Objective
In adults, one of the major determinants of leukocyte telomere length (LTL), a predictor of age-related diseases and mortality, is cumulative psychosocial stress exposure. More recently we reported that exposure to maternal psychosocial stress during intrauterine life is associated with LTL in young adulthood. The objective of the present study was to determine how early in life this effect of stress on LTL is apparent by quantifying the association of maternal psychosocial stress during pregnancy with newborn telomere length.
Study Design
In a prospective study of N = 27 mother-newborn dyads maternal pregnancy-specific stress was assessed in early gestation and cord blood peripheral blood mononuclear cells were subsequently collected and analyzed for LTL measurement.
Results
After accounting for the effects of potential determinants of newborn LTL (gestational age at birth, weight, sex, and exposure to antepartum obstetric complications), there was a significant, independent, linear effect of pregnancy-specific stress on newborn LTL that accounted for 25% of the variance in adjusted LTL (β = −0.099; P = .04).
Conclusion
Our finding provides the first preliminary evidence in human beings that maternal psychological stress during pregnancy may exert a “programming” effect on the developing telomere biology system that is already apparent at birth, as reflected by the setting of newborn LTL.
The contribution of various forms of stress including psychological stress to the origin and progression of many complex, common aging-related health disorders that represent the major global burden of disease is well established. The effects of exposure to (maternal) psychological stress during intrauterine life appear to be particularly salient, with important consequences for not only adverse birth and neonatal outcomes, but also for subsequent health and disease risk-related outcomes over the lifespan, including metabolic, endocrine, immune, and cognitive processes (ie, the concept of fetal or developmental programming of health). The elucidation of the molecular mechanism(s) underlying the link between stress and disease risk is an area of considerable interest and effort. In recent years telomere biology has emerged as a particularly attractive candidate mechanism in this context. Telomeres, noncoding double-stranded repeats of guanine-rich tandem DNA sequences, and shelterin protein structures that cap the ends of linear chromosomes play a central, critical role in maintaining the integrity (stability) of the genome and cell. They protect chromosomes from recognition by the DNA damage-repair system as DNA breaks. Because DNA polymerase is unable to fully replicate the 3′ end of the DNA strand (the so-called “end-replication problem”) telomeres shorten with each cell division. Eventually telomeres reach a critical short length, which, in turn, leads to cellular senescence or apoptosis. Telomerase, a cellular enzyme, provides maintenance of telomeres and can counteract shortening and its functional consequences by adding telomeric DNA to shortened telomeres. In just the past few years, telomere biology has emerged as a phenomenon of exceptional interest, moving well beyond its previously recognized role as a biomarker of cellular aging and senescence to one that appears to play a causal and far more pervasive role in regeneration of cells and tissues, physiological function, and aging.
Telomere maintenance has relevance for long-term health. Shortened leukocyte telomere length (LTL) in human beings and/or reduced telomerase activity has been consistently associated with earlier mortality, and age-associated disease risk (eg, cardiovascular disease, hypertension, atherosclerosis, heart failure, type 2 diabetes). Interindividual variation in adult LTL is a function of the initial (newborn) setting of telomere length (TL) and subsequent attrition rate. With respect to the determinants of the initial setting of TL, despite the relatively high heritability of TL known genetic variants (from candidate gene as well as genome wide association studies approaches) account for only a small proportion of the variance in LTL, highlighting the potential importance of intrauterine effects in the setting of TL at birth.
The initial observation linking psychological stress in adults with the activity of the telomere biology system has subsequently been replicated in several independent studies. More recently we made the first observation linking maternal psychological stress exposure during intrauterine life with shorter LTL in young adulthood. At the present time it is unknown whether the effect of maternal stress during pregnancy on offspring TL is already evident at birth. Based on observed stress-related alterations in maternal-placental-fetal immune and endocrine processes during gestation and on the putative effects of these processes on telomere biology we hypothesized that exposure to excess maternal stress during pregnancy may impact the fetal telomere biology system in utero, and this effect may already be apparent at birth. Hence, the objective of the present study was to examine the magnitude of the association between maternal psychosocial stress during pregnancy and that proportion of the variance in newborn LTL not accounted for by other potential determinants, ie, newborn gestational age at birth, weight, sex, and exposure to antepartum obstetric complications.
