Schizophreniaforum - Print Discussion

David Talmage		Amanda J. Law		View article

On Tuesday, 24 February, at 12 noon Eastern U.S. time, we hosted a live discussion of the state of the neuregulin 1 research as it may relate to schizophrenia. Amanda J. Law of the University of Oxford and National Institutes of Mental Health in Bethesda, Maryland, and David Talmage of the State University of New York, Stony Brook, touched on the dizzying array of isoforms and the many functions NRG1 is involved in from development through adult neurobiology.

In particular, Law and Talmage recommended these three papers for background reading:

Harrison PJ, Law AJ. Neuregulin 1 and schizophrenia: genetics, gene expression, and neurobiology. Biol Psychiatry. 2006 Jul 15;60(2):132-40. Abstract

Role LW, Talmage DA. Neurobiology: new order for thought disorders. Nature. 2007 Jul 19;448(7151):263-5. Abstract

Mei L, Xiong WC. Neuregulin 1 in neural development, synaptic plasticity and schizophrenia. Nat Rev Neurosci. 2008 Jun 1;9(6):437-52. Abstract

You can download the PDF of Harrison and Law (2006) directly from the link beneath the photo above, courtesy of Biological Psychiatry and Elsevier.

The association between genetic variation in NRG1 and schizophrenia (Stefansson et al., 2002) was greeted with substantial fanfare because of the known role of NRG1 signaling in the development of normal neuronal connectivity (see Harrison and Law, 2006; Role and Talmage, 2007). The NRG1 gene is complex: differential splicing coupled to utilization of at least six promoters generates mRNAs with the capacity to encode as many as 30 different proteins (Falls, 2003; Mei and Xiong, 2008). Different members of this family of signaling proteins utilize distinct and overlapping strategies for directing local cellular responses. As yet we do not have a clear picture of which aspects of NRG1 signaling are affected in the development of neuropsychiatric disorders such as schizophrenia, but emerging postmortem studies point towards altered levels of specific isoforms, perhaps resulting in aberrant ErbB4 activation and an imbalance between paracrine and juxtacrine signaling.

Of specific relevance to this discussion are recent findings linking changes in NRG1 signaling with perturbations of synaptic transmission, myelination, and the survival of particular sets of neurons and glia (Bjarnadottir et al., 2007; Michailov et al., 2004; Wolpowitz et al., 2000; Zhong et al., 2008). NRG1-ErbB signaling modulates activity dependent synaptic plasticity by regulating NMDA receptor levels and phosphorylation, by regulating AMPA and nicotinic acetylcholine receptor trafficking, and by regulating GABA release (Li et al., 2007; Woo et al., 2007; Chen et al., 2008). More recently NRG1 has been associated with dopaminergic signaling (Kwon et al., 2008). A fine balance of NRG1-ErbB signaling appears to be required: both deficient and excessive signaling interfere with synaptic plasticity. Given the apparent sensitivity of the NRG1-ErbB signaling to perturbations, it is becoming clearer how subtle changes in the levels and types of NRG1-ErbB interactions could alter the ability of key brain circuits to withstand additional genetic and environmental insults and contribute to disease.

Transcriptional regulation of NRG1 expression is likely to be important but is a largely underexplored area. Changes in NRG1 expression occur in pathological conditions, including injury to the nervous system. In addition, few of the DNA sequence changes in the NRG1 gene that are believed to confer susceptibility for schizophrenia alter protein structure. These non-coding polymorphisms are likely to alter the overall level of expression, the pattern and timing of expression, or the relative expression of different isoforms of NRG1, with the resulting changes in NRG1 protein levels contributing to the etiology of disease. Data consistent with this hypothesis include the finding of increases in the levels of type I mRNAs in postmortem brains from schizophrenia patients compared to controls (Hashimoto et al., 2004; Law et al., 2006) and of association of a functional promoter risk variant with type IV NRG1 levels (Law et al., 2006; Tan et al., 2007). The additional demonstration that reduced expression of subsets of Nrg1 isoforms in mice perturbs circuits underlying sensorimotor gating, memory performance, and synaptic functions relating to dopamine, GABA, glutamate, and acetylcholine underscores the point that altering the expression of NRG1 can contribute to pathology.

Based on this platform, the following points should be considered for this discussion:

2. Using animal and cellular models to investigate the biological role of NRG1 signaling during neurodevelopment and its role in adult brain function and pathology. The role of NRG1 signaling in the regulation of GABA, glutamate and cholinergic development and function.

3. Molecular mechanisms of NRG1 association. What the genetics has taught us about aberrant NRG1 gene regulation in schizophrenia.

4. Moving beyond NRG1. Thinking of NRG1 as part of a “disease pathway” based on association of other molecules in the signaling pathway to schizophrenia, i.e., ErbB receptors.

