Infant HPA axis as a potential mechanism linking maternal mental health and infant telomere length

Psychoneuroendocrinology 88 (2018) 38–46

Contents lists available at ScienceDirect

Psychoneuroendocrinology journal homepage: www.elsevier.com/locate/psyneuen

Infant HPA axis as a potential mechanism linking maternal mental health and infant telomere length☆

T

⁎

Benjamin W. Nelsona, , Nicholas B. Allena, Heidemarie Laurenta,b a b

Department of Psychology, University of Oregon, Eugene, OR, USA Department of Psychology, University of Illinois Urbana-Champaign, Champaign, IL, USA

A R T I C L E I N F O

A B S T R A C T

Keywords: Cortisol Depressive symptoms Infant Maternal mental health Mindfulness Telomere

Maternal depression has been suggested to be an independent risk factor for both dysregulated hypothalamicpituitary-adrenal axis (HPA) functioning and shorter telomere length in offspring. In contrast, research suggests that individual differences in mindfulness may act as a protective factor against one’s own telomere degradation. Currently, research has yet to investigate the association between longitudinal changes in maternal mental health (depressive symptoms and mindfulness) and salivary infant telomere length, and whether such changes might be mediated by alterations in infant cortisol response. In 48 mother-infant dyads, we investigated whether the changes in maternal mental health, when infants were 6–12 months of age, predicted change in infant cortisol reactivity and recovery over this period. We also investigated whether these changes in infant HPA functioning predicted subsequent infant salivary telomere length at 18 months of age. Furthermore, we investigated whether change in infant HPA functioning provided a potential pathway between changes in maternal mental health factors and infant salivary telomere length. Analyses revealed that increases in maternal depressive symptoms over that six-month period indirectly related to subsequent shorter infant telomere length through increased infant cortisol reactivity. Implications for the ways in which maternal mental health can impact offspring stress mechanisms related to aging and disease trajectories are discussed.

1. Introduction Shorter telomere length (TL) is associated with early life stress and negative health trajectories across the lifespan (Drury et al., 2012; Shalev, 2012; Shalev et al., 2013). Therefore, it is important to identify early life factors that are associated with TL to better understand the emergence of divergent health trajectories. Indeed, the mechanistic origins of adult disease and psychopathology can often be traced back to early developmental disturbances that are both psychological and biological in nature (Shonkoff et al., 2009). Research has begun to elucidate how maternal mental health and offspring hypothalamic-pituitary-adrenal axis (HPA) stress response are independently associated with adolescent offspring TL (Gotlib et al., 2015). Importantly, questions remain as to when these divergent health trajectories begin, as well as the etiological pathways that mediate the effects of maternal mental health factors on the biological aging of infants in early life. Telomeres are repeated nucleotide sequences (TTAGGG) at the end of eukaryotic chromosomes that protect chromosomes from

deterioration and enable cellular integrity (Blackburn and Epel, 2012). During somatic cell division DNA polymerase is not able to fully replicate the 3′ end of linear DNA resulting in a progressive loss of telomeric repeats (Blackburn, 1991). TL is affected by age and genetics, indicating that telomere length is a “biological clock”, which is also influenced by environmental factors, such as cumulative exposure to lifetime stressors (Aviv, 2008; Olovnikov, 1996). Shorter TL is associated with stressful life events, as well as a range of negative health outcomes (i.e., cardiovascular disease, dementia, diabetes, cancer, obesity, and early mortality; see Blackburn and Epel, 2012), while longer TL length is associated with both positive social relationships (Uchino et al., 2012) and mindfulness (Blackburn and Epel, 2012; Hoge et al., 2013; Jacobs et al., 2011). Cortisol, the end product of neuroendocrine activation, has been proposed to be one mechanism that is associated with TL and cell senescence (Shalev, 2012) as research has demonstrated that human T lymphocytes exposed to cortisol show reduced telomerase, the enzyme responsible for telomere maintenance (Choi et al., 2008), and reduced TL (Vartak et al., 2014). In addition,

Abbreviations: BMI, body mass index; CES-D, center for epidemiological studies of depression scale; HPA axis, hypothalamic pituitary adrenal axis; SES, socioeconomic status; TL, telomere length ☆ Open Science Framework Citation: osf.io/rc824 ⁎ Corresponding author: Department of Psychology, 1227 University of Oregon, Eugene, OR, 97403, USA. E-mail address: [email protected] (B.W. Nelson). https://doi.org/10.1016/j.psyneuen.2017.11.008 Received 10 June 2017; Received in revised form 13 October 2017; Accepted 15 November 2017 0306-4530/ © 2017 Elsevier Ltd. All rights reserved.

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

whether these patterns of cortisol reactivity and recovery mediated the relationship between changes in maternal mental health (depressive symptoms and mindfulness) during this time and subsequent infant TL at 18 months of age. First, we predicted that increased maternal depressive symptoms and decreased dispositional mindfulness would be associated with increased cortisol reactivity and incomplete cortisol recovery (i.e., greater return of cortisol to baseline levels after stress exposure) in infants. Second, we hypothesized that increased infant cortisol reactivity would predict shorter TL, while more complete infant cortisol recovery would predict longer TL. Finally, we hypothesized that change in maternal depressive symptoms and/or mindfulness would indirectly relate to subsequent infant TL due to change in infant cortisol response. In particular, mothers that showed increased depressive symptoms and/or decreased mindfulness from 6 to 12 months would have infants with increased cortisol reactivity and impaired cortisol recovery, and shorter telomeres at 18 months. This study was pre-registered with Open Science Framework (osf.io/rc824).

greater cortisol response to a laboratory stressor in young children (Kroenke et al., 2011) and adults (Tomiyama et al., 2012) is associated with shorter TL. Furthermore, adolescents of depressed mothers show both greater cortisol reactivity and shorter TL (Gotlib et al., 2015). Telomeres thus represent an important biomarker for the study of health risk and resilience processes. 1.1. Maternal mental health and infant stress response Early relationships have implications for health outcomes across the lifespan (Shonkoff et al., 2009). Maternal psychological characteristics can serve as risk or protective factors for infant outcomes. For example, in terms of risk factors, maternal stress during pregnancy is associated with shorter infant TL (Send et al., 2017) and infants of depressed mothers tend to have greater cortisol reactivity to stress (Azar et al., 2007; Brennan et al., 2008). Furthermore, it is important to study changes across time in maternal mental health during infancy as these patterns have differential associations with infant cortisol response dynamics over the first 18 months of life (Laurent et al., 2011). Similarly, it is important to study changes in infant cortisol responses over time as research has shown that from 2 to 15 months of age infants tend to dampen their HPA responses to stressors even though they may show signs of behavioral distress (Gunnar et al., 1996). On the protective side, individual differences in mindfulness, the psychological capacity to maintain nonjudgmental awareness of the present moment (KabatZinn and Hanh, 1990), may help with regulating one’s own and one’s infant’s stress. For example, women with greater levels of dispositional mindfulness tend to have more flexible HPA axis responses (i.e., quicker post-stress cortisol recovery; see Laurent et al., 2015), and mindfulness in the parenting relationship predicts both quicker maternal cortisol recovery and lower infant cortisol levels during stress (Laurent et al., 2017). Therefore, maternal depressive symptoms and mindfulness may be hypothesized as risk and protective factors, respectively, that influence infant stress responses – with risk factors increasing cortisol reactivity and slowing recovery, and protective factors lowering cortisol reactivity and quickening recovery. Research has yet to fully elucidate how maternal mental health impacts not only the stress response of their infants, but also their infant’s biological aging as reflected in TL.

