Eating Disorders Genetics in Asia

Abstract

A complete understanding of the genetic architecture of eating disorders requires adequately large sample sizes from individuals of all ancestries. Failure to include representative samples truncates understanding and may even exacerbate health disparities. Several countries in Asia have made rich contributions in psychiatric genetics; however, the eating disorders field requires concerted global efforts to increase representation from Asian ancestry populations to ensure that our global efforts accurately reflect the true distribution of eating disorders around the world and across ancestries.

Current state of global progress in eating disorders genomics

The past decade has witnessed considerable acceleration in the study of eating disorders genomics. The emergence of analytic approaches such as the genome-wide association study (GWAS) has enabled large-scale studies of individuals with a disorder or trait compared to ancestrally matched controls to identify regions on the human genome (or loci) that contain risk alleles for the target condition. GWAS of anorexia nervosa (AN) have been published and parallel studies of bulimia nervosa (BN) and binge-eating disorder (BED) are underway.

The largest AN GWAS to date included samples from all prior AN GWAS along with new samples from the Anorexia Nervosa Genetics initiative (ANGI) (Watson et al., 2019). This GWAS identified 8 loci and both psychiatric and metabolic/anthropometric genetic correlations. However, that study included 33 datasets comprising 16,992 cases and 55,525 controls from 17 countries—all of European ancestry. Over-representation of European ancestry samples is the norm for psychiatric genomics (Peterson et al., 2019). Fortunately, funding agencies such as the National Institutes of Mental Health and consortia like the Psychiatric Genomics Consortium (PGC) have prioritized genetic studies of diverse populations, leading to an increase in non-European samples over the past several years, with the largest increase coming from East Asian populations (Duncan et al., 2019). Similar efforts are necessary to increase samples from Asian samples for GWAS of eating disorders.

History of eating disorders genetics in Asia

Importantly, Asian countries were prominent contributors to early studies of the genetics of eating disorders, including several candidate gene studies of AN and bulimia nervosa (BN) [e.g., (Ando et al., 2014)] led by The Japanese Genetic Research Group For Eating Disorders (JGRED)—a multisite collaborative study group that was organized for the systematic recruitment of patients with an eating disorder for the purpose of genetic study in Japan. The study of isolated candidate genes has lost favor in the field with the advent of GWAS. In fact, the first genome-wide study of AN was published by this group in 2009 implicating a region on Chromosome 1 and marking a transition to genome-wide approaches in the field (Nakabayashi et al., 2009).

In the Wellcome Trust/Genetic Consortium for Anorexia Nervosa GWAS (Boraska et al., 2014), 458 cases and 421 controls from Japan were included as replication samples. The inclusion of these samples in the replication contributed to the optimism that subsequent larger GWAS of AN would be likely to identify significant variants as there was significant evidence of SNP effect sizes in the replication data in the same direction as the discovery set. This work very much encouraged the continued study of genetic contributions to AN etiology.

Technical advances necessary for diversifying samples

Several factors contributed to the lack of diversity in psychiatric genetic samples, including eating disorders. Not unlike what we have seen in our field for males, often if the number of samples available from non-European ancestries is small, they have frequently been excluded from analysis. Significant advances are being made and disseminated and within psychiatry much of this work is being undertaken by the Cross-Population Special Interest Group of the PGC (https://www.med.unc.edu/pgc/cross-population). One reason why clarity about ancestry is so important is that individuals with similar ancestries (e.g., East Asians, Africans, Europeans) have shared genomic signatures due to migration history, mutations and recombination, genetic drift, and natural selection (Peterson et al., 2019). These differences must be accounted for in analyses to ensure that observed differences between cases and controls are due to genes related to the target disease and not just ancestry.

On a technical level, the arrays used for GWAS still tend to work best in European samples (reflecting linkage disequilibrium structure and allele frequency of the European population). Newer multi-ancestry arrays are emerging that provide greater coverage for diverse samples or that perform better in other ancestral groups (e.g., the OmniZhongHua or H3Africa arrays). Moreover, GWAS arrays only genotype a subset of common genetic variants, with the remainder being imputed from large external reference samples. These panels must also capture the diversity of common variation across ancestries in order to be maximally informative. Optimal approaches to imputation in diverse GWAS samples are advancing as is the quality of available reference panels.

