22 September 2008. Four new mouse studies sketch how molecular alterations involving neuregulin1 (Nrg1) might lead to neurological and behavioral deficits in schizophrenia. In one, Bart De Strooper of the University of Leuven and Flanders Institute of Biotechnology, in Leuven, Belgium, and colleagues followed leads gained from their work on Alzheimer’s disease. Their July 15 PNAS report hints that turning off a gene that helps split proteins, including Nrg1, in the cell membrane produces cognitive deficits and neuropharmacological responses that parallel some of those seen in schizophrenia.
The other three studies zeroed in on a type of Nrg1 expressed mostly or totally in the nervous system. In the July 2 issue of the Journal of Neuroscience, a team led by David Talmage and Lorna Role, both of whom moved recently from Columbia University in New York City to Stony Brook University in Stony Brook, New York, conclude that type III Nrg1 plays crucial roles in brain structures involved in filtering sensory information and using short-term memory. Talmage's group also has a paper in the May 5 Journal of Cell Biology, suggesting that type III Nrg1 controls expression of α7 nicotinic acetylcholine receptors (nAChRs) along sensory neuron axons by acting as a receptor itself, instead of playing its better-known ligand role. Finally, in the September 10 issue of the Journal of Neuroscience, Role and colleagues link aspects of both these lines of research in describing a critical role for type III Nrg1, functioning as a receptor, in α7nAChR-mediated neurotransmission between hippocampus and striatum.
Complex molecular scissors
The gene that encodes Nrg1 has emerged as a leading schizophrenia candidate gene, notwithstanding the less-than-encouraging findings from the recent SchizophreniaGene meta-analysis (see SRF related news story). The protein facilitates processes that range from neural development to adult synaptic functioning, but only if split into the proper segments by a series of enzymes. One step involves snipping off a fragment that hangs outside the cell; this becomes the ligand for the ErbB4 receptor. Another occurs when the “stub” that remains lodged within the cell membrane subsequently undergoes splitting by γ-secretase, releasing it back into the cell.
The intramembrane protease γ-secretase may be best known for processing amyloid precursor protein into the amyloid-β that forms plaques in Alzheimer’s disease. In humans, the large γ-secretase complex includes one of the two forms of the protein Aph1, either the A or B variant. Mice carry an additional gene that encodes Aph1C, a twin of Aph1B; the two genes are typically referred to jointly as Aph1B/C.
De Strooper’s laboratory had been examining the role of the Aph1B/C-containing γ-secretase in the processing of amyloid precursor protein (Serneels et al., 2005). He and his colleagues wondered whether it also plays a part in the cleaving of Nrg1 and, hence, contributes to schizophrenia.
The causes of snipping snafus
In their new study, first author Tim Dejaegere and his coauthors examined the processing of Nrg1 in Aph1B/C knockout mice. They wanted to learn what Aph1B/C does; unlike Aph1A, it plays no crucial role in survival, so Aph1B/C KO mice survive to adulthood with no overt ill effects.
Dejaegere and colleagues found abnormal buildup of Nrg1 fragments in a number of brain areas in the knockout mice, suggesting that Aph1B/C is indeed critical for the proper cleaving of Nrg1. They identified the fragments not as the peptide that signals through ErbB4 receptors, but rather as the membrane-bound stub that should have been cleaved by Aph1B/C-γ-secretase.
In an interesting piece of supporting investigation, the researchers examined how deficits in Nrg1, as opposed to the γ-secretase complex, might lead to the same end. They focused on a particular single nucleotide polymorphism (SNP), in the coding region of the Nrg1 gene, reported to increase schizophrenia risk (Waiss-Bass et al., 2006). This SNP causes a valine-to-leucine substitution in the transmembrane region, where it might disrupt γ-secretase cleavage of the Nrg1 membrane-bound stub. Indeed, when Dejaegere and colleagues introduced this mutation to type III Nrg1 in otherwise normal cells in culture, they found above-normal accumulation of the stub.
