Neuregulin 1 (NRG1) is a trophic factor that can be released presynaptically in a soluble form, and the post-synaptic erbB4 receptor tyrosine kinase is thought to be the predominant receptor for NRG1. NRG1 binds directly to erbB4, and this binding stimulates the intrinsic tyrosine kinase activity of the erbB4 receptor (Corfas et al. 2004; Mei and Xiong 2008). The biological functions of the NRG1 and ERBB4 genes have received much recent attention owing to several studies showing associations between these genes and schizophrenia (Harrison and Law 2006; Buonanno 2010). Nevertheless, the biological functions of NRG1 and erbB4 are incompletely understood.
Almost all NRG1 isoforms are initially trans-membrane-associated proteins termed pro-NRG1s (Mei and Xiong 2008). Proteolytic cleavage of pro-NRG1s causes shedding of the extracellular ecto-domain segment of NRG1 (Wang et al. 2001). On the extracellular side of pro-NRG1s lays the epidermal growth factor-like domain proximal to the membrane, and this EGF-like domain is necessary and sufficient for erbB receptor binding and activation (Mei and Xiong 2008; Buonanno 2010). NRG1 proteins also contain other discrete domains such as an immunoglobulin domain, which in most isoforms lies between the EGF-like domain and the extreme N-terminus (Mei and Xiong 2008).
The target of cleaved, soluble NRG1s is the erythroblastic leukemia viral oncogene (erbB) receptor tyrosine kinase receptors. There are four erbB receptors expressed in the brain, erbB1 (epidermal growth factor receptor), and erbB2-4. Erbb4 is the only receptor isoform that can both directly bind NRG1 and is catalytically active. Given this autonomous function of erbB4 as well as the association of erbB4 with schizophrenia, this receptor isoform has been the most extensively studied (Buonanno 2010). ErbB4 can homodimerize or can form a heterodimer with erbB2; however, unlike erbB4, erbB2 does not directly bind NRG1 (Tzahar et al. 1996; Mei and Xiong 2008).
Historically, the role for acute NRG1 and erbB4 activity in regulating neuronal function has received much attention, and studies have shown that NRG1/erbB4 impede synaptic plasticity in pyramidal neurons. NRG1 can reverse long-term potentiation (LTP) at CA1 hippocampal synapses when applied 20-min after theta burst stimulation (Kwon et al. 2005). In addition to reversing LTP, NRG1 suppresses LTP induction at the Schaffer collateral-CA1 synapse (Chen et al. 2010). NRG1 has been shown to inhibit spontaneous firing rates in prefrontal cortex neurons, and also decreases the number of action potentials resulting from a 300 ms current injection (Wen et al. 2010). Most of the effects of NRG1 on regulating neuronal function are erbB4 dependent (Woo et al. 2007; Chen et al. 2010; Wen et al. 2010).
Long-term NRG1 activity, on the other hand, promotes plasticity, particularly the morphogenesis of dendritic spines on pyramidal neurons, the sites of most excitatory synapses in the brain. Notably, multi-day NRG1 treatment increases spine density and size in cultured forebrain neurons (Barros et al. 2009), and mice lacking NRG1 type III show a reduction in pyramidal neuron spine density (Chen et al. 2008). ErbB4 also has an established role in promoting spine morphogenesis as mice lacking erbB2/B4 show a reduction in spine density in the CA1 hippocampal field and in the prefrontal cortex (Barros et al. 2009). Knocking down erbB4 with a viral RNA interference (RNAi) in the CA1 hippocampal field reduces spine density and area, while the over-expression of erbB4 in pyramidal neurons increases spine size (Li et al. 2007).
Given the links of NRG1 and erbB4 to schizophrenia, and because schizophrenia is characterized by alterations in forebrain spine density, a better understanding of the precise roles for these molecules in regulating spine morphogenesis remains an important question and could shed light on the contribution of these molecules to schizophrenia pathogenesis. Here, we examine the role for NRG1 in regulating spines and determine the mechanisms important for these effects. Major regulators of spine morphogenesis are Rac1 guanine nucleotide exchange factors (GEFs). The kalirin-7 GEF plays a key role in regulating structural and functional plasticity at excitatory synapses (Penzes and Jones 2008), and kalirin has been functionally and genetically implicated in the pathogenesis of schizophrenia, including altered expression levels as well as genetic associations (Hill et al. 2006; Deo et al. 2012; Kushima et al. 2012; Rubio et al. 2012). Kalirin-7 interacts with erbB4, and is a critical regulator of NRG1-mediated interneuronal dendritic growth (Cahill et al. 2012). Here, we wanted to determine the contribution of kalirin to NRG1's effects on dendritic spines, and we show that NRG1 promotes spine morphogenesis in cortical pyramidal neurons and that kalirin is necessary for these effects. NRG1 also enhanced the expression of α-amino-3-hydroxy-5-methylisoxazole-4-propionate receptors in spines. Our data further suggest that erbB4 expression in pyramidal neurons is dispensable for NRG1's effects on spine density, but surprisingly, not for spine area. Finally, because little is known about the means through which erbB4 expression is regulated, we examined how environmental enrichment, which is known to regulate spine morphogenesis, affects erbB4 expression. Overall, these findings help clarify the relationship between NRG1 activity and spine morphogenesis.
