Schizophrenia susceptibility pathway neuregulin 1–ErbB4suppresses Src upregulation of NMDA receptors

Graham M Pitcher1,2, Lorraine V Kalia1,3,5, David Ng1,5, Nathalie M Goodfellow2, Kathleen TYee4, Evelyn K Lambe2, and Michael W Salter1,2,3

1Program in Neurosciences and Mental Health, The Hospital for Sick Children, Toronto, Ontario,Canada2Department of Physiology, University of Toronto, Ontario, Canada3Institute of Medical Sciences, University of Toronto, Ontario, Canada4Department of Anatomy and Cellular Biology, Tufts University School of Medicine, Boston,Massachusetts, USA

AbstractHypofunction of the N-methyl D-aspartate subtype of glutamate receptor (NMDAR) ishypothesized to be a mechanism underlying cognitive dysfunction in individuals withschizophrenia. For the schizophrenia-linked genes NRG1 and ERBB4, NMDAR hypofunction isthus considered a key detrimental consequence of the excessive NRG1-ErbB4 signaling found inpeople with schizophrenia. However, we show here that neuregulin 1β–ErbB4 (NRG1β-ErbB4)signaling does not cause general hypofunction of NMDARs. Rather, we find that, in thehippocampus and prefrontal cortex, NRG1β-ErbB4 signaling suppresses the enhancement ofsynaptic NMDAR currents by the nonreceptor tyrosine kinase Src. NRG1β-ErbB4 signalingprevented induction of long-term potentiation at hippocampal Schaffer collateral–CA1 synapsesand suppressed Src-dependent enhancement of NMDAR responses during theta-burst stimulation.Moreover, NRG1β-ErbB4 signaling prevented theta burst–induced phosphorylation of GluN2B byinhibiting Src kinase activity. We propose that NRG1-ErbB4 signaling participates in cognitivedysfunction in schizophrenia by aberrantly suppressing Src-mediated enhancement of synapticNMDAR function.

The NMDAR, a major excitatory ligand-gated ion channel in the central nervous system anda principal mediator of synaptic plasticity, is a multiprotein complex whose core is aheterotetramer comprising two obligate GluN1 subunits and two GluN2(A–D) subunits1,2.Proteins associated with the core NMDAR have key roles in trafficking, stability, subunit

© 2011 Nature America, Inc. All rights reserved.Correspondence should be addressed to M.W.S. (mike.salter@utoronto.ca).5These authors contributed equally to this work.Note: Supplementary information is available on the Nature Medicine website.AUTHOR CONTRIBUTIONSG.M.P. designed the project, conducted the hippocampal experiments, analyzed the data and wrote the manuscript. L.V.K. and D.N.carried out and analyzed biochemical experiments. E.K.L. oversaw and analyzed the prefrontal cortex experiments. N.M.G. carriedout and analyzed the prefrontal cortex experiments. K.T.Y. maintained, housed and provided ErbB4 knockout mice. All authorsparticipated in revising the manuscript and agreed to the final version. M.W.S. conceived the study, analyzed data, supervised theoverall project and wrote the manuscript.COMPETING FINANCIAL INTERESTSThe authors declare no competing financial interests.Reprints and permissions information is available online at http://npg.nature.com/reprintsandpermissions/.

PubMed Central CANADAAuthor Manuscript / Manuscrit d'auteurNat Med. Author manuscript; available in PMC 2012 January 23.

Published in final edited form as:Nat Med. 2011 April ; 17(4): 470–478. doi:10.1038/nm.2315.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

composition and function of NMDARs3. A key process regulating NMDAR activity isphosphorylation by NMDAR-associated kinases. Certain NMDAR-dependent functions,such as ventilation and locomotion, may be independent of NMDAR phosphorylation4,whereas phosphorylation-induced upregulation of NMDARs is critical for synapticplasticity5–7.

A prominent hypothesis for the pathophysiology of schizophrenia is that core symptomssuch as hallucinations and cognitive deficits may be due to hypofunction of theNMDAR8–10. In support of this hypothesis, administering NMDAR-blocking drugs tootherwise normal adults elicits hallucinations and cognitive deficits11. However, NMDAR-blocking drugs do not discriminate between the phosphorylation-dependent andphosphorylation-independent states of the NMDAR. Thus, the pathophysiology ofschizophrenia may involve reduction of basal NMDAR function or of phosphorylation-mediated enhancement of NMDAR function. In support of the latter, tyrosinephosphorylation of NMDAR subunit proteins is reduced in schizophrenic subjects ascompared with controls12. The predominant tyrosine kinase mediating NMDAR tyrosinephosphorylation is Src, which enhances NMDAR channel activity13. These findings suggestthe possibility that, in schizophrenia, Src-mediated enhancement of NMDAR function issuppressed specifically, rather than there being general hypofunction of NMDARs.

Here we investigated this possibility by determining whether there is a link betweenglutamatergic dysfunction and the candidate schizophrenia genes, Nrg1 and Erbb414–24,which encode the ligand-receptor pair neuregulin 1 (NRG1) and ErbB4, respectively. Inmouse models, behavioral signs of schizophrenia are found in mice heterozygous for Nrg1or Erbb4 (refs. 14,25–27) and also in mice overexpressing NRG1 selectively in thebrain28,29. In studies of schizophrenic individuals, NRG1 expression is increased in both thecortex30 and hippocampus31, where NRG1-ErbB4 signaling is excessive12,31. As NRG1βblocks NMDAR-dependent long-term potentiation (LTP) at hip-pocampal Schaffercollateral–CA1 synapses32–37, a form of LTP also dependent on Src activity5,38,39, wedetermined the effect of NRG1β-ErbB4 signaling on Src-mediated enhancement ofNMDAR function, tyrosine phosphorylation of NMDARs and the resultant potentiation ofsynaptic transmission. We examined neuronal responses not only in the hippocampus butalso in the prefrontal cortex (PFC); both of these brain regions are crucial in thepathobiology of cognitive dysfunction in schizophrenia21,40–43.

RESULTSNRG1β-ErbB4 blocks Src enhancement of NMDAR EPSCs in CA1

To determine whether NRG1-ErbB4 signaling affects Src-mediated enhancement ofNMDAR function, we made whole-cell recordings from visually identified neurons in theCA1 pyramidal layer in acute hippocampal slices from adult animals. We evoked synapticresponses by stimulating the Schaffer collateral afferent input to CA1, and wepharmacologically isolated NMDAR-mediated excitatory postsynaptic currents (EPSCs) bybathing the slices with extracellular solution containing the AMPA receptor antagonistCNQX (10 μM). To prevent potential effects on GABAA-mediated inhibition, the GABAAreceptor antagonist bicuculline methochloride (10 μM) was present in all experiments. Weactivated Src by using the phosphopeptide EPQ(pY)EEIPIA, which binds the SH2 domainof the kinase, preventing the binding of the C-terminal inhibitory phosphotyrosine5,13.

