ZNF804A was the first gene that reached genome-wide significance in a large GWAS,13 and this finding has been replicated. The function of this novel gene largely is unknown. ZNF804A is widely expressed in the brain, especially in the developing hippocampus and the cortex as well as in the adult cerebellum. Recent studies found that ZNF804A is a putative transcription factor, upregulating expression of catechol-O-methyltransferase while downregulating dopamine D2 receptors in animal studies.14 The minor allele of SNP rs1344706 was associated with impaired brain functional connectivity in a human study.15 More work is needed to understand how this gene increases schizophrenia susceptibility.
The MHC region on chromosome 6p22.1,1 also was significant in schizophrenia GWAS,16,17 and this may be the most replicated schizophrenia GWAS finding. This region is a recombination hotspot and harbors many genetic variants. Many immune-related genes previously were associated with autoimmune and infectious disorders, which may suggest that the immunologic system plays a role in schizophrenia pathogenesis. These genes also may involve neurodevelopment, synaptic plasticity, and other neuronal processes.18 However, the complex gene composition in the region makes it difficult to pinpoint the exact signal to schizophrenia pathophysiology.
The most recent finding from the largest GWAS is MIR137,19 coding for microRNA 137, which was associated with schizophrenia at P=1.6×10-11 in 17,836 patients and 33,859 controls. MicroRNAs are small, noncoding RNA fragments that are involved in post-transcriptional regulation of messenger RNAs. MIR137 plays important roles in neuron maturation and adult neurogenesis by acting at the level of dendritic morphogenesis and spine development.20 More interestingly, the other 4 loci achieving genome-wide significance in the same GWAS (TCF4, CACNA1C, CSMD1, and C10orf26) contain predicted target sites of MIR137. This suggests MIR137-mediated dysregulation may be an etiologic mechanism in schizophrenia.
Limitations of these findings. The effect sizes of these genetic variants are small, explaining only 1% to 2% of genetic risks of schizophrenia. However, this is not unique to schizophrenia or psychiatry. “Missing heritability” is puzzling in other branches of medicine.21 Future research will focus on gene-environment interactions as well as gene-gene interactions in relation to schizophrenia’s neurodevelopmental processes.
In addition, many top hits in GWAS are SNPs that are not functional or located in intergenic regions with unknown functions. They may be proxies of causal variants that truly play causal roles in pathogenesis of diseases but were not genotyped in those studies. Recently, researchers have grown increasingly interested in copy number variations (CNVs) in the etiology of complex diseases. Compared with SNPs, CNVs usually are much larger changes in the DNA sequence, including deletions and duplications of a large chunk of DNA segments. Disease-causing CNVs are rare but have large effect sizes. Recent studies have examined the role of CNVs in schizophrenia.22,23
Although genes such as DISC1 and CACNA1C are linked to schizophrenia, they are neither necessary nor sufficient for developing the disorder, and also are linked equally, if not more strongly, to other neuropsychiatric disorders, including BD and autism. Therefore, they are not “schizophrenia genes.” Variations in multiple genes likely cause slight deviations in neurodevelopment that interact with environmental variables and lead to development of schizophrenia.
Nevertheless, these schizophrenia GWAS findings provide insight into this complex disorder. Much work is needed to move from these association signals to understanding the function and regulation of these genes to turn basic biologic knowledge into targets for new drugs or other interventions.
Antipsychotic pharmacogenetics
Genetic research of schizophrenia also contributes to our knowledge of how to best use existing drugs. Medications for treating schizophrenia often need to be changed because patients experience lack of efficacy or intolerable side effects, which may lead them to discontinue treatment. Clinical predictors of which medication would work for an individual patient are lacking. Pharmacogenetics may be able to fulfill the promise of personalized medicine in psychiatry by using genetic information to guide drug selection to maximize therapeutic efficacy and minimize drug-induced side effects.
Researchers first attempted to find genetic predictors of antipsychotic efficacy in the early 1990s. One replicated finding is that DRD2, the gene coding for dopamine receptor D2, is associated with antipsychotic efficacy. This may not be surprising because D2 receptor antagonism is a common and necessary drug action mechanism for all antipsychotics. One SNP, -141C Ins/Del (rs1799732), represents a deletion (vs insertion) of cytosine at position -141, located in the 5’ promoter region of DRD2. Pre-clinical studies showed that this SNP might modulate DRD2 gene expression and influence D2 receptor density in the brain. Del allele carriers had poor response to clozapine among a treatment-refractory sample24 and took longer to respond to olanzapine and risperidone among first-episode schizophrenia patients.25 A 2010 meta-analysis of approximately 700 patients26 showed that the -141C Ins/Del polymorphism is significantly associated with antipsychotic response. Patients who carry 1 or 2 Del alleles tend to have a less favorable antipsychotic response than patients with the Ins/Ins genotype. Patients with the Ins/Ins genotype are 54% more likely to respond to antipsychotics than those with ≥1 copy of the Del allele.