Schizophrenia Research Forum - A Catalyst for Creative Thinking

Fragile X, MHC Proteins Shape Synapses

Adapted from a story that originally appeared on the Alzheimer Research Forum.

19 April 2007. The ability of synapses to strengthen or weaken, form anew, or disappear entirely underlies the brain’s capacity to establish functional neuronal circuits during development and to learn and remember later on in life. Two new papers probe the roles of distinct proteins that regulate this process. Work on the Fragile X syndrome protein FMRP, and the immune receptor/neural regulator MHC Class I reveal new insights into how these proteins act within the complex ensemble of players that regulate synapse dynamics.

Neither protein has been tied to schizophrenia, but the development of proper circuitry (Weinberger, 1987; Murray and Lewis, 1987) and the function of mature synapses (for review, see Harrison and Weinberger, 2005; Owen et al., 2005) are certainly major foci of interest among schizophrenia researchers.

In a paper published online April 8 in Nature Neuroscience, Claudia Bagni from the Universita “Tor Vergata” in Rome, Italy, and colleagues from Scotland and England describe an additional target for FMRP, the messenger RNA for the postsynaptic density 95 (PSD-95) protein. In contrast to other mRNAs whose translation FMRP regulates, the new data shows that FMRP increases the stability of PSD-95 mRNA. Led by first authors Francesca Zalfa, Boris Eleuteri, and Kirsten Dickson, the researchers show that the stabilization is further enhanced by glutamate receptor activation. In mice lacking the FMRP protein, PSD-95 mRNA and protein levels are decreased in the hippocampus.

Although a large number of potential FMRP targets have been identified in recent years, only a few are known to play a role in regulating synapse structure and function, the authors write. The addition of PSD-95 to this short list supports the idea that FMRP is important for establishing proper synaptic structure, function, and plasticity. Moreover, the dysregulation of PSD-95 could be a factor in the cognitive impairment seen in Fragile X syndrome. There have been tentative links between PSD-95 and schizophrenia, including a finding of decreased expression in postmortem tissue (Clinton and Meador-Woodruff, 2004).

A second report, from Carla Shatz and colleagues at the Harvard Medical School in Boston, Massachusetts, and Stanford University in Palo Alto, California, shows how another set of proteins function from a postsynaptic location near PSD-95 to regulate plasticity in response to synaptic activity. MHC Class I (MHCI) proteins are a well-known player in the immune system, and several years ago Shatz showed their unexpected involvement in activity-dependent synaptic remodeling during development and in the adult brain (see SRF related news story).

In the new work, out April 9 in PNAS online, Alex Goddard, Daniel Butts, and Shatz first showed by immunostaining that MHCI is localized in postsynaptic areas of cultured hippocampal neurons, with a distribution that largely overlaps with that of PSD-95. Then, the scientists found several abnormalities in the synaptic structure and function of mice lacking MHCI. Cultured neurons or brain slices from knockout mice showed increased frequency of mini-EPSCs, suggesting that basal transmission was abnormal. The enhanced presynaptic activity was associated with a modest increase in the size of presynaptic boutons, and slightly elevated numbers of synaptic vesicles.

These functional and structural changes in knockout mice resembled the changes seen in neurons where synaptic activity is blocked using tetrodotoxin (TTX). When the researchers treated MCHI knockout neurons with TTX, the neurons showed no further changes in synaptic activity or structure, consistent with their having already adopted a TTX-treated synaptic phenotype. The researchers reasoned that down regulation of MHCI could be responsible for TTX-induced synaptic changes. Indeed, Shatz and colleagues first identified MHCI in the nervous system because its expression was reduced in the brain of TTX-treated cats (Corriveau et al., 1998). In the current work, they found that treating cultured neurons from wild-type mice treated with TTX resulted in the expected synaptic changes and lower expression of MHCI mRNA and protein.

“A major finding of this study is that MHCI is part of the molecular machinery regulating synaptic morphology and function under basal conditions and following action potential blockade,” the authors write. They note that postsynaptic MHCI also appears to act across the synapse to change presynaptic structures in response to activity.—Pat McCaffrey.

