Evaluating the dopamine hypothesis of schizophrenia in a large-scale genome-wide association study

Abstract

Background

The dopamine hypothesis, which posits that dysregulation of the dopaminergic system is etiologic for schizophrenia, is among the most enduring biological theories in psychiatry. Although variation within genes related to dopaminergic functioning has been associated with schizophrenia, an aggregate test of variation, using the largest publicly available schizophrenia dataset, has not previously been conducted.

Methods

We first identified a core set of 11 genes involved in the synthesis, metabolism, and neurotransmission of dopamine. We then extracted summary statistics of markers falling within, or flanking, these genes from the Psychiatric Genomics Consortium's most recent schizophrenia mega-analysis results. We conducted aggregate tests for enrichment of dopamine-related pathways for association with schizophrenia.

Results

We did not detect significant enrichment of signals across the core set of dopamine-related genes. However, we did observe modest to strong enrichment of genetic signals within the DRD2 locus.

Conclusions

Within the limits of available power, common sequence variation within core genes of the dopaminergic system is not related to risk of schizophrenia. This does not preclude a role of dopamine, or dopamine-related genes, in the clinical presentation of schizophrenia or in treatment response. However, it does suggest that the genetic risk for schizophrenia is not substantially affected by common variation in those genes which, collectively, critically impact dopaminergic functioning.

Introduction

The first widely cited articulation of the dopamine (DA) hypothesis of schizophrenia (DHS) was by Matthysse in 1973 (Kendler and Schaffner, 2011a, Matthysse, 1973) when he suggested that schizophrenia might result from an “over-activity of dopaminergic transmission” (Matthysse, 1973). For several decades in the late 20th century, the DHS was the leading etiologic theory in psychiatry and generated an immense amount of both basic and clinical research.

Two prominent themes emerge from an analysis of the history of the DHS (Kendler and Schaffner, 2011a, Kendler and Schaffner, 2011b). First, the empirical track-record of the theory has been spotty, with a failure to verify robustly most—albeit probably not all—of its key empirical predictions. Second, its persistence over four decades is probably related to the protean nature of the theory undergoing a number of substantial revisions and re-interpretations. For example, Davis and colleagues proposed, in 1991, a major revision of the theory suggesting that “… schizophrenia can be characterized by hypodopaminergia in mesocortical and hyperdopaminergia in mesolimbic dopamine neurons …” (Davis et al., 1991). More recently, Howes and Kapur proposed another substantial modification of the DHS focusing on DA striatal dysfunction as a “final common pathway” for the etiopathogenesis of schizophrenia (Howes and Kapur, 2009). In addition, a review by Howes et al. (2015) described the extent of evidence implicating the dopaminergic system, ranging from antipsychotic efficacy to in vivo imaging. Furthermore, the authors elaborated on the interplay among glutamatergic and dopaminergic systems and their impact of schizophrenia etiology, demonstrating that the effects of dopamine do not act in isolation.

Given the central role of genetic factors in the etiology of schizophrenia, demonstrated both by classical genetic-epidemiologic methods (Sullivan et al., 2003) and by newer analytic methods applied to molecular data (Cross-Disorder Group of the Psychiatric Genomics et al., 2013), a plausible prediction of the DHS would be that variation in at least some genes critical to DA function would impact on risk for schizophrenia. Indeed, a large and inconclusive candidate gene literature emerged in the last several decades examining DA receptor genes and genes involved in the uptake, synthesis and metabolism of DA (Kendler and Schaffner, 2011a). As depicted in Table 1, quite large numbers of individual association studies have been reported for most of the core DA related genes. The percentage of positive reports for a number of them substantially exceeds the 5% expected under the null hypothesis although meta-analyses have not provided consistent positive evidence for association for any gene with the possible exception of DRD4. It is now clear from recent advances in complex disease genetics, as we have clarified the typical effect size of variants impacting multifactorial biomedical disorders, that these earlier candidate gene studies were unpowered which might explain the variability and inconsistency in findings. Furthermore, the potential for publication bias cannot be excluded, raising the possibility that more null results exist than are in the extant literature.

Advances in the field of schizophrenia genetics have now provided us with a much more powerful way to test the genetic predictions of the DHS. As a result of massive efforts, the Schizophrenia Working Group of the Psychiatric Genomics Consortium (PGC) has collected and made publically available GWAS results on 34,241 cases of schizophrenia and 45,604 controls (Schizophrenia Working Group of the Psychiatric Genomics Consortium, 2014). The PGC recently reported strong evidence of an association between variation in DRD2 and schizophrenia case-control status. In addition, they found signal enrichment in KEGG's dopaminergic synapse gene set. However, that category consists of > 100 genes (Kanehisa et al., 2016) whose functions are quite diverse and in many cases only indirectly related to dopamine function. Thus, in some ways the category reflects the diffuse nature of the DHS itself and may not represent a valid test of the core underlying theory.

In this paper, we selected, a priori, 11 genes directly related to DA function which have all been subject to prior candidate gene studies motivated by the DHS (Table 1). We ask a single, simple question: do common molecular variants in these genes, examined as a group, differ significantly in their frequency in cases and controls? A positive response to this question would provide confirmatory evidence for the DHS. By contrast, a negative response – in which the distributions of these variants in the two groups were consistent with chance effects – would provide evidence, although far from conclusive, against the importance of the etiologic role of DA dysfunction in schizophrenia.