References
Stefansson H, Sigurdsson E, Steinthorsdottir V, Bjornsdottir S, Sigmundsson T, Ghosh S, Brynjolfsson J, Gunnarsdottir S, Ivarsson O, Chou TT, Hjaltason O, Birgisdottir B, Jonsson H, Gudnadottir VG, Gudmundsdottir E, Bjornsson A, Ingvarsson B, Ingason A, Sigfusson S, Hardardottir H, Harvey RP, Lai D, Zhou M, Brunner D, Mutel V, Gonzalo A, Lemke G, Sainz J, Johannesson G, Andresson T, Gudbjartsson D, Manolescu A, Frigge ML, Gurney ME, Kong A, Gulcher JR, Petursson H, Stefansson K. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet. 2002 Oct 1;71(4):877-92. Abstract

Harrison PJ, Law AJ. Neuregulin 1 and schizophrenia: genetics, gene expression, and neurobiology. Biol Psychiatry. 2006 Jul 15;60(2):132-40. Abstract

Role LW, Talmage DA. Neurobiology: new order for thought disorders. Nature. 2007 Jul 19;448(7151):263-5. Abstract

Falls DL. Neuregulins: functions, forms, and signaling strategies. Exp Cell Res. 2003 Mar 10;284(1):14-30. Abstract

Mei L, Xiong WC. Neuregulin 1 in neural development, synaptic plasticity and schizophrenia. Nat Rev Neurosci. 2008 Jun 1;9(6):437-52. Abstract

Bjarnadottir M, Misner DL, Haverfield-Gross S, Bruun S, Helgason VG, Stefansson H, Sigmundsson A, Firth DR, Nielsen B, Stefansdottir R, Novak TJ, Stefansson K, Gurney ME, Andresson T. Neuregulin1 (NRG1) signaling through Fyn modulates NMDA receptor phosphorylation: differential synaptic function in NRG1+/- knock-outs compared with wild-type mice. J Neurosci. 2007 Apr 25;27(17):4519-29. Abstract

Michailov GV, Sereda MW, Brinkmann BG, Fischer TM, Haug B, Birchmeier C, Role L, Lai C, Schwab MH, Nave KA. Axonal neuregulin-1 regulates myelin sheath thickness. Science. 2004 Apr 30;304(5671):700-3. Abstract

Wolpowitz D, Mason TB, Dietrich P, Mendelsohn M, Talmage DA, Role LW. Cysteine-rich domain isoforms of the neuregulin-1 gene are required for maintenance of peripheral synapses. Neuron. 2000 Jan 1;25(1):79-91. 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

Li B, Woo RS, Mei L, Malinow R. The neuregulin-1 receptor erbB4 controls glutamatergic synapse maturation and plasticity. Neuron. 2007 May 24;54(4):583-97. Abstract

Woo RS, Li XM, Tao Y, Carpenter-Hyland E, Huang YZ, Weber J, Neiswender H, Dong XP, Wu J, Gassmann M, Lai C, Xiong WC, Gao TM, Mei L. Neuregulin-1 enhances depolarization-induced GABA release. Neuron. 2007 May 24;54(4):599-610. Abstract

Chen YJ, Johnson MA, Lieberman MD, Goodchild RE, Schobel S, Lewandowski N, Rosoklija G, Liu RC, 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 Jul 2;28(27):6872-83. Abstract

Kwon OB, Paredes D, Gonzalez CM, Neddens J, Hernandez L, Vullhorst D, Buonanno A. Neuregulin-1 regulates LTP at CA1 hippocampal synapses through activation of dopamine D4 receptors. Proc Natl Acad Sci U S A. 2008 Oct 7;105(40):15587-92. Abstract

Hashimoto R, Straub RE, Weickert CS, Hyde TM, Kleinman JE, Weinberger DR. Expression analysis of neuregulin-1 in the dorsolateral prefrontal cortex in schizophrenia. Mol Psychiatry. 2004 Mar 1;9(3):299-307. Abstract

Law AJ, Lipska BK, Weickert CS, Hyde TM, Straub RE, Hashimoto R, Harrison PJ, Kleinman JE, Weinberger DR. Neuregulin 1 transcripts are differentially expressed in schizophrenia and regulated by 5' SNPs associated with the disease. Proc Natl Acad Sci U S A. 2006 Apr 25;103(17):6747-52. Abstract

Tan W, Wang Y, Gold B, Chen J, Dean M, Harrison PJ, Weinberger DR, Law AJ. Molecular cloning of a brain-specific, developmentally regulated neuregulin 1 (NRG1) isoform and identification of a functional promoter variant associated with schizophrenia. J Biol Chem. 2007 Aug 17;282(33):24343-51. Abstract

Nick Brandon, Wyeth Research
Yachi Chen, SUNY, Stony Brook
Tim Crow, SANE Prince of Wales International Centre
Inga Deakin, Oxford University
Angela Epshtein, Schizophrenia Research Forum
Amelia Gallitano, University of Arizona College of Medicine, Phoenix
Jeremy Hall, Edinburgh University
Paul Harrison, University of Oxford
Hakon Heimer, Schizophrenia Research Forum
Sloka Iyengar, University of South Carolina School of Medicine
Anne Kirtley, Cardiff University School of Biosciences
Carsten Korth, University of Duesseldorf, Germany
Amanda Law, University of Oxford and NIH
Julie Markham
Maryland Psychiatric Research Center, Baltimore
Maureen Martin, University of California, Irvine
David O’Hanlon
Clare Paterson, University of Glasgow
Lorna Role, SUNY, Stony Brook
Chris Ross, Johns Hopkins University
John Sinnamon, SUNY, Stony Brook
David Talmage, SUNY, Stony Brook
Sara Taylor, Maryland Psychiatric Research Center, University of Maryland
Orna Tighe, Royal College of Surgeons in Ireland
Liz Tunbridge, Oxford Department of Psychiatry
Daniel Weinberger, NIMH
Chongbo Zhong, SUNY, Stony Brook

Hakon Heimer
I think we can start with the preliminaries. Let's start off by having all the participants in the "room" introduce themselves: I'm Hakon Heimer, editor of the Schizophrenia Research Forum. While you are doing that, I would like to introduce and thank our chat leaders, David Talmage and Amanda Law.