During recordings in which EPQ(pY)EEIPIA was administered intracellularly, we foundthat NMDAR EPSC amplitude progressively increased over 10–15 min to reach 218 ± 16%(mean ± s.e.m.) of the initial level, whereas NMDAR EPSCs were stable during recordingswithout EPQ(pY)EEIPIA (117 ± 7% of the initial level; P < 0.01 compared with

Pitcher et al. Page 2

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

EPQ(pY)EEIPIA; Fig. 1a,b). However, when we bath-applied a soluble form of NRG1,NRG1β (2 nM), 20 min before recordings in which EPQ(pY)EEIPIA was intracellularlyadministered, NMDAR EPSCs did not increase during 30 min of whole-cell recording (Fig.1a,b). EPQ(pY)EEIPIA potentiated NMDAR EPSCs in neurons from wild-type (WT) mice(Src+/+) but not in neurons from Src-null mutant mice (Src−/−; Supplementary Fig. 1a).Thus, NRG1β prevented the enhancement of NMDAR currents by Src.

NRG1β might have caused a change in the blockade of NMDARs by Mg2+ or reversalpotential that could have masked an increase in amplitude of NMDAR EPSCs if examinedonly at a single holding potential. We found, however, that in neurons recorded withEPQ(pY)EEIPIA and treated with NRG1β, the current-voltage relationship was not differentfrom that of neurons recorded with EPQ(pY)EEIPIA without NRG1β or of control neuronswithout EPQ(pY)EEIPIA (Fig. 1c). Thus, NRG1β prevented the EPQ(pY)EEIPIA-inducedincrease in NMDAR EPSCs but did not affect the voltage-dependent blockade of NMDARsby extracellular Mg2+ or the NMDAR EPSC reversal potential.

To determine whether the suppression of the Src-mediated enhancement of NMDARcurrents by NRG1β requires ErbB4, we investigated the effect of NRG1β on theEPQ(pY)EEIPIA-induced potentiation of NMDAR EPSCs in hippocampal neurons lackingErbB4. We used mutant mice in which Erbb4 had been deleted but in which human ErbB4expression was driven in the heart by the α-myosin heavy chain promoter(Erbb4−/−HER4heart), which rescued the otherwise lethal cardiac defect44. In NRG1β-treated hippocampal slices from adult Erbb4−/−HER4heart mice, we found thatEPQ(pY)EEIPIA-induced potentiation of NMDAR EPSCs reached 241 ± 49%, whereasNMDAR EPSCs remained stable in NRG1β-treated hippocampal slices from WT mice (115± 14% of the initial level; P < 0.001 compared with Erbb4−/−HER4heart; Fig. 1d,e). Thus,Erbb4 is necessary for suppression of Src-dependent enhancement of synaptic NMDARcurrents by NRG1β.

We also determined the effect of acute ErbB4 inhibition on Src-dependent enhancement ofsynaptic NMDAR currents in adult WT neurons to find whether the lack of effect of NRG1βon the enhancement in Erbb4−/−HER4heart neurons was due to the absence of ErbB4signaling in the adult or during development. To this end, we used an ErbB4 inhibitor,AG1478, which prevents ErbB4 signaling45,46. We found that bath-applied AG1478prevented NRG1β-mediated suppression of the Src enhancement of synaptic NMDARcurrents (Fig. 1d,e). Considering our findings together, we conclude that ErbB4 mediates theblockade of Src-induced enhancement of NMDAR EPSCs by NRG1β.

We next applied NRG1β in recordings with EPQ(pY)EEIPIA at a time point when theenhancement of NMDAR EPSCs had developed and found that the amplitude of the currentsreturned toward the initial level (Supplementary Fig. 1b), indicating that EPQ(pY)EEIPIA-induced enhancement of NMDAR EPSCs could be reversed by NRG1β. In contrast, wefound that NRG1β had no effect on the amplitude (Fig. 2a), time course (Fig. 2b) or voltagedependence (Supplementary Fig. 1d) of NMDAR ESPCs in neurons in whichEPQ(pY)EEIPIA was not administered. Moreover, in such recordings NMDAR EPSCs werenot affected by bath-applying a broad-spectrum ErbB kinase inhibitor, PD158780 (Fig. 2c).Thus, basal NMDAR synaptic responses were not affected by NRG1β-ErbB4 signaling,consistent with evidence that synaptic NMDARs are not tonically upregulated by Src inCA1 neurons5,47,48. Our findings indicate that NRG1β prevents and reverses theenhancement of synaptic NMDAR currents by Src but does not affect basal NMDARfunction in CA1 neurons.

Pitcher et al. Page 3

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

NRG1β blocks Src enhancement of NMDAR EPSCs in PFCTo investigate whether NRG1β affects Src-induced enhancement of synaptic NMDARs inanother brain region implicated in the pathogenesis of schizophrenia, we studied theprefrontal cortex43,49,50. In acute slices from the medial prefrontal cortex, we evokedNMDAR EPSCs in layer V pyramidal neurons by stimulating the corticocortical neuronsand afferents in layers II and III while pharmacologically blocking AMPA and GABAreceptors. During recordings in which EPQ(pY)EEIPIA was administered intra-cellularly,NMDAR EPSC amplitude progressively increased over 10–15 min to reach 144 ± 19% ofthe initial level, whereas NMDAR EPSCs were stable during recordings withoutEPQ(pY)EEIPIA (Fig. 2d,e). However, we found that in the presence of NRG1β (6 nM),intracellular administration of EPQ(pY)EEIPIA did not cause an increase in NMDAREPSCs. In layer V pyramidal neurons treated with NRG1β, but in which EPQ(pY)EEIPIAwas not administered, NMDAR EPSCs were stable (Fig. 2d,e). Therefore, similarly tohippocampal CA1 neurons, layer V pyramidal cells in prefrontal cortex show a Src-mediatedenhancement of synaptic NMDAR currents that is prevented by NRG1β.

Src potentiation of EPSPs is prevented by NRG1βSrc-enhanced NMDAR currents initiate Ca2+-mediated potentiation of AMPAR synapticresponses5,7,51. Here we found that during recordings in which EPQ(pY)EEIPIA wasdelivered into CA1 neurons through a patch pipette, the slope of the excitatory postsynapticpotential (EPSP) gradually increased to 163 ± 14% of the initial level (Fig. 3a,b). However,the enhancement of EPSPs by EPQ(pY)EEIPIA was suppressed in a concentration-dependent manner when NRG1β was administered 20 min before whole-cell recording. Incontrast to the effect of NRG1β administered before EPQ(pY)EEIPIA, there was nosubsequent change in EPSP slope when NRG1β (2 nM) was administered after thepotentiation by EPQ(pY)EEIPIA had been established (Fig. 3c,d).

