Introduction

Telomere is a nucleoprotein complex at the chromosome end, which maintains genomic DNA stability.1 Telomere progressively shortens with each cell division and eventually reaches a critical length that triggers cellular senescence.1 Because telomere attrition is a key mechanism of cellular senescence, telomere length (TL) has long been considered as a marker of biological aging. A substantial body of studies indicated that shortened TL in adulthood is associated with shorter lifespan and age-related diseases.2,3 TLs are highly variable across individuals. It was reported that this inter-individual variation was originated early in life, while the lifestyle or environment in adulthood exerts a minor impact.4 Therefore, understanding the determinants of initial TL (TL of newborns) might provide insights into possible mechanisms underlying relationships between prenatal exposures and later health outcomes.

Cigarette smoking is a major risk factor of age-related diseases,5 and is associated with intensified oxidative stress and inflammation.6,7 Both oxidative stress and inflammation are related to telomere shortening,8,9 leading to the hypothesis that cigarette smoking accelerates telomere attrition. It has been well established that active smoking is associated with shorter TL in adults.10 Compared with active smoking, second-hand smoke (SHS) exposure is a more common and avoidable public health threat. Findings about the relationship between SHS exposure and TL among adults are inconsistent.11,12,13 Fetuses are more vulnerable to environmental threats than adults. However, the association between prenatal SHS exposure and newborn TL was poorly understood.

In China, the active smoking rate for women of childbearing age is very low (0.7–1.6%),14 while researches showed that there were 38.9 to 75.1% non-smoking women exposed to SHS during pregnancy.15 The high prevalence of SHS exposure during pregnancy indicates an urgent need for the illustration of adverse impact of SHS exposure on fetus. In this study, we investigated the association between prenatal SHS exposure and newborn TL, with the hypothesis that prenatal SHS exposure was associated with shorter newborn TL.

Materials and methods

Participants

We recruited 762 mother–newborn pairs from Wuhan Children’s Hospital (Wuhan Maternal and Child Healthcare Hospital), a maternity hospital in Wuhan city, Hubei province, China, between November 2013 and March 2015. The following inclusion criteria were applied: (a) during the first trimester of pregnancy (<16 weeks); (b) singleton pregnancy; (c) residents of Wuhan City during pregnancy; (d) intention to have prenatal care and give birth at Wuhan Children’s Hospital; (e) consent for participation in this study. After excluding 16 participants with DNA sample unavailable for TL measurement, there were 746 mother–newborn pairs included in the analysis.

The study protocol of this study was approved by the Medical Ethics Committee of Tongji Medical College and the Wuhan Children’s Hospital. Signed informed consent was obtained from all participants.

Data collection

Maternal SHS exposure was collected via questionnaire, by asking “do you have SHS exposure during pregnancy?” SHS exposure was defined as non-smokers’ exposure to the tobacco smoke for an average of at least 10 min per day. Active maternal smoking status was collected via questionnaires, by asking “do you have active cigarette smoking six months before pregnancy” and “do you have active cigarette smoking during pregnancy.” Demographic information on age at delivery, education background (high school or below, college or above), and annual household income were collected via questionnaire. All the participants were interviewed face-to-face by trained nurses during the period of institutional delivery. Self-reported pre-pregnancy weight was collected at the first antenatal care visit (in the first trimester). Standing height was obtained by medical examination. Pre-pregnancy body mass index (BMI) was calculated as weight (kg) divided by height squared (m2). The information on parity, gestational diabetes, and hypertensive disorders in pregnancy were obtained from medical records. Information on newborn sex, birth weight, and date of delivery were obtained from hospital delivery records. Low birth weight infant was defined as infant with birth weight <2500 g. Information on gestational age was estimated by women’s last menstrual period, which was obtained from medical records. The preterm birth infant was defined as an infant with gestational age <37 weeks.