2. Methods and materials 2.1. Participants and recruitment Mothers were recruited from the Women Infants Children program and other community agencies serving low-income families in a midsized city in the pacific northwest of the United States. To be eligible, mothers had to speak English, have a ≪ 12-week-old infant, and anticipate remaining in the area until this target child was 18 months old. Table 1 gives demographic information about the sample at the first assessment. Of the 91 mother-infant dyads who began the study at time 1, 48 dyads (53%) participated at all four assessment times and provided a saliva samples for telomere assay, resulting in the final sample size. Compared to non-completers, study completers tended to be older (M = 28.40 vs. 25.38, F[1,88] = 7.44, p = 0.01), in a longer-term romantic relationship (M = 3.11 years, SD = 0.88 vs. 2.19 years, SD = 0.60 F[1,38] = 15.09, p ≪ 0.001), have more biological children (M = 2.94, SD = 0.93 vs. M = 2.55, SD = 0.83), and reported higher household income, χ2(7) = 14.36, p = 0.045. There were no differences in infant sex, racial/ethnic group identification, likelihood of being in a relationship with the target child’s biological father or degree of contact with the father, education, or employment status. Of the mental health-related variables reported at time 1, the only difference that emerged was for current depressive symptoms (M = 7.68 for completers vs. 11.70 for non-completers, F(1, 86) = 5.33, p = 0.02), indicating that mothers that experienced the highest levels of depressive symptoms may have been lost in this study due to attrition. As discussed in the limitations section, this unfortunately restricts the range of depressive symptomatology in our sample to a largely nonclinical range (CES-D at T2 mean = 8.83, SD = 6.75; CES-D at T3 mean = 10.47, SD = 7.55).

1.2. HPA axis and telomere length Research indicates that oxidative stress via dysregulated HPA axis activation may be one mechanism that connects the exposure to early life stress and later TL (Shalev, 2012). Indeed, as mentioned above, T lymphocyte exposure to cortisol is associated with reduced activity of telomerase, an enzyme responsible for maintaining TL (Choi et al., 2008), and reduced TL (Vartak et al., 2014). Research further demonstrates independent associations between early life stress and increased cortisol reactivity in infants of depressed mothers (Azar et al., 2007; Brennan et al., 2008) and shortened TL (Drury et al., 2012; Price et al., 2013). This trend seems to persist into adolescence as healthy adolescent offspring of mothers with recurrent episodes of depression are at increased risk for altered HPA axis response to stress (i.e., greater cortisol reactivity) and shorter TL (Gotlib et al., 2015). However, research has yet to investigate these connections in an infant sample and investigate whether changes in infant HPA axis activity mediate the relationship between changes in maternal mental health and later infant TL. In other words, the conceptual model of maternal depressive symptoms acting as a stressor to influence infant cortisol response, which subsequently impacts infant TL has yet to be tested.

2.2. Procedure Prior to study participation, mothers gave written informed consent to all study procedures, which had been approved by the University of Oregon Institutional Review Board. Mothers completed study assessments at four times postnatally: time 1 (T1) at 3 months, time 2 (T2) at 6 months, time 3 (T3) at 12 months, and time 4 (T4) at 18 months. The rate of participation, measures collected, infant age and sex, and maternal age at each wave are outlined in Fig. 1. At T1, participants completed a home visit that involved a diagnostic interview, and at times 2–4 participants completed laboratory sessions with their infant. Each laboratory session included a developmentally appropriate interpersonal stressor involving maternal unavailability and/or confrontation with a strange adult. At T2, the Still Face Experiment was used, which is a procedure

1.3. Current study The current study investigated whether changes in maternal mental health predict changes in infant cortisol reactivity and recovery between 6 months and 12 months of age. Moreover, we investigated 39

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

range of emotions and emotion regulation in 1–3 year-old children (Goldsmith and Rothbart, 1999). During the Maternal Separation episode, the infant was left alone in the room for 30s. Next, during the Stranger Approach episode, the infant was exposed to an unfamiliar male research assistant who gradually approached, spoke to, and picked up the infant over the course of 1 min.

Table 1 Sample Descriptives. Variable

Number

Percent of Sample

Race/Ethnic Identification. Caucasian Latina Asian American Native American

38 5 3 2

79.2 10.4 6.3 4.2

Infant Gender Female Male

29 19

60.4 39.6

2 2 17

4.2 4.2 35.4

23 4

47.9 8.3

Relationship Length < 1 year 1–2 years 2–5 years 5–10 years ≫ > 10 years Missing

1 2 11 4 1 29

2.1 4.2 22.9 8.3 2.1 60.4

Education High school Vocational/technical school (2-year) Some college College graduate (4-year) Master’s degree Other

11 6 20 6 4 1

22.9 12.5 41.7 12.5 8.3 2.1

Employment Self-employed Part-time paid work Full-time paid work On leave Unemployed Full-time homemaker Student

4 6 6 6 7 16 3

8.3 12.5 12.5 12.5 14.6 33.3 6.3

Household Income < $4,999 $5,000–$9,000 $10,000–$19,999 $20,000–$29,999 $30,000–$39,999 $40,000–$49,999 $50,000–$74,999 $75,000–$99,999

12 4 3 9 6 5 7 2

25.0 8.3 6.3 18.8 12.5 10.4 14.6 4.2

Relationship Status Single Dating Living with Someone (not a legal domestic partnership) Married Legal/Registered Domestic Partnership

2.2.1. Cortisol At each laboratory session, four saliva samples were collected from both the mother (passive drool, Salimetrics Saliva Collection Aid) and infant (Salimetrics Infant’s or Child’s Swab as appropriate for age). Sessions were conducted in the afternoon, a time when diurnal fluctuations in cortisol should be less marked. The first sample (baseline stress) was collected after arrival at the lab when the mother had answered questions about factors that could impact salivary cortisol. If any of several conditions had been violated—if the mother indicated that she or the infant had eaten recently or were sick with a fever—the participants were rescheduled. The second sample was collected immediately following the peak stressor (anticipatory stress), the third sample was collected 20 mins after the start of the peak stressor (peak stress), and the fourth sample was taken 30 mins after the preceding one (stress recovery). See supplementary material for cortisol trajectories. Due to budgetary limits we were unable to collect a second follow-up assessment to index recovery more fully as has been done in some research protocols. Samples were stored at −20° C until shipped on dry ice for assay. Here we defined cortisol reactivity as the change of cortisol in response to the laboratory stressor (i.e., change from baseline sample to our peak stress sample) and we defined cortisol recovery as the response of the system after exposure to the laboratory stressor and not to imply that the system was necessarily deactivating after the stressor (i.e., change from peak stress sample to recovery sample). 2.2.2. Telomere At laboratory session 4, one saliva sample was collected at the start of the session from both the mother (passive drool using DNA Genotek Oragene DISCOVER (OGR-500) collection devices) and infant (assisted collection using DNA Genotek Oragene DISCOVER (OGR-575) collection devices as appropriate for age) and then stored at room temperature. 2.3. Measures 2.3.1. Maternal depressive symptoms The Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977) is a 20-item self-report questionnaire that was designed to measure depressive symptoms in the general population. Total scores were used for analysis (T2 mean = 8.83, SD = 6.75; T3 mean = 10.47, SD = 7.55). Internal reliability was good for CES-D at both time points used, T2 CES-D (α = 0.881) and T3 CES-D (α = 0.864).

