Early life complications, placental genomics, and risk for neurodevelopmental disorders in offspring
Pasquale Di Carlo, Giovanna Punzi, and Gianluca Ursini
Many severe mental illnesses have been recognized as disorders of neurodevelopment (ND), as reported in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). They include autism spectrum disorders, attention deficit hyperactivity disorder (ADHD), intellectual disability, global developmental delay, communication disorders, ND motor disorders, and specific learning disorders (1). Schizophrenia is considered the most severe disorder in psychiatry. Even though epitomized in a separate chapter of the DSM-5, it has been largely accepted and supported as a disorder of the ND (2–4). Both schizophrenia and other disorders of ND present with “developmental deficits that produce impairments of personal, social, academic, or occupational functioning” (1); they reflect developmental processes that manifest early in life, and whose pathogenesis may imply events that concern placenta biology, such as complications during pregnancy, labor/delivery, and early neonatal life (early life complications [ELCs]). In this chapter, we discuss the link between ELCs, ND disorders, and placenta genomics; although our main focus will be schizophrenia, our analysis can be extended to other disorders of ND.
Early life complications
The theory of the “continuum of reproductive casualty” first proposed that the adversities experienced during prenatal and perinatal life have long-lasting consequences persisting during adulthood (5), in that pregnancy and birth complications lead to a gradient of injury extending from fetal and neonatal death through cerebral palsy, epilepsy, mental deficiency, and behavior disorder (6). In this scenario, central nervous system (CNS) development may be considered the cornerstone of the whole human ontogenesis, being highly sensitive to adversities in early life. ND starts during embryogenesis and continues until young adulthood with the completion of myelination (7). The high demand of CNS metabolism and its relevance for morphogenesis make pregnancy a critical window of susceptibility, where diverse internal (maternal and fetal) and/or external factors have the potential to deviate the normal growth of the most complex human organ.
However, prevalence of ELCs in the general population is fairly high, and we do not know what steers the transition from ELCs toward the outcome of health or of disease. About 25%–30% of births involve at least one complication (8,9); 62% of pregnancies already positive for at least one risk factor may be at high risk for further unexpected complications; among low-risk pregnancies (about 4 million surveyed births), 29% have an unexpected complication that would require nonroutine obstetric or neonatal care (10). Moreover, a recent investigation of U.S. natality data from 10,458,616 pregnancies found that 46% of newborns have at least one ELC (10). However, ELCs refer to a vast number of potentially adverse events, lacking both anatomo-pathological and clinical unity, and each having variable outcomes. As a consequence, studies linking placenta and pregnancy pathology to brain disorders benefit from epidemiological investigations aiming at a high specific assessment of at-risk pregnancies.
In studying why some individuals exposed to ELCs have a normal outcome, while others develop a brain disorder, a key aspect to consider is how serious the ELCs were, since low or high severity weighs on risk in a different manner, as discriminated by dedicated tools. The McNeil–Sjöström scale (11) provides a systematic evaluation and weighting of several specific somatic conditions and events during pregnancy, labor-delivery, and neonatal periods. Compared to other scales (12,13), the McNeil–Sjöström scale functions optimally in representing the amount, severity, and timing of hundreds of ELCs. Each item has been categorized according to a six-point scale (level 1 = not harmful, level 6 = great harm or deviation), whose severity level reflects the potential impact on somatic damage, especially on the CNS, in the offspring (11). Nonetheless, a more careful definition of exposures, such as through prenatal dosage of maternal antibodies, and the adoption of quantitative measures as predictors and outcomes, including birth weight or head circumference, are likely to show larger and more consistent effects (14).
Early life complications and neurodevelopmental disorders
Many studies associated a history of ELCs with a higher risk of disorders of brain development.
Several ELCs, including infections, malnutrition, and intrauterine growth retardation (IUGR) are risk factors linked to schizophrenia, autism, ADHD, and cerebral palsy in the offspring (14–25). Placenta vascular pathologies, such as chorionic vessel thrombi, villous edemas, and necrosis, are prevalent in cerebral palsy (26), while the ELCs associated with autism seem to concern more immune dysregulation (27–29) and hypoxia (30). Of all ELCs, preterm birth affects 5%–18% of pregnancies and is associated with many CNS alterations, such as cerebral palsy, intellectual disabilities, and vision and hearing impairments, in addition to schizophrenia, autism, and ADHD (31–34).