Umbilical cord blood collection and TL measurement

At birth, umbilical cord blood samples were drawn in EDTA tubes. DNA was extracted from stored umbilical cord blood samples using the Wizard® Genomic DNA Purification (Promega Corporation, Madison, WI) following the manufacturer’s instructions. Quantitative real-time polymerase chain reaction (qPCR) was used for TL measurement. This method measured the relative TL by determining the T/S ratio of telomere repeat copy number (T) and single-copy nuclear gene copy number (human β-globin) (S), and the T/S ratio of the experimental sample was standardized with the reference pooled sample. The standard DNA pool consists of 50 DNA samples that were randomly selected from the 746 samples of this study. The four primers 5′−3′ were as follows: telomere forward, 5′-ACACTAAGGTTTGGGTTTGGGTTTGGGTTTGGGTTAGTGT-3′; telomere reverse, 5′-TGTTAGGTATCCCTATCCCTATCCCTATCCCTATCCCTAACA-3′; reference gene forward, 5′-GTGCACCTGACTCCTGAGGAGA-3′; reference gene reverse, 5′-CCTTGATACCAACCTGCCCAG-3′. The reagents in the 10 μL PCR were 1 μL of extracted DNA (10 ng/μL), 5 μL of KAPA SYBR® FAST Master MIX 1× (KAPA Biosystems), and RNase free water. The final telomere primer concentrations were: telomere forward, 270 nM; telomere reverse, 900 nM. The final single copy gene primer concentrations were: HGB forward, 200 nM; HGB reverse, 200 nM. Samples were run in triplicate in a 384-well plate with the use of ViiA™ 7 Dx Real-Time PCR System (Applied Biosystems). The acceptable standard deviation (SD) for three Ct values was set at 0.3. If the SD of three Ct value was out of acceptable range, the measurement was repeated in a new 384-well plate. The thermal cycling profile were: 50 °C for 2 min, 95 °C for 3 min, followed by 40 cycles of 95 °C for 3 s, and 60 °C for 30 s. A melting-curve analysis was performed at the end of each run to ensure the amplification specificity and absence of primer dimer. For the standard curve, the standard DNA pool was four-fold serial diluted to generate five solutions with concentrations ranging from 0.4 to 104 ng/μL (R2 ≥ 0.99). In order to assess qPCR amplification efficiency, we run a five-point serial dilution of standard reference genomic DNA sample on each run. The qPCR amplification efficiency was 108% for telomere and 103% for single-copy gene. We calculated coefficient of variation (CV) to assess the variability. The CV was calculated using the following formula: (σ ÷ μ) × 100%, where σ and μ are standard deviation and mean value of the Ct values of single-copy gene of standard reference sample, respectively. The inter-run CV was calculated using Ct values of single-copy gene of standard reference sample in all qPCR assays. And the intra-run CV was calculated using Ct values of three replicates for single-copy gene of the standard reference genomic DNA sample in each 384-well sample plate. The inter-run CV and mean intra-run CV were 4.1% and 3.0%, respectively.

Statistical analysis

Categorical variables were expressed as proportions (%), and continuous variables were expressed as means ± SD)for normal distribution, or geometric mean (GM) and 95% confidence interval (CI) for skewed distribution. Differences in means and proportions between two categories of prenatal SHS exposure were tested by the analysis of variance or χ2 test. Linear regression models were employed to investigate the relationship between prenatal SHS exposure and ln-transformed T/S ratio. Newborns without prenatal SHS exposure were treated as the reference group, and the percent difference in newborn TL was estimated as eβ – 1 × 100%. An unadjusted model was used to assess the association between SHS and newborn TL without adjustment for any covariate. We selected covariates based on a priori knowledge, including maternal age, pre-pregnancy BMI, education background, annual household income, parity, gestational diabetes, hypertensive disorders in pregnancy, infant sex, gestational age, birth weight, and date of delivery.

Because previous studies have suggested sex-specific relationship between prenatal environmental pollutants exposure and newborn TL,16 we inserted newborn sex × SHS as an interaction term in the regression models to obtain P value for interaction. The stratified analyses were employed to show the specific differences in percent changes of TL between boys and girls. To evaluate the robustness of our study, we performed sensitivity analyses excluding newborns with preterm birth and low birth weight, women ≥35 years old, women with gestational diabetes and hypertensive disorders in pregnancy, and women with active cigarette smoking before or during pregnancy. All statistical analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC). Statistical significance was set at P < 0.05.

Results

Characteristics of the participants according to prenatal SHS exposure are shown in Table 1. Of 746 eligible women, 244 (32.7%) reported SHS exposure during pregnancy. Compared with women who had no exposure to SHS during pregnancy, women with SHS exposure were younger (P = 0.002) and had lower annual household income (P = 0.024). Compared with newborns without prenatal SHS exposure, those with SHS exposure had shorter umbilical cord blood TL (P < 0.001).

Table 1 Characteristics of the study participants according to prenatal second-hand smoke exposure
Full size table

The distribution of fetal TL in two groups is shown in Fig. 1. The violin plot and box plot revealed that newborn TL was shorter in the non-SHS group, and the two groups had similar distribution of fetal TL.

Fig. 1
figure 1

The violin plot and box plot in distribution of newborn telomere length between two groups

Full size image

We used two logistic regression models to assess the associations between prenatal SHS exposure and newborn TL (Table 2). In the unadjusted model, prenatal SHS exposure was associated with 10.5% shorter newborn TL (percent difference: −10.5%; 95% CI: −15.7, −5.0). After adjustment for potential confounders, including maternal age, pre-pregnancy BMI, education status, annual household income, parity, gestational diabetes, hypertension disorders in pregnancy, infant sex, birth weight, gestational age, and date of delivery, this association was slightly attenuated, but still significant (percent difference: −9.7%; 95% CI: −15.0, −4.0).