designed to observe how infants respond to a stressful situation involving maternal non-responsiveness (Toda and Fogel, 1993). During the first of three 2-min episodes, the mother and infant engaged in playful face-to-face interaction. During the second episode, the mother was asked to “show an expressionless or blank face to your baby, and try not to touch or talk to your baby” (still face). During the final episode, the mother was allowed to freely reengage with her infant. At T3, the Strange Situation was used, which is a procedure that involves a series of increasingly stressful separations from the mother or other attachment figure (Ainsworth and Wittig, 1969). During the first of seven 3-min episodes, the infant and mother engaged in play with toys. Next, an unfamiliar female research assistant (“stranger”) joined them, and after a period with both adults, the mother left the room (first separation). Following a reunion with the mother, the infant was left alone in the room (second separation). During the final portion of the procedure, the stranger and then the mother re-entered the room to engage with the infant. At T4, the Laboratory Temperament Assessment Battery (LabTAB) locomotor version procedure was used, which is designed to elicit a

2.3.2. Maternal dispositional mindfulness The Five Facet Mindfulness Questionnaire (FFMQ; Baer et al., 2006) is a 39-item self-report questionnaire designed to measure dispositional mindfulness. Total scores were used for analysis (T2 mean = 3.66, SD = 0.51; T3 mean = 3.64, SD = 0.46). Internal reliability was good for FFMQ at both time points used, T2 FFMQ (α = 0.895) and T3 FFMQ (α = 0.886). 2.3.3. Cortisol assay Mothers’ and infants’ saliva samples were assayed in duplicate with the commercially available Salivary Cortisol Enzyme Immunoassay (Salimetrics, Carlsbad, CA) without modification to the manufacturer’s recommended protocol. The test uses 25 μl saliva, has a lower limit of sensitivity of 0.007 μg/dl, standard curve range 0.012 μg/dl-3.0 μg/dl. The intra-assay coefficient of variation was on average ≪10%, and the inter-assay coefficient of variation was on average ≪15%. Infant 40

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

Fig. 1. Flowchart of participation across Times 1 through 4 (T1-T4, respectively). Note: at T1 average infant age was slightly younger than 3 months as mothers completed questionnaires prior to the in home session.

mix contains 20 mM Tris-HCl, pH 8.4; 50 mM KCl; 200 mM each dNTP; 1% DMSO; 0.4 x Syber Green I; 22 ng E. coli DNA per reaction; 0.4 Units of Platinum Taq DNA polymerase (Invitrogen Inc.) per 11 microliter reaction; 7 ng of genomic DNA. Tubes containing 26, 8.75, 2.9, 0.97, 0.324 and 0.108 ng of a reference DNA (from Hela cancer cells) are included in each PCR run so that the quantity of targeted templates in each research sample can be determined relative to the reference DNA sample by the standard curve method. The same reference DNA was used for all PCR runs. To control for inter-assay variability, 8 control DNA samples are included in each run. In each batch, the T/S ratio (i.e., the average telomere length, relative to the reference standard) of each control DNA is divided by the average T/S for the same DNA from 10 runs to get a normalizing factor. This was done for all 8 samples and the average normalizing factor for all 8 samples was used to correct the participant DNA samples to get the final T/S ratio. The T/S ratio for each sample was measured twice. When the duplicate T/S value and the initial value varied by more than 7%, the sample was run the third time and the two closest values were reported. Using this method, the typical average CV for this study was 3–4%. Covariates We collected infant age as a potential covariate as one’s own age has known associations with TL (Müezzinler et al., 2013). In addition, we collected maternal and paternal age as paternal age has been associated with infant TL at birth (De Meyer et al., 2007). We collected infant sex as this has known associations with cortisol stress response (Davis and Granger, 2009), although research indicates no difference in TL based on infant sex (Okuda et al., 2002). We also collected any current or past clinical diagnosis of depressive or anxiety disorders in mothers as maternal depression has known associations with adolescent cortisol response and TL (Gotlib et al., 2015). In addition we collected variables on race as there may be race related differences in TL at birth (Rewak et al., 2014), although this has not been found in all studies (Okuda et al., 2002). All saliva samples were collected in the afternoon during which we collected recent tooth brushing, sickness, and wake times. Finally, we collected maternal education, employment, and income as

cortisol responses were used in the current study. 2.3.4. Telomere assay Salivary telomere assay was performed by The Blackburn Lab at University of California San Francisco. The TL measurement assay is adapted from the published original method by Cawthon (Cawthon, 2002; Lin et al., 2010). Their methods for telomere assay are as follows. Genomic DNA was purified from 500 ul of saliva with Agencourt DNAdvance kit (cat# A48705, Beckman Coulter Genomics Inc. Brea CA) according to manufacturer’s instruction. DNA was quantified by Quant-iT™ PicoGreen® dsDNA Assay Kit (cat# P7589, Life Techonologies, Grand Island, NY) and ran on 0.8% agarose gels to check the integrity. No DNA samples were degraded and therefore they were all included in TL analysis. Salivary samples contain approximately 70% white blood cells and 30% buccal epithelial cells. While research has yet to look at T/S correlations between saliva and other tissues, research shows a strong positive correlation between leukocyte TL and salivary TL (Mitchell et al., 2014) as well as significant positive correlations between TL of leukocytes, skin, skeletal muscle, and subcutaneous fat (Daniali et al., 2013). In addition, research shows correlations between relative TL across different tissues (Reichert et al., 2013). Cycling for T(telomic) PCR: Denature at 96 °C for 1 min; denature at 96 °C for 1 s, anneal/extend at 54 °C for 60 s, with fluorescence data collection, 30 cycles. Cycling for S (single copy gene) PCR: Denature at 96 °C for 1 min; denature at 95 °C for 15 s, anneal at 58 °C for 1 s, extend at 72 °C for 20 s, 8 cycles; followed by denature at 96 °C for 1 s, anneal at 58 °C for 1 s, extend at 72 °C for 20 s, hold at 83 °C for 5 s with data collection, 35 cycles. The primers for the telomere PCR are tel1b [5′CGGTTT(GTTTGG)5GTT-3′], used at a final concentration of 100 nM, and tel2b [5′-GGCTTG(CCTTAC)5CCT-3′], used at a final concentration of 900 nM. The primers for the single-copy gene (human beta-globin) PCR are hbg1 [5′ GCTTCTGACACAACTGTGTTCACTAGC-3′], used at a final concentration of 300 nM, and hbg2 [5′-CACCAACTTCATCCACGT TCACC-3′], used at a final concentration of 700 nM. The final reaction 41

