rs2269726 is located in intron 5 within the GNB1L gene on chromosome 22q11. In males, the homozygous genotype, irrespective of whether it is the ancestral (C) or derived (T) allele, is associated with higher risk for schizophrenia than the heterozygous. A 2008 paper [PMID 18003636] showed the association (P = 1.8 × 10-5 (significant in this study); OR = 1.46) in a candidate gene study using a combined German-British sample of 1,408 cases and 2,746 controls.
Previous evidence linking GNB1L to schizophrenia includes higher incidence of schizophrenia among those with 22q11 deletions and genetic linkage studies. A well-documented ~3 Mb deletion on chromosome 22q11 is characteristic among those with a variety of psychiatric disorders—together termed 22q11 deletion syndrome (22q11DS)—including Velo-Cardio-Facial syndrome, DiGeorge syndrome, and conotruncal anomaly face syndrome [PMID 11715041]. Those with 22q11DS often manifest symptoms of schizophrenia (an estimated 25 percent) and bipolar disorder, as well as a host of other non-psychiatric effects such as immunological and cardiovascular defects [PMID 1308357]. Chromosome 22q is also one of only two genomic regions implicated in two major meta-analyses of schizophrenia linkage studies [PMID 11986984] [PMID 12802786]. Located in this region, GNB1L encodes a G-protein beta-subunit-like polypeptide that does not display homology to any known proteins [PMID 11350118]. Nonetheless, GNB1L is expressed in the brain during mouse adolescence, which might offer insight into its role in psychiatric diseases.
The 2008 paper [PMID 18003636] identifying Rs2269726 failed to arrive at a satisfactory model explaining either the male-only effects or the heterozygote-only association with decreased risk of schizophrenia. The authors suggest that the male-female difference might be due to a disparity in dosage sensitivity, given that the “T” allele is linked to a 20 percent average increase in gene expression relative to the “C” allele. This is a plausible explanation if males and females are each sensitive to a different, but very specific band of expression levels. While the increased expression associated with the “T” allele could also potentially offer insight to the heterozygote-only effect, similarly suggesting a mechanism that involves a very specific dose of GNB1L required to confer the decreased risk of schizophrenia, the authors show that this is not the case. Because GNB1L resides on chromosome 22q11, one would expect those with 22q11DS and schizophrenia would have alleles conferring low expression of GNB1L. Instead, they found that 22q11DS males often had alleles associated with increased expression, invalidating a dosage sensitivity model. As an alternative, they suggest, albeit “speculatively,” that hemizygosity—given a copy of 22q11 is deleted—for alleles associated with low expression triggers compensation mechanisms to increase expression whereas hemizygosity for those associated with high expression does not.