We further did the stratified analysis by infant sex. After adjustment for potential confounders, the estimate for boys was lower (percent difference: −10.9%; 95% CI: −18.6, −2.5) than that for girls (percent difference: −8.5%; 95% CI: −15.8, −0.5) (Table 2), but the interaction term between newborn sex and prenatal SHS was not significant (P = 0.751).

Table 2 Percent difference in newborn telomere length by prenatal second-hand smoke exposure
Full size table

To evaluate the robustness of our study, we performed sensitivity analyses excluding newborns with preterm birth and low birth weight, women ≥35 years, women with gestational diabetes and hypertensive disorders in pregnancy, and women with active cigarette smoking before or during pregnancy (Table 3). The associations between prenatal SHS exposure and umbilical cord blood TL did not materially changed.

Table 3 Sensitivity analysis
Full size table

Discussion

In this study, we observed that compared with newborns without prenatal SHS exposure, those with SHS exposure had shorter umbilical cord blood TL. Besides, although the interaction term between newborn sex and prenatal SHS was not significant, the estimate for boys was lower than that for girls.

The association between active smoking and shorter TL has been well established in adults (Table 3). A meta-analysis included 30 studies have suggested an inverse relationship between pack-years of smoking and TL.10 So far, there was only one study regarding the impact of prenatal tobacco exposure on cord blood TL. Salihu et al.17 conducted a study among 86 mother–newborn pairs, and they found a dose–response pattern that newborn TL for nonsmokers was longest, which is followed by that for passive smokers, and the shortest was that for active smokers. It was noteworthy that Salihu et al.17 did not control any potential confounders in their study. Besides, the small sample size, underprivileged study subjects with a high active smoking rate of Salihu’s study might reduce the extension of these results to other population. There were two studies which investigated the association between prenatal tobacco exposure and early life TL among children population. Thall et al.18 conducted a cross-sectional study among 104 children aged 4 to 14 years old, which suggested that children with prenatal tobacco exposure (maternal active and passive smoking during pregnancy) had shorter telomere than those without prenatal tobacco exposure. Consistent with Thall’s study, a case–control study, which conducted among 196 Hong Kong Chinese children aged <15 years old, reported a negative dose–response relationship between the buccal epithelial cells TL and prenatal tobacco exposure duration.19 Because TL in later life is mainly determined by initial TL,20 these two studies might support our findings.

Oxidative stress might be a potential mechanism underlying the associations between prenatal SHS exposure and newborn TL. Because of the high guanine content, telomeres are vulnerable to oxidative stress. Oxidative stress induce DNA breakage and accelerate the rate of telomere shorten with each cell division.21 SHS is a complex mixture containing various known carcinogens, teratogens, and toxins.22 Some of these chemical contaminants, such as nicotine, CO, benzo(a)pyrene, lead, and cadmium, readily access to the fetus via placenta.23,24,25,26,27 The exposure to these toxic chemicals in cigarette smoke would induce oxidative stress, which accelerate telomere shortening in the fetus. In addition, endothelial dysfunction caused by cigarette smoke would intensify the oxidative stress, resulting in shortened telomeres.28

In this study, we also observed that the estimate for boys was lower than that for girls. However, it should be noted that the interaction between newborn sex and prenatal SHS was not significant, which indicated that this sex-specific association between prenatal SHS and newborn TL might be attributed to chance. Compared with female fetuses, male fetuses have lower level estrogen.29 Estrogen upregulates superoxide dismutase and glutathione peroxidase gene expression, resulting significantly higher antioxidant enzymes level in female fetus.30 Therefore, male fetuses might be more susceptible to oxidative stress caused by prenatal SHS exposure.

There were several limitations in our study. First, one major limitation is that prenatal SHS exposure was self-reported, which might introduce recall bias. However, previous studies had shown good reliability and validity of self-reported exposure to SHS among adults.31 Second, although extensive potential confounders were controlled in this study, we could not eliminate the unmeasured confounding caused by variables, such as maternal stress, nutritional status during pregnancy, and paternal age. These factors were reported to be related to offspring TL,32,33,34 and might confound the relationship between SHS and TL. Further studies with these informations are needed to confirm the finding of this study.

In this study, we found out prenatal SHS exposure was related with shorter newborn TL. Because the effect of early-life TL persists over the life course, the prenatal SHS exposure might have a lasting harmful effect on offspring’s health outcomes in childhood and adulthood.20 China is the world’s largest producer and consumer of tobacco. According Global Adult Tobacco Survey estimation in 2010, 52.9% Chinese men were current smokers, whereas only 2.4% women were current smokers.14 The higher level of household smoking, poor ventilation, and overcrowding lead to widespread SHS exposure among Chinese women.14 The findings of our study adds to the growing body of evidence that prenatal exposure to SHS lead to adverse health impact on fetus. Effective intervention programs, which prevent SHS exposure during pregnancy, might improve longevity and life quality of offspring.