Heritability of subcortical brain measures: A perspective for future genome-wide association studies

Highlights

• The volumes of subcortical brain structures are highly heritable.

• Automatically segmented subcortical volumes are stable over a 5-year time period.

• Mean differences are observed between men and women.

• However, no sex by genotype interactions for subcortical volumes were found.

Abstract

Several large imaging-genetics consortia aim to identify genetic variants influencing subcortical brain volumes. We investigated the extent to which genetic variation accounts for the variation in subcortical volumes, including thalamus, amygdala, putamen, caudate nucleus, globus pallidus and nucleus accumbens and obtained the stability of these brain volumes over a five-year period. The heritability estimates for all subcortical regions were high, with the highest heritability estimates observed for the thalamus (.80) and caudate nucleus (.88) and lowest for the left nucleus accumbens (.44). Five-year stability was substantial and higher for larger [e.g., thalamus (.88), putamen (.86), caudate nucleus (.87)] compared to smaller [nucleus accumbens (.45)] subcortical structures. These results provide additional evidence that subcortical structures are promising starting points for identifying genetic variants that influence brain structure.

Introduction

Over the past few years the first genetic variants influencing variation in human brain structures have been identified based on genome-wide association (GWA) meta-analyses. Such findings are expected to be of great importance for understanding the biological mechanisms underlying cognition and neuropsychiatric disorders. In order to accomplish genome wide significance, large samples are needed. For this, several imaging genomics groups have been working collaboratively together (e.g., Enhancing Neuro Imaging Genetics through Meta-Analysis (ENIGMA), the Cohorts of Heart and Aging Research in Genomic Epidemiology (CHARGE), and the Early Growth Genetics (EGG) consortium). These studies together provided evidence for a significant association between hippocampal volume and rs7294919 on chromosome 12q24 (located between HRK and FBXW8 which is associated with expression of the TESC gene, involved in cell proliferation and differentiation) (Bis et al., 2012, Stein et al., 2012) and between intracranial volume (ICV) and rs10784502 on chromosome 12q14 (associated with expression of the HMGA2 gene, implicated in human growth) (Stein et al., 2012), rs4273712 on chromosome 6q22 (associated with adult height) (Ikram et al., 2012) and rs9915547 on chromosome 17q21 (associated with early brain development) (Ikram et al., 2012). Interestingly, the variant on 17q21 was also associated with head circumference (Taal et al., 2012). In view of these promising results, in the next series of steps the ENIGMA consortium aims to identify genetic variants influencing other subcortical brain structures, including the thalamus, amygdala, putamen, caudate nucleus, globus pallidus and nucleus accumbens. These subcortical brain regions have been found to be affected in several neuropsychiatric disorders (e.g., schizophrenia, attention deficit hyperactivity disorder, autism, anxiety disorders, depression (Holzschneider and Mulert, 2011, Koolschijn et al., 2009, Shepherd et al., 2012, Taurines et al., 2012)), and it is therefore of importance to disentangle the sources (genetic and environmental) that could explain the variation in these phenotypes. As a first step for identifying not only genetic variants, but also environmental factors, associated with the phenotypic variation, the extent to which the genome and/or the environment account for the variation in these volumes should be investigated. Since these large consortia make use of fully automated segmentation protocols, it would be furthermore of interest to assess the stability of these volume measurements, which could serve as an indicator of measurement accuracy. Although the heritabilities of global (intracranial, total brain, gray matter, and white matter volumes) and regional (cortical) brain structures have been extensively studied, as reviewed in Blokland et al. (2012), and Peper et al. (2007), only a small number of studies have reported on the heritability of the volumes of subcortical brain areas (Kremen et al., 2010, Stein et al., 2011, Wallace et al., 2006, Wright et al., 2002, Yoon et al., 2011). A meta-analysis that calculated a weighted average of the proportion of variance accounted for by genes (A), shared environment (C), and unshared environment (E) yielded fairly high heritability estimates for all subcortical structures, ranging from 52% for the right thalamus to 82% for the right putamen (Blokland et al., 2012). These high estimates indicate that the investigated subcortical structures could serve as endophenotypes and targets for GWA studies. However, because the number of individuals in which the heritability of subcortical volumes was calculated tends to be small, the confidence intervals in this meta-analysis are still wide, for example ranging between 42–80% and 36–69% for the left and right Thalamus, respectively (Blokland et al., 2012). Therefore, further replication of these findings across independent samples and demographic groups is desirable. Also sex differences in heritability have not been systematically investigated.

Here, we estimated the heritability for thalamus, caudate nucleus, putamen, globus pallidus, hippocampus, amygdala and nucleus accumbens in a sample of 176 monozygotic (MZ) twin pairs (107 female pairs/69 male pairs) and 88 dizygotic (DZ) twin pairs (24 female–female pairs, 23 male–male pairs and 41 male–female pairs) aged 11–56 years [mean (SD) = 29.10 (10.07), of which 13.6% were aged between 11 and 17]. Image processing was done following the streamlined ENIGMA protocol. We also report the retest stability for a subsample of 161 subjects who were scanned twice with a 5-year interval, and test if heritability estimates differ in men and women.