Materials and Methods
Study participants
The study sample was a subsample of women from a larger cohort of women attending prenatal care at a university-based clinic in Pittsburgh, PA, and participating in a prospective, longitudinal research study from early gestation through birth. This subsample comprised subjects in whom additional measures of maternal stress were administered, including the measure of pregnancy-specific stress used in the current report. As depicted in the Appendix (Supplemental Table) , this subsample is representative of the larger cohort. The study population was a predominantly Medicaid-insured, low-income population that was approximately 55% black and 44% white. Eligible women had singleton pregnancies and had no known prepregnancy major medical conditions or fetal anomalies. Enrollment took place at <16 wk of gestation (mean gestational age: 9.5 wk). The study was approved by the University of Pittsburgh Institutional Review Board and informed, written consent was obtained from all participants. At enrollment, women completed a structured interview to provide data on sociodemographic characteristics, medical history, and maternal behaviors. The subsample of N = 27 mother-infant pairs in whom data were also obtained about psychosocial stress during pregnancy comprises the study sample for the current report. The sociodemographic, pregnancy, and newborn birth outcome characteristics of this population are presented in Table 1 .
| Maternal characteristics | Mother-newborn dyads, n = 27 |
|---|---|
| Sociodemographic | |
| Age (y), mean ± SD | 24.0 ± 3.5 |
| Race/ethnicity | |
| Non-Hispanic white | 37% (n = 10) |
| African American | 63% (n = 17) |
| Annual family income | |
| <$10,000 | 44% (n = 12) |
| $10,000-25,000 | 30% (n = 8) |
| >$25,000 | 26% (n = 7) |
| Pregnancy | |
| Presence of obstetric risk condition | 33% (n = 9) |
| Multiparous | 100% (n = 27) |
| Newborn characteristics | |
| Sex (female) | 52% (n = 14) |
| Gestational age at birth (wk), mean ± SD | 38.8 ± 1.4 |
| Birthweight (g), mean ± SD | 3290 ± 460 |
Measures
Pregnancy-specific stress
At enrollment (on average at 9.2 weeks’ gestation), a 4-item pregnancy-specific stress scale was administered. This scale was adapted from a 10-item version that was specifically developed for use in pregnancy research. It assesses the major constructs that encompass pregnancy-specific stress, ie, feelings about being pregnant and concerns about the health of the unborn baby and about labor and delivery. Responses were provided on a 3-point scale (0-2); thus the total score could range from 0–8. This pregnancy-specific stress measure has previously been shown to be a better predictor of adverse birth and child developmental outcomes than other nonpregnancy-specific (or general) stress measures.
Obstetric risk, length of gestation, birthweight, and infant sex
Obstetric risk was defined as the presence of major medical complications in the index pregnancy, ie, gestational diabetes, vaginal bleeding, placenta abruptio, pregnancy-induced hypertension, preeclampsia, or infection. Risk conditions were ascertained by medical chart review and coded as a binary variable (presence or absence of obstetric risk), as previously described. Gestational age was determined by best obstetric estimate with a combination of last menstrual period and early uterine size, and was confirmed by obstetric ultrasonographic biometry <20 weeks using standard clinical criteria. Birthweight and infant sex were abstracted from the newborn medical record. Birthweight was adjusted for gestational age at birth before entering into the regression model.
Cord blood TL
Peripheral blood mononuclear cells (PBMCs) were isolated from cord blood using Ficoll (Sigma-Aldrich, St. Louis, MO) gradient centrifugation following established protocols. Whole DNA was isolated from cord blood PBMC samples using ethanol precipitation. Time from collection to processing the blood was <30 minutes. Aliquots of isolated DNA were shipped to the Blackburn Laboratory at the University of California, San Francisco, where TL assays were performed. Measurement of relative TL (telomere repeat copy number to single gene copy number [T/S] ratio) by quantitative polymerase chain reaction were adapted from a validated published method and performed as described previously. The conversion from T/S ratio to base pairs (bp) was calculated based on the mean telomeric restriction fragment length from Southern blot analysis and the slope of the plot of mean telomeric restriction fragment length vs T/S for these samples. This was expressed as the following formula: bp = 3274 + 2413 * (T/S).
Statistical analyses
Based on the Shapiro-Wilk test, TL data (T/S ratio) were normally distributed (W [Shapiro-Wilk test statistic] = 0.9676, P = .5400 for null hypothesis).
