Pilot Study of Absolute Telomere Lengths in Preterm Infants

Abstract

Background:

Annually, approximately 15 million babies are born preterm (< 37 weeks gestational age) globally. In the neonatal intensive care unit (NICU) environment, infants are exposed to repeated stressful or painful procedures as part of routine lifesaving care. These procedures have been associated with epigenetic alterations that may lead to an increased risk of neurodevelopmental disorders. Telomere length has been negatively associated with adverse life experiences in studies of adults.

Objectives:

This pilot study aimed to describe telomere length in a sample of preterm infants at NICU discharge and examine any associations with pain, feeding method, and neurodevelopment.

Methods:

This descriptive pilot study sample includes baseline absolute telomere length (aTL) of 36 preterm infants immediately prior to discharge. Quantitative polymerase chain reaction (qPCR) was used to determine absolute telomere length. Infant demographics, pain/stress, type of feeding, antibiotic use, neurodevelopment, and buccal swab data were collected. Descriptive data analysis was used to describe the telomere length using graphs.

Results:

Among our preterm infant samples, the mean absolute telomere length was far greater than the average adult telomere length. While no significant associations were found between absolute telomere length and pain, feeding method, and neurodevelopment, a trend between sex was noted where male telomere lengths were shorter than females as they aged.

Discussion:

This is one of few studies to evaluate preterm infant telomere length. While other researchers have used relative telomere length, we used the more accurate absolute telomere length. We found nonsignificant shorter telomere lengths among males. Additional large-scale, longitudinal studies are needed to better identify the predictors of telomere length at the time of discharge from NICU.

Each year, an estimated 15 million infants are born preterm (< 37 weeks gestational age; Richardson et al., 2020) globally (World Health Organization [WHO], 2018). While technological advancements in the neonatal intensive care unit (NICU) have led to increased survival rates, the gains in survivorship have been coupled with repeated exposure to stress as part of routine lifesaving care. Exposure to repeated or prolonged stress is associated with telomere erosion and disease development as one ages (Epel & Prather, 2018).

Telomeres are made up of repeating DNA-protein complexes that serve as end caps of chromosomes; they play a critical role in cell reproduction and protect chromosomes from adhering to one another during mitosis. The length of these repeating complexes dictates the epigenetic state of telomeres (Blackburn et al., 2006). Telomere length—a marker of cellular aging and closely related to psychological stress—has emerged as a biological correlate of many medical and psychological disorders (Lindqvist et al., 2015). As telomere length erodes during the aging process, the expression of nearby genes increases. Thus, individuals who have exceptionally long telomeres, or telomeres with slower attrition rates due to genetic or environmental factors, may never show symptoms of certain diseases (Stadler et al., 2013). There seems to be a complex interaction between stress, telomere length, and psychopathology (Epel & Prather, 2018). To date, few investigators have studied telomere length in the preterm infant population (Hadchouel et al., 2015; Provenzi et al., 2018; Vasu et al., 2017), and to our knowledge, this is the first study examining preterm infant telomere length with a sample size greater than 10 in the United States.

This pilot study aimed to describe telomere length in a sample of NICU preterm infants at discharge and examine associations with pain, feeding method, and neurodevelopment.

Methods

This is a baseline pilot study using an exploratory, descriptive design. Participants included 36 infants, 50% female, and recruited at one of two Connecticut Children’s Medical Center NICU sites. Inclusion criteria: (a) infants born between 28–36 weeks gestational age; and (b) exposed to repeated painful/stressful procedures or treatment while in the NICU as assessed daily using the validated and reliable Neonatal Infant Stressor Scale (NISS; Newnham et al., 2009). Exclusion criteria included: (a) major brain lesions (intraventricular hemorrhage > Grade 2, periventricular leukomalacia); (b) neurosensorial deficits (retinopathy of prematurity > Stage 2); and (c) genetic syndromes and/or major malformations. Each study participant was given a subject identification number to ensure confidentiality.

This study was approved by the Connecticut Children’s Medical Center Institutional Review Board. Written parental consent was obtained within the first 2 weeks of birth by trained research coordinators who informed parents of the research aims and procedures and provided a packet containing said information.

Infant variables included demographics, length of stay in the NICU, and NISS (Newnham et al., 2009), which assess acutely painful procedures and/or treatment daily and chronic painful procedures and/or treatment over time. Neurodevelopment while in the hospital was measured using the NICU Network Neurobehavioral Scale (NNNS; Lester & Tronick, 2004). The NNNS measures infant neurobehavioral outcomes when the infant reaches 36–37 weeks postmenstrual age (PMA) and provides an assessment of neurological integrity and behavioral function of a normal or at-risk preterm infant (Sullivan et al., 2012). The three main sections of the NNNS are neurological (tone and reflexes), behavior (state, sensory, and interactive processes), and stress/abstinence. The telomere length distribution profile represents the range of telomere lengths in the cell population, typically described as relative telomere lengths, while absolute measurements using quantitative polymerase chain reaction (qPCR) and a standardization curve provide precise information regarding telomere length.