- Top of page
- Conflict of interest
- Supporting Information
Here, we show that NRG1 promotes spine morphogenesis in cortical pyramidal neurons and our findings indicate that kalirin is responsible for these effects. Our findings are further supportive of the idea that long-term chronic NRG1 stimulation has trophic effects on pyramidal neuron spines (Barros et al. 2009). Consistent with this, the long-term loss of NRG1 as seen in NRG1 mutant mice results in a reduction in forebrain spine density (Chen et al. 2008).
NRG1 signaling seems to have differential effects on excitatory synapses depending on the length of stimulation. Notably, acute NRG1 treatment (around 30-min) reduces pyramidal neuronal synaptic plasticity (Kwon et al. 2005). Many of these effects are mediated through interneurons as acute NRG1 treatment can stimulate GABA release from interneurons thereby inhibiting pyramidal neuron activity (Wen et al. 2009). That acute NRG1 treatment inhibits pyramidal neuronal activity, while long-term NRG1 promotes plasticity, could indicate that under different circumstances excessive acute NRG1 signaling and/or deficient chronic NRG1 signaling might both impede synaptic plasticity and be relevant to schizophrenia pathogenesis.
Within the forebrain, erbB4 expression is highest in interneurons, and despite evidence that NRG1 and erbB4 have trophic effects on spine morphogenesis, the expression profile of erbB4 in pyramidal neurons is a matter of ongoing debate. Previous studies have detected erbB4 protein expression in the dendritic spines of mature pyramidal neurons using immuno-electron microscopy (immuno-EM) (Mechawar et al. 2007), but the specificity of the antibody used in these EM studies has been questioned, and antibodies against different erbB4 epitopes have turned up negative results (Neddens and Buonanno 2009; Vullhorst et al. 2009). However, several studies using antibodies from different sources against various epitopes of the erbB4 protein have detected protein expression in mature pyramidal neurons, albeit at significantly lower levels than in interneurons (Okada and Corfas 2004; Bernstein et al. 2006). Furthermore, several in situ studies have detected ERBB4 mRNA expression in mature pyramidal neurons (Gerecke et al. 2001; Fox and Kornblum 2005). Complicating matters, single cell RT-PCR experiments have failed to detect ERBB4 in hippocampal pyramidal neurons (Vullhorst et al. 2009), making it uncertain if erbB4 is truly expressed in forebrain pyramidal neurons.
Interestingly, our results indicate that the effect of NRG1 on spine numbers is likely independent of erbB4 expression, as hypothesized by other groups (Vullhorst et al. 2009), suggesting distinct role for erbB4 in mediating NRG1's effects on spine area and density (Figure S4). Moreover, erbB4 over-expression is capable of inducing spine enlargement, but not new spines, in a cell-autonomous fashion. This suggests that erbB4 may not completely underlie NRG1-induced spine growth. Several explanations for the effects of long-term NRG1 activity on facilitating plasticity in pyramidal neurons have been offered. Namely, that the trophic effects are because of the direct activation of low-levels of erbB4 receptors on pyramidal neurons (Li et al. 2007), or conversely, that these effects are compensatory mechanisms to preserve excitatability in the face of increased GABAergic input to pyramidal neurons resulting from NRG1 (Buonanno 2010). Similar to other studies (Li et al. 2007), we found that erbB4 over-expression increases spine size. In addition, we found that the knockdown of erbB4 cell autonomously blocked NRG1-mediated increases in spine size, and a previous study found that erbB4 knockdown in pyramidal neurons reduced AMPAR-mediated transmission in a cell-autonomous manner (Li et al. 2007). This leaves open the possibility that low levels of pyramidal erbB4 expression mediate some of NRG1's effects on plasticity. It is also possible that NRG1's effects are mediated by receptors other than erbB4. In this later case, over-expression of erbB4 might engage their downstream signaling without actually being expressed in these cells. Another possibility is that environmental and therapeutic treatments enhance erbB4 expression, which is than able to further enhance the spine growth promoting effects of NRG1. In this case, up-regulation of erbB4 might be a mediator of some therapeutic approaches.