Liz Tunbridge
Hi, everyone. I’m Liz Tunbridge, based in Oxford Department of Psychiatry in the U.K.

Angela Epshtein
I'm Angela Epshtein, managing editor of the Schizophrenia Research Forum.

Sloka Iyengar
I am Sloka Iyengar, graduate student at University of South Carolina School of Medicine.

Amelia Gallitano
Amelia Gallitano, Assistant Professor of Basic Medical Sciences at the University of Arizona College of Medicine, Phoenix.

Julie Markham
I'm a postdoc at the Maryland Psychiatric Research Center in Baltimore.

Lorna Role
Lorna Role, also Stony Brook, Chair of Neurobiology and Behavior.

Anne Kirtley
I'm Anne Kirtley from Cardiff University School of Biosciences in final year of Ph.D.

Sara Taylor
Sara Taylor, graduate student, Maryland Psychiatric Research Center, University of Maryland.

Maureen Martin
Maureen Martin, postdoc from the University of California, Irvine.

Hakon Heimer
In the informal spirit of this event, and because I think most people here are at least acquaintances, I will only mention that Amanda is on permanent loan from Real Madrid to the Los Angeles Galaxy. Oh wait...that was another chat!! It's Oxford and NIMH! And David is at SUNY, Stony Brook, recently arrived from Columbia.

Hakon Heimer
The first topic for conversation is…The NRG1 gene: complexity of splicing and isoform regulation.

Hakon Heimer
Yes, Amanda, please pose a controversial statement or burning question.

Amanda Law
The NRG1 gene undergoes complex regulation through alternative promoter usage and splicing to give rise to up to six families of isoforms. Let’s start by trying to think about how we understand the role of each of these isoform families in relation to the pathobiology of schizophrenia and genetic risk.

David Talmage
The last time I did a quick calculation in my head, there are over 100 possible protein products of this gene—how many are really made we have very little information on.

Paul Harrison
Regarding isoforms: does anyone want to (dis)agree with the proposition that, vis-à-vis schizophrenia, type IV is the key isoform?

Daniel Weinberger
Yes, me...I would disagree, too. It appears to be “a” key isoform, related at least in part to variation in a region identified by deCode Genetics, in Iceland (see Stefansson et al., 2002).

David Talmage
I think we have the most relevant information about type IV, but that doesn’t explain a lot yet.

Amanda Law
I think the whole point about the type IV isoforms is that their regulation is linked to one region of association; it’s not the only NRG1 isoform implicated in the disorder, but it does relate to this region of association.

Sloka Iyengar
Is schizophrenia the only disease that NRG has been implicated in?

Sloka Iyengar
I am studying NRG in epilepsy and how it affects synaptic plasticity in temporal lobe epilepsy.

Amanda Law
NRG1 has been linked to multiple sclerosis, bipolar disorder (BPD), Alzheimer’s disease (AD), etc., not only schizophrenia.

Liz Tunbridge
Amanda/David, do you have a sense of whether there is likely to be any disease specificity in terms of the NRG isoforms involved in the different diseases?

David Talmage
Liz, I don’t. I don’t think we have enough overall information about polymorphisms and specific aspects of disease. From my readings, most of those types of studies have focused on only a few single nucleotide polymorphisms (SNPs).

Amanda Law
Liz, that’s a good question. I don't think we know yet. At least in terms of the genetics, the same region of association has been in the Icelandic haplotype region for BPD and AD, but it is not clear which isoforms are affected. We need to do more expression studies in these diseases.

Clare Paterson
A general question: Although there are possibilities that there are countless isoforms due to combinations of immunoglobulin (Ig)/cysteine-rich domains (CRD), etc., is it plausible to say not all of these isoforms exist and indeed are not functional?

David Talmage
Clare, plausible, and it is possible to whittle it down a bit, but without the appropriate reagents (which we desperately need!), we can’t know.

Amanda Law
Clare, I think we need more careful study of each isoform family, similar to what Lorna and David have done with type III and what we have done with type IV. But antibodies are a problem and we need more development in this area.

Yachi Chen
All, given so many isoforms, what type is highly expressed in the prefrontal cortex (PFC)?

Amelia Gallitano
All, I asked Kari Stefansson several years ago whether he had an explanation for how NRG1 polymorphisms may interact with environmental factors to take into account the ~50 percent risk of schizophrenia that is not due to genomic sequence. He had none at the time. Does anyone have an explanation for it now?

David Talmage
Amelia, I think that in order to address that question we need to know what the molecular/cellular consequences of the environmental insults are. It is easy to speculate, for example, that infection increases cytokines that could interfere with neuronal migration, and this could be exacerbated by changes in NRG1 signaling, but that is pure speculation.

