Schizophrenia Research Forum - A Catalyst for Creative Thinking

Unkind Cuts of NRG3 May Lead to Schizophrenia

15 September 2010. Neuregulin 3, the little-understood cousin of neuregulin 1, has been drawing attention as a schizophrenia candidate gene itself. Two recent studies back earlier hints that it plays some role in the disorder. In a study published June 15 in Molecular Psychiatry online, Assen Jablensky and colleagues find that single-nucleotide polymorphisms (SNPs) in NRG3 may contribute to a cognition-sparing form of schizophrenia, but not to schizophrenia with cognitive deficits. In the August 16 online version of the Proceedings of the National Academy of Sciences, Amanda Law and colleagues tie these same polymorphisms to schizophrenia and to the severity of delusions. They also reveal a plethora of new splice variants of the protein, some of which are altered in postmortem schizophrenia brain tissue, as well as evidence that the clinical risk SNPs can alter relative expression levels of certain splice variants.

Family ties
NRG3 and its relatives in the neuregulin gene family make cell-cell signaling proteins (see Falls, 2003; Taveggia et al., 2005; Zhang et al., 1998). These neuregulin proteins serve as ligands for receptor tyrosine kinases of the ERbB family. Like NRG1 (see SRF related news story; SRF news story; SRF news story), NRG3 binds and activates ErbB4, which is encoded by another schizophrenia candidate gene (see SRF related news story).

In the only species that has been characterized—the mouse—NRG3 expression occurs mainly in the central nervous system, during both development and adulthood (Zhang et al., 1997; see Entrez Gene entry), and the gene inhabits chromosome region 10q22-q23, which has been implicated in schizophrenia (Fallin et al., 2003). Research also fingers this region in cognitive impairment and autism, among others (Balciuniene et al., 2007).

Studies have tied NRG3 single-nucleotide polymorphisms (SNPs) to schizophrenia in Chinese (Wang et al., 2008) and Scottish (Benzel et al., 2007) populations (see SRF related news story; also see the SZGene entry on NRG3). However, a study by Chen and colleagues in subjects of Ashkenazi Jewish descent found no clear relationship between NRG3 gene markers and schizophrenia itself, although it did connect three SNPs—rs10883866, rs6584400, and rs10748842—to a quantitative measure of delusions (Chen et al., 2009).

A mixed bag
The findings from Chen and colleagues sparked the interest of Jablensky, first author Bharti Morar, and others at the University of Western Australia in Perth. They sought to confirm and extend the findings in a new sample. To do so, they genotyped 411 patients with schizophrenia and 223 healthy control subjects for rs10883866 and rs6584400, the SNPs most strongly related to delusions in the Chen study. Case-control analyses of these mostly Anglo-Irish descendents tied rs6584400 to schizophrenia, in contrast to the Ashkenazi study.

Subjects also completed neurocognitive tests, and the researchers plugged the results into a latent structure analysis. This uncovered three groups of patients: 1) 180 with pervasive cognitive deficits, 2) 148 with relatively spared cognition, and 3) 83 patients, age 45 and up, who showed mild cognitive impairment, mainly slow information processing.

Morar and colleagues compared control subjects with the first two groups of patients, skipping the third group because their deficits might be age-related. As it turned out, the association of rs6584400 with schizophrenia came only from the group of patients with relatively intact cognition. Prior studies portray such patients as having Schneiderian first-rank symptoms, florid delusions, and positive thought disorder, in contrast to those with greater cognitive impairment. The latter tend to exhibit negative symptoms and impaired social functioning, but not complex delusions (see Jablensky, 2006). Merging these findings with their own and those from the Chen study, Morar and colleagues suggest that NRG3 contributes “to a subtype of schizophrenia with relatively preserved cognitive function but prone to florid and complex delusions.”