The cleavage of NRG1 is critical for the feed-forward signaling of most isoforms. NRG1 processing occurs after delivery of NRG1 to the cell surface (Loeb et al. 1998), and numerous lines of evidence indicate that synaptic activity increases levels of NRG1 (Eilam et al. 1998; Loeb et al. 1998; Ozaki et al. 2004). It is thus likely that neurons are exposed to constitutive levels of NRG1, making it important to determine the effects on long-term NRG1 activity on neurons. Our findings, in combination with previous studies (Li et al. 2007), indicate that chronic NRG1 activity increases glutamate receptor levels in spines, which would cause increased synaptic transmission onto pyramidal neurons. This increased synaptic transmission would be expected to further amplify NRG1 release contributing to long-term and sustainable plasticity (Mei and Xiong 2008). These trophic effects of NRG1 on synaptic plasticity could contribute to the neuroprotective effects of chronic NRG1 in animal models of neurodegeneration (Shyu et al. 2004; Guo et al. 2006; Woo et al. 2012).
We also found that that kalirin-7, a key regulator of spine remodeling, is important for NRG1-dependent increases in spine morphogenesis. The importance of investigating interactions between kalirin-7 and schizophrenia-associated genes is underscored by a recent study showing that the disrupted-in-schizophrenia protein regulates spine morphogenesis through a signaling complex involving kalirin-7. Other studies have found a reduction in kalirin mRNA expression in the prefrontal cortex of schizophrenia patients (Hill et al. 2006; Narayan et al. 2008), and the protein expression of kalirin-7 is reduced in the schizophrenia dorsolateral prefrontal cortex and anterior cingulate cortex (Rubio et al. 2012). Furthermore, kalirin loss correlates strongly with prefrontal cortical spine loss in schizophrenia patients irrespective of anti-psychotic treatment (Hill et al. 2006). Interestingly, a reduction in cortical spine density has been detected in NRG1, erbB4, and kalirin mutant mice, and mice hypomorphic or with mutations in these genes show similar behavioral phenotypes, including locomotor hyperactivity and deficits in pre-pulse inhibition, spatial working memory, and social behavior (Gerlai et al. 2000; Stefansson et al. 2002; Golub et al. 2004; Clapcote et al. 2007; Pletnikov et al. 2008; Barros et al. 2009; Cahill et al. 2009).
We found that environmental enrichment robustly up-regulates the expression of erbB4 protein in both the cortex and hippocampus. Schizophrenia is thought to result from a complex interaction between genes and the environment (Le Strat et al. 2009), and understanding how environmental manipulations affect the expression of schizophrenia susceptibility molecules could provide a more complete understanding of potential disease processes. The effect of therapeutic approaches on the expression profile of erbB4 has received recent attention. Notably, treatment with the anti-psychotic clozapine increases forebrain erbB4 expression in rodents (Wang et al. 2008). Overall, this suggests that altered erbB4 levels might be related to certain treatment strategies. Enrichment has been shown to alleviate deficits in neuronal ultrastructure, including deficits in spine density, in numerous animal models of neuropsychiatric disorders (Nithianantharajah and Hannan 2006). In humans, early life participation in enrichment programs has been shown to pre-emptively reduce schizotypal personality behavior, including reduced cognitive disorganization (Raine et al. 2003). The expression profile of erbB4 in schizophrenia remains complex, as some studies suggest that over-expression of the receptor occurs in schizophrenia (Law et al. 2007), while other findings indicate that erbB4 activity is decreased in the disease (Walsh et al. 2008). Nevertheless, our findings would support enrichment as a potentially useful model for studying how alterations in rodent erbB4 expression levels impact brain structure and function, with potential disease implications.
- Top of page
- Conflict of interest
- Supporting Information
This study was supported by grants from NIH-NIMH (R01MH071316, R01MH097216), National Alliance for Research on Schizophrenia and Depression (NARSAD) to P.P., Ruth L. Kirschstein National Research Service Awards 1F31AG031621-01A2 to M.E.C and 1F31MH085362 to K.A.J, and research grants from NIH (R01 MH 071533, R01 AG027224) and from the VHA (I01 BX000452) to RAS. All experiments involving animals were done according to the Institutional Animal Care and Use Committee of Northwestern University. M.E.C, C.R, K.A.J., Z.X. performed research and analyzed results. M.E.C., C.R., and P.P. wrote the manuscript, and P.P. and R.A.S. contributed resources necessary to complete this study.