Amelia Gallitano
David, so that addresses the in utero exposure to infection via the role of NRG1 on neuronal migration. Plausible. Okay. Thanks.

Daniel Weinberger
I doubt that there is a specific environmental factor that is NRG1-specific, anymore than there is a brain-derived neurotrophic factor (BDNF)-specific environmental factor. There will be factors, including other genes and environmental events, that interact with the biological effects of NRG1 at critical times in the action of NRG1 on brain development that exaggerate those subtly altered effects.

David Talmage
Danny, I agree; I was just trying to cite one possibility, largely to make the point that I don’t think we can get very far, mechanistically, with the environmental part of the story without converting the epidemiological data into something that can be tested/examined, etc., at the cellular/circuit level.

Tim Crow
I am unconvinced NRG is linked to schizophrenia at all; see my post to SRF of 10 minutes ago.

Daniel Weinberger
Tim, “convincing” you will not occur during this chat.

Chris Ross
Tim and Danny, I do think addressing that concern is important, though. For me some of the most intriguing data are the correlative imaging/genetic studies.

Tim Crow
Danny, hi! I'm open to evidence, but it's got to be weighed against the big samples which are negative for linkage and association.

Amanda Law
Chris and Tim, this forum was not meant to be about the genetics; it is about the biology of NRG1. That will be for another time.

Lorna Role
Amanda, I am with you. We need to get at the issues related to biological plausibility. I can’t see that any given gene or isoform can be identified as definitive until we get at what mechanisms could be involved.

Amanda Law
Lorna, totally. And dysregulation of this gene is likely to affect many isoforms and the signaling capacity of these isoforms.

Amelia Gallitano
David, can you elaborate on this signaling interest? Do you think disruption of NRG1 signaling is the issue? What about other signal transduction molecules in the pathway? (I'm bringing in my own pet interest in looking at mechanism by investigating biological pathways....)

David Talmage
Amelia, I like signaling because I have been studying it for 30-odd years. But in this context, it is important to remember that it was from the signaling world that NRG1 emerged (research-wise), and I feel it is critical to consider the genetics, the postmortem data, and the imaging data with this in mind. We know a lot about what NRG1 can do at the subcellular level. This is being applied some in terms of consideration of other candidates (ErbB4, Akt, SRF related news story) but could/should be extended to molecules like DISC1 (see SRF related news story).

Amanda Law
Amelia, I think we cannot think of this any other way. This has to be about how NRG1 mediates its effects (i.e., signaling), and with other genes in this pathway associated with schizophrenia, like ErbB4, ErbB3, and NRG3, it amounts to thinking about this as a biological risk pathway which may have many entry points.

Chris Ross
Sorry, I agree with not getting too far off topic. I do think imaging can be a powerful way of getting at neurobiology, and genetics can help get at isoforms. Amanda, isn't that what some of your studies suggest? One question regarding signaling I think would be very interesting to study is, How does neuregulin interact with other risk genes, such as DISC1? Any others?

Amelia Gallitano
Chris, we have started investigating the role of other candidate genes that are already known to function, or be regulated, downstream of NRG1, for example NMDA receptors, calcineurin, and Egr3.

Nick Brandon
Chris/David, I think we'll very likely hear more about DISC1 and NRG1 in the next year. We're not working on it, but a couple of groups have data on the relationship of these two factors in the pipeline.

Chris Ross
Nick, I would be interested to hear more. This is something we are just beginning to work on, and don't have any results yet. We do have some indirect evidence that there may be an interaction from zebrafish (work of Jon Wood and Vincent Cunliffe, which will be presented at the Keystone meeting, and has been partly published; also see Wood et al., 2009). I would like to do the mouse cross experiment....

Amelia Gallitano
Chris, NRG1 has long been known to regulate expression of the Egr (early growth response genes) in cell culture (see Gallitano-Mendel et al., 2007), and the induction and maintenance of the muscle spindle (see Hippenmeyer et al., 2002). Calcineurin (CN) is activated in response to increased calcium in the cell, which can occur via NMDAR activation. CN, NRG1, and activation of NMDARS each regulate expression. And knockout mice for either NMDARs or CN have memory deficits and defects in synaptic plasticity in the form of LTD long-term depression (see Miyakawa et al., 2003 and Mohn et al., 1999). A recent paper was just published in Science, reviewed on SRF, demonstrating the link between NRG1 and NMDAR function in myelination (see Kao et al., 2009). My colleague Josh Gordon and I presented a mini-symposium on this at the SfN meeting in November and Hakon wrote a great review of it on SRF, which includes the diagrams I made (see SRF related news story). Basically, one pathway could be NRG1 -> NMDARs -> CN.

David O’Hanlon
Amanda, how long is it expected until specific proteins on NRG1 are isolated in the susceptibility to schizophrenia? Any guess?

Amanda Law
David, I think there is growing evidence that isoforms like type I and IV are involved at the molecular level and also emerging evidence for type III. I doubt we will pin it to one isoform, one SNP, or region, since there seem to be many ways to disrupt NRG1 signaling.

Hakon Heimer
Question 2 that Amanda and David wanted to discuss: Using animal and cellular models to investigate the biological role of NRG1 signaling during neurodevelopment and its role in adult brain function and pathology. The role of NRG1 signaling in the regulation of GABA, glutamate, and cholinergic development and function. Any thoughts on this?