To examine the association between maternal psychosocial stress during pregnancy and the variance in newborn LTL adjusted for the effects of other potential determinants, a regression model was employed that simultaneously included the following predictors: maternal psychological stress, newborn gestational age at birth, birthweight (adjusted for gestational age), sex, and exposure to antepartum obstetric complications. Except for sex, these variables were selected a priori based on a review of the published literature on the determinants of newborn telomere biology across different tissues. Sex was also included in the model because it has been shown to be associated with adult LTL. Other sociodemographic (eg, race/ethnicity) and obstetric/medical (eg, preterm birth, mode of delivery) variables were not included because there was either no a priori hypotheses based on the existing literature regarding how they would influence TL, or we did not find an association with TL in our study sample. The coefficients for each predictor indicate the estimated independent effects of that predictor (ie, adjusting or controlling for the effects of the other predictors). Hence, the unstandardized regression (β) coefficient for maternal stress from this model indicates the change in the adjusted T/S ratio associated with a 1-U change in the pregnancy-specific stress score. Or, in other words, it estimates the magnitude of the effect of maternal stress on newborn TL in newborns of the same gestational age, weight, sex, and exposure to antepartum obstetric complications. Next, to determine the proportion of explained variance a Pearson correlation coefficient was calculated between maternal pregnancy-specific stress and the T/S ratio residualized for newborn gestational age at birth, birthweight (adjusted for gestational age), sex, and exposure to antepartum obstetric complications. Finally, to express the effect size in units of SD change we divided the study sample in 2 groups based on a median split on the pregnancy-specific stress score and then calculated Cohen’s d statistic based on the difference between the 2 groups.
All statistical analyses were run using SPSS 18 (SPSS, Inc, Cary, NC), and the statistical significance level was set at α = 0.05.
Results
The mean cord blood TL was 2.50 ± 0.37 (mean ± SD) T/S ratio (equivalent to 9314 ± 887 bp), which is comparable to previous reports.
The mean pregnancy-specific stress score was 2.40 ± 1.90 (SD) and ranged between 0–7.
As hypothesized, after accounting for the effects of other potential determinants of newborn LTL (gestational age at birth, weight, sex, and exposure to antepartum obstetric complications), there was a significant, independent, linear effect of pregnancy-specific stress on newborn LTL (β = −0.099; P = .04) ( Table 2 ). This effect accounted for 25% (R 2 = 0.25) of the variance in adjusted LTL ( Figure ). Based on a median split by pregnancy-specific stress, there was a 540-bp difference in newborn TL between mothers in the high- vs low-stress groups, which translates to a difference of 0.62 SD units (Cohen’s d ). We note that the unadjusted relationship between pregnancy-specific stress and newborn TL was not statistically significant (β = −0.062, P = .10).
| Variable | Unstandardized β (95% CI) | SE of β | Significance ( P value) | Standardized β | VIF |
|---|---|---|---|---|---|
| Pregnancy-specific stress | −0.099 (−0.197 to −0.002) | 0.047 | .04 | −0.516 | 1.571 |
| Gestational age at birth | −0.008 (−0.124 to 0.107) | 0.056 | .88 | −0.033 | 1.268 |
| Birthweight (adjusted for gestational age) | −0.007 (−0.026 to 0.011) | 0.009 | .42 | −0.208 | 1.722 |
| Infant sex (0 = male, 1 = female) | 0.046 (−0.286 to 0.379) | 0.160 | .77 | 0.064 | 1.312 |
| Antepartum obstetric complications (0 = absent; 1 = present) | −0.218 (−0.554 to 0.118) | 0.162 | .19 | −0.285 | 1.190 |
Results
The mean cord blood TL was 2.50 ± 0.37 (mean ± SD) T/S ratio (equivalent to 9314 ± 887 bp), which is comparable to previous reports.
The mean pregnancy-specific stress score was 2.40 ± 1.90 (SD) and ranged between 0–7.
As hypothesized, after accounting for the effects of other potential determinants of newborn LTL (gestational age at birth, weight, sex, and exposure to antepartum obstetric complications), there was a significant, independent, linear effect of pregnancy-specific stress on newborn LTL (β = −0.099; P = .04) ( Table 2 ). This effect accounted for 25% (R 2 = 0.25) of the variance in adjusted LTL ( Figure ). Based on a median split by pregnancy-specific stress, there was a 540-bp difference in newborn TL between mothers in the high- vs low-stress groups, which translates to a difference of 0.62 SD units (Cohen’s d ). We note that the unadjusted relationship between pregnancy-specific stress and newborn TL was not statistically significant (β = −0.062, P = .10).