The researchers also checked the allelic association of the two SNPs to specific cognitive traits. For both rs10883866 and rs6584400, they found no connection to general intelligence or verbal memory, but these SNPs’ minor alleles did relate to performance on the Degraded Stimulus-Continuous Performance Task. This requires paying attention to a speedy presentation of a series of unclear digits.

The alleles seemed to affect patients and control subjects differently, with the minor alleles tied to better performance in subjects with schizophrenia and worse performance in healthy subjects. “This suggests that NRG3 may be modulating early attentional processes for perceptual sensitivity and vigilance, with opposing effects in affected individuals and healthy controls,” wrote Morar and colleagues.

Allelic flip-flop
Whether NRG3 promotes one form of schizophrenia or the disease overall, the mystery of how it might do so and what functions it serves prompted the study by Law and colleagues. Law, first author Wee-Tin Kao, and others at the National Institute of Mental Health, Bethesda, Maryland, started by checking for associations between NRG3 polymorphisms on the one hand, and schizophrenia or its symptoms on the other. They also explored possible explanations for how genetic variation in NRG3 might contribute to schizophrenia by altering the expression of different isoforms of the protein.

In family-based analyses, the team found associations between schizophrenia and 12 SNPs in a noncoding part of NRG3. That stretch includes the aforementioned rs6584400, rs10748842, and rs10883866. Deconstructing schizophrenia, Kao and colleagues verified earlier findings relating these three SNPs to delusions, measured in this study by the delusion severity subscale of the Positive and Negative Syndrome Scale. They also connected the three SNPs to the severity of positive symptoms and tied rs10748842 to negative symptom load.

However, the “replication” proved less than straightforward. In contrast to the Chen and Morar studies, which fingered the rarer alleles, the new study blamed the common ones. According to Kao and colleagues, ethnic differences in the populations studied might explain these contrasting results. Their own family-based analyses examined white Americans of Western European descent. "Flipping of disease alleles in different populations has been observed for a number of other diseases, such as autism (HTTLPR locus) and Alzheimer disease (apolipoprotein E ε4-related polymorphisms)," the authors write, adding that this "has been suggested to be a valid biological phenomenon as a result of heterogeneous effects of the same variant related to multi-locus interactions (i.e., effects of epistasis), environment, or LD, with causal variants that emerged on different genetic backgrounds."

They note, however, that the protective alleles in their study (at rs10883866 and rs6584400) are the ones associated with ”better” cognitive performance in patients with schizophrenia in the study by Morar and colleagues, "suggesting complex genetic association with the schizophrenia phenotype."

Many ways to splice it
To learn more about what the gene's products, Kao and colleagues sequenced NRG3 clones from adult human hippocampus and whole brain. They found that, through alternative splicing, NRG3 makes 15 different isoforms (see SRF related news story on similar findings for NRG1). In data that may tie some of these isoforms to disease, the researchers found greater expression of certain splice variants in dorsolateral prefrontal cortex tissue from subjects with schizophrenia versus healthy control subjects.

The researchers thought that rs10748842, a standout in the family-based analyses, might regulate NRG3 expression in the brain. Although it is a noncoding nucleotide, rs10748842 lies near a site that regulates transcription of a fetal form of NRG3, and mathematical modeling put it in the middle of a binding unit for a family of transcription factors. These clues led the researchers to try to connect allelic variation at rs10748842 to the expression of different NRG3 transcripts. They write that such variation "strongly predicts expression (P = 1.097E−12 to 1.445E−15) of specific, developmentally regulated NRG3 isoforms in the normal and developing human brain and in schizophrenia, whereby the ancestral (T) allele is associated with elevated expression." Again, the common allele seemed to bestow greater risk than the rare one.

This study adds flesh to the shadowy NRG3 and further builds the case for it as a schizophrenia candidate gene, and, with the work from the Morar team, supplies plenty of leads for interested researchers to untangle.—Victoria L. Wilcox.