Chris Ross
Lorna, Amanda, regarding isoforms and neurobiology, what do we have from genetically modified mice regarding isoforms?

Lorna Role
All, the mouse data have been striking in terms of developmental profile. It’s pretty clear that some "symptoms" don’t appear until adult.

Liz Tunbridge
Amanda/David, to what extent do you think that NRG is important during post-adolescent brain function, or do you think its critical involvement is during initial brain development, rather than later remodeling?

David Talmage
Liz, I think that NRG function is a moving target—by which I mean that even within limited brain regions (based on mouse/rat data), functions differ in early embryonic development to late embryonic development to young adults. This makes this issue of "which isoform" very difficult to address, because there is also the question of "when."

Amanda Law
David, I agree totally. NRG1 is part of a network, and thinking about NRG1’s effects in schizophrenia without considering crosstalk between and within pathways will limit our understanding of the mechanisms of association.

Paul Harrison
Liz, regarding the timing of NRG effects: The Edinburgh and Hungarian “high risk” samples imply that NRG1 may be important in conversion to psychosis in adolescents/young adults at high risk (and with an effect size/penetrance much higher than that observed in other designs). On the other hand, the largely fetal expression of type IV NRG1 speaks to an earlier time point...so I guess it’s going to be temporally as well as molecularly complex!

Lorna Role
Paul, exactly. Temporally and spatially dynamic expression and actions….

Daniel Weinberger
Paul, not sure that the conversion to psychosis in adolescence means that the effects of the gene are "turned on" in adolescence. All schizophrenia converts in adolescence and early adulthood. Those findings would be no different if they found that people in those samples with obstetrical complications were the ones who became schizophrenic at the time of life when people become schizophrenic.

Liz Tunbridge
All, has anyone produced an inducible transgenic that we could use to investigate the importance of NRG at different time points?

Chris Ross
I agree that an inducible transgenic would be key. We would love to cross with an inducible DISC1 model....

Amelia Gallitano
Chris, sure. But first, do you have a mechanism in mind for an interaction between NRG1 and DISC1, or is it simply the idea that both are strong candidates, so perhaps crossing mice would create an additive, or even synergistic, effect?

Amanda Law
I still think we need to be modeling the molecular signature associated with the illness in terms of animal models and isoforms. The knockout/knockdown data are compelling, and it may be ultimately that the inducible transgenic to model the molecular schizophrenia association is key to understanding the developmental effects of NRG1 modification.

Liz Tunbridge
Amanda, I agree that focusing on modeling the schizophrenia molecular signature is key. However, I think it's important to keep studying other models too, though, to try and understand the basic biology of NRG (and all other putative risk genes) as the schizophrenia findings are necessarily derived from a relatively late stage of the disease process (i.e., post-diagnosis) and we don't know what occurs at earlier time points. Am I making any sense?!

Orna Tighe
To all, what information is available about the expression of NRG isoforms IV, V, and VI in the mouse model?

David Talmage
Orna, I can’t answer directly, but having stared at sequences for a bit, these types are likely to be fundamentally different in non-primates. At least for type IV, from the sequence information I have looked at, only in primates does the type IV unique exon contain an open reading frame, raising the possibility that as coding exons, these are primate-specific. I don’t remember for types V and VI.

Amanda Law
Orna, we have done some work on this and I know others have including Jeremy’s group. As for type IV, we have not detected in the mouse endogenously (but it could be a sequence problem), but we have created a transgenic humanized version of the gene. It also has a poor Kozak consensus start sequence in mouse. Not sure about V and VI.

Jeremy Hall
Orna, we couldn't find type IV in rats, either, so it looks like it isn't there in rodents.

Clare Paterson
General question: There are several knockouts published (see Dean et al., 2008 and O’Tuathaigh et al., 2007) from these some convincing data at the molecular level; however, what are everyone’s views on the behavioral analyses of these knockouts? Do you think the testing that has been done on them showing a "schizophrenic" phenotype suffices?

David Talmage
Clare, I would put it that the knockouts show parallels in some very specific measures but would not argue for a schizophrenic model.

Hakon Heimer
Amanda/David's fourth point of discussion, which is a direction we've already moved in was: 4. Moving beyond NRG1. Thinking of NRG1 as part of a “disease pathway” based on association of other molecules in the signaling pathway to schizophrenia, i.e., ErbB receptors.

Daniel Weinberger
So, one of the other points to be considered is whether NRG1, because of its discovery in one of the better linkage signals, was an entry point into this pathway, but really this was luck, and that it has less of an impact on schizophrenia risk than, say, NRG3 or ErbB4. Amanda, you want to chime in on this one??

Amanda Law
There is a body of emerging evidence suggesting that other genes in the NRG1 pathways (ErbB4, ErbB3, NRG3, etc.) are also risk genes for schizophrenia, and even that interactions between them further increase risk for disease. Lorna and David, do you have any thoughts on how we might start to think about modeling these in a network-type scenario?

David Talmage
All, this seems to be a ripe area for some computational modeling, both at the genetic level, which is being done, but also at the cellular level. Can we model the expected outcome of small changes in signaling strength or type that would result from changes in NRG expression, ErbB4 splicing, and Akt levels?