Historically, recruitment has been challenging for genetic studies, often because they require the collection of blood samples. As -omics technology has improved, blood samples are no longer needed to conduct genetic assays. A recent study validated the use of minimally invasive buccal swab samples for genetic assays, making recruitment and sample collection both more appealing to parents and the nurses who will collect the specimens. In addition, buccal swabs display greater stability over time. Since epithelial cells are ectodermal in origin, they serve as a better proxy for the brain in place of invasive brain sampling (Blackburn et al., 2006; Lindqvist et al., 2015). We used existing biomarker samples that were already collected and housed in a repository at UConn Health.

The infant buccal epithelial tissues were collected prior to discharge from the NICU using a soft cotton swab and standard protocols (e.g., Qiagen Buccal Cell Kit). The swab was inserted into a 15 ml or 50 ml conical tube labeled with the subject’s identification number and placed into the NICU freezer at −80 °C until DNA could be extracted. Total genomic DNA (gDNA) and RNA were extracted from buccal swabs using the Gentra Puregene Buccal Cell Kit (Qiagen, Germantown, MD) and RNeasy Kit (Qiagen), respectively. Once the DNA was extracted, we followed the methodology presented by O’Callaghan (O’Callaghan & Fenech, 2011). The use of absolute telomere length (aTL) has until now not been used to quantify telomeres in preterm infants. When the same qPCR parameters and standard curve dilutions optimized for adult telomere length quantitation were applied to our preterm samples as outlined by O’Callaghan and Fenech (2011), no meaningful data were obtained. Given the drastic difference of preterm infant telomere content compared to adults, the samples in our study required a 125fold reduction in starting input compared to that of adult samples in order for unknown samples to fall on our standard curve. This was determined empirically by creating serial dilutions of the samples and running them on the real-time system to determine which concentration would fall on the standard curve. Initial qPCR reactions were carried out following the previously published data for measuring adult telomere length (10ng input). Standard curve starting points and subsequent serial dilutions for the telomere standard (synthesized oligomer repeat standard) and single-copy gene standard (364B gene) were optimized, and when assayed, using only 0.08ng/μl of total input into the qPCR reaction.

All statistical analyses were conducted using R software (R Core Team, 2013), and graphs were produced by the ggplot2 package (Wickham, 2009). The absolute telomere lengths were compared between different categorical variables by the Mann–Whitney U test, including babies’ gender, race, and ethnicity. Spearman correlation was calculated to associate the absolute telomere length with other continuous variables, including corrected gestational age, acute/chronic painful procedures the babies received in the first 4 weeks of their NICU stay, and their feeding proportion of mother’s breast milk. The multiple linear regression model was then used to fit the absolute telomere length to all predictors. The absolute telomere lengths were log transformed before fitting the model to satisfy the normality assumption, and the missing predictors were imputed by multiple imputation (MI) algorithm.

Results

Demographic information for infants and univariate analyses are shown in Table 1. The longitudinal features of the NISS acute and chronic pain scores are illustrated by a spaghetti plot shown in Figure 1. The aTL at the time of buccal collection, shown in Figure 2, revealed a nonsignificant trend of aTL associated with corrected gestational age (weeks) across gender, with males having shorter telomeres. The aTL values were inversely correlated with corrected age at the time of buccal swab collection over time. Among the 36 samples, 33 samples had reliable aTL measures. All reactions were run in triplicate on the same plate, and negative control reactions were also used for each primer set (also in triplicate). The triplicate values had a standard deviation of more than 50% for the remaining three samples, which is greater than the acceptable variability within replicates. Those samples were excluded from analysis. The mean aTL was38,628 kilobase (kb) pair per diploid genome (range 14,895–114,388 kb).

Table 1.

Demographics and Univariate Analysis

	Mean	SD	Min	Max	Spearman correlation with aTL
Gestational Age (weeks)	27.86	3.128	23.29	37.43	.11
Birth weight (g)	988.58	330.06	500	1670	−.076
Birth length (cm)	35.31	4.00	27	44	−.015
Birth head circumference (cm)	24.01	4.19	5	30	−.019
APGAR 1mina	5	2.52	1	8	.185
APGAR 5min	7.31	2.15	2	9	.085
SNAPPE-IIb	26.51	18.88	0	77	.104
	Category	N	Percentage (%)	p-values by Mann-Whitney test
Sex	Male	18	50	0.873
	Female	18	50
Race	White	24	66.67	0.830
	Non-White	12	33.33
Ethnicity	Hispanic	9	25	0.156
	Non-Hispanic	24	66.67
	Not available	3	8.33

Note.