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

Fig. 2. Correlation Matrix. Note: I_TL = infant telomere length, M_TL = maternal telomere length, ICortReact = change in infant cortisol reactivity, ICortRecov = change in infant cortisol recovery, MDep = change in maternal depressive symptomatology, MMindf = change in maternal dispositional mindfulness, IGender = infant gender, IAge = infant age at time 4, MAge = mother age at time 4, Fage = father age at birth, MDepDx = current or past depressive diagnosis, MAnxDx = current or past anxiety diagnosis, MBMI = mother body mass index at time 4

sample 1 from sample 3, while cortisol recovery was calculated by subtracting sample 4 from sample 3. This was done for cortisol samples at T2 and T3. Analyses revealed that at T2, 21 infants (43.75%) showed increase in cortisol from the baseline to peak stress, while 27 infants (56.25) showed no increase or a decrease in cortisol. In addition, At T3, 22 infants (45.83%) showed increase in cortisol from baseline to peak stress, while 26 (54.17%) showed no increase or a decrease in cortisol. Change in cortisol reactivity and recovery were calculated by regressing T1 reactivity/recovery variables onto T2 reactivity/recovery variables and creating an unstandardized residual that represents the change in cortisol reactivity and recovery across time. As described above, we defined cortisol reactivity as the change in cortisol in response to the laboratory stressor (i.e., change from baseline sample to peak stress sample) and we defined cortisol recovery as the response of the system after exposure to the laboratory stressor (i.e., change from peak stress sample to recovery sample). Data were missing for T2 maternal mindfulness from 2 participants (4.2%), T2 maternal CESD from one participant (2.1%), infant T2 sample 4 cortisol from 2 participants (4.2%), infant T3 sample 1 cortisol from 5 participants (10.4%), infant T3 sample 3 cortisol from 6 participants (12.5%), and infant T3 sample 4 cortisol from 6 participants (12.5%). To preserve statistical power lost through deletion methods, single imputation with the EM algorithm was used to estimate missing data (Little and Rublin, 1987). Little’s MCAR test indicated that the null hypothesis that the data were missing in a random fashion cannot be rejected, χ2 = 88.42 (df = 123; p = 0.99).

markers of socioeconomic status (SES), as lower SES in childhood is associated with cortisol (Clearfield et al., 2014) and shorter TL in adulthood (Cohen et al., 2013). While infant birth weight was collected and had no association with infant TL (r = 0.107, p = 0.469), we did not collect infant length, so we were not able to calculate infant body mass index (BMI), although previous research indicates that neither infant body size (Kajantie et al., 2012) nor weight (Okuda et al., 2002) are associated with TL. In order to preserve power and avoid overfitting, only covariates that were significantly associated with TL were included in analyses. None of these variables showed a consistent effect on mother or infant TL, so they were not included in further model testing (see Fig. 2). 2.4. Statistical analyses All statistical transformations and imputation were conducted with IBM SPSS Statistics, version 23 (SPSS Inc., Chicago, IL, USA). All statistical analyses were conducted with R, version 3.3.2. Statistical significance was defined using 95% confidence intervals. Exploratory statistics of histograms as well as skew and kurtosis statistics were run for each variable to check for normality. All variables were winsorized to +/− 3 SD to correct for outliers. Cortisol values were positively skewed, so these values were log transformed. Change scores were used for maternal depressive symptoms and mindfulness as well as infant cortisol reactivity and recovery to take advantage of the longitudinal dataset. Change in maternal depressive symptoms and mindfulness were calculated by regressing T2 variables onto T3 variables and creating an unstandardized residual that represents the change in depressive symptoms and mindfulness across time. Raw changes for maternal depression symptoms from T2 to T3 showed that the largest decrease was by 100%, while the largest increase was by 800%. In addition, raw changes for maternal mindfulness from T2 to T3 showed that the largest decrease in mindfulness was 13.95%, while the largest increase was 39.27%. Cortisol reactivity was calculated by subtracting

2.4.1. Main analyses: regression and mediation Multiple regression was used to test paths between maternal mental health, infant cortisol reactivity/recovery, and infant TL to test the theoretical model that increases in maternal depression/decreases in mindfulness would be associated with increased infant cortisol reactivity and decreased cortisol recovery, which would subsequently be associated with shorter TL. Indirect effect analyses were conducted 42

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

supplementary material for the range of the indirect, direct, and total effects based on 10,000 bootstrapping samples and for the distribution of the indirect effect based on 10,000 Monte Carlo simulations.

using the MBESS package (Kelley, 2017), which generates a 95th percentile confidence interval based on 10,000 bootstrap samples for the indirect (mediated) effect. Such techniques have been shown to be particularly helpful for small sample sizes, as low as 20 (Shrout and Bolger, 2002). Mediation models only involved maternal predictors that were found to relate to infant cortisol (reactivity and recovery) and only used infant cortisol indices found to relate to TL. In the mediation analysis we used change in maternal depressive symptoms as the predictor (x), change in infant cortisol reactivity as the mediator (m), and infant TL as the outcome (y). Note that even if the direct effect of the primary predictor (i.e., change in maternal depressive symptoms) on the outcome (i.e., infant telomere length) is not significant, this does not preclude the use of mediation techniques to detect an indirect effect when bootstrapping is used as the significance of the total effect path does not “determine or constrain” the significance of the indirect path (Hayes, 2009; Hayes and Rockwood, 2016; Shrout and Bolger, 2002).

