10 March 2006. Since schizophrenia and related psychiatric disorders were found to strongly segregate with a familial mutation that causes a break in the gene disrupted in schizophrenia 1 (DISC1), researchers have raced to see if more subtle DISC1 variations, such as single nucleotide polymorphisms (SNPs), might associate with mental illness in the general population. While some studies support this idea, the evidence is still considered inconclusive. Now, in an article in press, published online March 1 at Human Molecular Genetics, researchers report that DISC1 polymorphisms may actually have more of an impact on expression of the genes for NUDEL, Fez1, and Lis1. The finding suggests that these DISC1 binding partners may be more disrupted in schizophrenia than DISC1 itself.
Barbara Lipska and colleagues at the National Institutes of Health, Bethesda, Maryland, together with collaborators at Johns Hopkins University School of Medicine in Baltimore and Astellas Pharma Inc., in Tsukuba, Japan, set out to determine if brain levels of the DISC1 protein are altered in people with schizophrenia, especially those with polymorphisms that have been linked to disease. Lipska and colleagues tested postmortem brain tissue samples from 79 normal individuals and 43 patients. The results were a little surprising.
The researchers found that though expression of the DISC1 gene in the brain varies with age—right after birth expression spikes, but by teenage years it falls to below neonatal values and continues to decline thereafter—there is no difference between transcript levels in normal and schizophrenia samples. The authors measured DISC1 mRNA in both the hippocampus and the dorsolateral prefrontal cortex (DLPFC) areas of the brain that have been shown to be affected by the disease. Furthermore, when the authors genotyped the samples, they found that three SNPs (hCV219779, rs821597, and rs821616) they previously identified as being associated with schizophrenia (see Callicott et al., 2005) had no effect on DISC1 transcript levels either.
But despite the normal DISC1 transcript levels, the researchers did find that DISC1 protein levels are modestly elevated (~20 percent) in hippocampal samples from schizophrenia patients compared to controls, suggesting that the protein may be more stable. But surprisingly they also found that SNP rs821597, which lies in intron 10 and therefore does not affect the protein sequence, was associated with even higher levels of DISC1 protein in the schizophrenia samples. Why this might be is unclear. Elevated protein is either due to increased production or decreased turnover. Though the SNP could conceivably affect mRNA splicing, leading to an increase in some specific transcripts and protein isoforms, the latter would not be detected by the antibody the authors used, which was raised to the C-terminal part of the protein that is absent from all the known alternatively spliced isoforms. Also, given that total DISC1 transcript levels are not affected by the SNP, decreased protein degradation/removal seems a likely explanation for the observed increase in DISC1. But how an intronic SNP might affect protein turnover is a mystery.
What is even more mysterious is that SNP rs821597 affects expression of NUDEL, Fez1, and Lis1—three proteins that interact with DISC1 (see SRF related news story and SRF news story). When the authors quantitated mRNAs in the hippocampus, they found all three were significantly lower in schizophrenia, and they also correlated with the rs821597 genotype. In tissue from schizophrenia patients homozygous for the G allele of this SNP, levels of the three transcripts were lower still (about twofold) than in tissue with the G/A or A/A genotypes—Fez1 mRNA was also significantly lower (about threefold) in the DLPFC. Curiously, the rs821597 genotype had no effect on transcript levels in normal brain tissue.
How this intronic SNP affects DISC1 protein without affecting DISC1 expression, and also affects levels of NUDEL, Fez1, and Lis1 in schizophrenia patients only, is uncertain. But the authors report that the results remain statistically significant even after the data is normalized to housekeeping proteins and transcripts, such as β-glucuronidase. They also caution, however, that the use of postmortem samples is an obvious limitation, and they note that another study using the same DISC1 antibody failed to detect any increased DISC1 in schizophrenia patients (see Sawamura et al., 2005).—Tom Fagan.
Lipska BK, Peters T, Hyde TM, Halim N, Horowitz C, Mitkus S, Weickert CS, Matsumoto M, Sawa A, Straub R, Vakkalanka R, Herman MM, Weinberger DR, Kleinman JE. Expression of DISC1 binding partners is reduced in schizophrenia and associated with DISC1 SNPs. Human Molecular Genetics, advanced access, March 1, 2006. Abstract