The potentiation of EPSPs by EPQ(pY)EEIPIA was prevented by the NMDAR antagonistD-amino-phosphonovaleric acid (D-APV) and the potentiated EPSPs were blocked byCNQX, indicating that EPQ(pY)EEIPIA had initiated NMDAR-dependent potentiation ofsynaptic AMPAR responses (Supplementary Fig. 2). Moreover, EPSPs were potentiated byEPQ(pY)EEIPIA in neurons from Src+/+ mice but not in Src−/− neurons (SupplementaryFig. 2a,b). Considering our findings together, we conclude that by suppressing Src-mediatedenhancement of NMDAR currents, NRG1β prevented the potentiation of AMPAR-mediatedsynaptic transmission. In contrast to the enhancement of NMDAR synaptic responses, thepotentiation of AMPAR synaptic responses was not reversed by NRG1β, indicating that,once established, AMPAR potentiation became independent of regulation by NRG1βsignaling.

NRG1β prevents LTP induced by theta-burst stimulationSrc-mediated enhancement of NMDAR currents is crucial for inducing but not maintainingthe long-term potentiation of synaptic transmission induced by high-frequency activation ofSchaffer collateral inputs5. Here we used theta-burst patterned stimulation (TBS) to generateTBS-induced LTP (tbLTP) that was prevented by Src inhibitors (Supplementary Fig. 3). Wefound that bath-applying NRG1β starting 20 min before TBS suppressed the increase in theslope of field EPSPs (f EPSP) that characterized tbLTP (Fig. 4a). In contrast, NRG1β had noeffect on basal fEPSP slope before TBS (Fig. 4a) nor did it affect f EPSP slope when appliedstarting 30 min after TBS, a time when synaptic responses had already been potentiated(Fig. 4b). Thus, NRG1β suppressed the potentiation of AMPAR-mediated synaptictransmission induced by TBS, as well as that induced by the direct activation of Src.

Pitcher et al. Page 4

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

NRG1β suppression of tbLTP is prevented by lack of SrcAs our findings indicate that NRG1β suppresses Src-dependent enhancement of synaptictransmission, we considered that NRG1β might not have an effect on Src-independent formsof synaptic potentiation. Because Src-null mutant mice show residual LTP, which bydefinition is Src independent, we tested the effect of NRG1β on LTP induction inhippocampal slices from Src+/+ mice versus Src−/− mice. Basal synaptic transmission, input-output relationship and paired-pulse facilitation were not different between Src+/+ andSrc−/− slices, and NRG1β had no effect on basal synaptic transmission in either genotype(Supplementary Fig. 4a–c). However, NRG1β significantly decreased tbLTP in slices fromSrc+/+ mice but had no effect on tbLTP in slices from Src−/− mice (Fig. 4c,d). As the levelof tbLTP in Src−/− mice was reduced compared with that in Src+/+ mice, we conclude thatthe effect of NRG1β was occluded in Src-null mutant mice. The loss of effect of NRG1β inSrc−/− slices could not be accounted for by lack of ErbB4 because the amount of ErbB4 inthe hippocampus of Src−/− mice was not different from that in Src+/+ mice (Fig. 4e andSupplementary Fig. 4d). Moreover, Src−/− mice were not different from Src+/+ mice in theamount of the following prominent synaptic proteins: the NMDAR subunits GluN1,GluN2A and GluN2B; the AMPAR subunits GluA1 and GluA2 and/or GluA3 (the antibodyused detects both); postsynaptic density protein-95 (PSD-95); α-calcium calmodulin-dependent kinase II; or the Src-family kinase Fyn (Supplementary Fig. 4d,e). Thus, thesuppression of the induction of tbLTP by NRG1β requires Src. On the basis of our findings,we conclude that NRG1β suppresses activity-dependent potentiation of synaptictransmission in CA1 hippocampus by preventing Src-mediated enhancement of NMDARfunction.

NRG1β suppresses burst EPSPs during TBSOur findings that NRG1β prevents induction of tbLTP but does not affect basal synaptictransmission led us to consider that NRG1β might affect transmission during the period ofTBS itself. We therefore made current-clamp recordings from CA1 neurons and monitoredthe membrane potential throughout the TBS, which was composed of a train of 15 bursts offour stimuli delivered every 200 ms (Fig. 5). We observed that each of the four-stimulibursts elicited a characteristic change in membrane potential, a response we call the ‘burstEPSP’ (Fig. 5a): the membrane potential depolarized progressively with each stimulus,peaking after the fourth and then decaying toward the resting membrane potential until thenext burst in the train. In control cells, the peak amplitude of the first-burst EPSP was 67 ± 3mV, and the peak amplitude declined gradually to 20 ± 3 mV with the subsequent 14 burstsin the train (n = 28 cells; Fig. 5b,c). In addition, the membrane potential did not return to theresting membrane potential between the bursts but only did so ~400 ms after the final burstin the TBS train.

Before investigating the possible effects of NRG1β during TBS, we determined whethercomponents of the membrane potential response during the train depended on NMDARs orSrc. We found that D-APV reduced the peak amplitude of the burst EPSPs and acceleratedthe recovery of membrane potential at the end of the burst EPSPs (Fig. 5a andSupplementary Fig. 5a,d). Thus, a component of the burst EPSPs depends on NMDARs. ForEPSPs evoked by a single stimulus, we found that D-APV had no effect on the rising phaseor the peak amplitude but shortened the decay of the single EPSP (Fig. 5a andSupplementary Fig. 5f). Hence, both the single EPSPs and burst EPSPs had an NMDAR-dependent component. The burst EPSP amplitude was reduced in Src−/− neurons, ascompared with Src+/+, whereas the single EPSPs did not differ between the two genotypes(Fig. 5a and Supplementary Fig. 5b,e,g).

Pitcher et al. Page 5

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

We found that applying NRG1β did not alter the rising phase, peak amplitude or decay ofsingle-stimulus–evoked EPSPs (Fig. 5a and Supplementary Fig. 5h). In contrast, NRG1βcaused a reduction in the peak amplitude of the first-burst EPSP (54 ± 4 mV, P < 0.05compared with control without NRG1β; Fig. 5a–c). The NRG1β-induced reduction in thefirst-burst EPSP amplitude was less than that produced by D-APV (38 ± 4 mV; P < 0.01compared with NRG1β) but was not different from that in Src−/− mice (53 ± 3 mV; P > 0.5compared with NRG1β). Furthermore, we found that AG1478 had no effect on singlestimulus–evoked EPSPs (Fig. 5e and Supplementary Fig. 5i) but prevented the NRG1β-induced suppression of burst EPSPs (Fig. 5e,f and Supplementary Fig. 5c). Thus, NRG1β-ErbB4 signaling reduced responses of CA1 neurons during the period of TBS itself.Notably, although both single and burst EPSPs showed NMDAR-dependent components,the burst EPSPs but not the single EPSPs were reduced by NRG1β-ErbB4 signaling or bylack of Src.

NRG1β suppresses Src and GluN2B tyrosine phosphorylationNRG1β did not alter the level of Src within the NMDAR complex in CA1 hippocampus(Fig. 6a), but we found that Src activity in tissue from slices treated with NRG1β wassignificantly lower than that in untreated slices (Fig. 6b). In NRG1β-treated slices, AG1478increased Src activity (data not shown), indicating that the suppression of Src function byNRG1β required ErbB4 signaling.

