10 April 2008. Does a seven-year-old mouse model that mirrors schizophrenia traits bring good luck? The mouse, missing BACE1 (β-site amyloid precursor protein (APP) cleaving enzyme 1) and engineered to study Alzheimer’s disease, debuted in 2001 (see Cai et al., 2001), but in last week’s PNAS online, the lab that developed the mouse reported that it may also be a good model for endophenotypes of schizophrenia. Philip Wong and colleagues at the Johns Hopkins University, Baltimore, Maryland, show that the mouse exhibits behavioral traits and endophenotypes associated with this psychiatric disorder, and that some of these symptoms can be relieved with an atypical antipsychotic. The BACE1-negative mouse might prove valuable for studying the pathogenesis of schizophrenia and also for evaluating potential new treatments.
That an Alzheimer’s disease-related protease should be linked to schizophrenia is not so far-fetched. In addition to cleaving APP, BACE1 also processes neuregulin 1 (NRG1), a protein of intense interest in schizophrenia research. The NRG1 gene has popped up again and again as a risk factor in genetic association studies (see SRF related news story and SchizophreniaGene), and there is also evidence that this genetic risk is elevated when accompanied by genetic changes in the gene for ErbB4, one of neuregulin’s binding partners (see SRF related news story and SchizophreniaGene). That the NRG1 gene variations occur in non-coding regions suggests that it is a change in gene expression, rather than function, that may be causative. In fact, altered expression of some subtypes of NRG1 (there are four major isoforms) have been linked to genetic variants in both control and schizophrenia brains (see SRF related news story) and levels of both NRG1 and ErbB4 were recently reported to be increased in the prefrontal cortex in schizophrenia patients (see Chong et al., 2008).
The growing links between NRG1/ErbB4 signaling and schizophrenia suggest that any major change to this pathway, such as abolishing BACE1 cleavage of NRG1, might be pathological. To explore this idea, first author Alena Savonenko and colleagues examined their BACE1-null animals for a variety of psychological and behavioral traits that have parallels in schizophrenia patients. Though it is unlikely that any one mouse model might recapitulate the full range of positive and negative symptoms found in patients, the BACE1-negative animals have quite a few.
Savonenko and colleagues found that the animals had reduced prepulse inhibition (PPI), a phenomenon where a brief faint sound reduces the startle reflex to a loud noise that follows soon after. Many people with schizophrenia have impaired PPI. The mice were also hyperactive when presented with novel stimuli in various environments (open field, Y maze, plus maze), which may reflect the agitation that patients sometimes feel. BACE1-negative mice were also highly sensitive to low levels of the psychostimulant MK-801, suggesting they are on the cusp of psychosis. MK-801 is an antagonist of the NMDA-type glutamate receptor, and reduced glutamatergic transmission is potentially a major facet of this disease. The animals also had reduced cognitive skills, as exemplified by poor performance in tasks that rely on working memory, in this case, navigating a radial water maze. And they showed altered social behavior—they found it difficult to distinguish a new mouse from one with which they were already familiar. That these traits in mice model similar behavior in patients is supported by the effects of clozapine, an atypical antipsychotic prescribed for some people with schizophrenia. On the drug, BACE1-null mice were no longer hyperactive and they had normal PPI.
How does eliminating BACE1cause these schizophrenia-like traits? The protease is necessary for maturation of NRG1, which in turn binds to ErbB4. The authors found that in BACE1-nulls, the distribution of ErbB4 is not normal. Specifically, the amount of the kinase associated with the post-synapse is reduced, as judged by binding to post-synaptic density 95 (PSD95), a protein marker of post-synapses. Recent work from Sean P. Murphy and Keren Bielby-Clarke at the University of Nottingham, England, suggests that ErbB4-PSD95 binding is crucial for NRG1 signaling (see Murphy and Bielby-Clarke, 2008).
If BACE1 reduces NRG1 signaling, then how does this translate into schizophrenia-like traits? One possibility is that synaptic transmission is compromised. NRG1 has been implicated in myelination (see SRF related news story) and in the regulation of transmission through both glutamatergic and GABAergic neurons (see SRF related news story), which have both been implicated in this disorder. In fact, Savonenko and colleagues found that the density of dendritic spines in the hippocampal neurons from BACE1-nulls is significantly reduced compared to wild-type animals, tying in faulty NRG1 signaling with neurotransmission deficit.
“Our findings identify BACE1-/- mice as a rodent model that exhibits schizophrenia-like behavioral abnormalities and suggest that genetic or epigenetic alteration of BACE1 may participate in the development of some schizophrenic symptoms in individuals with this complex psychiatric disorder,” conclude the authors. That remains to be determined, as does the role of the various NRG1 isoforms. BACE1 is known to proteolytically cleave types I and III neuregulin, and the structure of the newly identified, type IV, which has been linked to schizophrenia (see SRF related news story), suggests it may be a BACE1 substrate, too. Time will tell.—Tom Fagan.
Savonenko AV, Melnikova T, Laird FM, Stewart KA, Price DL, Wong PC. Alteration of BACE1-dependent NRG1/ErbB4 signaling and schizophrenia-like phenotypes in BACE1-null mice. PNAS 2008 April 2, early edition. Abstract