This article was originally published in a journal published byElsevier, and the attached copy is provided by Elsevier for the

author’s benefit and for the benefit of the author’s institution, fornon-commercial research and educational use including without

limitation use in instruction at your institution, sending it to specificcolleagues that you know, and providing a copy to your institution’s

administrator.

All other uses, reproduction and distribution, including withoutlimitation commercial reprints, selling or licensing copies or access,

or posting on open internet sites, your personal or institution’swebsite or repository, are prohibited. For exceptions, permission

may be sought for such use through Elsevier’s permissions site at:

http://www.elsevier.com/locate/permissionusematerial

Autho

r's

pers

onal

co

py

Integrin-binding RGD peptides induce rapid intracellular calciumincreases and MAPK signaling in cortical neurons

P. Marc D. Watson,a,1 Martin J. Humphries,a Jane Relton,b Nancy J. Rothwell,a

Alex Verkhratsky,a and Rosemary M. Gibsona,⁎,1

aFaculty of Life Sciences, Michael Smith Building, University of Manchester, Manchester, M13 9PT, UKbBiogenIdec, 14 Cambridge Center, Cambridge, MA 02142, USA

Received 17 May 2006; revised 12 October 2006; accepted 17 October 2006Available online 5 December 2006

Integrins mediate cell adhesion to the extracellular matrix and initiateintracellular signaling. They play key roles in the central nervoussystem (CNS), participating in synaptogenesis, synaptic transmissionand memory formation, but their precise mechanism of action remainsunknown. Here we show that the integrin ligand-mimetic peptideGRGDSP induced NMDA receptor-dependent increases in intracel-lular calcium levels within seconds of presentation to primary corticalneurons. These were followed by transient activation and nucleartranslocation of the ERK1/2 mitogen-activated protein kinase. RGD-induced effects were reduced by the NMDA receptor antagonistMK801, and ERK1/2 signaling was specifically inhibited by ifenprodiland PP2, indicating a functional connection between integrins, Src andNR2B-containing NMDA receptors. GRGDSP peptides were notsignificantly neuroprotective against excitotoxic insults. These resultsdemonstrate a previously undescribed, extremely rapid effect of RGDpeptide binding to integrins on cortical neurons that implies a close,functionally relevant connection between adhesion receptors andsynaptic transmission.© 2006 Elsevier Inc. All rights reserved.

Keywords: Integrin; NMDA receptor; MAP kinase; ERK; Calcium;GRGDSP; Cortical neuron

Introduction

Cell adhesion processes play important roles in the centralnervous system (CNS) of developing and adult organisms (Milnerand Campbell, 2002; Dityatev and Schachner, 2003). Receptors ofthe integrin family mediate extracellular adhesion and bidirectionalsignaling in many cell types, modulating cell shape, differentiation,

migration and survival (Hynes, 2002; van der Flier and Sonnenberg,2001; Humphries, 2000). Integrins are critical for the developmentof the nervous system (Clegg et al., 2003), and emerging evidenceindicates multiple roles for integrins in adult neurons.

In the adult rodent brain, integrins are expressed differentiallyacross brain regions and within individual neurons (Chan et al.,2003; Bi et al., 2001; Pinkstaff et al., 1999; Grooms et al., 1993).Integrins are particularly enriched in synaptic regions (e.g. Kramaret al., 2002; Einheber et al., 2001; Rodriguez et al., 2000;Nishimura et al., 1998; Bahr et al., 1997; Capaldi et al., 1997),where they participate in synaptic development, maintenance(Karanian et al., 2005; Hama et al., 2004; Nikonenko et al., 2003;Chavis and Westbrook, 2001) and the cytoskeletal rearrangementsthat accompany synaptic activity (Smart et al., 2004; Bahr, 2000).Integrins have neuromodulatory effects in mature neurons: ligand-mimetic peptides reversibly increase the strength and duration offast AMPA receptor-dependent post-synaptic responses (Kramar etal., 2003) and modulate NMDA receptor subunit phosphorylationand currents (Bernard-Trifilo et al., 2005; Lin et al., 2003). AMPAreceptor stimulation also increases surface expression and signal-ing downstream of integrin α5β1 (Lin et al., 2005). In thehippocampus, integrins regulate stabilization of long-term poten-tiation (LTP) (Kramar and Lynch, 2003; LeBaron et al., 2003;Kramar et al., 2002; Chun et al., 2001; Staubli et al., 1998; Bahr etal., 1997), while mice with altered expression of integrins showneurotransmission and memory defects (Chan et al., 2003, 2006).