LTP-inducing tetanic stimulation increases tyrosine phosphorylation of the GluN2B subunitof the NMDAR in the hippocampus52,53. Here we found that TBS caused an increase inGluN2B tyrosine phosphorylation that depended on Src (Fig. 6c). TBS increased tyrosinephosphorylation in the GluN2B band in untreated slices but not in slices treated withNRG1β (Fig. 6d). Furthermore, the suppression of TBS-induced GluN2B tyrosinephosphorylation by NRG1β was prevented by AG1478 (Fig. 6e). Collectively, thesefindings indicate that NRG1β-ErbB4 signaling inhibits Src kinase activity, leading toblockade of TBS-induced tyrosine phosphorylation of GluN2B.

DISCUSSIONWe have discovered that NRG1β suppresses the enhancement of synaptic NMDAR currentsby Src kinase in CA1 hippocampus and in prefrontal cortex. The most parsimonious modelto account for our findings is that signaling by NRG1, through its cognate receptor ErbB4,leads to inhibition of the catalytic activity of Src kinase, thereby blocking the enhancementof NMDAR function. Consequently, NRG1-ErbB4 signaling prevents downstream eventsthat require Src-mediated enhancement of NMDARs; in the case of CA1 neurons, this is theprevention of long-term potentiation at Schaffer collateral synapses. Thus, our studyidentifies Src as a downstream target of NRG1-ErbB4 signaling and inhibition of Src as anessential step by which this signaling pathway regulates synaptic plasticity in thehippocampus. Because the dominant paradigm for signaling by the family of ErbB receptorsis that receptor stimulation leads to activation, rather than to suppression, of Src and otherSrc-family kinases54,55, we have discovered a function of ErbB receptors unexpected fromprevious work.

Our findings that basal NMDAR currents and voltage dependence are unaffected by NRG1βindicate that NRG1-ErbB4 signaling does not generally suppress NMDAR channel function.Nor does NRG1-ErbB4 signaling generally affect AMPAR function, as indicated by the lackof effect of NRG1 on basal EPSPs or on EPSPs potentiated after the administration ofEPQ(pY)EEIPIA or TBS. NRG1β also did not seem to affect presynaptic glutamate release,as paired-pulse responses were unaffected. Rather, our evidence that NRG1β prevents thepost-synaptic upregulation of synaptic NMDAR and AMPAR responses by

Pitcher et al. Page 6

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

EPQ(pY)EEIPIA, the tyrosine phosphorylation of NMDAR subunits and the LTP inducedby TBS indicates that NRG1-ErbB4 signaling interferes with Src, as demonstrated byNRG1β-induced reduction of Src activity. The effects of reduced Src activity were evidentduring the induction of LTP, as the Src- and NMDAR-dependent components of the burstEPSPs were reduced by NRG1β. Thus, NRG1-ErbB4 signaling preferentially suppressesSrc-dependent upregulation of NMDARs and, in turn, the induction of persistent synapticplasticity in CA1 without affecting basal synaptic transmission or short-term plasticity, bothof which are Src independent.

NRG1-ErbB4 signaling suppresses responses during TBS but does not have an effect onresponses to individual, isolated stimuli, indicating that CA1 neuron responses aredifferentially sensitive to patterned synaptic input. This differential sensitivity is apparenteven during the first-burst EPSP, indicating that responses to bursts of as few as three orfour stimuli are suppressed by NRG1-ErbB4 signaling, with the degree of suppressionincreasing with later bursts in the train. The stimulation train we used corresponds to thepattern of theta rhythm activity that normally occurs in the hippocampus56. Thus, ourfindings suggest that oscillatory network activity in the brain at theta frequency may besuppressed by NRG1-ErbB4 signaling, whereas nonoscillatory, irregular activity may beunaffected. Therefore, functions that depend on theta rhythm may be disrupted underconditions when NRG1-ErbB4 signaling is increased. NRG1 disruption of responses of CA1neurons to theta-patterned input could contribute to the alterations in oscillatory brainnetwork activity observed in schizophrenia56,57. Because alteration of oscillatory activitymay underlie disorders of thought56,57 and hallucinations57, our findings suggest thatsuppression of Src enhancement of NMDAR activity by excessive NRG1-ErbB4 signalingmay contribute to cognitive dysfunction and positive symptoms in individuals withschizophrenia.

Decreased Src catalytic activity by NRG-ErbB4 may be caused by suppressing activators, orby facilitating inhibitors, of kinase function. Within the NMDAR complex, potentialmediators of NRG1-ErbB4 signaling include the Src regulators Csk48 and protein tyrosinephosphatase-α58. Another potential mediator is PSD-95, which suppresses Src activity in theNMDAR complex through binding of a sequence in the PSD-95 N-terminal region to theSH2 domain of Src59.

Our finding that acute pharmacological inhibition of ErbB4 did not affect basal NMDARresponses suggests that NMDARs are not tonically suppressed by NRG1-ErbB4 signaling.In contrast, the magnitude of LTP is increased by acute ErbB4 inhibition, or in mice lackingErbB4 (ref. 36), indicating that a physiological role of NRG1-ErbB4 signaling is to suppressLTP at Schaffer collateral–CA1 synapses. The differential effect of NRG1-ErbB4 signalingon LTP induction but not on basal NMDAR currents is consistent with increasing neuronalactivity causing NRG1 release in the hippocampus60. Alternatively, there might be tonicrelease of NRG1, which suppresses Src-enhanced but not basal NMDAR currents. Ourfindings indicate that NRG1-ErbB4 functions physiologically to inhibit Src-mediatedupregulation of NMDAR function. This physiological function of NRG1-ErbB4 is notnormally saturated because LTP can be further suppressed by increasing this signalingthrough the administration of NRG1β32,36. Thus, the normal physiological function could beexaggerated under conditions of increased expression of NRG1 or elevated NRG1-ErbB4signaling.

NMDAR EPSCs in layer V prefrontal cortex neurons were increased by intracellularadministration of EPQ(pY)EEIPIA, indicating that synaptic NMDAR-mediated responses inthese neurons are enhanced by the activation of Src kinase7. This enhancement of NMDARfunction in layer V prefrontal cortex neurons was suppressed by NRG1β, as in hippocampal

Pitcher et al. Page 7

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

CA1 neurons. Together, these findings show similarities in the effect of NRG1β on neuronsthat are major output pathways for the prefrontal cortex and the hippocampus, respectively.Suppressing Src enhancement of NMDAR activity in these pathways could thereby impairbrain functions that depend on these outputs, including learning, memory and executivefunction. Impairment of these functions is linked to dysfunction of cognition and positivesymptoms in schizophrenia. Overall, our work identifies a specific molecular mechanism,suppression of the upregulation of NMDARs by Src, that links gain of function of NRG1-ErbB4 signaling in schizophrenia and hypofunction of NMDARs. Our findings suggest thatstrategies to normalize Src-mediated enhancement of NMDARs could be new therapeuticapproaches in the treatment of schizophrenia.

METHODSMethods and any associated references are available in the online version of the paper athttp://www.nature.com/naturemedicine/.

Supplementary MaterialRefer to Web version on PubMed Central for supplementary material.