Comments on Related News

Related News: Neuregulin, ErbB4—Levels Normal but Signaling Strengthened in Schizophrenia

Comment by:  Patricia Estani
Submitted 22 June 2006
Posted 22 June 2006
  I recommend the Primary Papers

Related News: Neuregulin, ErbB4—Levels Normal but Signaling Strengthened in Schizophrenia

Comment by:  Cynthia Shannon Weickert, SRF AdvisorVictor Chong
Submitted 8 August 2006
Posted 8 August 2006

In contrast to its once barren form, the table of potential causative genes for schizophrenia is now stocked to feast level (Straub and Weinberger, 2006). In keeping with the culinary theme, we suggest that this recent paper by Chang-Gyu Hahn and Hoau-Yan Wang is “a full course meal”!

Appetizer: An Important Biological Problem
If one assumes that alterations in NRG-1 account for at least some of the liability to developing schizophrenia, then it is only reasonable to look to the NRG-1 receptors for clues as to how and where NRG-1 may be acting. However, there are three known NRG-1 receptors that mediate a myriad of biological functions, almost all of which could be argued to be relevant to schizophrenia pathology. This paper draws our attention to one NRG-1 receptor, ErbB4, showing this receptor to be a probable NRG-1 partner in mediating this pathology. Recent studies provide further support that ErbB4 may be integral to the development of schizophrenia by demonstrating its gene to be a potential susceptibility gene (Norton et al., 2006; Silberberg et al., 2006; Nicodemus et al., in press). So, genetic and neurobiological evidence suggest the authors selected their NRG-1 receptor wisely.

Main Course: A New Approach
The novel postmortem-stimulation approach used by Hahn and colleagues represents an important advance in the field of schizophrenia research. Through extensive validation of this protocol, this research group has paved the way for future experimentation into the molecular activation of proteins within the schizophrenic brain. More specifically, while previous studies have only been able to draw conclusions about the static state of the schizophrenic brain, this article has introduced a novel method for examining dynamic signaling systems in postmortem brains of patients with schizophrenia. For example, based on the finding that certain splice variant ErbB4 mRNAs are elevated in the prefrontal cortex of these individuals (Silberberg et al., 2006), one would assume that ErbB4 protein should also be elevated in these patients. But Hahn et al. demonstrate that schizophrenic individuals show only marginal increases in prefrontal cortical ErbB4 protein levels, which could suggest that ErbB4 protein plays little role in the pathology of schizophrenia. However, using the more dynamic postmortem-stimulation approach, the authors showed that ErbB4 signaling is, in fact, greatly enhanced in the prefrontal cortex of patients with this disease, leading to the alternative interpretation that ErbB4 protein may play significant roles in schizophrenia. In other words, this postmortem-stimulation protocol extends the examination of human postmortem brain protein from quantification to the functional level. We view this method as a powerful approach that will be important in translating genetic susceptibility into molecular mechanisms of the disease process. The postmortem-stimulation approach also gave rise to the observation that schizophrenic patients exhibit reduced prefrontal cortical NMDA receptor signaling capacity. This finding is highly significant because it is the first evidence directly linking reduced prefrontal cortical NMDA receptor function to schizophrenia. However, whether NRG-1-ErbB4 signaling is a major contributor to NMDA receptor hypofunction is debatable since the attenuation of NMDA receptor phosphorylation by NRG-1 appears proportionally similar between controls and schizophrenic patients.