Approximately half of all integrin heterodimers recognise theaspartate within the RGD peptide sequence (Hynes, 1992). Reagentsbased on the RGD motif have therefore been used as antagonists ofintegrin function in vivo (Humphries et al., 1986) and havetherapeutic potential for treating many diseases (e.g. Leclerc,2002). The ligand-mimetic nature of these peptides however alsoendows them with agonistic properties and the ability to stimulateintegrin-dependent signaling (Humphries, 2000).

Soluble RGD peptides have been used to investigate the effectsof integrin ligation on neuronal function. However, the immediateneuronal signaling responses to integrin ligation, or the conse-quences of antagonising integrin function, are currently unknown.

www.elsevier.com/locate/ymcneMol. Cell. Neurosci. 34 (2007) 147–154

⁎ Corresponding author.E-mail address: rosemary.gibson@hsl.gov.uk (R.M. Gibson).

1 Current addresses: MRC Laboratory for Molecular Cell Biology,University College London, Gordon Street, London, WC1E 6BT, UK(PMDW); Health and Safety Laboratory, Harpur Hill, Buxton, SK17 9JN,UK (RMG).

Available online on ScienceDirect (www.sciencedirect.com).

1044-7431/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.mcn.2006.10.007

Autho

r's

pers

onal

co

py

Such information is crucial for understanding how integrins affectlonger-term, stable alterations in neurons such as changes insynaptic efficiency or memory formation in vivo. We hypothesisedthat changes in integrin ligation may affect intracellular signalingevents in neurons. The aim of this study was therefore to analyseintegrin signaling in primary cortical neurons in response toaddition of the integrin ligand GRGDSP.

Results

GRGDSP causes very rapid NMDA receptor-mediatedintracellular calcium increases in cortical neurons

Real-time ratiometric calcium imaging was used to examinewhether the integrin-binding ligand GRGDSP affects intracellularfree calcium concentrations ([Ca2+]i) in primary cortical neurons.The authentic peptide GRGDSP or a negative control GRADSP,which binds to integrins with a much lower affinity than GRGDSP(both at 2.5 mM), was delivered by bath application. Thisconcentration of peptide is consistent with the millimolar affinityof the peptide for isolated integrins (Akiyama et al., 1985) and itsuppresses neurite outgrowth in cortical neurons cultured onfibronectin (J.D. Moore and R.M. Gibson, unpublished observa-tions). GRGDSP peptide induced a rapid, spontaneous increase inintracellular calcium levels, peaking within 1–2 s, which thendecayed to a steady-state level; [Ca2+]i remained elevated through-out the remainder of the experiment (up to 1 h; Fig. 1a). The controlGRADSP peptide induced a small change in intracellular calciumlevels which was significantly different to the active peptide (Fig. 1a

and d: peak change in [Ca2+]i: 21±7% (GRADSP) versus 65±6%(GRGDSP); P<0.01).

Increases in neuronal [Ca2+]i can be generated via entry throughthe calcium-permeable NMDA subtype of glutamate receptor (MoriandMishina, 1995). To test whether the intracellular calcium changesobserved on peptide addition were due to opening of NMDA receptorchannels, GRGDSP was applied in the presence of the NMDAreceptor antagonist MK-801 (Huettner and Bean, 1998). MK-801abolished the GRGDSP-induced calcium increase (Fig. 1b and d:peak change in [Ca2+]i: 7±2% (GRGDSP+MK801) versus 65±6%(GRGDSP); P<0.001), indicating that calcium influx depends on theactivity of NMDA receptors. NMDA receptors are multimericchannels, containing the obligate NR1 subunit with combinations ofNR2A-D and NR3A/B subunits (Cull-Candy et al., 2001). Antago-nists such as ifenprodil that bind to the polyamine site on the NR2Bsubunit can be used to analyse the contribution of NMDA receptorscontaining this subunit. The peak change in [Ca2+]i induced byGRGDSP in the presence of ifenprodil was not significantly different tothat with peptide alone (Fig. 1c and d: peak change in [Ca2+]i: 40±5%(GRGDSP+ifenprodil) versus 65±6% (GRGDSP)). GRGDSP-in-duced increases in [Ca2+]i were nonetheless altered by ifenprodil,suggesting that these receptors may partially contribute to the calciumchanges observed.

GRGDSP peptides induce rapid phosphorylation and nucleartranslocation of ERK1/2 MAP kinase

Neurons were exposed to 2.5 mM GRGDSP or GRADSPpeptide and phosphorylation of ERK1/2 and other key signaling

Fig. 1. Integrin-binding GRGDSP peptides induce rapid increases in intracellular calcium in cortical neurons. Intracellular calcium changes in cortical neurons(12–14 DIV) were analysed by ratiometric fura-2 imaging after treatment with (a) 2.5 mM GRGDSP or GRADSP; (b) 2.5 mM GRGDSP or GRGDSP plus10 μM MK-801; (c) 2.5 mM GRGDSP or GRGDSP plus 10 μM ifenprodil. NMDA antagonists were added to the bath 10 min before the peptide. Horizontalblack lines on each trace represent the duration of peptide treatment. Representative traces showing the average responses of at least 50 individual neurons perexperiment are shown and superimposed traces are averages from the same sister culture. (d) Calcium responses from all experiments were quantified as thedifference between the average normalized baseline (100%) fluorescence values subtracted from the average peak fluorescence value, for each stimulation (eachexperiment measuring at least 50 cells). Data are presented as mean±SEM (n=3). Statistical differences were calculated using one-way ANOVAwith Tukey posthoc test: **P<0.01, ***P<0.001 versus GRGDSP alone.

