5 January 2009. Despite the failures of some recent studies to support the genes for neuregulin or its receptor ErbB4 as schizophrenia susceptibility factors (e.g., Sanders et al., 2008), many geneticists still refer to NRG1 as a top candidate based on the strength of the original finding (Stefansson et al., 2002) and the low probability that subsequent replications could be due to chance. As with other questionable candidate genes identified by their biological plausibility and location in top linkage regions, it is proposed that different SNPs within Nrg1 could underlie the disease in different populations, or even that Nrg1 might be a schizophrenia susceptibility gene in some populations and not others.
Because of the continuing interest in NRG1 from both geneticists and basic scientists, Joshua A Gordon, of Columbia University in New York City, and Amelia L. Gallitano, of the University of Arizona in Phoenix, organized a mini-symposium on "Integrative Approaches to Candidate Gene Models of Schizophrenia: Exploring the Neuregulin/ErbB4 Pathway," on Monday, October 17, at the Neuroscience 2009 meeting.
Characterizing cellular pathways that incorporate multiple schizophrenia candidate genes may help elucidate the pathophysiology of this devastating illness. Multiple components of the signaling pathway acting downstream of the candidate gene, neuregulin1, and its receptor, ErbB4, have been linked to schizophrenia. This mini-symposium will explore the consequences of manipulating individual elements of this pathway, addressing their functional and clinical relevance.
Figure 1. Pictorial overview of symposium (by Amelia Gallitano; based in part on Mei and Xiong, 2008). (View larger image.)
In her introductory remarks, Gallitano noted some of the fundamental aspects of this ligand receptor duo: Nrg1 is a neurotrophic factor that begins its functional life after it has been inserted as a transmembrane protein. The extracellular domain can be cleaved to interact with the ErbB family of receptor tyrosine kinases on postsynaptic membranes, but there is also evidence that Nrg1 can "back-signal" to its home cell.
Bo Li of Cold Spring Harbor Laboratory in New York gave the first talk: "The schizophrenia-linked protein neuregulin1 and its receptor ErbB4 control glutamatergic synapse maturation and plasticity." Li reported evidence that synaptic activity in glutamatergic synapses regulates the distribution of ErbB4 receptors on the surface of dendritic spines, and that Nrg1 is the ligand for this activity (see SRF related news story). He and his colleagues have also found evidence that the relationship is a two-way street—Nrg1-ErbB4 signaling in turn regulates synaptic signaling. RNAi knockdown experiments of ErbB4 indicated that without Nrg1-ErbB4 signaling, synapses do not mature properly, there are changes in dendritic spine morphology, and long-term potentiation is reduced. The stabilization of AMPA receptors appears to be one mechanism through which this regulation is achieved, according to Li.
In the next talk, Lin Mei of the Medical College of Georgia in Augusta discussed "Neuregulin1/ErbB4 in neurotransmission and synaptic plasticity." Mei first surveyed some of the very complex signaling of Nrg1 isoforms (for review, see Mei and Xiong, 2008), and noted that the candidate neuregulin risk SNPs are in the promoter and suggested that any resultant changes in expression would lead to a gain of function in Nrg1-ErbB4 signaling overall.
Mei then turned to his group's work in inhibitory GABAergic neurons in prefrontal cortex, where they have identified a role for Nrg1-ErbB4 signaling in controlling the release of GABA (some of this work is described in an SRF related news story). The synaptic terminals of these inhibitory neurons are enriched in ErbB4, and Mei's group has found that Nrg1 enhances the release of GABA via activation of ErbB4.
Moving their analysis over to the other side of the inhibitory synapse, Mei and colleagues have found that Nrg1 regulates pyramidal cell activity, leading them to hypothesize that the apparent hypofunction of glutamatergic pathways in schizophrenia could be in part due to an increase in Nrg1 activity leading to increased GABA release onto pyramidal neurons. In support of this, Mei and colleagues' experiments indicate that the major target of Nrg1 in prefrontal cortex is parvalbumin-positive interneurons (see SRF interview with David Lewis). When ErbB4 is knocked out specifically in these neurons, mice show putative schizophrenic phenotypes.