Also of interest, in normal mice, the researchers found high levels of expression of Aph1B/C in the cerebral cortex, hippocampus, olfactory bulb, and cerebellum, but not the striatum. “Aph1B/C expression is enriched in brain areas with relevance for schizophrenia,” they write. They note that the areas with the most expression in the prefrontal cortex (PFC) and hippocampus, layer 5 and CA1, respectively, include those that send excitatory glutamatergic signals to the ventral striatum and other brain areas (for information on glutamate’s possible role in schizophrenia, see SRF current hypothesis by B. Moghaddam, as well as paper by D. Javitt cited below).
The consequences of snipping snafus
To illuminate the functions of Aph1B/C, the researchers compared knockout mice with their normal littermates using paradigms that often stand in for behavioral aspects of schizophrenia—specifically, tests of prepulse inhibition (PPI) and working memory. PPI, a common experimental paradigm for sensory gating, occurs when a weak stimulus lessens the startle response to a subsequent stronger stimulus. Studies have found PPI deficits in some patients with schizophrenia, suggesting that they cannot filter sensory information well. PPI tasks are often used to test putative animal models of schizophrenia pathology.
In the study by Dejaegere and colleagues, the Aph1B/C knockout mice showed PPI impairments, which disappeared after injection of haloperidol, a typical antipsychotic, or clozapine, an atypical antipsychotic. Since all current antipsychotic drugs block dopamine D2 receptors to alleviate psychosis, the team assessed dopaminergic functioning by administering amphetamine. This drug not only induces psychosis in otherwise healthy people, it worsens the symptoms of people with schizophrenia, apparently by promoting dopamine release in the ventral striatum.
The scientists found that the amphetamine revved up the Aph1B/C knockout mice more than their normal siblings. Furthermore, the knockouts showed evidence of excessive dopamine turnover in tissue from the ventral striatum. These findings dovetail with a prominent theory that puts dopamine hyperactivity in the mesolimbic pathway at the root of schizophrenia (see SRF current hypothesis). This pathway links the ventral tegmentum in the midbrain to the ventral striatum.
Dejaegere and colleagues also examined other possible neurotransmitter disturbances. “A hyperdopaminergic state of the ventral striatum can be a consequence of a primary PFC dysfunction of glutamatergic signaling,” they note. Since tests of working memory can gauge PFC disturbances, the researchers subjected the rodents to a working memory test in which recently presented information would help them navigate a maze. The knockout mice performed worse than their wild-type siblings.
To further explore glutamate’s role in the knockout mice’s behavior, the researchers injected both groups of mice with MK-801. This drug, used in a standard animal model of schizophrenia, blocks N-methyl-D-aspartate (NMDA) receptors. These glutamate receptors probably play a key role in memory, attention, and other cognitive functions (see SRF current hypothesis). Consistent with past research, MK-801 impaired PPI in normal mice, but it weakened it even more in the knockouts, to the point of nearly eliminating it. These findings relate the flawed filtering of sensory information to disturbed glutamatergic signaling involving NMDA receptors.
The investigators conclude, “Dysregulation of intramembrane proteolysis of Nrg1 could increase risk for schizophrenia and related disorders” by unleashing ripples through the glutamatergic and dopaminergic pathways. They view their data as supporting both the glutamate and dopamine hypotheses of schizophrenia.
In related work, some members of the research team have been exploring whether silencing Aph1B/C improves symptoms in a mouse model of Alzheimer’s disease. Their schizophrenia study, although it did not look at amyloid processing, warns that treating Alzheimer’s in such a way could raise schizophrenia risk. Hedging their bets, Dejaegere and colleagues write, “Because a wealth of data supports the contention that the development of schizophrenia depends on a neuronal insult early in life, it is possible that suppression of Aph1B/C activity in the brain in adulthood only will have no severe side effects at all.”
Portrait of a mouse with too little type III
Different gene promoters and ways of splicing result in over 15 isoforms of Nrg1. Enzyme-mediated cleavage snips most forms from the cell surface, enabling them to diffuse and affect distant targets. Only type III stays bound to the cell membrane, limiting its signaling to bordering cells ("juxtacrine" signaling”).
The uniqueness of type III, combined with its neuronal-specific expression, has intrigued the two closely collaborating groups of Talmage and Role for some time. In their recent study in the Journal of Neuroscience, led by first author Ying-Jiun Chen, currently of the University of California at San Francisco, they extended their research on its effects on neural circuits and behaviors that relate to schizophrenia. In addition, they sought to learn what regulates Nrg1 expression.