Paul Harrison
Presumably our NRG networks will also be isoform (and developmentally and spatially) variable, so we'll have to make tough decisions on what to prioritize.

Lorna Role
Paul, again I agree with your point totally. I think a lot of cross-fertilization of experiments and approaches and data-sharing is the key.

David Talmage
Liz/Paul et al., one area that needs more information is the timing and pattern of isoform expression across species—something the NIMH people have really made progress with in humans—so that we can better design rodent experiments.

Liz Tunbridge
Amanda/David, do you think we currently have a good sense of what the NRG network is, or is there still work to be done in this area?

Liz Tunbridge
David, what do you think the best approach to this is likely to be? (I'm thinking particularly since some of the key NRG isoforms don't seem to be expressed in rodents.)

David Talmage
Liz, I am not sure what the significance is in terms of no type IV in rodents. We have no data (that I know of) that would argue for a significant difference in protein function between, say, type IV and type I. If there isn’t, then it becomes an issue of time and place.

Amanda Law
I agree, David; we need to demonstrate functional differences between the Ig isoforms before we can think about whether absence of type IV is relevant. Although I think they have a different 5' leader exon for some reason (trafficking, processing, etc). We need more insight in this area.

Hakon Heimer
Since we're down to the last 10 minutes or so, and Tim is itching to be bombarded with the heterogeneity of populations argument, I'll quote the salient point from his recent comment: The only conclusion that one can reach is that the Stefannson et al. finding is not confirmed with sample sizes of 30 and 300 times greater.

Tim Crow
All, supposing Stefansson's result is not replicable? Where does that leave us?

Daniel Weinberger
Tim, not wanting to open this Pandora's box, but there have been many replications. The issue is not really replications, which are plentiful for NRG1; the issue is how many failures to replicate mean the finding is not real. This is impossible to argue about, because the number of positive reports and the positive meta-analyses are way beyond what would be expected by chance. The issue of heterogeneity at all levels (clinical and biological) is the point that leads to failures to replicate in the context of so many positive reports. We now have evidence in our own datasets, and in the GAIN datasets, of interactions within the 5' and 3' risk regions of NRG1, and some of the literature inconsistency likely involves this level of heterogeneity within the gene itself. The nature of complex genetics is such that rules from Mendelian genetics are misplaced and inappropriate.

Jeremy Hall
Amanda, returning to a point you made earlier in the discussion, even if type IV is very functionally similar to type I, altering levels could impact on the balance of signaling between different isoforms and perhaps especially between Ig and CRD isoforms.

David Talmage
Jeremy, I absolutely agree, and studying possible consequences of this is approachable in non-primate models.

Paul Harrison
Lorna/Jeremy, and the strikingly distinct developmental expression profile of type IV speaks to this.

Amanda Law
Jeremy, excellent point, and Lorna and I talked about this possibility in the very beginning, and it is discussed in her and David’s review. It is possible that by increasing type I or IV, this interferes with signaling via type IIIs. Very possible.

Liz Tunbridge
Presumably, there must be some specialization of each of the different isoforms, though. Otherwise, why produce so many?

Amanda Law
I agree, Liz. This is why we are trying to understand if there are any signaling differences in culture models, especially between type I and IV. I think these genes make these isoforms for a reason....

David Talmage
All, a basic scientist side step of the Stefansson question: If we learn something fundamental about the disease process from studying NRG1 as a minor contributor, aren’t we closer to understanding possible therapeutic approaches? Not ones that target NRG1 per se, but processes that are common to NRG1 and the "real causative genes."

Liz Tunbridge
David, I agree. I think we'd be stupid to rule out any genes at the moment on the basis of genetic evidence. I don't think we know enough about how the brain works to do that.

Hakon Heimer
All, I'm going to throw the last bone into the feeding frenzy. If anyone wants to start a thread, please jump in: 5: Where next for NRG1? Potential for therapeutic intervention?

Clare Paterson
General point: Can I complicate this discussion even more and drop in the subject of ErbB receptor isoforms, i.e., different NRG1 isoforms acting on different isoforms of ErbB receptors (ErbB4, in particular) and the significance of this to possible pathophysiology in schizophrenia?

Lorna Role
Clare, yes, important point and likely a very real complexity of the scenario.

David Talmage
Clare, as well, this is an area where there is a history of data from the signaling literature (mostly breast cancer-related) showing different cellular responses elicited by different ligands acting on the same ErbBs. There is so much more that needs to be looked at here.

Amanda Law
Clare, yes definitely. The evidence that specific isoforms of ErbB4 may be related to the disease (cyt1) and the recent data on the ErbB4 interactome and signaling in cancer suggest that definitely this is an area of important investigation in trying to link this with NRG1 isoform alterations in the disease (see, e.g., Takahashi et al., 2008).

Paul Harrison
Amanda, your point, of course, raises the cancer versus schizophrenia debate, which may impact on the attractiveness of NRG therapies.

Amanda Law
Paul, I agree, and this will have impact on the approach to therapies, but...I think as we delve deeper into these pathways, we may find targets that circumvent targeting the ligand or receptor. Clearly you would stay away from that if you could.

