Neuregulin 1: genetic support for schizophrenia subtypes

SIR—The deficit syndrome is a form of chronic schizophrenia characterized by enduring, idiopathic negative symptoms.1 Patients with this clinically distinct disease subtype have been suggested to differ in several biological aspects from patients with a more benign course with exacerbations and remissions.2 We hypothesized that part of the heterogeneity of disease course in schizophrenia is related to different contributing genetic factors. The neuregulin 1 gene (NRG1), assumed to be involved in glutamate neurotransmission and neurodevelopment, was recently reported to be associated with schizophrenia in several, but not all populations.3, 4, 5, 6, 7, 8, 9 We have investigated the association of NRG1 with schizophrenia in patients with and without the deficit syndrome. In order to enrich the sample for patients with the deficit syndrome, 282 schizophrenia patients were mainly recruited from psychiatric hospitals. Patients had at least three Dutch-born Caucasian grandparents. DSM-IV diagnosis of schizophrenia, excluding schizoaffective disorder, was made using the Comprehensive Assessment of Symptoms and History10 and information from medical records. The Schedule for the Deficit Syndrome (SDS)11 was completed for 260 patients (92.2%), 130 of whom met deficit criteria. In 29 patients with severe negative symptoms, it could not be ruled out that these symptoms were secondary to factors such as substance abuse. Following the SDS, they were classified as nondeficit schizophrenia. The Medical Ethical Committee of the UMC Utrecht approved the study and all patients gave written informed consent. The control panel (n=585) consisted of 472 DNA samples from random Dutch individuals, obtained from the Immunogenetics and Transplantation Immunology Section of the Department of Immunohematology and Blood Transfusion, LUMC, Leiden and 113 healthy controls from our department. Under optimal conditions, these samples have approximately 90% power to detect a locus with a relative risk of 1.5, and still 75% to detect a locus with an RR of 1.25. All five single-nucleotide polymorphisms (SNPs) from the previously reported at-risk haplotype5 were first screened in DNA pools from an unselected subset of patients (n=208) and independent controls (n=179), using the SNaPshot technique (Applied Biosystems, Foster City, CA, USA). SNP8NRG221533 was genotyped individually on a 7900HT TaqMan system (Applied Biosystems). Microsatellites 478B14-642 and 478B14-8485) were analyzed on an ABI 3700 sequencer (Applied Biosystems) and genotyped by two independent raters. Genepop12 software was used to verify Hardy–Weinberg equilibrium (HWE). Haplotypes, linkage disequilibrium (LD) and likelihood ratios were calculated using UNPHASED software.13 Alleles and haplotypes with frequencies <1% were combined. Global P-values were calculated using n-1 degrees of freedom (n=number of alleles or haplotypes). No correction of P-values for testing several markers and haplotypes was applied, because the markers were in LD and tests are therefore not independent. SNP8NRG221533, which was the most strongly associated single marker in previous studies, showed a significantly higher frequency of the T allele in the schizophrenia pool than in controls (P<0.01). This marker was then genotyped individually in the extended sample, as was microsatellite 478B14-848, which is part of the reported at-risk haplotype.4 Before testing for association, we calculated LD between the markers, and found it to be low (D'=0.20), which is in agreement with a recent study.8 Therefore, we genotyped microsatellite 478B14-642, instead of the even more distant third microsatellite from the reported at-risk haplotype. This marker is located much closer to the SNP, and was recently found to be in a different LD block than 478B14-848.8 Indeed, in our sample, 478B14-642 was in considerable LD with SNP8NRG221533 (D'=0.60), but not with 478B14-848. All three markers were in HWE. Interestingly, SNP8NRG221533 was associated with nondeficit schizophrenia (P=0.004), but not with deficit schizophrenia (P=0.542). Likewise, haplotypes of SNP8NRG221533 and 478B14-642 were significantly associated with nondeficit schizophrenia only (Table 1). It is not very likely that population stratification caused the association, because there were no differences in allele frequencies between patients originating from different areas of the Netherlands or between the two control samples. Given the relatively low prevalence of deficit schizophrenia (15–30% of patients2), it is perhaps not surprising that we found NRG1, first identified in samples unselected for deficit schizophrenia, to be associated with nondeficit schizophrenia. The putative functions of NRG1 in gliogenesis, neuronal migration and synaptic plasticity5 seem to indicate a neurodevelopmental cause in nondeficit schizophrenia. Our results support previous suggestions that deficit schizophrenia may represent a distinct biological entity.2 The rapid and persistent clinical deterioration seen in many deficit patients could be the result of a degenerative process, dominated by other than the recently suggested gene systems, and environmental factors. The identification of susceptibility genes for deficit schizophrenia may be very valuable, since negative symptoms are a cause of much disability, and often resistant to the current treatment. Compared with the findings in Iceland, different SNP alleles and haplotypes are associated in the Dutch population. Different associated NRG1 haplotypes have been described6, 9 and, as pointed out, it is not unexpected that mutations in different populations have a different haplotype background.14 Likewise, different haplotypes were found to be associated for dysbindin (DTNBP1), another schizophrenia susceptibility gene.15 Additional fine-mapping studies are required to identify the causative variation(s) and to establish if NRG1 or a nearby gene is involved.9 In conclusion, our results further support the evidence for NRG1 as a schizophrenia susceptibility gene across populations, and suggest that the course of illness in schizophrenia is influenced by different sets of genes.