Side Dish: Dealing with Antipsychotic Drugs
Since most patients with schizophrenia have received antipsychotic drugs and these agents can have profound impact on brain systems, it is essential to determine whether changes observed in the brains of patients with schizophrenia are secondary to antipsychotic drug exposure. To address this issue, the authors took two important steps. Firstly, Hahn et al. examined whether antipsychotic drug exposure affected prefrontal cortical ErbB4 expression or signaling in their human study group and found no correlation between antipsychotic drug treatment and either of these measured variables. Secondly, the authors examined antipsychotic drug effects on prefrontal cortical ErbB4 signaling in mice implanted with a haloperidol-containing bioabsorbable polymer, which has a number of advantages. For example, it allows for long-term treatment of the animals (12 weeks) while minimizing handling. This duration of exposure is arguably more appropriate than some schedules used to examine chronic effects of antipsychotic drugs in rodents. Remarkably, haloperidol treatment caused a reduction in ErbB4 signaling in the mice, suggesting that a decrease in ErbB4 signaling is associated with the therapeutic effects of antipsychotic agents. What may have been more informative is to show whether haloperidol had any effect on ErbB4 protein levels without NRG-1 treatment. In addition, the authors could have considered examining antipsychotic drug effect in mice whose ages were more reflective of those of the investigated human cohort, which consisted of elderly individuals (65-92 years). Furthermore, while their analysis of antipsychotic drugs on ErbB4 expression and signaling in postmortem brain was noteworthy, the authors only examined the effects of antipsychotic drugs taken in the final month before death in a very aged sample population. Thus, it is difficult to ascertain whether ErbB4 expression or signaling is not affected by lifetime antipsychotic drug treatment, which can result in cellular and molecular consequences that can remain long after termination of therapy.

Dessert: Challenging the Field
Of course, the first thing the field needs to do is attempt to replicate these findings in another cohort of patients with schizophrenia compared to controls. Careful attention to matching for age, PMI, and gender, etc., as was done in this study, is critical. We suggest that using a young cohort of patients would help rule out potential confounds such as associated dementia and interaction with the aging process. However, it is recognized that many other potential confounds will still remain in most studies comparing schizophrenics to unaffected controls. These confounds include suffering from years of an unremitting illness that compromises normal social and environmental stimulation, increased incidence of cigarette smoking among patients with schizophrenia, and years of antipsychotic drug exposure. When the finding of schizophrenia-associated increased ErbB4 signaling capacity is replicated, then the task at hand will be to determine how possible genetic changes in the DNA at the NRG-1 or ErbB4 locus (representing one etiological route) could lead to a “hyperactivatable” ErbB4.

Doggie Bag: Nagging Questions
One of the caveats we would like to raise in attempting to link molecular neurobiological changes found in schizophrenic brain tissue with possible changes in DNA is that causative variants in any one susceptibility gene are expected to occur only in a minority of schizophrenic patients. Most measures performed on postmortem schizophrenic brains are made on small sample sizes, which likely show much heterogeneity in terms of etiology. In other words, only a handful of patients in this study would be expected to have a faulty NRG-1 gene; yet this subpopulation shows alterations in ErbB4 signaling as a group. The logical extension of this observation may be that there are multiple routes by which ErbB4 could be “hyperactivatable” (i.e., not solely through NRG-1 genetic liability). To sort this out, we need to work from the gene forward, and thus there is a need to identify causative variants in susceptibility genes and to use these as starting points for basic mechanistic molecular and cellular studies.

View all comments by Cynthia Shannon Weickert
View all comments by Victor Chong

Related News: Glutamate Receptor Trafficking—AMPARs Rooted by Stargazin

Comment by:  Monica Beneyto
Submitted 8 March 2007
Posted 10 March 2007
  I recommend the Primary Papers

These two very interesting papers, a welcome article testing previous models proposed for the role of stargazin in AMPA receptor trafficking and a mini-review about TARPs, represent a much needed update on AMPA receptor trafficking and offer some insight into the possible effect disturbances of these proteins have in schizophrenia.