4. Discussion The current study was designed to investigate the connection between changes in maternal risk (depressive symptoms) and protective (mindfulness) factors and concurrent changes in infants’ cortisol reactivity and recovery, and to test whether these coupled changes in maternal and infant health were associated with subsequent infant TL. This study is the first to examine the mediating role of infant HPA dynamics in the relationship between alterations in maternal mental health and infant biological aging as measured by TL. The results of the current study provide partial support for the hypotheses. Consistent with our hypotheses, increases in maternal depressive symptoms across a 6-month period predicted increased infant cortisol reactivity. In addition, increases in infant cortisol reactivity across this time predicted subsequent shorter infant TL, while more complete infant cortisol recovery (i.e., greater return of cortisol to baseline levels after stress exposure) predicted longer infant TL. Lastly, there was a small, but significant indirect effect of increased maternal depressive symptoms on later infant TL via increased infant cortisol reactivity. It is important to note that telomeres were not collected in a repeated measures fashion, which raises questions about whether the current findings represent a (static) shared risk versus a developmental process. We considered the possibility that mothers who are more prone to depression may have infants with shorter TL from the beginning of life. Post-hoc exploratory analyses revealed that absolute levels of maternal symptoms at each time did not relate to infant TL either directly or indirectly, supporting the idea that the path found between increases in maternal depressive symptoms and shorter infant telomere length represented a developmental process. During the six months between T2 and T3, 13 mothers (27%) showed increased depressive symptoms and 9 mothers (18.75%) had clinically elevated depressive symptoms (CES-D score of 16 or higher). The small indirect effect is not strong and needs to be replicated, but may be due to the fact that mothers who completed the study reported significantly less depressive symptoms than mothers lost to attrition. If this were the case, it is possible that the effect could be stronger for mothers exhibiting higher clinical levels of depressive symptomatology. It is also important to note a potential alternative explanation for the association between greater cortisol reactivity in infants and shorter TL. As mentioned previously, between 2–15 months of age infants tend to have a dampening of their HPA stress responses across time (Gunnar et al., 1996). Within this context, it is possible that rather than greater cortisol reactivity mediating the relationship between increasing maternal depressive symptoms and shorter TL, it may be the case that it is the lack of age-appropriate decreases in cortisol reactivity across infancy, that may in fact mediate this relationship. The observed relationship between increases in maternal depressive symptoms and increased infant cortisol reactivity, as well as increases in infant cortisol reactivity and shorter infant TL, extends previous research to a younger phase of development, which indicates that divergent health trajectories likely begin in the first 18 months of life. Research has found that healthy adolescent daughters of depressed mothers tend to have greater cortisol reactivity and that cortisol reactivity was associated with shorter TL (Gotlib et al., 2015), although a mediation model was not tested. One possible theoretical explanation

3. Results 3.1. Preliminary analyses Correlation analyses using the corrplot R package (Wei and Simko, 2016) revealed that no covariates were associated with TL (Fig. 2 for correlation matrix among study variables). Infant and mother TL were significantly related (r = 0.350, 95% CI [0.073, 0.577], p = 0.015) and mother TL was significantly shorter than infant TL as expected, t(47) = 13.17, 95% CI [0.56, 0.76], p ≪ 0.001 (see Supplementary Material Fig. S1). 3.2. Main analyses: direct paths Increase in infant cortisol reactivity between 6–12 months of age significantly predicted shorter TL at 18 months (β = −0.316, SE = 0.128, 95% CI [−0.574, −0.057], p = 0.018), while increase in infant cortisol recovery over this period significantly predicted longer TL at 18 months (β = 0.167, SE = 0.057, 95% CI [.052, 0.281], p = 0.005). Moreover, increase in maternal depressive symptoms from 6–12 months was associated with increased infant cortisol reactivity over the same period (β = 0.019, SE = 0.009, 95% CI [.001, 0.037], p = 0.040), but did not predict changes in infant cortisol recovery over that period (β = 0.008, SE = 0.021, 95% CI [−0.034, 0.049], p = 0.706). In contrast, increase in maternal dispositional mindfulness did not significantly predict infant cortisol reactivity (β = −0.087, SE = 0.192, 95% CI [−0.474, 0.301], p = 0.654) or recovery (β = 0.135, SE = 0.424, 95% CI [−0.719, 0.988], p = 0.752). Finally, increase in maternal depressive symptoms (β = 0.001, SE = 0.009, 95% CI [−0.017, 0.018], p = 0.939) and increase in maternal dispositional mindfulness (β = 0.129, SE = 0.178, 95% CI [−0.228, 0.487], p = 0.470) did not significantly predict infant TL. 3.3. Main analyses: indirect path As a result of testing the significance of the direct paths of our mediation model above, we decided to run one mediation analysis, with the change in infant cortisol reactivity providing an indirect effect that explained how increases in maternal depressive symptoms were associated with infant TL. According to the 95% CI, increase in infant cortisol reactivity significantly mediated the relationship between increase in maternal depressive symptoms and shorter infant TL (indirect effect = −0.007, 95% CI [−0.027, −0.001]). Specifically, increasing maternal depressive symptoms between 6 and 12 months was associated with increasing infant cortisol reactivity over the same period, which was then associated with shorter infant TL at 18 months.1 See 1

(footnote continued) based confidence interval based on 10,000 bootstrap samples for the indirect (mediated) effect. The indirect effect was significant (effect = −0.007, SE = 0.006, 95% CI [−0.027,−0.001]. We also used the Monte Carlo method in R using 10,000 simulation repetitions for calculating the indirect effect of the mediation. The indirect effect was significant (95% CI [−.0162, −0.0001]).

The SPSS PROCESS Macro (Hayes, 2013) was also used to generate a 95% percentile-

43

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

symptoms to clinically elevated levels of depressive symptoms and/or diagnosed depressive disorders. We did not collect repeated measures of TL, which precluded this study from implementing a pre-post design. In addition, the assessment of change in maternal depressive symptoms and infant cortisol response to a laboratory stressor occurred concurrently, which limited the temporal sequencing of our mediation model. Future studies should use repeated measures of TL in order to index rate of telomere degradation and include further longitudinal measures of maternal symptoms/infant cortisol to better test temporally sequenced mediation. We also did not collect any biological markers of stress during pregnancy, which have been associated with infant TL (Send et al., 2017). Future studies could use hair cortisol or other measures of chronic activation of the HPA system to index maternal stress exposure across pregnancy, as stress during pregnancy has been shown to correlate with TL at birth (Entringer et al., 2013) and in young adulthood (Entringer et al., 2011). Another potentially relevant factor we did not examine is the effect of maternal-infant relationship factors such as attachment. Future studies should address variables representing maternal-infant relational functioning. While infant body weight at birth was collected, infant length at birth was not, so we were unable to calculate BMI. While some studies have shown that infant body size (Kajantie et al., 2012) and weight (Okuda et al., 2002) are not associated with TL as was found in our sample (r = 0.107, p = 0.469), it is less clear whether body weight or BMI changes during infancy may predict shorter TL as some studies have proposed that increased energy demands and growth likely lead to accelerated aging (Belsky and Shalev, 2016; Lemaître et al., 2015). In addition, research has shown BMI to relate to cortisol reactivity and maternal pre-pregnancy BMI to relate to infant TL (Martens et al., 2016). Future studies should control for these potential confounds. We used different developmentally appropriate stressors at each laboratory session. While this had the benefit of preventing possible practice effects, desensitization, and lack of age appropriate stress that may have occurred if the same stress paradigm was used across all laboratory sessions, changing the stress paradigm may have added additional variance that we did not account for in modeling cortisol change from 6 to 12 months of age. Future research might attempt to replicate these findings using the same stress paradigm at each laboratory session. In addition, it is possible that there is potential third variable (e.g., stressful environment, shared genetic/epigenetic factors) that we did not measure that may underlie and explain these findings. For example, there is the possibility that infant stress sensitivity might make parenting more stressful for the mother, increasing her depression, which then leads to harsher parenting and therefore shorter infant TL. Lastly, we would like to underscore the preliminary nature of the current findings and call for replication studies in order to provide more evidence for the current findings.