Patients with schizophrenia are more likely to have a history of ELCs, compared with patients with other psychiatric disorders (35,36), and with the general population (14). In particular, several ELCs have been robustly associated with risk for this disease (14), e.g., rhesus variables (comprising rhesus incompatibility, rhesus-negative mother, rhesus antibodies), diabetes in pregnancy, bleeding in pregnancy, preeclampsia, uterine atony, emergency cesarean section, congenital malformations, asphyxia, and birth weight <2500 g. Such events appear to group into three main categories: complications of pregnancy (bleeding, preeclampsia, diabetes, and rhesus incompatibility); abnormal fetal growth and development (low birth weight, congenital malformations, and small head circumference); and complications of delivery (asphyxia, uterine atony, and emergency cesarean section). Meta-analyses on the subject have found that such ELCs increase the risk for schizophrenia by 1.5- to 2-fold (14), with the possible mechanism of risk being related to their effect on altered blood flow resulting in fetal chronic hypoxia or malnutrition (37). Individuals with three or more hypoxia-related ELCs were more than five times as likely to develop schizophrenia than individuals with no hypoxia-related ELCs (38). Some studies suggest that a history of ELCs could contribute to stratifying patients with schizophrenia: specifically, a history of ELCs seems to characterize patients with early onset schizophrenia (39), and it may predict treatment response (40,41).
Despite these studies, and many others, stressing the role of perinatal development in affecting risk for schizophrenia (14,19–21,35,36), as well as autism (22,23,27–30), the mechanisms by which ELCs affect risk for later life diseases remains elusive. At the individual level, the potential outcome of ELCs can be heterogeneous: many individuals with a history of ELCs are healthy adults, while others develop various disorders, including autism (22,23,27–30), ADHD (31–34), schizophrenia (14,19–21,35,36), diabetes (42), obesity (43), hypertension, various cardiovascular disorders (44,45), and cancer (46); moreover, an alternative possible outcome of ELCs is stillbirth (47,48). The investigation of the factors mediating the relationship between ELCs and developmental outcomes is still needed to define etiopathogenesis, and to design specific strategies of prevention and treatment.
Interaction between genetic risk factors and EARLY LIFE COMPLICATIONS
Studies on ELCs in high-risk individuals (that is, offspring of parents affected with schizophrenia) suggest that genetic background can interact with ELCs in affecting risk for the disorder (as reviewed in ), which has been supported by preliminary evidence of a relationship between ELCs, hypoxia-related genes, and risk for schizophrenia (14,49,50).
In the post–Human Genome Project era (51), genetic vulnerability for schizophrenia (and other brain disorders) can be defined, thanks to genome-wide association studies (GWAS), as the cumulative effect of thousands of genetic common variants that are more frequent in patients compared to healthy individuals (52). This opens to the possibility to analyze whether the relationship between ELCs and ND outcomes is modulated by genetic risk. Genetic risk can be summarized by the polygenic risk score (PRS), which is calculated for each individual as the sum of risk alleles weighted for the odds ratio of the association with the disorder (53,54). By this approach, it has been shown that genetic risk for schizophrenia, as measured with PRSs calculated from the most GWAS-significant loci, interacts with serious ELCs in affecting risk for schizophrenia, so that the liability of schizophrenia explained by genetic risk is fivefold higher in individuals with a history of ELCs compared with individuals without ELCs records (55). Consistently, the genes mapping to the schizophrenia risk loci, and interacting with ELCs, are highly expressed in placenta and differentially expressed in placentae from complicated pregnancies compared to normal ones. Moreover, these genes are upregulated in placentae from male compared with female offspring. Such findings suggest a link between genetic risk for schizophrenia and placenta pathophysiology, together with a sex-biased role for the placenta in expressing genetic risk for schizophrenia (55).
ELCs-exposed placenta, as an early context influencing cumulative genetic vulnerability for schizophrenia, represents a compelling example of Gene × Environment (G × E) interaction in the etiology of a complex disease (55). Interestingly, effects of ELCs and defective placentation support a sex bias toward a greater male vulnerability compared to female, with long-lasting influence on multiple developing systems, known as the “selective male affliction” (56). Moreover, ELCs are more likely to occur in mothers carrying male fetuses, which configures a sort of “double jeopardy” for male conceptuses, suggesting sex as a further moderating factor for outcomes (56,57). Specifically, preterm birth is more frequent in males compared with females, and a poorer neurocognitive outcome is more likely in preterm males compared to preterm females (58–61). Such sex bias emerges also from epidemiological studies on schizophrenia, which shows greater incidence (62), earlier age of onset (63,64), and worse prognosis in male patients (65).
