23 November 2011. Attention DISC1-lovers: the latest issue of Neuron carries a twofer on schizophrenia’s most famous gene. Published November 17, one study from Li-Huei Tsai’s lab at the Massachusetts Institute of Technology connects human DISC1 variants to disruptions in Wnt signaling and brain development. The other, led by Guo-li Ming and Hongjun Song of Johns Hopkins School of Medicine highlights an interaction between DISC1 and a protein called FEZ1 in regulating dendrite growth of newborn neurons in the adult brain, and finds evidence that the genes encoding these proteins interact to increase risk for schizophrenia in humans. Together, the studies dissect the multiple pathways emanating from DISC1, and lay the groundwork for obtaining functional insights about the genetic data streaming in for schizophrenia and other mental disorders.
“These complementary studies elegantly bridge the gap between genetic and cellular studies of schizophrenia, providing a level of functional validation that is often lacking in the field,” write Eric Wexler and Daniel Geschwind of University of California in Los Angeles in an accompanying perspective.
Since identifying DISC1 as a risk factor in schizophrenia and other major mental illnesses, researchers have industriously studied DISC1’s function (Brandon and Sawa, 2011). As a scaffold protein, DISC1 interacts with many other proteins, and takes on multiple roles in brain development and function, including neurogenesis (see SRF related news story), neuronal migration (see SRF related news story), dendrite and axon growth, synapse formation (Wang et al., 2011), and dopamine neuron function (Niwa et al., 2010).
The trouble lies in picking out which of these roles is disrupted in psychiatric disease. One clue comes from DISC1’s inhibition of glycogen synthase kinase 3β (GSK-3β), also a target of the antidepressant lithium (Mao et al., 2009 or SRF related news story). GSK-3β acts within the canonical Wnt pathway, which regulates cell proliferation and fate early in development. With data from humans, mice, and zebrafish, Tsai’s study argues that DISC1 variants are loss of function, and negatively impact either neural proliferation or migration in the developing brain—something fitting the idea that schizophrenia stems from neurodevelopmental aberrations.
In the adult brain, Song and Ming’s study examines an interaction between DISC1 and FEZ1 (fasciculation and elongation protein zeta-1), one of the earliest identified binding partners for DISC1 (Miyoshi et al., 2003). Despite limited genetic evidence for FEZ1 in schizophrenia, postmortem work finds a significant reduction in FEZ1 mRNA in schizophrenia brain, and these levels have been associated with risk DISC1 polymorphisms (Lipska et al., 2006). This suggested some kind of DISC1-FEZ1 interaction, for which the new study finds evidence in newborn neurons integrating themselves into the adult mouse hippocampus. This finding prompted a genetic association study in humans, which discovered an interaction, or epistasis, between certain single nucleotide polymorphisms (SNPs) within the DISC1 and FEZ1 genes that increased risk for schizophrenia.
Real! Live! DISC1 variants!
First author Karun Singh and colleagues culled human DISC1 variants from exome sequences of over 700 individuals with schizophrenia, bipolar disorder, or healthy controls. They focused on protein-altering SNPs: three common (R264Q, L607F, S704C) and one rare (A83V). When tested in vitro in a human non-neural cell line, all variants but S704C reduced Wnt signaling, as assayed by activity of TCF-LEF transcription factors, which constitute a terminus of the Wnt pathway. Mirroring this finding, these same variants could not rescue knockdown of endogenous DISC1 in a mouse cell line, except for S704C. The Wnt-inhibiting effects of these variants translated into a reduction of cell proliferation measured in vitro, and seemed to reflect reduced binding between DISC1 and GSK-3β.
The team then tested the significance of these findings in vivo using in-utero electroporation to introduce DISC1 constructs in developing mouse embryos. Co-transfecting cells with one of the DISC1 variants along with DISC1 shRNA to decrease endogenous DISC1 revealed that the same three variants with subpar stimulation of the Wnt pathway—R264Q, L607F, and A83V—could not rescue the DISC1 knockdown, and the resulting brains exhibited reduced neural proliferation. This suggests that these variants are loss of function. A similar experiment was done in zebrafish, which display a number of abnormalities when the zebrafish version of DISC1 is knocked down, including misshapen brains, disorganized muscle structure, and disrupted axon tract formation. Human wild-type DISC1 and the S704C variant could rescue these malformations, but the other two common variants tried—R264Q and L607F—could not.
To extend these results to human tissue, the researchers then turned to human lymphoblast cell lines (LCLs) from healthy and bipolar patients. Sorting the cells by DISC1 genotype, the researchers found allele-dependent differences in Wnt signaling in the cells, as assayed by the introduced TCF-LEF luciferase reporter. For example, cells homozygous for RR264 (major allele) had nearly three times more Wnt pathway activity than cells homozygous for the minor allele (264QQ) did. What’s more, among cells with the same R264Q genotype, the Wnt activity in cells from bipolar patients was significantly reduced compared to that in controls, suggesting other interacting factors.