for the association between increases in cortisol reactivity from 6 to 12 months predicting shorter telomere length at 18 months is that infants that have greater cortisol reactivity may be under greater allostatic load and the exposure to cortisol during this time may down-regulate telomerase, leading to shorter telomeres as research has demonstrated that human T lymphocytes exposed to cortisol show reduced telomerase (Choi et al., 2008). Another possible explanation is that increased cortisol reactivity during this developmental phase might indicate that there is increased growth rate and energy demands, which has been proposed to contribute to more rapid cellular division (Belsky and Shalev, 2016; Lemaître et al., 2015), although our lack of data on body growth precludes us from being able to test this latter possible explanation. The indirect effect of changes in maternal depressive symptoms on infant TL via cortisol reactivity detected here is an important and novel finding that indicates divergent trajectories in health likely start within the first 18 months of life. There are a number of possible reasons for the null direct effect between changes in maternal depressive symptoms and infant TL. First, this lack of association may be due to the presence of a third variable that is functioning as a suppressor variable that may be working in the opposite direction to counteract the path between changes in maternal depressive symptoms and shorter infant TL via increased cortisol reactivity, but more data would be needed to uncover this possibility. Second, this null finding may be due to the relatively short period of time infants would have experienced their mother’s depressive symptoms in the first 18 months of life. In other words, negative early experiences, such as maternal depressive symptoms, may take years before these experiences are expressed at the level of shorter telomeres or as disease (Shonkoff et al., 2009). In contrast, biological factors, such as changes in HPA functioning, may have a more immediate impact on infant TL as was shown in the current sample. The null findings for the role of maternal mindfulness on infant cortisol reactivity and recovery parallels previous research that shows that while maternal depression is associated with infant cortisol secretion over the first month of life, other putatively protective factors, such as maternal sensitivity, were not associated with infant cortisol secretion (Murray et al., 2010). Another possible explanation is that more general maternal mindfulness may have less of an impact on infant cortisol functioning than mindfulness that is specific to parenting (see Laurent et al., 2017). Lastly, while the FFMQ assesses trait mindfulness, the CES-D assesses state depressive symptoms. These variables were significantly associated at both T2 (r = −0.408, p = 0.004) and T3 (r = −0.452, p = 0.001), which indicates that mothers that tended to have higher levels of depression also tended to be less mindful. However, it appears that the variance in depressive symptoms that was related to infant outcomes was independent of the variance it shares with mindfulness, which may explain the null finding for maternal mindfulness.

4.2. Conclusion 4.1. Limitations and future directions This study is the first to address the effect of coupled changes in maternal mental health and infant stress reactivity on subsequent infant biological health. Overall, our findings indicate that health disparities may begin within the first 18 months of life. Findings support more effective maternal mental health treatment and early childhood policies and practices (Shonkoff, 2012), which may be a particularly effective way to confront the etiology of health disparities that emerge across the lifespan (Shonkoff et al., 2009).

While this study leveraged a longitudinal design with a difficult to recruit population (i.e., low income) and provided novel insights into the biological mechanism through which maternal psychological factors may influence infant biological aging, it is important to note a number of limitations. Our study sample of 48 mother-infant dyads from lower income families only allows a small-scale test of hypothesized paths between maternal mental health and infant biological health, which may not be generally representative. Future studies should attempt to replicate this study with a larger sample size and with families that have varying levels of income. Study completers had significantly lower levels of depressive symptoms than did those that were lost to attrition, which may explain the small effect observed for the mediation and lack of significance of the direct effect between change in maternal depressive symptoms and infant TL. Future research should attempt to recruit a range of mothers with non-clinical depressive

Conflict of interest All authors declare no conflicts of interest. Acknowledgments This research was supported by grants from The Mind and Life 44

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

Institute (1440 Research Award 2015-1440-Nelson) awarded to the first author and The Society for Research in Child Development Victoria Levin Award and the University of Oregon College of Arts and Sciences issued to the last author. The funding sources had no role in the study design, data collection and analysis, or submission process. We thank Elissa Epel, Ph.D. for her consultation on our study design and Jue Lin, Ph.D. and Elizabeth Blackburn, Ph.D. for their expertise in assaying the telomere samples.