Daniel Weinberger
I also agree with David et al. on this. There is very compelling evidence that the NRG1/ErbB4 signaling system is implicated in psychosis and cognition related to schizophrenia, and the more this pathway is studied, the more relevant it appears; to wit, five positive genes at least. This is a very impressive story that begs for extensive scientific inquiry. There has been much more devoted to clues to CNS disorders that led to promising new treatments with much less scientific evidence than this.

Clare Paterson
Daniel/Amanda, do you think there is a feasible way to dissect the ErbB4/NRG1 pathway; i.e., how can we determine differential effects seen by NRG1 activating ErbB4 homodimers or ErbB4/3/2 heterodimers?

David Talmage
Clare, yes in engineered systems this is certainly possible. The other approach would be using specific inhibitors, if any are ever made, of the different ErbBs.

Paul Harrison
Amanda, sure, molecular specificity will be the therapeutic key—and at least it’s the one advantage of the frustrating complexity of the whole thing!

Daniel Weinberger
Another thing to remember from the Kety adoption studies and the twin studies of cognition is that psychosis, per se, is not clearly what is genetic about schizophrenia (for a review, see Ingraham and Kety, 2000).

Jeremy Hall
Danny, I agree; this pathway is clearly the best bet we have for new treatments, but what to target (ErbB4)? What markers to use in preclinical studies? When to treat? It ain’t going to be easy!

Daniel Weinberger
Jeremy, not sure really what is easy and what isn't. It all depends on where you look. It's much easier having a pathway to study than looking for another therapeutic accident. I bet that as we explore this pathway in various experimental model systems, under the right temporal and spatial conditions, there will be “druggable” targets that appear.

Jeremy Hall
Danny, I agree that detailed study of the NRG1 pathway is very likely to identify "druggable" targets. I think the difficulty will come (as ever) in identifying suitable preclinical models in which to screen targets for potential efficacy before moving forward to Phase 2 and 3.

Amelia Gallitano
All, regarding mice and schizophrenia, my opinion is that two things are of considerable interest: 1. Reversibility of behavioral abnormalities with antipsychotic drugs, and 2. Synaptic plasticity abnormalities.

Liz Tunbridge
Amelia, I agree, except that not all of the symptoms respond to antipsychotics in patients.

Amelia Gallitano
Liz, I agree completely. But, if we see a positive effect of antipsychotics that helps support a potential role of that gene in neurobiology that could be relevant to schizophrenia.

Lorna Role
Hakon/Amelia, I am with David that learning about the biology is key and developing testable "models," even if they are limited, is the best approach/hope for developing therapeutics.

Amanda Law
Amelia, I think the mice data regarding synaptic development and LTP. I mentioned LTD above, but it is the less-commonly studied form of synaptic plasticity. and NMDA signaling, etc., are key to understanding what happens when NRG1 is perturbed during development. I think the antipsychotics could be less clear in terms of the actual mouse response.

David Talmage
Amelia, I have some issue with the antipsychotics in mice question (maybe from lack of understanding). First, I thought these drugs had minimal effect on negative/cognitive symptoms (am I getting this correct?), yet that is presumably all we can semi-confidently assess in rodent behavior models. Second, I see little data to argue for specificity in these drug-treated rodents.

Hakon Heimer
All, we're just few minutes from official finish time. If anyone has to go to lunch/dinner/back to bed (anyone from Asia?), now is the time for final questions for Amanda/David, or sweeping closing statements.

Amanda Law
David, we still need to think about what the state of the system is in schizophrenia. We may not be searching for inhibitors in some contexts.

David Talmage
Amanda, I meant in the inhibitors in the context of in vitro dissection; I wouldn’t go there in vivo at this point (even my in vivo = mice).

Amelia Gallitano
Amanda, I realize that the antipsychotic responses may be variable. But the advantage of such a response in the mouse is a clear link with the disease in humans, since obviously human symptoms respond to these drugs. I am tremendously excited about synaptic plasticity and the NMDA signaling, and this is supported by the NMDAR hypofunction hypothesis as well as cognitive deficits in schizophrenia, but I still feel we are speculating more about the connection between synaptic plasticity and schizophrenia than we are about the connection between the drugs and the disease.

Liz Tunbridge
Amanda, I think an understanding of the roles of different NRG isoforms during development might be key for therapeutics. Presumably, the earlier something happens, the more difficult it is likely to be to reverse it.

Amelia Gallitano
David, you are right regarding the drugs working on negative symptoms, but, actually, the second-generation drugs appear to target negative symptoms better. Truthfully, I think we are still learning about the etiology of mouse behaviors. For example, hyperactivity is one of the classic responses to psychotomimetic drugs, and can be reversed with antipsychotics. The frequency of concordance between this behavior and apparent memory or plasticity deficits in rodents makes me think there may be a connection. Not sure if this addresses your issue....

Hakon Heimer
Tim, I think even Danny will like that one! I like my herring with Erbs.

Daniel Weinberger
I like red herrings. I spent much of my career after some of Tim's.

Daniel Weinberger
Final profound comment: This has been a terrific chat. My thanks to David and Amanda, and Hakon and all the rest.

Hakon Heimer
All, we're officially in repose now, but feel free to rush the speakers.