AcknowledgmentsWe thank J.F. MacDonald, P.A. Frankland, L.Y. Wang, M.F. Jackson and M.H. Pitcher for critical reading of themanuscript, and Genentech and L. Mei for providing NRG1β. Src knockout mice were obtained as a gift from B.F.Boyce, University of Rochester Medical Center. This study was supported by grants from the Canadian Institutes ofHealth Research (CIHR) to M.W.S. (MT-12682) and to E.K.L. (MOP-89825), and from the Deafness ResearchFoundation to K.T.Y. M.W.S. holds a Canada research chair (tier I) in neuroplasticity and pain, and is anInternational Research Scholar of the Howard Hughes Medical Institute. E.K.L. holds a Canada research chair (tierII) in developmental cortical physiology. G.M.P. was a CIHR postdoctoral fellow and was supported by a MerckFrosst Fellowship award. D.N. held a CIHR doctoral research award, L.V.K. was a CIHR MD-PhD student, andN.M.G. holds a CIHR Canada Graduate Scholarship doctoral award. We thank D. Wong, J. Hicks and S. Singhroyfor technical support.

References1. Dingledine R, Borges K, Bowie D, Traynelis SF. The glutamate receptor ion channels. Pharmacol

Rev. 1999; 51:7–61. [PubMed: 10049997]2. Paoletti P, Neyton J. NMDA receptor subunits: function and pharmacology. Curr Opin Pharmacol.

2007; 7:39–47. [PubMed: 17088105]3. Grant SG, Blackstock WP. Proteomics in neuroscience: from protein to network. J Neurosci. 2001;

21:8315–8318. [PubMed: 11606617]4. Liu XJ, et al. Treatment of inflammatory and neuropathic pain by uncoupling Src from the NMDA

receptor complex. Nat Med. 2008; 14:1325–1332. [PubMed: 19011637]5. Lu YM, Roder JC, Davidow J, Salter MW. Src activation in the induction of long-term potentiation

in CA1 hippocampal neurons. Science. 1998; 279:1363–1367. [PubMed: 9478899]6. MacDonald JF, Kotecha SA, Lu WY, Jackson MF. Convergence of PKC-dependent kinase signal

cascades on NMDA receptors. Curr Drug Targets. 2001; 2:299–312. [PubMed: 11554554]7. Salter MW, Kalia LV. Src kinases: a hub for NMDA receptor regulation. Nat Rev Neurosci. 2004;

5:317–328. [PubMed: 15034556]8. Coyle JT, Tsai G, Goff D. Converging evidence of NMDA receptor hypofunction in the

pathophysiology of schizophrenia. Ann NY Acad Sci. 2003; 1003:318–327. [PubMed: 14684455]9. Pilowsky LS, et al. First in vivo evidence of an NMDA receptor deficit in medication-free

schizophrenic patients. Mol Psychiatry. 2006; 11:118–119. [PubMed: 16189506]10. Balu DT, Coyle JT. Neuroplasticity signaling pathways linked to the pathophysiology of

schizophrenia. Neurosci Biobehav Rev. 2011; 35:848–870. [PubMed: 20951727]

Pitcher et al. Page 8

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

11. Lindsley CW, et al. Progress towards validating the NMDA receptor hypofunction hypothesis ofschizophrenia. Curr Top Med Chem. 2006; 6:771–785. [PubMed: 16719816]

12. Hahn CG, et al. Altered neuregulin 1-erbB4 signaling contributes to NMDA receptor hypofunctionin schizophrenia. Nat Med. 2006; 12:824–828. [PubMed: 16767099]

13. Yu XM, Askalan R, Keil GJ, Salter MW. NMDA channel regulation by channel-associated proteintyrosine kinase Src. Science. 1997; 275:674–678. [PubMed: 9005855]

14. Stefansson H, et al. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet. 2002;71:877–892. [PubMed: 12145742]

15. Yang JZ, et al. Association study of neuregulin 1 gene with schizophrenia. Mol Psychiatry. 2003;8:706–709. [PubMed: 12874607]

16. Hashimoto R, et al. Expression analysis of neuregulin-1 in the dorsolateral prefrontal cortex inschizophrenia. Mol Psychiatry. 2004; 9:299–307. [PubMed: 14569272]

17. Harrison PJ, Law AJ. Neuregulin 1 and schizophrenia: genetics, gene expression, andneurobiology. Biol Psychiatry. 2006; 60:132–140. [PubMed: 16442083]

18. Nicodemus KK, et al. Further evidence for association between ErbB4 and schizophrenia andinfluence on cognitive intermediate phenotypes in healthy controls. Mol Psychiatry. 2006;11:1062–1065. [PubMed: 17130882]

19. Norton N, et al. Evidence that interaction between neuregulin 1 and its receptor erbB4 increasessusceptibility to schizophrenia. Am J Med Genet B Neuropsychiatr Genet. 2006; 141B:96–101.[PubMed: 16249994]

20. Silberberg G, Darvasi A, Pinkas-Kramarski R, Navon R. The involvement of ErbB4 withschizophrenia: association and expression studies. Am J Med Genet B Neuropsychiatr Genet.2006; 141B:142–148. [PubMed: 16402353]

21. Law AJ, Kleinman JE, Weinberger DR, Weickert CS. Disease-associated intronic variants in theErbB4 gene are related to altered ErbB4 splice-variant expression in the brain in schizophrenia.Hum Mol Genet. 2007; 16:129–141. [PubMed: 17164265]

22. Tan W, et al. 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; 282:24343–24351. [PubMed: 17565985]

23. Colantuoni C, et al. Age-related changes in the expression of schizophrenia susceptibility genes inthe human prefrontal cortex. Brain Struct Funct. 2008; 213:255–271. [PubMed: 18470533]

24. Nicodemus KK, et al. A 5′ promoter region SNP in NRG1 is associated with schizophrenia riskand type III isoform expression. Mol Psychiatry. 2009; 14:741–743. [PubMed: 19626024]

25. Chen YJ, et al. Type III neuregulin-1 is required for normal sensorimotor gating, memory-relatedbehaviors, and corticostriatal circuit components. J Neurosci. 2008; 28:6872–6883. [PubMed:18596162]

26. O’Tuathaigh CM, et al. Disruption to social dyadic interactions but not emotional/anxiety-relatedbehaviour in mice with heterozygous ‘knockout’ of the schizophrenia risk gene neuregulin-1. ProgNeuropsychopharmacol Biol Psychiatry. 2008; 32:462–466. [PubMed: 17980471]

27. Duffy L, Cappas E, Lai D, Boucher AA, Karl T. Cognition in transmembrane domain neuregulin 1mutant mice. Neuroscience. 2010; 170:800–807. [PubMed: 20678553]

28. Deakin IH, et al. Behavioural characterization of neuregulin 1 type I overexpressing transgenicmice. Neuroreport. 2009; 20:1523–1528. [PubMed: 19829162]

29. Kato T, et al. Phenotypic characterization of transgenic mice overexpressing neuregulin-1. PLoSONE. 2010; 5:e14185. [PubMed: 21151609]