Finally, the researchers turned to the S704C variant, which acted like wild-type DISC1 in all of their assays of Wnt signaling. Did this variant impact one of the other functions for DISC1? Tipped off by the fact that this variant lies in a region of the protein that interacts with neural migration factors NDEL1 and Dixdc1 (see SRF related news story), the researchers tested its effects on migration using their in-utero electroporation paradigm. With DISC1 knockdown arresting neural migration, the S706C variant could not rescue migration, leaving neurons stuck below the cortical plate. This likely involves the ERK pathway, and together these findings point to DISC1 variants that impact different pathways with distinct roles in development.
Another partnership for DISC1
Focusing on the adult brain, first author Eunchai Kang and colleagues examined DISC1’s role in newborn cells in the mouse hippocampus. When FEZ1 expression was decreased in these cells via retrovirus-introduced shRNAs, newborn neurons had increased soma size and more extensive dendritic trees two weeks later. Co-expressing FEZ1 along with the shRNA could restore the soma and dendrite overgrowth to their correct proportions. The phenotype resulting from FEZ1 knockdown matches some features found in DISC1 knockdown experiments (Duan et al., 2007), but not others, such as inappropriate positioning or ectopic dendrites. This suggested that FEZ1 might help carry out a subset of DISC1’s functions. To look for a functional interaction between these two genes, the researchers knocked down both FEZ1 and DISC1, and found a worse phenotype with respect to dendrite overgrowth. This synergy suggests that the dendrite overgrowth involves both FEZ1-dependent and FEZ1-independent pathways.
To look more directly at the DISC1-FEZ1 interaction, the researchers developed a blocking peptide to bind to the DISC1 region FEZ1 normally binds to. When introduced in adult neural progenitors, this blocking peptide reduced interaction between endogenous DISC1 and FEZ1 without disrupting DISC1’s interaction with NDEL. This fully copied the FEZ1 knockdown results, with dendrite overgrowth and increased soma size; these effects were independent from the DISC1-NDEL1 interaction, which regulated positioning of neurons and their processes. The DISC1-FEZ1 interaction was further distinguished from the DISC1-NDEL1 one in co-immunoprecipitation experiments that found that, though NDEL1 and FEZ1 form a complex with DISC1, they did not directly interact with each other without DISC1.
The researchers then brought this cell biology-derived insight of a DISC1-FEZ1 interaction to a genetic association study involving 279 cases of schizophrenia and 249 controls. When four SNPs within the FEZ1 gene were examined alone, none were significantly associated with schizophrenia. But when the FEZ1 SNPs were considered in conjunction with a DISC1 SNP (S704C), evidence for a genetic interaction, or epistasis, turned up. Specifically, carriers of the C allele at one of the FEZ1 SNPs had a 2.5-fold increased risk of schizophrenia only when they were also homozygous for the S DISC1 allele (p = 0.029). A replication cohort from the GAIN sample set confirmed FEZ1-DISC1 epistasis, but the details of the genotypes differed.
The researchers end by suggesting that this kind of interaction could reconcile the common-versus-rare argument of genetic susceptibility for disorders like schizophrenia (see SRF related news story). Disrupting DISC1, a node of multiple, distinct pathways, would confer greater risk than disrupting a single pathway leading away from DISC1. The study also shows how biology may inform genetic studies—not just the other way around—and both studies illustrate that a functional emphasis may help make sense of the very long list of genes implicated so far in schizophrenia.—Michele Solis.
Wexler EM, Geschwind DH. DISC1: A Schizophrenia Gene with Multiple Personalities. Neuron. 2011 Nov 17; 72: 501-503. Abstract
Singh KK, De Rienzo G, Drane L, Mao Y, Flood Z, Madison J, Ferreira M, Bergen S, King C, Sklar P, Sive H, Tsai LH. Common DISC1 Polymorphisms Disrupt Wnt/GSK-3β Signaling and Brain Development. Neuron. 2011 Nov 17; 72: 545-558. Abstract
Kang E, Burdick KE, Kim JY, Duan X, Guo JU, Sailor KA, Jung DE, Ganesan S, Choi S, Pradhan D, Lu B, Avramopoulos D, Christian K, Malhotra AK, Song H, Ming G. Interaction between FEZ1 and DISC1 in Regulation of Neuronal Development and Risk for Schizophrenia. Neuron. 2011 Nov 17; 72: 559-571. Abstract