148 P.M.D. Watson et al. / Mol. Cell. Neurosci. 34 (2007) 147–154

Autho

r's

pers

onal

co

py

enzymes was analysed by immunoblotting. GRGDSP inducedtransient ERK1/2 phosphorylation with a distinct time course (Fig.2a): phosphorylation was induced within 5 min of GRGDSPapplication and returned to basal levels after 30 min. The controlGRADSP peptide did not cause significant changes in ERKphosphorylation (Fig. 2b). Phosphorylation of ERK1/2 was furtheranalysed by immunostaining. In resting cells or cells treated withGRADSP, there was only weak immunoreactivity for phosphory-lated ERK1/2 (Fig. 2c and d, respectively). Following treatmentwith GRGDSP peptide for 5 min, neurons showed strongimmunostaining for phosphorylated ERK1/2 throughout the cellbody and dendrites, with some neurons displaying intensephosphorylated ERK1/2 staining in the nucleus (Fig. 2e).

In contrast, GRGDSP had no significant effect on thephosphorylation state of either the pro-survival enzyme AKT orjun N-terminal kinase (JNK) at any time point (Figs. 3a and b,respectively). Similarly, GRGDSP had no significant effect onphosphorylation of a key regulatory residue (Tyr 397) of focaladhesion kinase (FAK; Fig 3c), which is expressed in neurons andan important mediator of integrin signaling (Parsons, 2003). Theeffect of GRGDSP on intracellular signaling in cortical neuronsthus appears to be selective for the ERK1/2 pathway.

GRGDSP-induced ERK1/2 signaling is specifically linked to theNR2B subtype of NMDA receptor

Since GRGDSP-induced increases in [Ca2+]i were blocked bythe NMDA receptor antagonist MK801, the contribution of NMDAreceptors to GRGDSP-induced ERK1/2 phosphorylation wasassessed. MK-801 abolished ERK1/2 phosphorylation in responseto GRGDSP peptide (Fig. 4a; mean fold change in phosphorylation:3.3 ± 0.7 (GRGDSP) vs. 0.28 ± 0.1 (GRGDSP+MK-801);P<0.001). Ifenprodil also abolished the ERK1/2 phosphorylationin response to GRDGSP (Fig. 4a; mean fold change in phosphor-ylation: 3.3±0.7 (GRGDSP) vs. 0.97±0.3 (GRGDSP+ifenprodil);P<0.05). These results suggest that although NR2B-containingNMDA receptors make only a small contribution to the NMDAreceptor-dependent increases in intracellular calcium in response toGRGDSP, they are specifically linked to ERK1/2 phosphorylation.

To analyse which enzymes lie upstream of GRGDSP-inducedERK1/2 phosphorylation, the response to the peptide was tested in thepresence of various signaling inhibitors.MEK1 is directly upstreamofERK1/2 and the MEK1/2 antagonist U0126 significantly reducedERK phosphorylation relative to treatment with peptide alone (Fig.4b; mean fold change in phosphorylation: 1.9±0.2 (GRGDSP) vs.

Fig. 2. GRGDSP induces ERK1/2 phosphorylation and translocation of phosphorylated ERK1/2 to the neuronal nucleus. Cortical neurons (12–14 DIV) weretreatedwith either 2.5mMGRGDSP or GRADSP. Total and phosphorylated ERK1/2 (p42/p44, phosphorylated on residues Thr 202 andTyr 204)were detected by(a and b) Western blotting of equal amounts of total protein from cells lysed after 5, 15, 30 or 60 min. Representative blots from three independent experiments onseparate cultures are shown. Cortical neurons that were either untreated (c), or treated for 5 min with GRADSP (d), or GRGDSP (e) were also fixed andimmunostained with phospho-ERK antibodies. White arrows highlight strong immunoreactivity where phosphorylated ERK1/2 has translocated to the neuronalnucleus.

