclusion, our computational study allowed for the identification of a panel of clock genes and circadian miRNAs which can be suggested as biomarkers of susceptibility, useful for identifying groups of night shift workers at higher risk of developing BC. In particular, PER1, TEF, and CLOCK may be recommended as predictive markers for BC susceptibility. This study has a number of strengths, but also limitations. One of the main strengths is the comprehensive analysis of clock gene expression across multiple datasets, including paired BC and normal tissues, as well as single-cell RNA-sequencing data. Also, leveraging large-scale datasets from NHS I and II adds robustness to our findings, given the extensive follow-up and detailed exposure assessment made in these cohorts over time.

However, there are also limitations. Firstly, the study relies on existing gene expression datasets; hence, while we identified significant associations between specific clock genes’ expression and BC risk, the cross-sectional nature of the data limits our ability to infer causality. Moreover, the lack of rhythmicity observed in clock genes from the PBMCs of healthy night shift workers highlights the need for tissue-specific investigations, as peripheral tissues may not accurately reflect changes occurring in the breast tissue. Furthermore, this study relies on pre-existing gene expression datasets, which limits the ability to control for potential confounding factors associated with such data. For example, variability in sample handling could introduce biases. Additionally, the absence of experimental validation across diverse tissue types constitutes a significant limitation. Future research should address this by employing experimental approaches such as tissue-specific in vitro assays or animal models to confirm the findings and account for tissue-specific differences. Finally, these in silico findings will require further validation through longitudinal studies to confirm the clinical utility of the identified biomarkers.

Overall, while this study provides important insights into the molecular mechanisms linking NSW to BC risk, our future research should focus on validating these biomarkers in diverse populations and exploring their role in early detection and prevention strategies. In particular, validation data at different timepoints in both saliva and/or oral mucosal cell samples from NSWs will be needed to establish a more precise temporal profile of gene and miRNA expression changes associated with circadian disruption. To facilitate the clinical application of our findings, future studies should prioritize the development of standardized protocols for the quantification of clock gene-related biomarkers in clinical and primary care settings. This includes determining optimal sampling times to account for circadian variability, utilizing non-invasive sampling techniques such as peripheral blood or saliva, and employing high-throughput, sensitive methodologies such as RT-qPCR and next-generation sequencing for the precise measurement of gene expression and miRNA profiles.

Such validation might pave the way for early detection strategies and personalized risk assessment in women at increased risk of BC due to prolonged exposure to NSW, with significant implications for occupational health, as it offers a new avenue for monitoring the health of employees exposed to circadian disruption, potentially leading to the implementation of personalized intervention strategies to mitigate BC risk. Once validated, these genetic and epigenetic biomarkers might be useful to enhance screening programmes, enabling more targeted prevention efforts. Integrating circadian biology into occupational health practises could lead to the development of useful guidelines that minimize circadian disruption, thereby reducing both the short- and long-term health effects linked with night shift work.

The following supporting information can be downloaded at https://www.mdpi.com/article/10.3390/ijms25168659/s1↗.

Conceptualization, C.F. and S.V.; methodology, S.V.; validation, S.V., G.S., C.C., F.G. and C.F.; formal analysis, S.V.; data curation, C.F. and S.V.; writing—original draft preparation, C.F. and S.V.; writing—review and editing, S.V., G.S., C.C., F.G. and C.F.; visualization, S.V., G.S., C.C. and F.G. All authors have read and agreed to the published version of the manuscript.

Not applicable.

Not applicable.

The data generated in the present study may be requested from the corresponding author.

The authors declare no conflicts of interest.

This research received no external funding.