^aNamed after Virginia Apgar, the Apgar is a quick test performed on a baby at 1 min and 5 min after birth. The 1-min score determines how well the baby tolerated the birthing process. The 5-min score tells the health care provider how well the baby is doing outside the mother’s womb. In rare cases, the test will be done 10 min after birth.

^bScore for Neonatal Acute Physiology with Perinatal Extension is an illness severity score and predictor of mortality and morbidity in intensive care units.

Figure 1.

NISS acute and chronic pain scores over time

Figure 2.

Absolute Telomere Length at Time of Buccal Collection (Using Corrected Age)

The results for the multiple linear regression are shown in Table 2. The predictors include demographic characteristics and clinical characteristics depending on their first 4 weeks stay in NICU (e.g., acute and chronic painful procedures and different types of feedings). However, none of the associations of those characteristics with telomere length reached a 5% significant level.

Table 2.

Multiple Linear Regression Results for Modeling the Log-Transformed Absolute Telomere Length

	Scaled B-coefficient	95% CI	p-value
Intercept	7.826	(2.622, 13.029)	0.006
Gender, Male [ref. Female]	1.418	(−4.053, 6.888)	0.592
Corrected gestational age (weeks) [ref. Female]	0.006	(−0.017, 0.029)	0.572
Corrected gestational age (weeks), Male [ref. Female]	−0.004	(−0.024, 0.016)	0.670
Race, White [ref. Non-white]	0.119	(−0.379, 0.616)	0.623
Ethnicity, Hispanic [ref. Non-Hispanic]	−0.409	(−0.933, 0.115)	0.119
APGAR 1min score at birth	0.063	(−0.030, 0.157)	0.174
SNAPPE II score at birth	0.000	(−0.016, 0.017)	0.955
Acute painful procedures	0.000	(−0.016, 0.016)	0.968
Chronic painful procedures	−0.002	(−0.064, 0.061)	0.958
Mother’s breast milk	0.455	(−0.397, 1.307)	0.277
Body weight at buccal swab	0.000	(−0.001, 0.001)	0.497

Discussion

A focus on adequate stress management for preterm infants during their NICU hospitalizations is an urgent clinical and research priority. In this pilot study, we examined telomere length as a potential indicator of stress in this population and found that male telomere lengths eroded faster than females as they aged. Telomeres are noncoding, repeating protein sequences found at the ends of chromosomes (Moyzis et al., 1988). They serve as a protective cap that preserves the genetic material during cell division (O’Sullivan & Karlseder, 2010). Research has indicated that telomeres shorten progressively with age (Lindsey et al., 1991). Chromosomes with critically short telomeres are identified as impaired DNA, leading to cell senescence, increasing the risk of diseases related to decreased cell proliferation and tissue degradation (Aviv & Shay, 2018; di Fagagna et al., 2003).

Our results supported the work of Provenzi et al. (2018) as well as that of Vasu et al. (2017). In preterm infants, telomere length is greater than in term infants (Provenzi et al., 2018; Vasu et al., 2017). In the Provenzi et al. (2018) sample, a significant difference emerged at birth between very preterm infants and full-term infants, confirming that telomere erosion might be at least partially related to physiological fetal development, which is then disrupted by preterm birth.

However, numerous factors might influence telomere length in preterm infants, including heritable genes, the NICU environment (loud, disturbed sleep, impaired parental bonding), the presence of infection or excess inflammatory biomarkers in the placenta (Eisenberg et al., 2017), and birth weight (Friedrich et al., 2001). The first study to focus on full-term newborn sex in association with telomere length showed that at birth, male telomeres were shorter than those of females, although the difference was not significant. Thus TL differences may arise from different rates of attrition in extrauterine life (Okuda et al., 2002).

Limitation

One limitation of this baseline telomere study is the small sample size and wide range of gestational ages, which precludes us from controlling confounding variables. However, these pilot study findings serve as a baseline for gaining a greater understanding of the underlying molecular and biological influences of stress and sex on telomere length.

Conclusion

Telomere length is an essential indicator of biological and psychological stress, with stress eroding telomeres at a more rapid rate and thus potentially leading to early cell senescence and psychological diseases. This study indicates an association between sex and telomere erosion over time. Additional large-scale, longitudinal studies are needed to identify the predictors of telomere length over time. These future studies have the potential to lead to interventions that might mitigate telomere erosion and thus improve both quality and length of life.