Mothers’ Global Psychological Health and Sex-specific Expression in Newborns

Stefanie R. Pilkay1*, Terri Combs-Orme2, Frances Tylavsky3, Nicole Bush4, Alicia K. Smith5 1Falk College, School of Social Work, Syracuse University, Syracuse, New York 2College of Social Work, University of Tennessee, Knoxville, Tennessee 3Department of Preventive Medicine, University of Tennesse Health and Science Center, e, Memphis, Tennessee 4Department of Psychiatry, University of California San Francisco School of Medicine, San Francisco, California 5Department of Gynecology and Obstetrics, Department of Psychiatry, Emory University School of Medicine, Atlanta, Georgia


Introduction
Prenatal development is sensitive to maternal stress which can have lasting effects on neonate health 1,2 . Furthermore, males and females have shown differences in prenatal stress programming that could represent different health and developmental risks according to neonate sex 3 . Maternal psychological health is one maternal stress factor linked to prenatal development 4 in a sex-specific manner 5 that may represent sex-specific biological embedding. Maternal psychological health has been measured in different ways showing varying associations with neonate health. For example, mothers with greater affect intensity showed greater fetal motor activity during pregnancy compared to mothers with more stable affect 6 . Moreover, depression or anxiety during pregnancy associated with more difficult neonate temperaments 7 . The links between maternal psychological health and neonate health can best be explained by the process of experience influencing developmental processes, termed "biological embedding" 8 .
burden of psychological symptoms can affect the fetus' developing brain. Previous findings showed mother and newborn hypothalamus-pituitary-adrenal (HPA) axis functioning was linked to cortisol activity and suggested the HPA axis had been sensitized to stress and stimulation in neonates 10 through the process of fetal programming 11 . Chronic stress activation can influence a sensitized HPA axis which can affect the immune system and metabolism 12 . These physiological adaptations increase a child's vulnerability to environmental insults during neurodevelopment. Collectively the research suggests that women who suffer psychological health symptoms are at increased risk to have children with a greater vulnerability to stress.
Emerging evidence suggests that the mothers' psychological health during pregnancy, can impose fetal stress that influences gene expression and fetal development [13][14][15] . Moreover, there are developmental differences between male and female embryos resulting in varying sensitivities to the prenatal environment. For example, male fetuses undergo an epigenetic process during pregnancy to masculinize the brain 16 . The potential for sex-specific variation in sensitivity to the prenatal environment underscores the need to examine sex-specific effects within the context of maternal psychological health. Sex-specific differences in gene expression may help explain sex-specific risk and resilience within childhood psychopathology 17 as well as affective disorders found in adulthood, i.e. anxiety, bipolar disorder, depression, and post-traumatic stress disorder 18 .
This study aimed to determine if: 1) maternal psychological health during pregnancy will associate with gene expression in newborns, and 2) the association between maternal psychological health and newborn gene expression is sex-specific.

Full Cohort Associations
PsyH did not associate with newborn gene expression in the full newborn cohort while controlling for sex, race and cell composition. However, gene expression did associate with newborn sex for 17 genes, although all were located on the sex chromosomes. Plots of the effect sizes and t-statistics for the gene expression probes according to newborn sex suggests male and female newborns responded differently to PsyH (see Figure 3). Therefore, we conducted a sex-stratified analysis to investigate possible gene expression associations with PsyH unique to male or female newborns.

Sex-stratified Cohort Associations
Males exhibited significant gene expression of 157 genes related to increases in PsyH scores during pregnancy (Supplemental Table 1). The 157 genes showed positive (38%) and negative (62%) associations, as modeled in a volcano plot (Figure 4), suggesting a complex interaction with the prenatal environment. These genes were enriched for a variety of metabolic functions (Table 1). For example, expression of the eukaryotic translation initiation factor 4E binding protein 2 (EIF4EBP2), a gene represented among multiple enriched biological processes is lower in males whose mothers have higher levels of PsyH symptoms a b ( Figure 5). Female newborns showed no gene expression associations with PsyH after controlling for covariates and adjusting for multiple comparisons (FDR > .05).

