Evaluate the impact of maternal smoking during pregnancy on sleep duration in female offspring

Background Sleep is a fundamental state of decreased activity and alertness conserved in human life. Maternal smoking during pregnancy may affect sleep duration female offspring, however the exact connection and mechanism still remain unclear. Methods Based on the UK Biobank data, sleep duration was categorized as short sleep (< 7h/day) and long sleep ( ≥ 9 h/day). Logistic regression analysis was established to evaluate the potential impact of maternal smoking during pregnancy on the short sleep and long sleep of female offspring. Subsequently, genome-wide gene-environment interaction study (GWGEIS) of sleep duration was performed using maternal smoking during pregnancy as an environmental factor. Results We found that maternal smoking during pregnancy was related to short sleep and long sleep duration of female offspring (P < 0.001). GWGEIS analysis detected a signicant association at ANKLE2 (rs78408032, P = 1.33×10 − 8 ) and a suggestive association at MYT1L (rs76688230, P = 8.85×10 − 7 ) in female subjects. Our study results suggested a signicant effect of maternal smoking during pregnancy on short sleep of female offspring.


Introduction
Sleep occupies a third of the people's lives. The quality and duration of sleep could affect both psychological and physical health of children and adults, especially in cognition and execution. Proper and highly quality sleep enables person having su cient ability to cope with the problems in life and work. Sleep disturbance includes short and long sleep duration, insomnia, signs of daytime dysfunction and di culty awakening in the morning, which are recognized as the risk factor of human health. For example, long-term sleep duration is associated with increasing the risk of stroke and breast cancer [1,2], and inadequate sleep duration is considered as the risk factor of cognition, memory and neurologic diseases, which decreases vigour and heightens fatigue [3,4].
Sleep problems are generally linked to both environmental and genetic factors. Environmental factors involve alcohol, skin warming and social issues. For instance, previous study has shown that proper skin temperature facilitates the onset of sleep [5]. It has been commonly known that genetic factors contributed greatly to the regulation of sleep duration. Y Dauvilliers et al and M Partinen et al. have proved that sleep length can be heritable [6,7]. For example, it was suggested that sleep problems were related to FXR1 and DCE2 gene mutation [8,9].
Cigarette smoke contains approximately 4700 substances, many of which are carcinogenic. Cigarettes are one of the most important causes of lung cancer, and the cancer for recent research have established a causal related to tobacco smoking are oral, nasal cavities, larynx, liver, oesophagus, pharynx, stomach, kidney, urinary tract and cervix [10]. As the number 1 and the number 1 preventable cause of death, cigarette smoking contributes to 6 million deaths every year, and secondhand smoke is responsible for the death of more than 600,000 individuals worldwide [11]. Unfortunately, pregnant women are often exposed to tobacco in their daily lives, either passively or actively.
Recently, it has been commonly known that maternal smoking during pregnancy (MSDP) takes considerable risk for baby. Compelling correlative evidence claims that ingredients in cigarette smoke could greatly harm the newborn via placenta [12]. Some studies suggested that maternal smoking during pregnancy may change the glutamatergic system in the hippocampus, and increase the risk of abnormal neurobehavioral outcomes in offspring [13]. And, maternal smoking is associated with altered total brain volumes and cortical gray matter volumes, which indicated that exposure to cigarette smoke environment in utero may impact the brain function and result in neurodevelopmental abnormalities [14]. Sleep disturbance is associated with exposure to high levels of tobacco smoking [15]. Prenatal smoking exposure could be concerned with altering neurodevelopment, which affects offspring sleep [16]. It's reported that maternal smoking during pregnancy affects offspring signi cantly greater in female than in males [17], therefore, we would like to pay more attention to sex-speci c risk of maternal smoking during pregnancy on the sleep duration in female offspring.
In the study, we performed observational analyses and GWGEIS to investigate the association between maternal smoking during pregnancy and sleep duration in female offspring, based on UK Biobank samples. We aimed to explore the relationship between maternal smoking during pregnancy and sleep duration, and detect novel candidate genes that affect sleep duration by interacting with maternal smoking during pregnancy in females.

Methods And Materials UK Biobank
UK Biobank is a large scale prospective, this cohort contained about 500,000 subjects aged between 37 and 76 years (99.5% were aged 40-69 years). All involved subjects provided a sequence of information on demographics, health status, and lifestyle get through questionnaires and interviews [18]. The UK Biobank has obtained informed consent of all participants and ethical approval from the Northwest Multi-Center Research Ethics Committee. Obtain the informed consent of all participants.

Phenotype de nition
Sleep duration was recorded as number of reported hours by asking 'About how many hours sleep do you get in every 24h? (include naps). And then we categorized the sleep duration as short sleep (<7h/day), normal (7-8h/day), and long (≥9 h/day) consistent with previous studies [19]. Smoking phenotype is de ned as "Did your mother smoke regularly around the time when you were born?". If the answer is yes, it is de ned as smoking, and individuals were classed as never smoke if answered to question "No".
Individuals who answered to either question "Do not know", "Prefer not to answer" were excluded in this study (UK Biobank code: 1787). The age and gender data of study subjects was derived from the baseline characteristics of the UK Biobank.

Genotyping, imputation and quality control
In UK Biobank Affymetrix UK BiLEVE Axiom and Affymetrix UK Biobank Axiom was used for SNP genotyping. Sample-based and label-based quality control were performed. Imputation was conduction by IMPUTE4. Principal components (PC) were computed to account for population structure. We restricted study subjects to the 'White British' group based on self-reported ethnic background (UK Biobank data eld: 21000). Genetically related individuals were excluded according to the kinship coe cient calculated by the KING software. Details about imputation, genotype and quality control are available in previous studies [20].