Kao W-T, Wang Y, Kleinman JE, Lipska BK, Hyde TM, Weinberger DR, Law AJ. Common genetic variation in Neuregulin 3 (NRG3) influences risk for schizophrenia and impacts NRG3 expression in human brain. Proc Natl Acad Sci U S A. 2010 Aug 16. Abstract

Morar B, Dragović M, Waters FAV, Chandler D, Kalaydjieva L, Jablensky A. Neuregulin 3 (NRG3) as a susceptibility gene in a schizophrenia subtype with florid delusions and relatively spared cognition. Mol Psychiatry. 2010 Jun 15. Abstract

Comments on News and Primary Papers
Comment by:  Assen Jablensky
Submitted 15 September 2010
Posted 15 September 2010

Common or rare genetic variation in NRG3 influences risk for schizophrenia?
Emerging evidence implicating NRG3 as a likely susceptibility gene in population samples as diverse as the Ashkenazi Jews, Han Chinese, Australians of Anglo-Irish ancestry, and white Americans is certainly a “noteworthy” occurrence in schizophrenia genetics. The latest addition to the evidence (Kao et al., 2010) provides considerable support to earlier (Fallin et al., 2003; Wang et al., 2008) and recent findings of association of several polymorphisms (rs10883866, rs6584400, rs10748842) within a conserved linkage disequilibrium (LD) block in intron 1 of the NRG3 gene with a delusion-laden factor and a neurocognitive quantitative trait in the schizophrenia phenotype (Chen et al., 2009; Morar et al., 2010).

A fundamental contribution of the present study is the cloning and detailed characterization of full-length NRG3 transcripts from postmortem fetal, child, adolescent, and adult brain samples (whole brain, hippocampus, and dorsolateral prefrontal cortex). Sequencing of the cDNA clones and expression analysis revealed a complex picture of alternative splicing, abundance of developmentally regulated transcripts in schizophrenia brains, and, notably, increased expression of a fetal brain-derived clone (hFBNRG3), which introduces a premature stop codon resulting in a truncated protein and a possibly destabilized NRG3-ErbB4 signalling pathway. In their clinical collections (a family-based sample and a partially independent case-control sample), the authors report significant associations of rs10748842 (representing 12 SNPs located in the LD block within intron 1) with schizophrenia, with the PANSS (Positive and Negative Syndrome Scale) subscale score on delusion severity, as well as with the PANSS negative symptom load.

Overall, the findings from this investigation and the earlier studies appear to be in a broad agreement, converging on a plausible role of NRG3 in schizophrenia pathogenesis. However, there is a fly in the ointment: The associations found in the present study exhibit a risk allele reversal compared to previously reported results; namely, all significant associations are with the major, common alleles, rather than with the minor alleles, as in Chen et al. (2009) and Morar et al. (2010). While many reasons for genuine allele flipping can be invoked (multi-locus interactions, variation in local patterns of LD, environmental exposures, ethnic background differences—see Clarke and Cardon, 2010), the explanation for the flip in this particular context is not obvious, and NRG3 should remain on the examination bench. Even in the GWAS era, studies proceeding from biologically and clinically anchored hypotheses remain rewarding and potentially productive.


Chen PL, Avramopoulos D, Lasseter VK, McGrath JA, Fallin MD, Liang K-Y, Nestadt G, Feng N, Steel G, Cutting AS, Wolyniec P, Pulver AE, Valle D. Fine mapping on chromosome 10q22-q23 implicates Neuregulin 3 in schizophrenia. Am J Hum Genet. 2009;84:21-34. Abstract

Clarke GM, Cardon LR. Aspects of observing and claiming allele flips in association studies. Genet Epidemiol. 2010;34:266-74. Abstract

Fallin MD, Lasseter VK, Wolyniec PS, McGrath JA, Nestadt G, Valle D, Liang KY, Pulver AE. Genomewide linkage scan for schizophrenia susceptibility loci among Ashkenazi Jewish families shows evidence of linkage on chromosome 10q22. Am J Hum Genet. 2003;73:601-11. Abstract