responses during infancy: normative changes and individual differences. Child Dev. 67 (3), 877–889. Hayes, A.F., Rockwood, N.J., 2016. Regression-based statistical mediation and moderation analysis in clinical research: observations, recommendations, and implementation. Behav. Res. Ther. 39–57. http://dx.doi.org/10.1016/j.brat.2016.11.001. Hayes, A.F., 2009. Beyond baron and kenny: statistical mediation analysis in the new millennium. Commun. Monogr. 76 (4), 408–420. Hayes, A.F., 2013. Introduction to Mediation, Moderation, and Conditional Process Analysis. Guilford, New York, NY, pp. 3–4 https://doi.org/978-1-60918-230-4. Hoge, E.A., Chen, M.M., Orr, E., Metcalf, C.A., Fischer, L.E., Pollack, M.H., De Vivo, I., Simon, N.M., 2013. Loving-Kindness Meditation practice associated with longer telomeres in women. Brain Behav. Immun. 32, 159–163. http://dx.doi.org/10.1016/j. bbi.2013.04.005. Jacobs, T.L., Epel, E.S., Lin, J., Blackburn, E.H., Wolkowitz, O.M., Bridwell, D.A., Zanesco, A.P., Aichele, S.R., Sahdra, B.K., MacLean, K.A., King, B.G., Shaver, P.R., Rosenberg, E.L., Ferrer, E., Wallace, B.A., Saron, C.D., 2011. Intensive meditation training, immune cell telomerase activity, and psychological mediators. Psychoneuroendocrinology 36 (5), 664–681. http://dx.doi.org/10.1016/j.psyneuen. 2010.09.010. Kabat-Zinn, J., Hanh, T.N., 1990. Full Catastrophe Living: Using the Wisdom of Your Body and Mind to Face Stress, Pain, and Illness. Delta Trade, New York, New York. Kajantie, E., Pietiläinen, K.H., Wehkalampi, K., Kananen, L., Räikkönen, K., Rissanen, A., Hovi, P., Kaprio, J., Andersson, S., Eriksson, J.G., Hovatta, I., 2012. No association between body size at birth and leucocyte telomere length in adult life-evidence from three cohort studies. Int. J. Epidemiol. 41 (5), 1400–1408. http://dx.doi.org/10. 1093/ije/dys127. Kelley, K. (2017). MBESS: The MBESS R Package. Kroenke, C.H., Epel, E., Adler, N., Bush, N.R., Obradovic, J., Lin, J., Blackburn, E., Stamperdahl, J.L., Boyce, W.T., 2011. Autonomic and adrenocortical reactivity and buccal cell telomere length in kindergarten children. Psychosom. Med. 73 (7), 533–540. http://dx.doi.org/10.1097/PSY.0b013e318229acfc. Laurent, H.K., Ablow, J.C., Measelle, J., 2011. Risky shifts: how the timing and course of mothers’ depressive symptoms across the perinatal period shape their own and infant’s stress response profiles. Dev. Psychopathol. 23 (2), 521–538. http://dx.doi.org/ 10.1017/S0954579411000083. Laurent, H.K., Laurent, S.M., Nelson, B., Wright, D.B., Sanchez, M.A.D.A., 2015. Dispositional mindfulness moderates the effect of a brief mindfulness induction on physiological stress responses. Mindfulness 6 (5), 1192–1200. http://dx.doi.org/10. 1037/dev0000258. Laurent, H.K., Duncan, L.G., Lightcap, A., Khan, F., 2017. Mindful parenting predicts mothers’ and infants’ hypothalamic-pituitary-adrenal activity during a dyadic stressor. Dev. Psychol. 53 (3), 417–424. http://dx.doi.org/10.1037/dev0000258. Lemaître, J.-F., Berger, V., Bonenfant, C., Douhard, M., Gamelon, M., Plard, F., Gaillard, J.-M., 2015. Early-late life trade-offs and the evolution of ageing in the wild. Proc. Biol. Sci. 282 (1806). http://dx.doi.org/10.1098/rspb.2015.0209. p. 20150209. Lin, J., Epel, E.S., Cheon, J., Kroenke, C.H., Sinclair, E., Bigos, M., Wolkowitz, O., Blackburn, E., 2010. Analyses and comparisons of telomerase activity and telomere length in human T and B cells: insights for epidemiology of telomere maintenance. J. Immunol. Methods 352 (1–2), 71–80. http://dx.doi.org/10.1016/j.jim.2009.09.012. Little, R.J.A., Rublin, D., 1987. Statistical Analysis with Missing Data. Wiley, New York. http://dx.doi.org/10.1002/9781119013563. Müezzinler, A., Zaineddin, A.K., Brenner, H., 2013. A systematic review of leukocyte telomere length and age in adults. Ageing Res. Rev. http://dx.doi.org/10.1016/j.arr. 2013.01.003. Martens, D.S., Plusquin, M., Gyselaers, W., De Vivo, I., Nawrot, T.S., 2016. Maternal prepregnancy body mass index and newborn telomere length. BMC Med. 1–10. http://dx.doi.org/10.1186/s1291601606890. Mitchell, C., Hobcraft, J., McLanahan, S.S., Siegel, S.R., Berg, A., Brooks-Gunn, J., Garfinkel, I., Notterman, D., 2014. Social disadvantage, genetic sensitivity, and children’s telomere length. PNAS 111 (16), 5944–5949. http://dx.doi.org/10.1073/ pnas.1404293111. Murray, L., Halligan, S.L., Goodyer, I., Herbert, J., 2010. Disturbances in early parenting of depressed mothers and cortisol secretion in offspring: a preliminary study. J. Affect. Disord. 122 (3), 218–223. http://dx.doi.org/10.1016/j.jad.2009.06.034. Okuda, K., Bardeguez, A., Gardner, J.P., Rodriguez, P., Ganesh, V., Kimura, M., Skurnick, J., Awad, G., Aviv, A., 2002. Telomere length in the newborn. Pediatr. Res. 52 (3), 377–381. http://dx.doi.org/10.1203/01.PDR.0000022341.72856.72. Olovnikov, A.M., 1996. Telomeres, telomerase, and aging: origin of the theory. Exp. Gerontol. 31 (4), 443–448. http://dx.doi.org/10.1016/0531-5565(96)00005-8. Price, L.H., Kao, H.-T.T., Burgers, D.E., Carpenter, L.L., Tyrka, A.R., 2013. Telomeres and early-life stress: an overview. Biol. Psychiatry 73 (1), 15–23. http://dx.doi.org/10. 1016/j.biopsych.2012.06.025. Radloff, L.S., 1977. The CES-D scale: a self-report depression scale for research in the general population. Appl. Psychol. Meas. 1 (3), 385–401. http://dx.doi.org/10.1177/ 014662167700100306. Reichert, S., Criscuolo, F., Verinaud, E., Zahn, S., Massemin, S., 2013. Telomere length correlations among somatic tissues in adult zebra finches. PLoS One 8 (12), 1–7. http://dx.doi.org/10.1371/journal.pone.0081496. Rewak, M., Buka, S., Prescott, J., De Vivo, I., Loucks, E.B., Kawachi, I., Non, A.L., Kubzansky, L.D., 2014. Race-related health disparities and biological aging: does rate of telomere shortening differ across blacks and whites? Biol. Psychol. 99 (1), 92–99. http://dx.doi.org/10.1016/j.biopsycho.2014.03.007. Send, T., Gilles, M., Codd, V., Wolf, I.A.C., Bardtke, S., Streit, F., Strohmaier, J., Frank, J., Schendel, D., Sütterlin, M.W., Denniff, M., Laucht, M., Samani, N.J., Deuschle, M., Rietschel, M., Witt, S.H., 2017. Telomere length in newborns is related to maternal stress during pregnancy. Neuropsychopharmacology. http://dx.doi.org/10.1038/