30. Chong VZ, et al. Elevated neuregulin-1 and ErbB4 protein in the prefrontal cortex of schizophrenicpatients. Schizophr Res. 2008; 100:270–280. [PubMed: 18243664]

31. Law AJ, et al. Neuregulin 1 transcripts are differentially expressed in schizophrenia and regulatedby 5′SNPs associated with the disease. Proc Natl Acad Sci USA. 2006; 103:6747–6752. [PubMed:16618933]

32. Huang YZ, et al. Regulation of neuregulin signaling by PSD-95 interacting with ErbB4 at CNSsynapses. Neuron. 2000; 26:443–455. [PubMed: 10839362]

Pitcher et al. Page 9

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

33. Ma L, et al. Ligand-dependent recruitment of the ErbB4 signaling complex into neuronal lipidrafts. J Neurosci. 2003; 23:3164–3175. [PubMed: 12716924]

34. Bjarnadottir M, et al. Neuregulin1 (NRG1) signaling through Fyn modulates NMDA receptorphosphorylation: differential synaptic function in Nrg1+/− knock-outs compared with wild-typemice. J Neurosci. 2007; 27:4519–4529. [PubMed: 17460065]

35. Kwon OB, Longart M, Vullhorst D, Hoffman DA, Buonanno A. Neuregulin-1 reverses long-termpotentiation at CA1 hippocampal synapses. J Neurosci. 2005; 25:9378–9383. [PubMed:16221846]

36. Pitcher GM, Beggs S, Woo RS, Mei L, Salter MW. ErbB4 is a suppressor of long-termpotentiation in the adult hippocampus. Neuroreport. 2008; 19:139–143. [PubMed: 18185097]

37. Iyengar SS, Mott DD. Neuregulin blocks synaptic strengthening after epileptiform activity in therat hippocampus. Brain Res. 2008; 1208:67–73. [PubMed: 18387600]

38. Huang Y, et al. CAKβ/Pyk2 kinase is a signaling link for induction of long-term potentiation inCA1 hippocampus. Neuron. 2001; 29:485–496. [PubMed: 11239437]

39. Pelkey KA, et al. Tyrosine phosphatase STEP is a tonic brake on induction of long-termpotentiation. Neuron. 2002; 34:127–138. [PubMed: 11931747]

40. Gao XM, et al. Ionotropic glutamate receptors and expression of N-methyl-D-aspartate receptorsubunits in subregions of human hippocampus: effects of schizophrenia. Am J Psychiatry. 2000;157:1141–1149. [PubMed: 10873924]

41. Siekmeier PJ, Hasselmo ME, Howard MW, Coyle J. Modeling of context-dependent retrieval inhippocampal region CA1: implications for cognitive function in schizophrenia. Schizophr Res.2007; 89:177–190. [PubMed: 17055702]

42. Zierhut K, et al. The role of hippocampus dysfunction in deficient memory encoding and positivesymptoms in schizophrenia. Psychiatry Res. 2010; 183:187–194. [PubMed: 20702070]

43. Benetti S, et al. Functional integration between the posterior hippocampus and prefrontal cortex isimpaired in both first episode schizophrenia and the at risk mental state. Brain. 2009; 132:2426–2436. [PubMed: 19420091]

44. Tidcombe H, et al. Neural and mammary gland defects in ErbB4 knockout mice geneticallyrescued from embryonic lethality. Proc Natl Acad Sci USA. 2003; 100:8281–8286. [PubMed:12824469]

45. Fukazawa R, et al. Neuregulin-1 protects ventricular myocytes from anthracycline-inducedapoptosis via erbB4-dependent activation of PI3-kinase/Akt. J Mol Cell Cardiol. 2003; 35:1473–1479. [PubMed: 14654373]

46. Woo RS, et al. Neuregulin-1 enhances depolarization-induced GABA release. Neuron. 2007;54:599–610. [PubMed: 17521572]

47. Macdonald DS, et al. Modulation of NMDA receptors by pituitary adenylate cyclase activatingpeptide in CA1 neurons requires Gαq, protein kinase C and activation of Src. J Neurosci. 2005;25:11374–11384. [PubMed: 16339032]

48. Xu J, et al. Control of excitatory synaptic transmission by C-terminal Src kinase. J Biol Chem.2008; 283:17503–17514. [PubMed: 18445593]

49. Gu Z, Jiang Q, Fu AK, Ip NY, Yan Z. Regulation of NMDA receptors by neuregulin signaling inprefrontal cortex. J Neurosci. 2005; 25:4974–4984. [PubMed: 15901778]

50. Eisenberg DP, Berman KF. Executive function, neural circuitry and genetic mechanisms inschizophrenia. Neuropsychopharmacology. 2010; 35:258–277. [PubMed: 19693005]

51. Merrill MA, Chen Y, Strack S, Hell JW. Activity-driven postsynaptic translocation of CaMKII.Trends Pharmacol Sci. 2005; 26:645–653. [PubMed: 16253351]

52. Rostas JA, et al. Enhanced tyrosine phosphorylation of the 2B subunit of the N-methyl-D-aspartatereceptor in long-term potentiation. Proc Natl Acad Sci USA. 1996; 93:10452–10456. [PubMed:8816821]

53. Rosenblum K, Dudai Y, Richter-Levin G. Long-term potentiation increases tyrosinephosphorylation of the N-methyl-D-aspartate receptor subunit 2B in rat dentate gyrus in vivo. ProcNatl Acad Sci USA. 1996; 93:10457–10460. [PubMed: 8816822]

Pitcher et al. Page 10

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

54. Kim H, et al. The c-Src tyrosine kinase associates with the catalytic domain of ErbB-2:implications for ErbB-2 mediated signaling and transformation. Oncogene. 2005; 24:7599–7607.[PubMed: 16170374]

55. Olayioye MA, Beuvink I, Horsch K, Daly JM, Hynes NE. ErbB receptor–induced activation of stattranscription factors is mediated by Src tyrosine kinases. J Biol Chem. 1999; 274:17209–17218.[PubMed: 10358079]

56. Lisman J, Buzsaki G. A neural coding scheme formed by the combined function of gamma andtheta oscillations. Schizophr Bull. 2008; 34:974–980. [PubMed: 18559405]

57. Uhlhaas PJ, Haenschel C, Nikolic D, Singer W. The role of oscillations and synchrony in corticalnetworks and their putative relevance for the pathophysiology of schizophrenia. Schizophr Bull.2008; 34:927–943. [PubMed: 18562344]

58. Lei G, et al. Gain control of N-methyl-D-aspartate receptor activity by receptor-like proteintyrosine phosphatase alpha. EMBO J. 2002; 21:2977–2989. [PubMed: 12065411]

59. Kalia LV, Pitcher GM, Pelkey KA, Salter MW. PSD-95 is a negative regulator of the tyrosinekinase Src in the NMDA receptor complex. EMBO J. 2006; 25:4971–4982. [PubMed: 16990796]