Discussion
Our findings suggest that maternal psychological health can influence the prenatal environment with gene expression associations unique to male newborns. These results suggest males may be more sensitive to maternal psychological health during pregnancy compared to females. Male newborns showed differential expression of 157 genes associated with greater detrimental maternal psychological health, and these genes were enriched for fundamental processes of metabolism. Moreover, a PubMed search of the 157 genes identified EIF4EBP2 as a gene of particular interest. EIF4EBP2 is involved in inhibiting translation initiation, which can affect the amount of protein produced.
Interestingly, EIF4EBP2 has been implicated in the mechanisms of Autism Spectrum Disorders in rodent models with increased ASD behaviors in mice with the gene knocked out 19 . Moreover, newborn expression of EIF4EBP2 has been shown to associate with maternal inflammatory responses during pregnancy 20 that have been linked to maternal psychological health 21 . Although investigations are ongoing, EIF4EBP2 clearly plays an important role in health and the transmission of health from mother to child. The unique association for male newborns in relation to maternal psychological health, therefore, could be an indication of a sex-specific neonatal developmental response to the prenatal environment. At a minimum, our findings support previous research proposing mothers' psychological health is associated with gene expression in male newborns that could predispose risk for or resilience to metabolic disorders 22,23 , enduring cognitive deficits 14,15 , and increased stress sensitivity 5 .
Maternal psychological health functioning during pregnancy should be investigated further. It is possible that some psychological health symptoms exert greater influence than others, or associate with male and female newborns differently. Future research would benefit from the inclusion of different psychological symptoms such as depression, anxiety, somatization, or combinations thereof to provide an examination of differences among symptom types. Furthermore, psychological health can influence other things such as nutrition, income, and social support which could each affect the prenatal environment.
Continuing exploration into gene expression patterns associated with prenatal variables has the potential to inform professionals seeking to foster resilience in vulnerable populations through innovative approaches to reduce risk. For example, prenatal screenings could include assessment for variables that increase the risk for adverse behavioural outcomes in offspring. Therefore, identified environmental interventions for the pregnant mother such as diet and stress reduction could be applied to foster resilience in the fetus, and later the newborn. There are many avenues to promote good psychological health through stress reduction techniques like mindfulness 24 , breathing exercises 25 , physical exercise 26 , and social connectedness 27 to name a few. Moreover, behaviors like healthy sleep patterns 28 and playing a musical instrument 29 have been shown to promote positive emotions and emotion regulation that promote good maternal psychological health.
These study findings, in conjunction with previous findings 5, 14,15 , suggest prenatal exposure to maternal psychological illness is one link for intergenerational patterns of risk, and risks may be sex-specific. Future research may deepen our understanding of mechanisms involved in these intergenerational patterns, and thereby provide new opportunities to intervene for improved quality of life. For example, this avenue of research could bolster arguments for the ongoing need for universal prenatal care, proper affordable nutrition, and other services to support mothers and families.  There are some limitations to this study. The sample is from one urban population in west Tennessee and is not generalizable to other populations. However, our replication of an association between maternal psychological health and male newborn gene expression provides additional context with previous findings 5, 14,15 . The sample size is modest for transcriptome-wide investigations, but the transcriptome-wide analysis is an excellent tool to discover unknown or unsuspected relationships. However, analysis of the whole transcriptome significantly increases the number of conducted tests. To account for those tests, and the decreased probability of committing type I errors, we controlled for multiple comparisons with the False Discovery Rate and calculated bootstrap 95% confidence intervals to show the accuracy of estimated relationships. Lastly, we investigated these relationships in umbilical cord blood, and there may be relationships in other tissues (i.e., placenta).
The findings from our study, in conjunction with previous research 30,31 , suggest that social problems such as stress and psychological illness have potential to exert influence on human development across generations. Therefore, the design of interventions will benefit from biopsychosocial research, like gene expression studies, that will help illuminate pathways to risk and resilience. Knowledge like this could one day be used in screening measures to aid intervention and prevention strategies and in the design and targeting of services to those who are most in need.

Sample and Procedures
The data for this study is from the Conditions Affecting Neurocognitive Development and Learning in Early Childhood investigation (CANDLE), and the University of Tennessee Health Science Center Institutional Review Board approved all measures and procedures. This study was carried out in accordance with the Belmont Report ethical principles and guidelines for human subjects research. The sample (111 mother/infant pairs) consists of healthy mothers aged 16-40 years solicited in prenatal settings in Shelby County, Tennessee. Announcements and brochures containing information about the study were provided to all local gynecology clinics. Interested women contacted study personnel by telephone and were then screened for eligibility. Women meeting eligibility criteria were asked to visit one of two research clinics utilized for the study. Forty percent of interested women met eligibility criteria and were invited to participate. All participants signed informed consent documents, and participants under the age of eighteen years provided a parent signed informed consent. The participants provided umbilical cord blood samples immediately after birth for biological measures including gene expression. Umbilical cord blood has been used to measure newborn gene expression in multiple studies [32][33][34][35] .

Measures
Mothers' psychological health (PsyH) status variable was created using the global severity index (GSI) summary score from the Brief Symptom Inventory (BSI) (Derogatis & Melisaratos, 1983). The Brief Symptom Inventory is a shortened version of the SCL-90 psychological health assessment and has been found to be sensitive to psychopathology and psychological distress 36 . The global severity index score has been investigated and found to be a more accurate assessment of overall psychological health functioning than the positive symptom total score in the BSI measure 37 .

Gene Expression
Gene expression is assessed by measuring the RNA transcript levels 38 . The Illumina Human WG-6 expression array was used to measure RNA transcription. Samples with less than 10% of the gene probes detected were eliminated, as well as probes with less than 10% of the samples detected within each dataset. No sample was lost based on these QC criteria. We performed quantile normalization, scaled the data, and performed a log2 transformation. Prior to analysis, updated annotation files for the Illumina Human WG-6 array were consulted for accurate gene expression measurement information 39 . A total of 10,821 expressed genes passed QC in these umbilical cord blood samples.

Statistical Analyses
Descriptive statistics were calculated to determine sample characteristics. We examined the association between PsyH and newborn transcriptome-wide gene expression as measured with the Illumina HumanWG-6 BeadChip. We performed multiple regression to conduct the newborn gene expression analyses and controlled for child sex, child race, and cell composition. Cell composition was estimated for each sample as previously described 40 . We statistically analyzed the cohort first and then conducted sex-stratified analyses controlling for child race and cell composition. Gene ontology analysis was performed on genes identified as significant using GATHER 41 . Bootstrap confidence intervals were generated to address potential non-normality of the data. As is standard in genetics research, we implemented the False Discovery Rate control for multiple comparisons in all gene expression analyses 42 .