Prenatal particulate air pollution and newborn telomere length: Effect modification by maternal antioxidant intakes and infant sex

Abstract

Background: Evidence links gestational exposure to particulate matter with an aerodynamic diameter of less than 2.5 μm (PM_2.5) with changes in leukocyte telomere length in cord blood with some studies showing sex-specific effects. PM_2.5 exposure in utero increases oxidative stress, which can impact telomere biology. Thus, maternal antioxidant intakes may also modify the particulate air pollution effects.

Methods: We examined associations among prenatal PM_2.5 exposure and newborn relative leukocyte telomere length (rLTL), and the modifying effects of maternal antioxidant intake and infant sex. We estimated daily PM_2.5 exposures over gestation using a validated spatiotemporally resolved satellite-based model. Maternal dietary and supplemental antioxidant intakes over the prior three months were ascertained during the second trimester using the modified Block98 food frequency questionnaire; high and low antioxidant intakes were categorized based on a median split. We employed Bayesian distributed lag interaction models (BDLIMs) to identify both sensitive windows of exposure and cumulative effect estimates for prenatal PM_2.5 exposure on newborn rLTL, and to examine effect modification by maternal antioxidant intakes. A 3-way interaction between PM_2.5, maternal antioxidant intake and infant sex was also explored.

Results: For the main effect of PM_2.5, BDLIMs identified a sensitive window at 12-20 weeks gestation for the association between increased prenatal PM_2.5 exposure and shorter newborn rLTL and a cumulative effect of PM_2.5 over gestation on newborn telomere length [cumulative effect estimate (CEE) = -0.29 (95% CI -0.49 to -0.10) per 1μg/m³ increase in PM_2.5]. In models examining maternal antioxidant intake effects, BDLIMs found that children born to mothers reporting low antioxidant intakes were most vulnerable [CEE of low maternal antioxidant intake = -0.31 (95% CI -0.55 to -0.06) vs high maternal antioxidant intake = -0.07 (95% CI -0.34 to 0.17) per 1μg/m³ increase in PM_2.5]. In exploratory models examining effect modification by both maternal antioxidant intakes and infant sex, the cumulative effect remained significant only in boys whose mothers reported low antioxidant intakes [CEE = -0.38 (95% CI -0.80 to -0.004)]; no sensitive windows were identified in any group.

Conclusions: Prenatal PM_2.5 exposure in mid-gestation was associated with reduced infant telomere length. Higher maternal antioxidant intakes mitigated these effects.

Keywords: Antioxidant intakes; Particulate air pollution; Prenatal; Sex-specific effects; Telomere length.

Figure 1.

Scatterplot of the relationship between maternal total antioxidant intakes and newborn rLTL.

Figure 2.. Associations between weekly PM_2.5 levels over gestation and newborn relative leukocyte telomere length (rLTL).

This figure demonstrates the association between PM_2.5 exposures over pregnancy and rLTL using a BDLIM assuming week-specific effects for the overall sample. The model was adjusted for maternal age, self-reported ethnicity, marital status, education level, maternal lifetime stress, antioxidant intake and infant sex. The y-axis represents the change in rLTL per 1 μg/m³ increase in PM_2.5. The x-axis represents gestational age in weeks. The solid line shows the estimated change in rLTL while gray areas indicate 95% credible intervals (CIs). A sensitive window is identified for weeks 12-20 of gestation, where the estimated pointwise 95% CI does not include zero. BDLIMs also identified a cumulative estimated effect of PM_2.5 over pregnancy on rLTL of −0.29 (95% CI −0.49 to −0.10) per 1μg/m³ increase in PM_2.5.

Figure 3.. Maternal antioxidant intake-specific associations between PM_2.5 exposure over gestation and newborn relative telomere length (rLTL).

This figure demonstrates the maternal antioxidant intake-specific associations between PM_2.5 exposures over pregnancy and rLTL using a BDLIM assuming week-specific effects for children born to mothers reporting low versus high antioxidant intake during pregnancy. Models were adjusted for maternal age, self-reported ethnicity, marital status, education level, LSC-R, and infant sex. Panel A demonstrates the cumulative effect on newborn rLTL per 1 μg/m³ increase in PM_2.5 by maternal antioxidant intake [cumulative effect estimate: low maternal antioxidant intake = −0.31 (95% CI −0.55 to −0.06) vs high maternal antioxidant intake = −0.07 (95% CI −0.34 to 0.17) per 1μg/m³ increase in PM_2.5.5]. In panel B, the y-axis represents the change in rLTL per 1 μg/m³ increase in PM_2.5. The x-axis represents gestational age in weeks. The solid line shows the estimated change in rLTL while gray areas indicate 95% credible intervals (CIs). No sensitive window is identified.

Figure 4.. Sex-specific associations between weekly PM_2.5 levels over gestation and newborn relative telomere length (rLTL).

This figure demonstrates the sex-specific associations between PM^2.5 exposures over pregnancy and rLTL using a BDLIM assuming week-specific effects for the girls and boys. Models were adjusted for maternal age, self-reported ethnicity, marital status, education level, LSC-R, and antioxidant intake. Panel A demonstrates the cumulative effect on newborn rLTL per 1 μg/m³ increase in PM_2.5 by infant sex [cumulative effect estimate: girls = −0.15 (95% CI −0.40 to −0.05) vs boys = −0.30 (95% CI −0.58 to −0.01) per 1μg/m³ increase in PM_2.5.5]. In panel B, the y-axis represents the change in rLTL per 1 μg/m³ increase in PM_2.5. The x-axis represents gestational age in weeks, for girls and boys separately. The solid line shows the estimated change in rLTL while gray areas indicate 95% credible intervals (CIs). No sensitive window is identified in girls whereas a sensitive window is identified in boys from 16-19 weeks gestation where the estimated pointwise 95% CI does not include zero.

Figure 5.. Sex- and maternal antioxidant (AO) intake-specific associations between cumulative prenatal PM_2.5 levels and newborn relative telomere length (rLTL).

This figure demonstrates the group-specific cumulative estimated effect and 95% credible intervals (CIs) between PM_2.5 exposures over pregnancy and rLTL using a BDLIM, specifically for girls and boys by maternal total AO intakes dichotomized around the median. Models were adjusted for maternal age, selfreported race/ethnicity, marital status, education level and lifetime stress. By group, the cumulative estimated effects were: boys with high maternal AO intake (N=36) = −0.04 (95% CI −0.45 to 0.33); girls with high maternal AO intake (N=40) = 0.02 (−0.26 to 0.30); boys with low maternal AO intake (N=36) = −0.38 (95% CI −0.80 to −0.004); and girls with low maternal AO intake (N=40) = −0.26 (95% CI −0.59 to 0.02), per 1μg/m³ increase in PM_2.5.