60. Eilam R, Pinkas-Kramarski R, Ratzkin BJ, Segal M, Yarden Y. Activity-dependent regulation ofNeu differentiation factor/neuregulin expression in rat brain. Proc Natl Acad Sci USA. 1998;95:1888–1893. [PubMed: 9465112]

Pitcher et al. Page 11

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

Figure 1.NRG1β-ErbB4 signaling prevents endogenous Src activation–induced potentiation ofNMDAR-mediated synaptic responses in the hippocampal CA1. (a) Scatter plot of NMDAREPSC peak amplitude over time (min) from three rat CA1 neurons recorded with controlintracellular solution (ICS), ICS containing EPQ(pY)EEIPIA or ICS containingEPQ(pY)EEIPIA and NRG1β (2 nM) bath-applied beginning 20 min before whole-cellrecording (black bar). Top, average NMDAR EPSC traces from the three neurons (scalebars, 45 ms; 50 pA) at the times indicated (1 and 2). (b) Summary scatter plot of NMDAREPSC peak amplitude with control ICS (n = 20), EPQ(pY)EEIPIA (n = 13) orEPQ(pY)EEIPIA with NRG1β in ACSF (n = 7; *P < 0.05 or **P < 0.01 versusEPQ(pY)EEIPIA). (c) Current-voltage (I-V) relationship for pharmacologically isolatedNMDARs with control solution, EPQ(pY)EEIPIA or EPQ(pY)EEIPIA with NRG1β in

Pitcher et al. Page 12

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

ACSF. Right, superimposed NMDAR EPSC traces at membrane potentials from −80 to +60mV. Scale bars, 150 ms; 250 pA. (d) Summary scatter plot of normalized peak NMDAREPSC amplitude from CA1 neurons from WT mice during intracellular administration ofICS containing EPQ(pY)EEIPIA with NRG1β (2 nM) in ACSF (n = 18), from neurons fromErbb4−/−HER4heart mice during intracellular administration of EPQ(pY)EEIPIA withNRG1β (2 nM) (n = 7) or from neurons from WT mice during intracellular administration ofEPQ(pY)EEIPIA with bath-applied NRG1β (2 nM) and AG1478 (n = 16). Top,representative average NMDAR EPSCs at the indicated times (1 and 2) and after D-APV(scale bars, 50 ms; 50 pA). (e) Summary histogram of EPQ(pY)EEIPIA-induced increase inNMDAR EPSC amplitude at 30 min of recording from WT neurons with ICS containingEPQ(pY)EEIPIA (n = 13), ICS containing EPQ(pY)EEIPIA with NRG1β present in ACSFbeginning 20 min before whole-cell recording (n = 18), from CA1 pyramidal neurons fromErbb4−/−HER4heart mice with ICS containing EPQ(pY)EEIPIA with NRG1β in ACSF (n =7) or from WT CA1 pyramidal neurons with ICS containing EPQ(pY)EEIPIA with NRG1βand AG1478 in ACSF (n = 16). ***P < 0.001 versus wild-type. In all figures, group data aremean ± s.e.m. In voltage-clamp experiments the holding potential was −60 mV, exceptwhere otherwise indicated.

Pitcher et al. Page 13

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

Figure 2.NRG1β has no effect on basal NMDAR-mediated synaptic responses in hippocampal CA1or in prefrontal cortex but prevents endogenous Src activation-induced potentiation ofNMDAR EPSCs at prefrontal cortex synapses. (a) Summary scatter plot of peak NMDAREPSC amplitude over time from mouse CA1 neurons recorded with control ICS (n = 10) andNRG1β (2 nM) present in ACSF beginning at 10 min (black bar). Top, average NMDAREPSCs from a representative neuron (scale bars, 50 ms; 50 pA) at the times indicated (1 and2) and after D-APV. (b) Summary histogram of NMDAR EPSC decay from the CA1neurons recorded in a at the 10-min time point immediately before NRG1β administrationand at the 40-min time point during NRG1β perfusion. Results are percentage of NMDAREPSC decay with mean decay during first 2 min of recording before NRG1β administrationnormalized to 100% (dotted line). (c) Summary histogram of pharmacologically isolatedNMDAR fEPSP responses (n = 14) before or during PD158780 application. Results arepercentage of NMDAR fEPSP slope at the beginning of the recording normalized to 100%(dotted line). (d) Scatter plot of NMDAR EPSC peak amplitude from four prefrontal cortexneurons: with control ICS, with control ICS with NRG1β (6 nM) in ACSF, with ICScontaining EPQ(pY)EEIPIA or with ICS containing EPQ(pY)EEIPIA with NRG1β in ACSF(black bar). Top, average NMDAR EPSC traces from the four neurons (scale bars, 100 ms;40 pA) at the times indicated (1 and 2). (e) Summary histogram of normalized NMDAREPSC amplitude at 30 min of recording from prefrontal cortex neurons with control ICS (n =11), ICS containing EPQ(pY)EEIPIA (n = 9), control ICS with NRG1β present in ACSF (n= 7) or ICS containing EPQ(pY)EEIPIA with NRG1β in ACSF (n = 5). ***P < 0.001.

Pitcher et al. Page 14

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

Figure 3.NRG1β prevents but does not reverse endogenous Src-induced synaptic potentiation. (a)Scatter plot of EPSP slope over time from three rat CA1 neurons recorded with control ICS,ICS containing EPQ(pY)EEIPIA or ICS containing EPQ(pY)EEIPIA with NRG1β (2 nM)in the ACSF beginning 20 min before whole-cell recording (black bar). Right, average EPSPtraces from the three neurons (scale bars, 50 ms; 10 mV). Bottom, scatter plots ofsimultaneously recorded fEPSPs from CA1 stratum radiatum. Right, average fEPSPs (scalebars, 30 ms; 0.5 mV) at the times indicated (1 and 2). (b) Summary scatter plot of EPSPslope with control ICS (n = 16), ICS containing EPQ(pY)EEIPIA (n = 9) or ICS containingEPQ(pY)EEIPIA with NRG1β in the ACSF (n = 6; *P < 0.05, **P < 0.01 or ***P < 0.001versus EPQ(pY)EEIPIA). Bottom, averaged slope of fEPSPs recorded simultaneouslyduring experiments when whole-cell recordings were carried out. (c) Scatter plot of EPSPslope over time from two WT mouse CA1 neurons recorded with ICS containingEPQ(pY)EEIPIA with NRG1β (2 nM) or vehicle administered to the ACSF duringEPQ(pY)EEIPIA-induced potentiation of EPSP responses (black bar). Right, average EPSPsfrom the two neurons (scale bars, 100 ms; 15 mV). Bottom, scatter plots of simultaneouslyrecorded fEPSPs from CA1. Right, average fEPSPs (scale bars, 10 ms; 0.5 mV) at the timesindicated (1, 2 and 3). (d) Summary histogram of the EPQ(pY)EEIPIA-induced increase inEPSP slope at 30 min of recording from neurons with ICS containing EPQ(pY)EEIPIA (n =19), ICS containing EPQ(pY)EEIPIA with NRG1β present in ACSF beginning 20 minbefore whole-cell recording at 0.2 nM (n = 16; **P < 0.01 versus vehicle) or 2 nM (n = 6;***P < 0.001 versus vehicle) or ICS containing EPQ(pY)EEIPIA with NRG1β (2 nM)treatment during EPQ(pY)EEIPIA-induced potentiation of EPSPs (n = 12).

