As part of our ongoing coverage of DISC1 2010, held 3-6 September 2010, in Edinburgh, the United Kingdom, we bring you a meeting missive from Rosie Walker, a graduate student at the University of Edinburgh.
Opening the seventh session of the conference, the second in the theme Networks and Signaling, chair David Porteous introduced Akira Sawa of Johns Hopkins University, Baltimore, Maryland, who broached the sticky subject of DISC1 antibodies, with particular reference to the detection of DISC1 in the 129 mouse strain. It has been known for a few years that 129 mice carry a 25bp deletion in DISC1 exon 6, which introduces a premature stop codon in exon 7 (Clapcote and Roder, 2006; SRF related news story). Detection of DISC1 is these mice has produced intriguing results: Koike et al. (Koike et al., 2006) failed to detect either full-length DISC1 or the predicted C-terminal truncated protein; however, Ishizuka and colleagues (Ishizuka et al., 2007), using several antibodies against different epitopes detected indistinguishable levels of full-length DISC1 in 129S6/SvEv mice and C57BL/6J mice, with all antibodies except that used by Koike and colleagues. This led to the suggestion that the 25bp deletion might lead to the specific loss of an isoform of DISC1 that is uniquely detected by the antibody of Koike et al. (Koike et al., 2006), and is very similar in length to full-length DISC1.
This finding prompted Sawa to raise two questions: firstly, what is the effect of the 25bp deletion? And secondly, how can we define a “good antibody”? With reference to the first question, Sawa commented on the apparently paradoxical situation of a relatively mild phenotype in mice carrying the 25bp deletion compared to mice with point mutations in DISC1 (Koike et al., 2006; Clapcote et al., 2007). Sawa drew attention to the recent finding that shRNA-mediated knockdown of DISC1 in C57BL/6NCr mice, ICR mice, and 129X1/SvJ mice results in similar migration deficits (Kubo et al., 2010), suggesting that previous studies using RNAi in mice with the 25bp deletion are likely to have produced valid results. Nevertheless, despite the lack of a marked phenotype in 129 mice, Sawa emphasized the need to fully investigate the significance of this deletion, particularly as unpublished work from Sawa’s group has produced preliminary data suggesting that 40-70 percent of outbred mouse strains, such as Swiss-Webster and ICR, are contaminated by the 25bp deletion polymorphism.
Moving on to the theme of a system for determining “good DISC1 antibodies,” Sawa laid down a set of criteria that he felt should be met by all studies using antibodies. These were that every study should, firstly, use at least two different antibodies in some way; secondly, demonstrate clear knockdown of DISC1 by more than one RNAi; and thirdly, that the two antibodies should yield consistent Western blot and immunoprecipitation results. Sawa concluded by mentioning plans to create a webpage of up-to-date DISC1 RNAi and antibody information on the Johns Hopkins University website, with the hope that this database will facilitate better regulated studies in the future.
Second to speak was Sawa's Johns Hopkins colleague Saurav Seshadri, who discussed the function of DISC1 in interneurons. In a recent publication Seshadri and colleagues demonstrated a functional interaction among NRG1 and NRG2 and DISC1, mediated by ErbB2 and ErbB3 receptors (see SRF related news story). As interneurons are the predominant site of ErbB4, an NRG1 receptor, expression in the hippocampus and cortex (Fazzari et al., 2010), Seshadri asked whether DISC1 modulates NRG1/ErbB4 in interneurons. Expression of DISC1 in interneurons has been demonstrated previously (Meyer and Morris, 2008), and was confirmed by Seshadri, who demonstrated coexpression of the interneuron-marker GAD67, and DISC1 in rat primary cortical neurons. Co-immunoprecipitation assays confirmed an interaction between DISC1 and ErbB4. The DISC1 ErbB4 binding domain has been narrowed down to the C-terminus of DISC1, and, although still in the process of being refined, it is currently thought to overlap with the Dixdc1, PCM1, and Lis1 binding domains. Further supporting the interaction between NRG1/ErbB4 and DISC1, an increase in DISC1-ErbB4 binding was observed in cultured cells transfected with DISC1 and ErbB4 when treated with NRG1. Additionally, knocking down DISC1 increased baseline and NRG1-induced ErbB4 activation. Following on from the 2005 finding by Huang et al. (Huang et al., 2000) that ErbB4 binds PSD-95, and that PSD-95 regulates NRG1 signaling, Seshadri found that DISC1 affects ErbB4/PSD-95 binding, suggesting a pathway for the regulation of NRG1 by DISC1. Finally, Seshadri presented data supporting a role for interneuronal DISC1 in the regulation of pyramidal neuron function: increased spine density is observed in the pyramidal neurons of ErbB4 knockout mice, suggesting potentially altered function.