As reported by Tom Fagan, several lines of evidence have led to the current thinking that disruptions in any of the numerous mechanisms that influence the function of NMDA receptors, by either modifying the kinetics of the NMDA channel itself, or the function of proteins that link NMDA receptors to signal transduction pathways, can reproduce schizophrenic symptoms in normal subjects (see related SRF current hypothesis by Bita Mohaddam). The NMDA receptor dysfunction hypothesis of schizophrenia has subsequently been refined to include abnormalities of other glutamate receptor subtypes (AMPA, kainate, and metabotropic receptors), glutamate transporters, and glutamatergic enzymes. Glutamate has a central role in behaviors such as associative learning, working memory, and attention that are often impaired in this illness. Underlying these behaviors are electrophysiological events, such as long-term potentiation, that are mediated in many brain regions by NMDA and AMPA receptor colocalization, activation and signaling. AMPA receptor colocalization with NMDA receptors is crucial for NMDA receptor function, as sustained activation of nearby AMPA receptors is necessary to depolarize the postsynaptic cell and release the tonic Mg2+ blockade of the NMDA receptor, permitting NMDA receptor activation. It is in this scenario where the role of the protein stargazin seems to be crucial. Clinical studies have suggested that ampakines, AMPA receptor positive modulators, can improve cognitive function in schizophrenia, while enhancement of AMPA receptor-mediated currents by these compounds may potentiate the efficacy of antipsychotics (Coyle, 1996). These pharmacologic observations, taken together with a central role for AMPA receptor function in the facilitation of the molecular events that likely underlie learning, memory, and attention, suggest that AMPA receptors may be abnormal in schizophrenia.

AMPA receptor localization, cell surface expression, and activity-dependent modulation are regulated and maintained by a complex network of protein-protein interactions, including those involving stargazin, associated with targeting, anchoring, and spatially organizing synaptic proteins at the cell membrane. These proteins alter receptor sensitivity to glutamate and modulate signaling cascades associated with synaptic transmission by linking AMPA receptors to critical intracellular effector molecules. AMPA receptor trafficking as a critical aspect in the regulation of synaptic efficacy has been recently reviewed (Barry and Ziff, 2002; Malinow, 2003; Malinow and Malenka, 2002; Scannevin and Huganir, 2000; Sheng, 1997; Sheng, 2001; Sheng and Hyoung Lee, 2003; Sheng and Kim, 2002; Sheng and Lee, 2001; Sheng and Nakagawa, 2002; Sheng and Pak, 1999).

Trafficking of AMPA receptors in the postsynaptic cell occurs in several integrated pathways. The constitutive pathway consists of a basal continuous and rapid recycling of GluR2-containing AMPA receptors from non-synaptic to synaptic sites. It has been suggested that NSF [N-ethylmaleimide-sensitive factor] regulates this pool of AMPA receptors independently of NMDA activation by its specific binding to GluR2 (Song et al., 1998). The regulatory pathway, unlike the constitutive pathway, is activity-dependent, triggered by NMDA receptor activation, and driven by PDZ interactions between C-terminal regions of the GluR2 and GluR3 AMPA subunits with the intracellular proteins PICK1, GRIP1, and ABP (Chung et al., 2000; Dev et al., 1999a; Hirbec et al., 2002; Kim et al., 2001; Lu and Ziff, 2005; Osten et al., 2000; Perez et al., 2001; Terashima et al., 2004; Wyszynski et al., 1999; Xia et al., 1999). The new synthesis pathway is dependent on an increase in glutamate activation in the synapse and involves the GluR1 AMPA subunit. Finally, a fourth pathway is mediated by the protein stargazin, the star of these two recent Neuron papers, that binds to all four AMPA subunits and has been identified as a mediator of AMPA receptor function in a two-step trafficking model, in which it first conveys AMPA receptors to the neuronal surface and then sweeps them laterally into postsynaptic sites (Chen et al., 2000). As demonstrated by Bats and colleagues, this process requires an interaction of the carboxy terminus of stargazin with a PDZ anchoring protein such as PSD95, that regulates the diffusion and mediates the clustering of AMPA with NMDA receptors.