Analytic and technical advances are continuing to evolve and are removing any remaining barriers to diversification. For example, whole genome sequencing with low coverage (i.e., the current optimal combination of sequencing depth and cost) could emerge as the optimal technology for all humans. The responsibility now lies with researchers and funding bodies to invest resources and energy into ascertaining diverse samples.

Why are diverse samples important in psychiatric genomics?

As with any complex medical condition, we simply do not know if the underlying genetic contributions to etiology, course, and outcome of AN, BN, and BED are the same across individuals from different ancestral backgrounds. Only by including adequately large samples from diverse ancestries will we be able to determine the extent to which causal genetic factors are similarly operative across populations.

Another important factor underlying the need for diverse samples relates to the predictive ability of polygenic risk scores (PRS) and their utility in improving health outcomes (i.e. precision medicine). PRS are calculated from GWAS results and are individual-level metrics of genetic risk for a disease or trait. A PRS is a sum of risk alleles, weighted by effect sizes that reflects an individual’s inherited susceptibility to a specific disease or trait. For example, a PRS for AN can be derived from the Watson et al. (Watson et al., 2019) GWAS (discovery sample), and then applied to another sample to see if the PRS actually predicts who in that sample is most likely to have AN. PRS become more robust as sample sizes increase and have the potential to improve health outcomes by accelerating diagnosis and eventually matching patients to treatments tailored to underlying genetics. Importantly, PRS have not been shown to be equally predictive across ancestries. This means that PRS derived from European populations would confer preferential benefit to health outcomes in European populations, further perpetuating health disparities (Martin et al., 2019).

An excellent example comes from a GWAS of schizophrenia in individuals from East Asia (22,778 individuals with schizophrenia and 35,362 controls) (Lam et al., 2019). This study revealed a high genetic correlation between schizophrenia in East Asian and European populations (r_g=0.98) – remarkably, the common variant basis of schizophrenia in Europe and East Asia was effectively identical. However, the schizophrenia PRS derived from the European sample performed more poorly when applied to the East Asian sample, underscoring the importance of large diverse samples in GWAS especially related to health prediction and outcome.

Another potential important advantage of diverse samples is if a disease presents differently across cultures, and presumably across ancestries. For example, the literature suggests that non-fat phobic AN (i.e., AN in the absence of drive for thinness/fear of fatness) may be more common in Asian populations (Pike & Dunne, 2015). An adequately powered cross-ancestry GWAS could dissect whether specific genetic contributions to drive for thinness exist that are more common in individuals with AN from European than Asian ancestry. Greater sample diversity could allow us to understand genetic factors that contribute to phenotypic heterogeneity.

Diverse samples also have a potential role to play in reducing health disparities and matching patients to treatments (i.e., personalized medicine). One of the eventual goals of GWAS is to identify biological pathways that lend themselves to drug repositioning or development. It should not be assumed that a medication developed based on a European sample would necessarily be equally effective in individuals from other ancestry backgrounds. Having large diverse samples ensures that any pharmacogenomic advances made will consider the role of ancestry in drug response and directly address potential health disparities.

Diversity of genetic samples has the potential to shed light on the very pertinent role of environment, gene x environment interplay, and genetic and environmental protective factors in eating disorders. We have yet to answer the fundamental question of why nearly ubiquitous exposure to a thin-ideal culture does not lead to universal eating disorders. Exploring individuals across ancestries who are at high genetic risk for AN who do not have the illness could point to environmental of genetic factors that buffer against genetic risk.

Finally, it is simply scientifically responsible to have diverse representation in genetic samples to accurately reflect the true distribution of eating disorders around the world and across ancestries.

Expanding eating disorders genetics in Asia

The Eating Disorders Working Group of the PGC (PGC-ED) welcomes clinicians and researchers from all around the world to join the global effort to explicate the genetics of all eating disorders. The flagship Eating Disorders Genetics Initiative (EDGI, www.edgi.org) is underway globally and has developed streamlined procedures for new sites to join. The next AN GWAS will include both Korean and Japanese samples; however, it is important to vastly extend the network in Asia to ensure adequate representation. Plans are underway to develop EDGI-Taiwan and targeted recruitment of individuals with eating disorders of Asian ancestry who live in other EDGI countries (US, Australia, New Zealand, UK, Denmark, Sweden, the Netherlands, Mexico) will also expand our sample diversity. The PGC-ED welcomes individuals from across Asia who may have interest in joining the PGC-ED or developing an EDGI site in their region. Only via global collaboration will we be able to fully explicate the genetics of all eating disorders across all ancestries.