Kao WT, Wang Y, Kleinman JE, Lipska BK, Hyde TM, Weinberger DR, Law AJ. Common genetic variation in Neuregulin 3 (NRG3) influences risk for schizophrenia and impacts NRG3 expression in human brain. Proc Natl Acad Sci U S A. 2010 Aug 31;107(35):15619-24. Abstract

Morar B, Dragovic M, Waters FAV, Chandler D, Kalaydjieva L, Jablensky A. Neuregulin 3 (NRG3) as a susceptibility gene in a schizophrenia subtype with florid delusions and relatively spared cognition. Mol Psychiatry. 2010 June 15. Abstract

Wang YC, Chen JY, Chen ML, Chen CH, Lai IC, Chen TT, Hong CJ, Tsai SJ, Liou YL. Neuregulin 3 genetic variations and susceptibility to schizophrenia in a Chinese population. Biol Psychiatry. 2008;64:1093-6. Abstract

View all comments by Assen Jablensky

Comments on Related News

Related News: Neuregulin Partner ErbB4 Spices Up Genetic Associations

Comment by:  Amanda Jayne Law, SRF Advisor
Submitted 22 February 2006
Posted 22 February 2006
  I recommend the Primary Papers

The study of Ghashghaei and colleagues provides a remarkable insight into the function of neuregulin 1 (NRG1), and NRG2 in adult neurogenesis. The study demonstrates that NRG1(2)/ErbB4 signaling influences the proliferation, differentiation, organization, and migration of adult neural progenitor cells in the subventricular zone (SVZ) and rostral migratory stream (RMS), in a ligand- and cell-dependent fashion. Using immunohistochemistry, Ghashghaei and colleagues first demonstrate that NRG1, NRG2, and ErbB4 are expressed by distinct cell types in the SVZ and RMS, notably ErbB4 and NRG1 by polysialylated neural cell adhesion molecule positive (PSA-NCAM+) neuroblasts, and ErbB2/3/4 by a subset of GFAP+ cells. These observations extend the group's previous studies of NRG1 and ErbB4 in the RMS (Anton et al., 2004). In their current study, Ghashghaei went on to examine the effects of exogenous infusion of NRG1 and NRG2 on neurogenesis in the RMS of adult mice. Interestingly, NRG1 was shown to decrease the initiation of neuroblast migration from the SVZ to the RMS by inducing the rapid aggregation of cells in the SVZ. The consequence of this rise in NRG1 was a decrease in the number of PSA-NCAM+ cells in the RMS and GABA+ cells in the olfactory bulb, demonstrating that ectopic or elevated expression of NRG1 prevents differentiation and migration of neurons from the adult SVZ to the RMS.

The study is particularly interesting in terms of the role of NRG1/ErbB4 signaling in directional cell migration. Flames et al. (2004) recently reported that NRG1 (specifically the Ig containing family of isoforms, e.g., Types I, II and IV; for review, see Harrison and Law, 2006) functions as a long distance chemoattractant for ErbB4 positive GABAergic interneurons migrating from the medial ganglionic eminence to the developing cortex. The observation that NRG1 is a chemoattractant in other brain regions may appear somewhat contradictory to the findings of Ghashghaei, which suggest that in-vivo NRG1 actually inhibits migration of neurons from the SVZ (at least when introduced ectopically). However, it would seem that these two findings are actually consistent. Ghashghaei and colleagues ectopically infused NRG1 into the lateral ventricles of adult mice. The subsequent aggregation of cells in the SVZ demonstrates that NRG1 indeed acts as a chemoattractant, not in an obvious manner by inducing the cells to migrate away, but simply by "attracting" them to aggregate or "clump" where they are (subsequently preventing migration to the RMS). So in fact, both the studies of Flames and Ghashghaei show that NRG1 is chemotactic to specific populations of neurons and cells, whether it is expressed at a distance and cells preferentially migrate toward it, or in the immediate environment and cells are attracted to migrate to, or stay in its vicinity.