Liz Tunbridge
Thanks, everyone, This has been great. I have to go in a minute and feed my pet Kozak.

Daniel Weinberger
Bye, everyone. I'm off to try to pry loose some of that stimulus money.

Paul Harrison
Cheers, folks—have to dash home for my fish supper. Thanks, Amanda and David. Let’s do this again sometime.

Orna Tighe
Dear All. Looks like the first NRG1 conference I will attend will be good fun.

Hakon Heimer
All, if Tim will agree to serve some herrings, I hope we can have a follow-up round table with more discussion of development.

Jeremy Hall
Dear Amanda and David (and all), many thanks. That was great. The Edinburgh team will now retire for finest smoked salmon.



	Comments on Online Discussion



	Comment by: Richard Deth
	Submitted 28 January 2009	Posted 29 January 2009
	The alternative splicing of neuregulin-1 is only one of many examples in the brain (BDNF, etc.), and it appears that alternative splicing is more common in the brain than in other tissues. It may, therefore, be useful to also consider the mechanism of alternative mRNA splicing of neuregulin-1 as a candidate process which may underlie schizophrenia. In other words, if the factors regulating alternative splicing functioned abnormally, it would influence a broad number of gene products. SNPs or mutations in these genes would represent a second source of vulnerability. View all comments by Richard Deth

	Comment by: Malcolm Nason
	Submitted 23 February 2009	Posted 23 February 2009
	Considering the strong influence of puberty on the development of schizophrenia, how do you think that Neuregulin affects this process? I am particularly interested in the "chicken and egg" problem, so to speak, since Nrg has direct effects on development of tissues which induce and influence puberty itself. Could a better regulated puberty (if such a thing were possible, say through exogenous hormone treatments) reduce the influence of aberrant Nrg? View all comments by Malcolm Nason

	Comment by: Timothy Crow
	Submitted 24 February 2009	Posted 24 February 2009
	With respect to my colleagues in the Oxford Department, whose technical expertise I do not doubt, this question surely reflects the quagmire into which psychosis genetics has now fallen. 1. Neuregulin came into prominence in this field with the paper by Stefannson et al. (Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet 2002; 71: 877-892) who reported a multipoint lod score of 3.06 on chromosome 8p in 33 families. Much larger linkage surveys are now available and they do not support this finding. I reviewed three whole genome scans (the largest, each n >300 pairs) on a total of 1,144 sibling pairs with schizophrenia or schizoaffective disorder (Crow TJ. How and why genetic linkage has not solved the problem of psychosis: review and hypothesis. Am J Psychiatry 2007; 164: 13-21). Only one of these studies showed a lod peak on chromosome 8 above 1.5 (Suarez et al. Genome-wide linkage scan of 409 European-ancestry and African American families with schizophrenia: suggestive evidence of linkage at 8p23.3-p21.2 and 11p13.1-q14.1 in the combined sample. Am J Hum Genet 2006; 78(2) 315-333) and the authors of that study considered that the peak was not close to neuregulin. Sanders et al (No significant association of 14 candidate genes with schizophrenia in a large European ancestor sample: implications for psychiatric genetics. Am J Psychiatry 2008; 165(4): 497-506) conducted an intensive association study of candidate genes in a population of 1,870 controls and 2,002 patients with a mean of 45 markers across each of the candidate genes including neuregulin and found no evidence of linkage disequilibrium with psychosis. The only conclusion that one can reach is that the Stefannson et al finding is not confirmed with sample sizes of 30 and 300 times greater. If reassurance were needed no evidence of association at this locus is forthcoming from other whole genome association studies (Lencz et al. Converging evidence for a Pseudoautosomal cytokine receptor gene locus in schizophrenia. Mol Psychiatry 2007; 12: 572-580 and O’Donovan et al. Identification of novel schizophrenia loci by genome-wide association and follow-up. Nature Genet 2008; 40: 1053-1055). Interesting though the accumulation of findings shows this gene to be, it can now be asserted with confidence that neuregulin is unrelated to schizophrenia. View all comments by Timothy Crow

	Comment by: Brian Dean
	Submitted 25 February 2009	Posted 25 February 2009
	If we consider other disorders that were first diagnosed on symptom profile presentation, such as diabetes, renal diseases, and congenital heart diseases, it is very likely that the diagnosis of schizophrenia as it is currently used defines a syndrome of disorders. The fact that a significant proportion of subjects with schizophrenia are resistant to antipsychotic drug treatment and therefore probably do not have a dopamine-related disorder clearly supports the notion of schizophrenia as a syndrome. Therefore, in considering any gene as a candidate gene, it must be asked whether it is likely that any gene could strongly associate with a syndrome. This would seem very unlikely as syndromes are usually made up of different disorders that each have unique pathophysiologies. Debates on the roles of molecules should not get bogged down in whether or not the gene for that molecule is a "candidate" gene for schizophrenia. Rather, they should focus on whether the biology of the molecule could implicate it in the pathophysiology of the symptoms of schizophrenia. NRG1 would seem to definitely be involved in pathways that could precipitate the onset of the symptoms associated with schizophrenia. The challenge is therefore to unravel its complex biology to understand how it may act to precipitate such symptoms. View all comments by Brian Dean

Live Discussion: Neuregulin 1 Roundtable 2009