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.psyneuen.2017.11.008. References Ainsworth, M.D.S., Wittig, B.A., 1969. Attachment and exploratory behavior of 1-yearolds in a strange situation. In: Foss, B.M. (Ed.), Determinants of Infant Behavior IV. Methuen, pp. 111–136. Aviv, A., 2008. The epidemiology of human telomeres: faults and promises. J. Gerontol.: Med. Sci. 63A (9), 979–983. http://dx.doi.org/10.1093/gerona/63.9.979. Azar, R., Paquette, D., Zoccolillo, M., Baltzer, F., Tremblay, R.E., 2007. The association of major depression, conduct disorder, and maternal overcontrol with a failure to show a cortisol buffered response in 4-Month-Old infants of teenage mothers. Biol. Psychiatry 62 (6), 573–579. http://dx.doi.org/10.1016/j.biopsych.2006.11.009. Baer, R.A., Smith, G.T., Lykins, E., Button, D., Krietemeyer, J., Sauer, S., Walsh, E., Duggan, D., Williams, J.M., 2006. Construct validity of the five facet mindfulness questionnaire in meditating and nonmeditating samples. Assessment 15, 329–342. http://dx.doi.org/10.1177/1073191107313003. Belsky, J., Shalev, I., 2016. Contextual adversity, telomere erosion, pubertal development, and health: two models of accelerated aging, or one? Dev. Psychopathol. 28, 1–17. http://dx.doi.org/10.1017/S0954579416000900. Blackburn, E.H., Epel, E.S., 2012. Telomeres and adversity: too toxic to ignore. Nature 490 (7419), 169–171. http://dx.doi.org/10.1038/490169a. Blackburn, E.H., 1991. Structure and function of telomeres. Nature 350 (6319), 569–573. http://dx.doi.org/10.1038/350569a0. Brennan, P.A., Pargas, R., Walker, E.F., Green, P., Jeffrey Newport, D., Stowe, Z., 2008. Maternal depression and infant cortisol: influences of timing, comorbidity and treatment. J. Child Psychol. Psychiatry 49 (10), 1099–1107. http://dx.doi.org/10. 1111/j.1469-7610.2008.01914.x. Cawthon, R.M., 2002. Telomere measurement by quantitative PCR. Nucleic Acids Res. 30 (10), e47. http://dx.doi.org/10.1093/nar/30.10.e47. Choi, J., Fauce, S.R., Effros, R.B., 2008. Reduced telomerase activity in human T lymphocytes exposed to cortisol. Brain Behav. Immun. 22 (4), 600–605. http://dx.doi. org/10.1016/j.bbi.2007.12.004. Clearfield, M.W., Carter-Rodriguez, A., Merali, A.R., Shober, R., 2014. The effects of SES on infant and maternal diurnal salivary cortisol output. Infant Behav. Dev. 37 (3), 298–304. http://dx.doi.org/10.1016/j.infbeh.2014.04.008. Cohen, S., Janicki-Deverts, D., Turner, R.B., Marsland, A.L., Casselbrant, M.L., Li-Korotky, H.S., Epel, E.S., Doyle, W.J., 2013. Childhood socioeconomic status, telomere length, and susceptibility to upper respiratory infection. Brain Behav. Immun. 34, 31–38. http://dx.doi.org/10.1016/j.bbi.2013.06.009. Daniali, L., Benetos, A., Susser, E., Kark, J.D., Labat, C., Kimura, M., Desai, K., Granick, M., Aviv, A., 2013. Telomeres shorten at equivalent rates in somatic tissues of adults. Nat. Commun. 4, 1597. http://dx.doi.org/10.1038/ncomms2602. Davis, E.P., Granger, D.A., 2009. Developmental differences in infant salivary alphaamylase and cortisol responses to stress. Psychoneuroendocrinology 34 (6), 795–804. http://dx.doi.org/10.1016/j.psyneuen.2009.02.001. De Meyer, T., Rietzschel, E.R., De buyzere, M.L., De Bacquer, D., Van Criekinge, W., De Backer, G.G., Gillebert, T.C., Van Oostveldt, P., Bekaert, S., 2007. Paternal age at birth is an important determinant of offspring telomere length. Hum. Mol. Genet. 16 (24), 3097–3102. http://dx.doi.org/10.1093/hmg/ddm271. Drury, S.S., Theall, K., Gleason, M.M., Smyke, a T., De Vivo, I., Wong, J.Y., Fox, N.A., Zeanah, C.H., Nelson, C.A., 2012. Telomere length and early severe social deprivation: linking early adversity and cellular aging. Mol. Psychiatry 17 (7), 719–727. http://dx.doi.org/10.1038/mp.2011.53. Entringer, S., Epel, E.S., Kumsta, R., Lin, J., Hellhammer, D.H., Blackburn, E.H., Wüst, S., Wadhwa, P.D., 2011. Stress exposure in intrauterine life is associated with shorter telomere length in young adulthood. PNAS 108 (33), E513–E518. http://dx.doi.org/ 10.1073/pnas.1107759108. Entringer, S., Epel, E.S., Lin, J., Buss, C., Shahbaba, B., Blackburn, E.H., Simhan, H.N., Wadhwa, P.D., 2013. Maternal psychosocial stress during pregnancy is associated with newborn leukocyte telomere length. Am. J. Obstet. Gynecol. 208. http://dx.doi. org/10.1016/j.ajog.2012.11.033. Goldsmith, H.H., Rothbart, M.K., 1999. The Laboratory Temperament Assessment Battery (Locomotor Version 3.1). University of Wisconsin-Madison, Madison, WI. Gotlib, I.H., LeMoult, J., Colich, N.L., Foland-Ross, L.C., Hallmayer, J., Joormann, J., Lin, J., Wolkowitz, O.M., 2015. Telomere length and cortisol reactivity in children of depressed mothers. Mol. Psychiatry 20 (5), 615–620. http://dx.doi.org/10.1038/mp. 2014.119. Gunnar, M.R., Brodersen, L., Krueger, K., Rigatuso, J., 1996. Dampening of adrenocortical

45

Psychoneuroendocrinology 88 (2018) 38–46

B.W. Nelson et al.

Toda, S., Fogel, A., 1993. Infant response to the still-face situation at 3 and 6 months. Dev. Psychol. 29 (3), 532–538. http://dx.doi.org/10.1037/0012-1649.29.3.532. Tomiyama, A.J., O’Donovan, A., Lin, J., Puterman, E., Lazaro, A., Chan, J., Dhabhar, F.S., Wolkowitz, O., Blackburn, E., Epel, E., 2012. Does cellular aging relate to patterns of allostasis? An examination of basal and stress reactive HPA axis activity and telomere length. Physiol. Behav. 106 (1), 40–45. http://dx.doi.org/10.1016/j.physbeh.2011. 11.016. Uchino, B.N., Cawthon, R.M., Smith, T.W., Light, K.C., McKenzie, J., Carlisle, M., Gunn, H., Birmingham, W., Bowen, K., 2012. Social relationships and health: is feeling positive, negative, or both (ambivalent) about your social ties related to telomeres? Health Psychol. 31 (6), 789–796. http://dx.doi.org/10.1037/a0026836. Vartak, S., Deshpande, A., Barve, S., 2014. Reduction in the telomere length in human TLymphocytes on exposure to cortisol. Curr. Res. Med. Med. Sci. 4 (2), 20–25. http:// dx.doi.org/10.1016/j.bbi.2007.12.004. Wei, T., & Simko, V. (2016). corrplot: Visualization of a Correlation Matrix. Retrieved from https://cran.r-project.org/package=corrplot.

npp.2017.73. Shalev, I., Entringer, S., Wadhwa, P.D., Wolkowitz, O.M., Puterman, E., Lin, J., Epel, E.S., 2013. Stress and telomere biology: a lifespan perspective. Psychoneuroendocrinology 38 (9), 1835–1842. http://dx.doi.org/10.1016/j.psyneuen.2013.03.010. Shalev, I., 2012. Early life stress and telomere length: investigating the connection and possible mechanisms: a critical survey of the evidence base, research methodology and basic biology. Bioessays 34 (11), 943–952. http://dx.doi.org/10.1002/bies. 201200084. Shonkoff, J.P., Boyce, W.T., McEwen, B.S., 2009. Neuroscience. molecular biology, and the childhood roots of health disparities. JAMA 301, 2252–2259. http://dx.doi.org/ 10.1001/jama.2009.754. Shonkoff, J.P., 2012. Leveraging the biology of adversity to address the roots of disparities in health and development. PNAS 109 (Suppl. 2), 17302–17307. http://dx. doi.org/10.1073/pnas.1121259109. Shrout, P.E., Bolger, N., 2002. Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychol. Methods 7 (4), 422. http://dx.doi. org/10.1037//1082-989x.7.4.422.

46