Shifting the focus of the session to a genetic perspective, Liisa Tomppo of the National Public Health Institute, Finland, talked about her work investigating genetic variants that associate with anhedonia, a psychosis-related trait. In a previous study, Tomppo et al. (Tomppo et al., 2009) identified DISC1 variants that associate with performance on the Revised Physical Anhedonia Scale (RPAS) and the Revised Social Anhedonia Scale (RSAS) in the Northern Finland Birth Cohort of 1966. Tomppo has now performed a genomewide association study to look for variants that associate with RPAS and RSAS performance conditioned on the previously identified DISC1 variants. None of the SNPs assessed met the predefined threshold for genomewide significance; however, many SNPs showed suggestive evidence for association with RPAS. Ingenuity pathway analysis was performed on 18 loci of interest and this revealed three loci closely related to the DISC1 pathway. The genes located at these loci are CCDC141, which is known to directly interact with DISC1; the lactase gene, LCT; and the microRNA MIR620. Interestingly, Ingenuity pathway analysis of the genes targeted by MIR260 revealed enrichment for genes implicated in bipolar disorder, as well as central nervous system development and other relevant pathways. Furthermore, LCT expression has already been shown to be regulated by DISC1 (Hennah and Porteous, 2009). The findings by Tomppo et al. hint at the likely complexity of genetic interactions underlying schizophrenia-related traits, and highlight the potential for conditional association analyses to identify new candidate genes.
Last to speak in this session was Kathy Evans of the University of Edinburgh, U.K., who presented her work investigating neurexin1 (Nrxn1) and neurexin3 (Nrxn3) expression in L100P mice. These mice carry a point mutation in DISC1 amino acid 100, resulting in a leucine-to-proline substitution, and show a schizophrenia-like phenotype (see SRF related news story). Whole-genome expression profiling of hippocampal tissue from 12-week-old L100P mice revealed a significant change in the expression of several genes compared to wild-type mice. Of these genes, Nrxn1 and Nrxn3 were selected for follow-up. Neurexins are a family of genes encoding synaptic cell adhesion molecules, which are located pre-synaptically and interact with post-synaptic neuroligins to form a trans-synaptic complex (Südhof, 2008). Nrxn1 has been implicated in autism and schizophrenia (see O’Dushlaine et al., 2010 for references; SRF related news story; SRF news story), and Nrxn3 was found to be upregulated in a microarray study of the t(1;11) family (unpublished data). Developmental expression profiles of Nrxn1 and Nrxn3 in L100P mice were found to differ from wild-type most markedly at embryonic day 18, a period of synaptic formation and neuronal maturation, and postnatal day 7, when neurite outgrowth, myelination, apoptosis, and synaptic pruning occur, suggesting that altered expression could have deleterious consequences for brain development. The expression of both Nrxn1 and Nrxn3 has been shown to be altered by synaptic activity, leading Evans to suggest that mutant DISC1 could alter synaptic function, and thus the expression of Nrxn1 and Nrxn3. This, in turn, could alter the distribution of excitatory and inhibitory synapses, therefore offering a functional mechanism for the contribution of Nrxn1 and Nrxn3 variation to schizophrenia and autism pathogenesis.
The theme of DISC1 and cell adhesion was continued by Tsuyoshi Hattori of Osaka University, Japan, who presented a poster covering his recently published work on DISC1’s involvement in cell-cell and cell-matrix adhesion, via N-cadherin and β1 integrin, respectively (Hattori et al., 2010). Also looking at gene expression in L100P mice, Tatiana Lipina of Mount Sinai Hospital, Canada, presented a poster on the effects of valproate at the level of gene expression. Valproate, which can ameliorate the prepulse inhibition deficit and hyperactivity observed in L100P mice when administered chronically prior to the onset of behavioral abnormalities, was found to normalize the expression of six genes (Purb, Lcn2, Dusp1, Arhgap24, Igf1, and Cyr61), which were dysregulated in the mutant mice. Lipina proposes these genes as novel candidate drug targets for the treatment of schizophrenia.—Rosie Walker.