In the past few years, NRG1 and ErbB4 have both been identified as potential susceptibility genes for schizophrenia. The aim now is to determine the molecular and biological mechanisms by which the genes confer risk for the disease. In terms of schizophrenia, we have previously demonstrated that the Type I isoform of NRG1 is elevated in the hippocampus (and prefrontal cortex; see Hashimoto et al., 2004) in the disease and that expression of the novel Type IV isoform is related to disease-associated sequence variants within the NRG1 gene (Law et al., 2006). Furthermore, we have recently demonstrated that these changes are accompanied by altered expression of specific isoforms of the ErbB4 receptor, consistent with that of Silberberg et al., 2006 (Law et al., 2005). Ghashghaei and colleagues provide the first direct evidence that ectopic or elevated expression of NRG1 in the brain can perturb cell migration. In light of this and other evidence, our findings in schizophrenia may translate into altered neuronal migration, cortical development and possibly neurogenesis in the disease.

At present, the exact links between altered NRG1/ErbB4 signaling and the pathophysiology of schizophrenia are unknown and potentially numerous (i.e., synaptogenesis, neurotransmitter function, neuronal migration, differentiation, glia formation and function, myelination). Studies such as that of Ghashghaei et al. provide insight into the normal role of NRG1/ErbB4 signaling in neurodevelopment and the adult brain which is essential if we are to understand the pathogenic role of the NRG1 gene and its receptors in disease.


Anton ES, Ghashghaei HT, Weber JL, McCann C, Fischer TM, Cheung ID, Gassmann M, Messing A, Klein R, Schwab MH, Lloyd KC, Lai C. Receptor tyrosine kinase ErbB4 modulates neuroblast migration and placement in the adult forebrain. Nat Neurosci. 2004 Dec;7(12):1319-28. Epub 2004 Nov 7. Abstract

Flames N, Long JE, Garratt AN, Fischer TM, Gassmann M, Birchmeier C, Lai C, Rubenstein JL, Marin O. Short- and long-range attraction of cortical GABAergic interneurons by neuregulin-1. Neuron. 2004 Oct 14;44(2):251-61. Abstract

Hashimoto et al., 2004, Mol. Psychiatry 9, 299-307.

Law et al (a) 2006. Neuregulin 1 (NRG1) transcripts are differentially expressed in schizophrenia and regulated by 5’ SNPs associated with the disease. PNAS

Also See SfN 2005 SRF research news: Cortical Deficits in Schizophrenia: Have Genes, Will Hypothesize

Law 2005, SNPing away at NRG1 and ErbB4 gene expression in schizophrenia Neuropsychopharmacology, vol. 30, Supplement 1.

View all comments by Amanda Jayne Law

Related News: Second Neuregulin Gene Supported In Schizophrenia Genetic Studies

Comment by:  Lin Mei
Submitted 30 January 2009
Posted 30 January 2009

Diagnosis of schizophrenia is psychiatric and based on patients’ self-reported experience or behavior. There are positive symptoms, negative symptoms, and cognitive deficits. Evidently, schizophrenia is a complex disease. No wonder current antipsychotics are effective for some, but not for all symptoms (in particular, negative symptoms and cognitive deficits). It would be no surprise if schizophrenia is classified or subclassified into different disorders in future when tools are available.

The beauty of the current study is to take advantage of detailed clinical information of schizophrenia patients and classify them to generate nine factors. Remarkably, a strong association was identified by using the delusion factor as quantitative trait phenotype. The finding itself identifies NRG3 as a candidate gene of "schizophrenia." In addition, it could represent a first step toward subclassification of schizophrenia for better diagnosis and treatment.

View all comments by Lin Mei