The understanding of how ELCs, genetic risk, and sex-related factors may converge on the pathophysiology of ND disorders might benefit from high-throughput biotechnologies that enable large-scale genome-wide gene expression analysis. In particular, transcriptome-wide studies of the placenta may allow the identification of gene expression signatures mediating the relationship between ELCs, genetic risk, sex, and the disorders.
The placenta is the core regulator of fetal and maternal interactions during pregnancy. The placenta is a membranous, temporary organ, mediating all the biological functions needed by the developing embryo and fetus. The maternal-fetal interface resides at the decidual-trophoblast junction of the placenta and across the syncytiotrophoblast cells that form the boundary between maternal and fetal blood spaces. Rather than a static interhaemal barrier, it actively regulates transport of nutrients, gas exchange, waste elimination, thermoregulation, endocrine functions, and immune tolerance, assuring a correct development. Consequently, abnormalities in the structure and function of this complex organ are intertwined with the fate of pregnancy itself, and with outcomes later in life. The placenta is the place where a number of pathological insults, the ELCs, may intervene, potentially compromising the physiological development of the embryo and the fetus.
An important role of the placenta in the etiology of ND disorders, like schizophrenia, is consistent with heritability studies. Specifically, the 45%–60% concordance (66) of the disorder in monozygotic twins leaves plenty of room for environmental mechanisms acting early in life, since monozygotic twins share the same intrauterine environment. Notably, just one-third of monozygotic twins share also placenta and chorion (monochorionic), while the remaining two-thirds have separate placentae and chorions (dichorionic). Monochorionic twins show a much higher concordance rate for schizophrenia (60%) compared to dichorionic twins (10%) (67), emphasizing the role of the shared placenta environment in disease diathesis. Consistently, dizygotic twins, who also share the same intrauterine environment, have a concordance rate for schizophrenia twice that of siblings, despite both groups having 50% of segregating DNA variation in common (66).
The recent availability of high-throughput biotechnologies has allowed the sequencing, decoding, and quantification of every human molecule (nucleic acids, proteins, and metabolites), and the deep sequencing of DNA at the single base resolution (methylation, acetylation). This has led to the collection of large data sets (thousands of individuals) scanned for the over twenty thousand gene expression values, the hundreds of thousands methylation sites, up to the millions of single nucleotide polymorphisms (SNPs). Such a massive gathering of data plunges biology into the “big data” era of the “omics.” The power of the omics lies in the opportunity to investigate ab ovo the system-level organization of the living organisms, surpassing reductionism and allowing hypothesis testing on the emergent properties of complex systems. Studying placental omics has the potential to connect properties of this complex organ with developmental trajectories of risk for ND disorders. Therefore, the predictive value of placental omics may allow novel strategies of prevention.
The placental transcriptome may offer insight into the link between ELCs, genetic risk factors, and ND disorders, since both ELCs and genetic risk factors may lead to change in—potentially sex-specific—gene expression in placenta, which may in turn underlie pathophysiological processes relevant for brain development and disorders.
A preliminary expression quantitative trait loci (eQTL) analysis in placenta has detected an enrichment of SNPs associated with schizophrenia among the genetic variants that predict gene expression in placenta (68). Consistently, it has been found in multiple independent data sets that the genes in the schizophrenia risk loci tend to be differentially expressed, and specifically upregulated, in placentae from complicated pregnancies compared with controls, and in male compared with female placentae (55). In the same study, placental transcriptome data have been employed to calculate a measure of placenta genomic risk for schizophrenia (PlacPRS) based on schizophrenia risk loci containing genes highly and differentially expressed in placentae, and a measure of genomic risk based on the remaining loci (NonPlacPRS). Interestingly, the PlacPRSs specifically interact with ELCs in affecting risk for schizophrenia, while the NonPlacPRSs do not. These results suggest that the study of placental transcriptome may help to identify a specific component of genomic risk that is linked to etiopathogenetic mechanisms acting in placenta. Pathway analyses further revealed that the PlacPRS genes are involved in cellular stress response and are coexpressed with inflammatory/immune response genes, while the NonPlacPRS genes implicate orthogonal and more traditionally brain-related biological processes (e.g., synaptic function) (55).
This work suggests that genetic risk for ND disorders, ELCs, and sex may converge in affecting expression of specific genes in placenta; the altered expression of such genes may alter developmental trajectories, increasing the risk for the disorders (69