Observational analyses
Logistic regression analysis was used to evaluate the relationship between short/long sleep duration and maternal smoking during pregnancy in female offspring. Short sleep duration, long sleep duration and maternal smoking during pregnancy was set as binary variable. 10 principle components, age, daily frequency of smoking were used as covariates in the logistic regression models.

Genome-wide gene-environmental interaction study (GWGEIS) analyses
Maternal smoking during pregnancy was regarded as an environmental factor, we performed GWGEIS of female short/long sleep duration by PLINK2.0 [21]. This method considered the impact of SNPenvironment interaction from a regression model [22]. The GWAS P < 5.00×10 -8 and P < 1.00×10 -6 were used as signi cant and suggestive threshold, respectively. For quality control, we excluded the SNPs with low call rates (<0.90), low Hardy-Weinberg equilibrium exact test P values (<0.001), or low minor allele frequencies (MAFs<0.01) in this study.

Basic characteristics of study subjects
Our study includes a total of 175135 female samples. Smoking sample contains a total of 52599 participants and the mean (standard deviation) age was 55.96 (7.59). Never smoke sample contains a total of 122536 participants and the mean (standard deviation) age was 56.86 (7.99). The basic characteristics of female offspring sleep duration in smoking and never smoke during pregnancy were presented in Table 1.

Discussion
In the present study, to explore the potential in uence of maternal smoking during pregnancy on sleep duration in female offspring, we conducted an observation study and GEWIS in UK Biobank cohort. Our results showed that maternal smoking during pregnancy can effect sleep duration in offspring, which were consistent with the previous study [16,23].
We found that ankyrin repeat and LEM domain containing 2 (ANKLE2) achieved signi cant threshold in the GWGEIS for short sleep duration in females. ANKLE2 encodes a member of the LEM family of inner nuclear membrane proteins. The encoded protein functions as a mitotic regulator through postmitotic formation of the nuclear envelope. Previous research suggested that ANKLE2 regulates C. elegans nuclear envelope reassembly through barrier-to-autointegration factor (BAF) dehosphorylation [24]. Mutations in this gene cause BAF hyperphosphorylation, which is an important DNA-binding protein that interacts with a unique structural motif called LAP2-Emerin-MAN1 (LEM) domain to establish a connection between chromatin and a family of proteins containing LEM domains [25]. Now there is striking evidence that decreased number of glial cells and neuronal within the brain can lead to microcephaly [26,27], sleep is associated with decreased neurons [28]. And more recently studies have shown that ANKLE2 gene associated with Congenital microcephaly [29,30], and it has been proved that genetic mutations can cause short sleep phenotype. Based on our study results, ANKLE2 may contribute to the dysfunction of female sleep regulation.
Myelin transcripyion factor 1 like (MYT1L) is another identi ed candidate gene, which encodes a member of the zinc nger superfamily of transcription factors. The encoded protein is part of a novel class of cystein-cystein-histidine-cystein zinc nger proteins that function in the animal central nervous system. It associated with neuronal cells and via indirect methods exerts its pro-neuronal function [31]. MYT1L only expresses in the human brain, and plays a critical role in the developing central nervous system, which has been related to schizophrenia, intellectual disability, autism [32]. Our results suggest that MYT1L can be considered as a potential gene affecting sleep, consistent with previous study [33]. Enlarging the comparative analysis of MYT1L and other diseases, we found that several diseases regulated by MYT1L are related to sleep. For example, bromyalgia associated with sleep disturbance and gene MYT1L is candidate gene bromyalgia [33,34]. Serpin family E member 2 (SERPINE2) is a member of the serpin family, a group of proteins that inhibit serine proteases. It is a susceptibility gene for chronic obstructive pulmonary disease and human tumors [35,36]. It is well known that many chronic diseases are associated with sleep disorders, studies shown a signi cant association between short-term sleep and chronic obstructive pulmonary disease in female [37].
Until now, the biological mechanism of maternal smoking affecting offspring development remains largely known, but established studies could enhance our understanding of the relationship between maternal smoking during pregnancy and sleep duration in female offspring. On the one hand, smoking during pregnancy can have a long-term effect on offspring through hypothalamic-pituitary-adrenal axis [38,39], and effects developing brain's acetycholine neurotransmitter systems [40,41]. On the other hand, smoking during pregnancy might affect early morphological changes of the placenta, resulting in a decreased surface area and volume of fetal capillaries [42,43], inducing smaller head circumference at birth, and the identi ed gene ANKLE2 in our further con rmed this in a certain degree. These mechanisms may partly explain the effect of maternal smoking during pregnancy on the offspring's sleep duration.
With the development of science and technology in recent years, it has been commonly known that many diseases are affected by both the environment and genes. Own to the large sample and the rigorous design of UK Biobank, our results are credible and representative. Nevertheless, several limitations should be noticed. Firstly, though we could detect the in uence of maternal smoking during pregnancy on offspring sleep duration, the information of indirect exposure to smoking is missing, such as, second-hand smoke. Secondly, the data of maternal smoking during pregnancy were driven from online questionnaire surveys, retrospective measurements and which may increase the possibility of recall bias and measurement. In addition, UK Biobank subjects were European whites, it should be cautious when popularize current studies results to other-race.
In summary, our results suggested effect that maternal smoking during pregnancy has effects on sleep duration in offspring, especially on short-term sleep, which may provide an essential information with the pregnancy health, and highlight the detrimental effect of smoking in health from new perspective.
Declarations Figure 1 (a-b) Genomic regions interacting with maternal smoking during pregnancy for short sleep (a) and long sleep (b) duration in female offspring Note. From the center, the rst circle depicts the -log10 p-values of each variant due to double exposure. The second circle shows chromosome density. The plots were generated using the "CMplot" R script (https://github.com/YinLiLin/R-CMplot).