We and others have systematically examined the expression of the AMPA receptors, and other molecules associated with glutamate neurotransmission, in postmortem brain samples from persons with schizophrenia. We have previously examined both NMDA and AMPA receptors in multiple brain regions, and have typically found only modest changes in both transcript and protein levels (Clinton et al., 2003; Clinton and Meador-Woodruff, 2004a; Clinton and Meador-Woodruff, 2004b; Healy et al., 1998a; Ibrahim et al., 2000b; Kristiansen and Meador-Woodruff, 2005; Meador-Woodruff et al., 1996; Meador-Woodruff et al., 2003; Meador-Woodruff and Healy, 2000; Meador-Woodruff et al., 2001; Mueller and Meador-Woodruff, 2004). Given the strength of clinical and pharmacological data suggesting glutamate receptor dysfunction in schizophrenia, our focus then shifted to studying intracellular signaling pathways associated with the ionotropic glutamate receptors, postulating that a receptor abnormality might not be a problem with the receptor itself, but rather with associated intracellular trafficking mechanisms.

Several studies have been conducted in vitro to characterize these molecules and their interactions, but few studies have demonstrated their expression in brain, particularly in regions such as the hippocampus, cerebellum, and cortex (Puschel et al., 1994; Chen et al., 2000). Our report in the Journal of Comparative Neurology two years ago showed the distribution of transcripts encoding stargazin, and another five intracellular proteins related to the AMPA receptors in the macaque brain. In our study stargazin was highly expressed in some regions of the hippocampus, thalamus, pons, cerebellum and specially in the cerebral cortex (Beneyto and Meador-Woodruff, 2004).

In a recent paper published in Synapse we identified abnormalities in the expression of transcripts encoding the proteins PICK1 and stargazin in the dorsolateral prefrontal cortex (DLPFC) in schizophrenia (Beneyto and Meador-Woodruff, 2006). Results from in situ hybridization experiments showed lamina-specific decreased expression of PICK1 transcripts and increased expression of stargazin mRNA in DLPFC in layer III in schizophrenia. Recent reports have demonstrated PICK1 involvement in exocytosis of GluR2-containing receptors into the plasma membrane in cerebellar stellate cells (Gardner et al., 2005; Liu and Cull-Candy, 2005; Malinow and Malenka, 2002). Decreased expression of PICK1 suggests increased number of AMPA receptors available to bind to ABP/GRIP and thereby be sequestered in immobile vesicles inside the PSD not ready to be inserted at the postsynaptic membrane, resulting in decreased number of AMPA receptors at the cell surface in schizophrenia. In this scenario, the increased stargazin expression we found could be a compensation for the decreased number of AMPA receptors inserted at the postsynaptic membrane.

Stargazin is responsible for lateral transport of AMPA receptors already inserted in the membrane from extrasynaptic to synaptic sites and for their colocalization with NMDA receptors (Tomita et al., 2004, 2005; Turetsky et al., 2005). Perhaps more importantly, stargazin also acts as a positive allosteric modulator of AMPA receptor ion channel function. Recent data suggest that AMPA receptors complexed with stargazin at the postsynaptic membrane are significantly more responsive to synaptically released glutamate, compared with AMPA receptors lacking a stargazin interaction (Priel et al., 2005; Tomita et al., 2005; Yamazaki et al., 2004). Coexpression of stargazin with AMPA receptor subunits significantly reduces receptor desensitization in response to glutamate, slowing receptor deactivation rates and accelerating the recovery from desensitization. Structurally, based on this tight correlation between desensitization and the stability of the AMPA receptor intradimer interface, binding of stargazin has been suggested to be a stabilizer of receptor conformation (Priel et al., 2005). This fact, together with the AMPA receptor trafficking role of stargazin, suggests that the stargazin upregulation we found may be secondary to decreased PICK1 expression and possible reduced AMPA receptor membrane expression in DLPFC in schizophrenia.

Regulated insertion and removal of AMPA receptors at the synapse provides a mechanism for altering synaptic efficacy (Malenka and Nicoll, 1999; Malinow et al., 2000). Taken together, changes of AMPA subunit and related protein expression in the prefrontal cortex of schizophrenic patients might provide a molecular basis for both structural and neurochemical impairments in this illness.

View all comments by Monica Beneyto