Pitcher et al. Page 15

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

Figure 4.NRG1β prevents but does not reverse TBS-induced LTP in CA1 hippocampus and has noeffect in Src−/− mice. (a) Scatter plot of normalized fEPSP slope over time for a rathippocampal slice treated with NRG1β (2 nM) or another slice treated with vehicle (blackbar). Top, average fEPSP traces (scale bars, 10 ms; 0.5 mV) at the times indicated (1 and 2)from the two individual experiments. Bottom, summary scatter plot of grouped normalizedfEPSP slope from control slices (n = 23) or NRG1β-treated slices (n = 14). (b) Scatter plotof normalized fEPSP slope over time for a rat hippocampal slice treated with NRG1β (2 nM)present in ACSF beginning 30 min after TBS (black bar). Top, average fEPSP traces (scalebars, 15 ms; 0.5 mV) at the times indicated (1, 2, 3 and 4) from the experiment plotted.Bottom, summary scatter plot of grouped normalized fEPSP slope (n = 19). (c) Summaryscatter plot of grouped normalized fEPSP slope in slices from Src+/+ mice treated withvehicle (VEH, n = 13) or NRG1β (NRG, 2 nM; n = 14; P < 0.001 versus VEH) (black bar).Inset, average fEPSP traces at the times indicated (1 and 2) (scale bars, 10 ms; 0.5 mV) fromtwo representative experiments. Bottom, summary scatter plot of normalized fEPSP slope inslices from Src−/− mice treated with NRG1β (2 nM, n = 11) or vehicle (n = 14) (black bar).Inset, average fEPSP traces recorded at the times indicated (1 and 2) (scale bars, 10 ms; 0.5mV) from two representative experiments. (d) Summary histogram of TBS-induced increasein fEPSP slope 60 min after TBS in slices from Src+/+ and Src−/− mice with NRG1β (+) orvehicle (−) treatment. Data are percentage of tbLTP with vehicle normalized to 100%. ***P< 0.001 versus Src+/+ with vehicle. (e) Immunoblot analysis of ErbB4, GluN1 and GluN2, inCA1 from a Src+/+ mouse or a Src−/− littermate. GAPDH was loading control.

Pitcher et al. Page 16

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

Figure 5.NRG1β reduces depolarization of CA1 neurons during the period of TBS. (a) The first fourpulse-induced burst EPSP of TBS for control mouse (Src+/+) slices (n = 28), slices (fromSrc+/+ mice) pretreated with D-APV (n = 17), slices from Src−/− mice (n = 12) or slicesfrom Src+/+ mice pretreated with NRG1β (2 nM; n = 19). Traces are the mean membranepotential from all recordings in response to the first four stimuli. Gray region, ± s.e.m.Vertical arrows, start of stimulation for each of four pulses delivered at 100 Hz. x axis, time(200 ms). Bottom, average pre-TBS baseline single stimulus–evoked EPSPs (scale bars, 100ms; 10 mV). (b) Burst EPSPs for the entire duration of TBS from control slices (n = 28) orslices pretreated with NRG1β (n = 19). (c) Summary scatter plot of peak amplitude of burstEPSPs in response to each of the four stimuli-elicited bursts in control slices (n = 28) orslices pretreated with NRG1β (n = 19; *P < 0.05 or **P < 0.01 versus control). (d)Summary scatter plot of grouped normalized EPSP slope over time. (e) First four pulse-induced burst EPSP of TBS for neurons from control slices (from Src+/+mice), slices (fromSrc+/+ mice) pretreated with NRG1β (2 nM) or slices (from Src+/+ mice) pretreated withNRG1β (2 nM) and AG1478. x axis, time (200 ms). Inset, average representative pre-TBSbaseline single stimulus–evoked EPSPs from a control slice and a slice pretreated with

Pitcher et al. Page 17

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

NRG1β and AG1478 (scale bars, 100 ms; 8 mV). (f) Summary scatter plot of peakamplitude of burst EPSPs in response to each of the four stimuli-elicited bursts in slicespretreated with NRG1β (n = 19) or with NRG1β and AG1478 (n = 25; *P < 0.05, **P <0.01 or ‡P < 0.001 versus NRG1β).

Pitcher et al. Page 18

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

Figure 6.NRG1β does not alter Src association with the NMDAR but reduces Src tyrosine kinaseactivity and prevents TBS-induced GluN2B phosphorylation in hippocampal CA1. (a)Immunoprecipitation (IP) of GluN2 subunits carried out from hippocampal proteinsprepared from control (−) or NRG1β (2 nM)-treated (+) slices. Proteins thatimmunoprecipitated with antibody to GluN2B were probed with antibodies to GluN2B,PSD-95 or Src (left). In ‘input’ lanes, 60 μg of proteins without immunoprecipitation wereloaded. (b) Histogram of Src kinase activity in CA1 lysates without (n = 4) or with (n = 4)NRG1β (2 nM) pretreatment. Src kinase activity normalized to activity of Src in controllysates (**P < 0.01, t-test versus control). (c) Summary histogram of TBS-inducedpYGluN2B from CA1 region collected from Src+/+ (n = 6) or Src−/− (n = 6; **P < 0.01, t-test versus Src+/+) mice. (d) Top image, representative immunoblots of lysates from rat CA1region with TBS or without (baseline, BL) probed with antibody to phosphorylated tyrosine(pY). The same membrane was stripped and reprobed with GluN2B antibody. Top graph,four-point standard curve derived from serial dilutions of rat brain proteins. Bottom image,representative immunoblot of lysates from rat CA1 region with TBS or without (BL) andtreated with NRG1β (2 nM) probed sequentially with antibodies to pY and GluN2B. Bottomgraph, four-point standard curve derived from serial dilutions of rat brain lysate. Right,summary histogram of immunoblot analysis. Densiometric quantification was carried out forpY GluN2B from six immunoblot experiments for each condition. Band intensity wasquantified as mean gray value and the ratio of pY GluN2B to total GluN2B was calculated.Bars correspond to mean ratios normalized to ratios obtained for baseline (BL; *P < 0.05, t-test versus BL). (e) Top, four-point standard curve derived from serial dilutions of Src+/+

mouse brain lysate. Bottom, summary histogram of TBS-induced pY GluN2B from

Pitcher et al. Page 19

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript

hippocampal CA1 region without (n = 4) or with (n = 4) NRG1β (2 nM) treatment or withNRG1β (2 nM) and AG1478 treatment (n = 4). **P < 0.01, t-test versus control or versusNRG1β and AG1478.

Pitcher et al. Page 20

Nat Med. Author manuscript; available in PMC 2012 January 23.

PMC

Canada Author M

anuscriptPM

C C

anada Author Manuscript

PMC

Canada Author M

anuscript