SIR – The existence of an important genetic contribution to the etiology of schizophrenia is well established from genetic epidemiological studies. Recently, a study performed by Straub et al¹ showed that the gene of DTNBP1 (dystrobrevin-binding protein 1) located on 6p22.3, a high-susceptibility chromosome region of schizophrenia, was highly associated with schizophrenia in Irish high-density pedigrees. DTNBP1, the human ortholog of mouse dysbindin, appears to have a role in neuromuscular synapse formation and maintenance.² It is found in multiple locations within the central nervous system, so it may also have a possible function in signal transduction.^2,3 In the present study, to confirm and extend Straub's findings of DTNBP1 in independent samples, we investigated seven SNPs within 140-kb length of the gene DTNBP1 in 233 Han Chinese trios.

Parent–offspring trios used in this study containing 173 trios from Shanghai and 60 from Xi'an (131 male and 102 female probands with a mean age of 22.88, SD=0.57) were also involved in our previous study.⁴ The unrelated schizophrenic probands were collected from Shanghai Mental Health Center and Xi'an Mental Health Center, respectively. Clinical diagnosis was made according to DSM-IIIR criteria⁵ by two independent clinicians. A standard informed consent, which was reviewed and approved by the Shanghai Ethnical Committee of Human Genetic Resources, was given by the participated subjects after the nature of study had been fully explained. All subjects were Han Chinese in origin. Seven SNPs (P1328, P1287, P1655, P1635, P1763, P1578 and P1583) adopted in Straub's work were genotyped in all our trios samples. The genotyping assay combines kinetic (real-time quantitative) PCR with allele-specific amplification, which was described elsewhere.⁶ Kinetic PCR reactions were performed on an ABI PRISM 7900 Sequence Detection System (Applied Biosystems).

Our sample showed less heterozygosity in the seven SNPs of DTNBP1 gene compared with the data from Straub's study with Irish high-density pedigrees collected from the relatively homogeneous population of Ireland and Northern Ireland.⁷ Two markers, P1287 and P1635, were excluded in the later statistical analyses since their rare-minor allele frequencies were ≤ 5%. All the trios were grouped together in statistical analysis because no differences were found in distribution of genotype frequencies between trios from Shanghai and Xi'an. The markers used for further analyses were in Hardy-Weinberg equilibrium (all P-values ≥0.14) and had no Mendelian inheritance errors. Linkage disequilibrium (LD) between each pair of markers was estimated with software 2LD.⁸ Strong LD was observed among these markers but P1328 was not in tight LD with P1655 (D′=0.14, P=0.03) although it was in tight LD with other markers. ETDT program⁹ was used for transmission/disequilibrium test analysis of individual SNPs where all of the results were negative. However, multiple haplotype analysis of the five markers by TRANSMIT v2.5¹⁰ showed a very significant result (Table 1). Totally, 25 haplotypes were observed, but only five of them were present with probabilities greater than 3%, accounting for vast majority of the haplotype diversity (89.5%). The global P-value for these five haplotypes was 0.00072 (χ²=21.27, 5df), of which the most significant overtransmitted haplotype was the most common haplotype ‘TGTCA’ (frequency=37%, P=0.01). Owing to the lower LD extent of P1328 with P1655, we also analyzed haplotypes based on the other four markers, which showed excess transmission of haplotype ‘GTCA’ to the schizophrenia probands (P=0.00091).

According to the diagnostic definition of categories D1–D2 in Straub's work,¹ probands of the 233 Han Chinese trios used in the present study can be submitted in this category. In Straub's analysis, the marker P1635, which was identified by Straub et al in affected individuals from 6p22-linked families, was strongly associated with schizophrenia when diagnostic categories D1–D2 were used. However, the rare allele was not found in our pretested 92 samples. The highly significant three-marker haplotype of P1763, P1578 and P1583 in Straub's study showed no association with schizophrenia in 233 Han Chinese trios (the lowest P=0.072 for individual haplotypes). These differences between Straub's data and our result could be because of the different ethnical samples used. Nevertheless, multimarker haplotype analysis of DTNBP1 indicates global tansmission disequilibrium at this locus. Both Straub's results and our results suggest that DTNBP1 is likely a susceptibility gene of schizophrenia. These results warrant further analysis of DTNBP1 in large independent samples.