Chen and colleagues compared adult mutant mice that were heterozygous knockouts for type III Nrg1 with their wild-type siblings on structural, functional, and behavioral measures. They looked at heterozygotes because mice that produce no type III Nrg1 die at birth. The authors note that the heterozygotes had larger cerebral lateral ventricles than the wild-type mice. In addition, they had reduced density of dendritic spines, which receive excitatory input, on pyramidal neurons in the subiculum. These neurons convey processed information from the hippocampus to other brain areas such as the ventral striatum. The researchers liken these structural anomalies to those found in autopsy studies of humans with schizophrenia.
Further clues indicate that type III Nrg1 affects not only brain structure, but also function. In the heterozygotes, Chen and colleagues found evidence of low metabolism in hippocampal regions CA3 and CA1, the subiculum, and the medial prefrontal cortex. Such structural and functional changes might trigger abnormal behavior. On a maze test that required use of short-term memory and on an auditory PPI task, the heterozygotes performed worse than their wild-type peers. Interestingly, when the researchers had characterized the normal distribution of type III in the mouse brain, they found that it is "expressed in the medial prefrontal cortex, ventral hippocampus, and ventral subiculum, regions involved in the regulation of sensorimotor gating and short-term memory.”
In prior research, PPI deficits in mice and subjects with schizophrenia returned to normal after nicotine administration. Furthermore, many people with schizophrenia smoke, perhaps to self-medicate. After Chen and colleagues gave mice water laced with nicotine for six weeks, the heterozygotes performed as well as the wild-type mice.
Summing up their results, Chen and colleagues write, “Decreased expression of type III Nrg1 leads to structural, functional, and behavioral alterations that are related to schizophrenia.” This study, however, did not test whether the structural and functional changes directly affect behaviors related to schizophrenia in mice, let alone humans with the disease.
Backwards signaling targets nicotinic receptors
Complementing the work by Chen and associates, the remaining study from the Talmage and Role groups looked at the relationship between Nrg1 and nAChRs. Nicotine docks on these receptors, and past research hints that type III Nrg1 regulates expression of nAChR subunits, including α7. Research on animals, as well as people with schizophrenia and their relatives, links α7-containing nAChRs to sensory gating and to schizophrenia (for a literature review, see Martin and Freedman, 2007; see SRF related news story; SRF news story).
This other study, led by first author Melissa Hancock, investigated the molecular mechanisms by which type III Nrg1 affects presynaptic α7-containing nAChRs. When the researchers examined sensory neurons from embryos of type III Nrg1 knockout mice, they detected a significant reduction, relative to wild-type, of axonal α7nAChR surface expression. There was no overall reduction of this receptor subunit in the neurons, suggesting that trafficking to the cell membrane was specifically impaired.
The next challenge involved figuring out what role type III Nrg1 plays in α7 trafficking. In forward signaling, Nrg1 acts as a ligand for ErbB receptor tyrosine kinases (see SRF related news story). The Nrg1 fragment that is cleaved from the outside of the cell interacts with receptors on target cells some distance away. However, in this case, the researchers suspected that a different process, dubbed “back signaling” (Bao et al., 2003), might play a role. This presynaptic signaling occurs when the fragment that remains within the cell membrane acts as a receptor for ErbB, spurring the cell that expressed type III Nrg1 to fire.
To find out if type III Nrg1 controls nAChR expression by acting as a ligand or a receptor, Hancock and colleagues treated the neurons with extracellular domain from either ErbB4 (B4-ECD) or ErbB2 (B2-ECD). B4-ECD binds tightly with Nrg1; since B2-ECD does not, it served as a control substance. Only B4-ECD increased expression of α7 nAChRs along axons from the wild-type mice; neither substance did so on cells from the mutants.
When the researchers repeated the experiment with the addition of substances that hinder ErbB tyrosine kinase activity, α7 nAChRs still gathered along the axons. They inferred that type III controls the insertion of α7 nAChRs into the membrane by acting as a receptor for ErbB4 and not as a ligand that launches tyrosine kinase signaling.