The session shifted gears from electrophysiology to gene expression research, with a talk by Tsuyoshi Miyakawa, of Fujita Health University, Toyoake, Japan, on "Loss of α-CamKII results in dentate gyrus immaturity, a candidate endophenotype for schizophrenia." He and his colleagues have done large-scale behavioral testing of 90 strains of genetically engineered mice, and most show some behavioral variation. In earlier work with Susumu Tonegawa at MIT, he had found that mice with forebrain-specific calcineurin deficits displayed possible schizophrenia-like phenotypes (Miyakawa et al., 2003). This group subsequently reported an association of the calcineurin gene, PPP3CC, with schizophrenia susceptibility in humans (Gerber et al., 2003). These results have led Miyakawa et al. to take a closer look at the behavior of mice with mutations in calcineurin pathway genes, and they have found behavioral abnormalities with mutations in the gene for αCAMKII, a protein kinase that phosphorylates serine or threonine residues. The link to the theme of the symposium is that αCAMKII and calcineurin are NMDA receptor targets regulated by NRG1/ErbB4 signaling.
In a recently published study, the researchers report that mice with a heterozygous knockout of αCAMKII have a number of abnormalities in hippocampus, including apparent deficits in adult neurogenesis, evidence Miyakawa and colleagues have dubbed an "immature" dentate gyrus (Yamasaki et al., 2008; see also editorial: Frankland et al., 2008). In more recent work, they have looked for other mouse models with "immature" dentate gyri, finding similar phenotypes in mice lacking the gene for the zinc-finger protein Schnurri-2 (SHN2). Hippocampal gene expression patterns in these mice parallel those of the αCAMKII mice.
Is there any evidence for "immature" hippocampus in human disorders? In their recent paper, Miyakawa's group report that gene expression pattern analysis, predicated on the patterns seen in hippocampi of αCAMKII mutant mice, clustered 16 of 18 schizophrenic hippocampal brain samples together. Many of the abnormally expressed genes are involved in neurodevelopmental processes.
Session co-organizer Amelia Gallitano, followed with the talk, "Dysfunction of the immediate early gene transcription factor Egr3 is implicated in schizophrenia." In her introduction, Gallitano positioned her research within the quest for biological mediators of the environmental component of schizophrenia risk (e.g., in utero infections, traumatic birth, or stressful life events), which are estimated to account for about half the etiology of the disorder. Immediate early genes (activated within 30 min. of a stimulus) are a prime candidate, as they are known to be activated by environmental stressors.
Gallitano has been investigating the potential role of the transcription factor Egr3, which is activated by immediate early genes, in schizophrenia based on a number of different avenues of evidence for several years. Her interest in the gene received further support from the research reported by Tonegawa and colleagues on calcineurin, which regulates Egr3. Egr3 is also regulated by NMDA receptors and by Nrg1.
Gallitano has found that Egr3 knockout mice show some of the phenotypes proposed as animal models of schizophrenia psychopathology, including hyperactivity in a novel environment, an abnormal startle response, and working memory deficits, as well as hypersensitivity to stress. Some of these behavioral abnormalities are rescued by antipsychotic drugs; of particular interest is the finding that Egr3 knockout mice are relatively resistant to the sedating effects of clozapine, as are schizophrenia patients.
While hippocampal long-term potentiation is apparently normal in Egr3 knockout mice, long-term depression cannot be induced normally, a finding that may be traceable to the NR2B subunit of the NMDA receptor. These results and others, Gallitano suggested, may place Egr3 at "the convergence of stress and genetic susceptibility to schizophrenia."
In the next talk, Malcolm Nason, State University of New York, Stony Brook, discussed "Neuregulin-1 Type-III heterozygous mice as a model for schizophrenia: behavior, anatomy and physiology." He reviewed several recent studies from the laboratories of Lorna Role and David Talmage (several of which are described in detail in this SRF overview of recent Nrg1 studies). Briefly, these mice have behavioral deficits and regional neurobiological abnormalities. More recently, Nason has found evidence of a "disorganized" electrophysiologic response in the nucleus accumbens in response to input from the ventral hippocampus. Specifically, accumbens neuron activity is less synchronized to ventral hippocampal oscillatory input.
Session co-organizer Joshua Gordon finished the session with a talk on his collaborations with the Role and Talmage groups: "Neural correlates of cognitive dysfunction in Nrg1 heterozygous mice." These mice have deficits in spatial working memory, which requires an intact hippocampus. In awake behaving recordings from Nrg1 heterozygote and wild-type mice during working memory tasks, Gordon and colleagues have noted decreased gamma band oscillations in the nucleus accumbens in the mutants. However, in contrast to Nason's findings, the hippocampal modulation of the accumbens neurons appears to be normal. What the researchers do find is an increase in hippocampal activity, suggesting a compensatory hyperactivity in the hippocampus of Nrg1 heterozygous knockout mice that may be able to overcome the deficits in the hippocampal-accumbens circuit noted by Nason and colleagues.—Hakon Heimer.