Understanding the genetics of peripartum depression: Research

1 IntroductionPregnancy-associated depression (hereafter referred to as peripartum depression [PD]) is a potentially debilitating mood disorder and one of the most common complications of pregnancy and childbirth. PD partially shares symptoms with the nonpathologic postpartum “baby blues” but is a distinct disorder characterized by greater impairment and longevity. PD episodes impair daily functioning and can present with intense feelings of sadness, guilt, anxiety, anhedonia, disturbed sleep, significant decreases in energy and concentration, and thoughts of death, self-harm, suicide, and infanticide (Wisner et al., 2013). PD affects approximately one in seven women in the United States (US), (Wisner et al., 2013) but global prevalence estimates vary substantially depending on how PD is defined and measured and on population demographics, including socio-economic environment and cultural perceptions of mental health and motherhood (Halbreich and Karkun, 2006). Adverse outcomes associated with PD include increased risks for maternal mortality, poor maternal-offspring bonding, paternal depression, poor maternal self-care (e.g., noncompliance with medical guidance), poor infant feeding outcomes, and increased likelihood of behavioral issues in early childhood (Slomian et al., 2019). Early intervention and treatment can reduce the likelihood of negative outcomes, (Batt et al., 2020) but as many as 50% of PD cases may go unrecognized or undertreated (Payne and Maguire, 2019).PD genetic studies have the potential to uncover biological mechanisms involved in pathogenesis and advance precision psychiatry efforts by identifying individuals at high risk and those likely to respond to treatment (pharmacogenetics). However, several challenges need to be overcome before robust, generalizable findings can be produced. Researchers are optimistic that the creation and use of large-scale public biobanks and repositories will accelerate scientific discovery and offer solutions to longstanding issues in biomedical research like non-representative cohorts and low statistical power (Guintivano et al., 2019; Kimmel et al., 2020). While the shift to biobanks will alleviate some issues, it is not a panacea. Significant challenges will persist for PD research unless scientists capitalize on opportunities to promote rigorous, inclusive studies through improved study design and community engagement. This manuscript identifies challenges in PD genetics research, discusses how each may impact future meta-analytic efforts, and concludes with recommendations to advance the understanding of PD etiology.2 The genetic etiology of peripartum depressionTwin and family studies have consistently supported the notion that major depression (MD), and by extension, PD, are complex disorders influenced by both genetic and environmental factors (Sullivan et al., 2000). The largest family study of PD to date reported that additive genetic factors account for a higher proportion of variance in PD liability compared to that of non-perinatal depression (54% and 32%, respectively) (Viktorin et al., 2016). In general, sample sizes in PD twin and family studies have been modest, likely because a key requirement for a PD twin study is that both twins in a pair have been pregnant at least once reducing available samples when studying PD compared to MD.Most of the molecular genetic PD research has focused on candidate gene studies (Batt et al., 2020; Guintivano et al., 2018). Candidate gene studies select and test only a small number of genes or genetic variants (“candidates”) for an association with a phenotype. The candidates are chosen based on assumptions that the products of those genes influence the phenotype through a hypothesized biological mechanism. In PD research, hormone receptors for oxytocin (OXTR) and estrogen (ESR1) have been popular candidates (Payne and Maguire, 2019). The candidate gene approach has been largely unsuccessful because not enough is known about the human genome to make reasonably good guesses about which variants to test (Border et al., 2019). Given that candidate genes are no more predictive of MD than non-candidate genes, and robust replications have been sparse, it is likely that the majority of reported associations between PD and candidate gene variants represent false positives (Border et al., 2019).An alternative to testing only a few loci is to test variants across the entire genome via genome-wide association studies (GWAS). GWAS offers an agnostic approach for identifying trait-associated variants, and GWAS summary statistics can be used to estimate heritability, quantify shared genetic architecture across traits, and construct aggregated risk scores. In order to robustly detect the expected small effect sizes of common genetic variants, GWAS requires large sample sizes to achieve adequate statistical power (Wray et al., 2012). The largest published PD GWAS to date consisted of 3,804 cases and 6,134 controls, a sample size vastly underpowered to detect genome-wide effects (Kiewa et al., 2022; Wray et al., 2012) and approximately 100 times smaller than the largest published MD GWAS (Levey et al., 2021; Howard et al., 2019).In order to increase power to detect associations, whole-genome research has primarily relied on aggregate genetic methods to characterize genetic relationships between PD and other traits, often psychiatric disorders. This approach leverages summary statistics from GWAS of related disorders to calculate individual aggregate measures of genetic risk (e.g., summing number of risk variants), known as polygenic risk scores (PRS). The PRS are then applied to predict PD in an independent target sample with the purpose of describing the degree of genetic similarity between PD and the related disorder. While genetic correlation estimates more accurately assess the shared genetic etiology between two disorders, (Bulik-Sullivan et al., 2015) cross-trait PRS analysis does not require that well-powered GWAS are available for both traits of interest. Significant associations between PD and PRS for MD, (Kiewa et al., 2022; Bauer et al., 2019; Pouget et al., 2021; Rantalainen et al., 2020) bipolar disorder, (Byrne et al., 2014; Pouget et al., 2021) and schizophrenia have been reported (Rantalainen et al., 2020). While the estimated SNP-based heritability of PD has remained fairly consistent (h2 = 0.22), (Kiewa et al., 2022; Byrne et al., 2014) the magnitude of relationships between PD and PRS for other psychiatric outcomes has varied. For example, Byrne et al. (2014) found that bipolar disorder exhibited the greatest genetic overlap with PD (R2 = 1.64%), while other studies have either not replicated this association (R2 = 0.01%) (Bauer et al., 2019) or identified stronger relationships between PD and the PRS for MD (PRS-MD R2 = 7.6%) (Kiewa et al., 2022). These inconsistencies are likely influenced by cross-study differences in study design choices, including phenotyping, sample collection, and statistical power (Bauer et al., 2019).3 Defining and measuring peripartum depression3.1 Defining PDArguably one of the most challenging aspects of studying PD is the evolving nature of its clinical definition (Figure 1) which has complicated epidemiological efforts to estimate prevalence and study trends over time. Pregnancy-related depression was not distinguished by specifiers or a unique diagnosis in the Diagnostic and Statistical Manual (DSM) until 1994 when it appeared as a postpartum specifier to MD in the DSM-IV (American Psychiatric Association, 1994). The most recent edition (DSM-5) expands the definition of PD to include any major depressive episode that onsets either during pregnancy or within the first 4 weeks postpartum (American Psychiatric Association, 1994). The International Classification of Diseases code (ICD) recently underwent a similar expansion from ICD-10 and ICD-11, to include the prenatal period and up to 6 weeks postpartum (World Health Organization (WHO) 1993; World Health Organization (WHO) 2019). While this shift in clinical definition reflects improvements in prenatal care, the need to harmonize data across multiple diagnostic definitions complicates meta-analyses
Read More