The investigators found that the increased surface expression did not stem from ramped-up production of α7 nAChRs. Rather, back signaling appeared to shift existing nAChRs from caches inside the neuron to the axon membrane. It did so by activating a phosphatidylinositol 3-kinase signaling pathway that controls the movement of cell surface proteins. “These findings, in conjunction with prior results establishing that Type III Nrg1 back signaling controls gene transcription, demonstrate that Type III Nrg1 back signaling can regulate both short- and long-term changes in neuronal function,” Hancock and colleagues write.
Chimeras are different
The most recent installment in the Talmage-Role collaboration links aspects of both the Chen and Hancock papers. In a study published in the September 10 issue of the Journal of Neuroscience, Role, first author Chongbo Zhong, and colleagues examined the role of type III Nrg1 in synaptic transmission in a hippocampus-to-ventral striatum circuit, which they note, is "an important relay in sensory-motor gating."
The experimental method is a major technical challenge—creating a chimeric circuit by "hooking up" the output area of the hippocampus—ventral CA1 and subiculum—from a type III Nrg1 deficient mouse to the target region in the ventral striatum—the nucleus accumbens—from a normal mouse. This is all done in a cultured brain slice preparation, of course, where an excised piece of hippocampal tissue is allowed a day to accommodate itself to life in vitro before nucleus accumbens neurons from a different brain are added. Within a week in culture, the hippocampal neurons form synaptic connections with accumbal cells, and the researchers are able to record glutamatergic synaptic activity from the acumbal target cells in the chimeric circuits.
Whether the hippocampal cells came from mice with normal or reduced levels of type III Nrg1 made no difference in their effects on baseline postsynaptic activity of the nucleus accumbens cells. But when Zhong and colleagues added nicotine, they noticed a major difference. In wild-type-to-wild-type circuits, a single puff of nicotine boosted neurotransmission in the circuit for up to an hour. But when the hippocampal cells had reduced levels of type III Nrg1, nicotine produced only a brief enhancement of glutamatergic transmission in the circuit. "Alterations in this temporal profile might lead
to deficits in sensory gating by altering glutamatergic transmission in corticostriatal
circuits," the authors write.
Underlying this deficiency, the researchers report, is an 80 percent reduction in α7nACh receptors on the axons of the hippocampal cells. Zhong and colleagues were able to reverse this situation by incubating the hippocampal cells with the ErbB4 extracellular domain, supporting the evidence from the earlier paper by Hancock and colleagues that back signaling by type III Nrg1 controls the levels of α7nACh receptors on these axons.
These four new studies contribute to a framework that may eventually tie together the genes that encode Nrg1, ErbB4, and α7 as potential contributors to cognitive deficits in schizophrenia. They may also fuel suspicions that the pathways underlying schizophrenia and neurodegenerative diseases overlap.—Victoria L. Wilcox and Hakon Heimer.
Dejaegere T, Serneels L, Schäfer MK, Van Biervliet J, Horré K, Depboylu C, Alvarez-Fischer D, Herreman A, Willem M, Haass C, Höglinger GU, D’Hooge R, De Strooper B. Deficiency of Aph1B/C-gamma-secretase disturbs Nrg1 cleavage and sensorimotor gating that can be reversed with antipsychotic treatment. PNAS. 2008 July 15;105(28):9775-9780. Abstract
Hancock ML, Canetta SE, Role LW, Talmage DA. Presynaptic type III neuregulin 1-ErbB signaling targets alpha7 nicotinic acetylcholine receptors to axons. J Cell Biol. 2008 May 5;181(3):511-521. Abstract
Chen Y-J J, Johnson MA, Lieberman MD, Goodchild RE, Schobel S, Lewandowski N, Rosoklija G, Liu R-C, Gingrich JA, Small S, Moore H, Dwork AJ, Talmage DA, Role LW. Type III neuregulin-1 is required for normal sensorimotor gating, memory-related behaviors, and corticostriatal circuit components. J Neurosci. 2008 July 2;28(27):6872-6883. Abstract
Zhong C, Du C, Hancock M, Mertz M, Talmage DA, Role LW. Presynaptic type III neuregulin 1 is required for sustained enhancement of hippocampal transmission by nicotine and for axonal targeting of alpha7 nicotinic acetylcholine receptors. J Neurosci. 2008 Sep 10;28(37):9111-6. Abstract