149P.M.D. Watson et al. / Mol. Cell. Neurosci. 34 (2007) 147–154

Autho

r's

pers

onal

co

py0.64±0.07 (GRGDSP+U0126); P<0.05). Src is both a target ofintegrin signaling and an upstream regulator of MAPK in neuronaland non-neuronal cells (reviewed by Salter and Kalia, 2004). Srchas also been implicated in modifying AMPA and NMDA receptoractivity in response to integrin ligands in hippocampal slices(Bernard-Trifilo et al., 2005; Lin et al., 2003; Kramar et al., 2003).The Src inhibitor PP2 significantly reduced ERK1/2 phosphoryla-tion in response to GRGDSP peptide compared to peptide alone(Fig. 4b; mean fold change in phosphorylation: 1.9±0.2 (GRGDSP)vs. 0.93±0.2 (GRGDSP+PP2); P<0.05). Inhibition of phosphati-dylinositol 3-kinase (PI 3-K) with LY294002 or calmodulin kinaseII (CaMKII) with KN62 did not significantly alter ERK1/2phosphorylation (Fig. 4b).

GRGDSP peptide has no effect on excitotoxic neuronal death

Integrin binding to the extracellular matrix is important forsurvival of many cell types (Stupack and Cheresh, 2002). Integrinactivation increases neuronal precursor cell survival (Gibson et al.,2005), and neuronal apoptosis in response to glutamate is reducedeither by culturing hippocampal neurons on laminin or applying alaminin peptide mimetic (Gary et al., 2003; Gary and Mattson,2001). We therefore investigated whether GRGDSP modulatescortical neuronal survival in response to excitotoxic stimuli.Neither GRGDSP nor the control peptide GRADSP significantlyaltered basal cortical neuronal viability, although high doses ofglutamate induced extensive neuronal death (Fig. 5a). When co-applied with NMDA (50 μM) or glutamate (10 μM), which induce

Fig. 3. GRGDSP does not affect phosphorylation of AKT, JNK or FAK.Cortical neurons were lysed after treatment with GRGDSP (2.5 mM) for 5,15, 30 or 60 min. Phosphorylation of signaling enzymes was analysed byimmunoblotting equal amounts of total protein and probing with antibodiesspecific for phospho-S473 AKT (a), phospho-JNK (p46/p54; T183, Y185)(b) and phospho-Y397 FAK (c). Representative Western blot images fromthree independent experiments on separate cultures are shown (n=3).

Fig. 4. GRGDSP-induced ERK1/2 phosphorylation requires MEK1, Src,and NR2B-containing NMDA receptors. Cortical neurons were pre-treated(10 min) with MK-801 or ifenprodil (a), or inhibitors of signaling enzymes(b): UO126 (MEK1), LY294002 (PI 3-K), PP2 (Src) or KN62 (CaMKII), allat 10 μM. After 5 min treatment with 2.5 mMGRGDSP, cells were lysed andERK1/2 phosphorylation was analysed by immunoblotting equal amounts oftotal protein. Representative blots from 3 to 6 independent experiments onseparate cultures are shown, and densitometry was used to calculateincreases in phospho-ERK1/2 band intensity relative to total ERK1/2. Alldata are presented as mean±SEM. Statistical differences were calculatedusing one-way ANOVAwith Tukey post hoc test: ##P<0.01 GRGDSP aloneversus control, *P<0.05 GRGDSP plus inhibitor versus GRGDSP alone,***P<0.001 GRGDSP plus inhibitor versus GRGDSP alone.

150 P.M.D. Watson et al. / Mol. Cell. Neurosci. 34 (2007) 147–154

Autho

r's

pers

onal

co

pyapproximately 50% cell death after 24 h (data not shown),GRGDSP did not significantly alter excitotoxic cell death (Fig.5b). Pre-treating neurons with GRGDSP for up to 24 h beforeaddition of glutamate similarly did not modulate cell death (Fig.5c). Finally, to explore whether GRGDSP is neuroprotectiveagainst an acute application of excitotoxin, cortical neurons wereexposed to a higher dose of glutamate for 1 h in the presence ofGRGDSP or GRADSP. However, GRGDSP had no effect onneuronal survival under these treatment conditions (Fig. 5d).

Discussion

The contribution of integrins to neuronal function, most notablysynaptic transmission, is increasingly apparent (Gall and Lynch,2004; Milner and Campbell, 2002). Several studies have indicatedthat integrin-binding peptides containing the RGD sequence caninfluence neuronal excitability (Bernard-Trifilo et al., 2005;Kramar et al., 2003, Lin et al., 2003). Published studies howeverall demonstrate effects of GRGDSP on neuronal behaviour over atime course of minutes. We believe that the data we present hereare the first to show that RGD peptides induce very rapid effects in

neurons, within seconds of peptide application. These effects,which include NMDA receptor-dependent intracellular calciumincreases (probably due to direct activation of the receptor andgeneration of plasmalemmal Ca2+ influx) and ERK1/2 signaling,are likely to be upstream of longer lasting integrin-dependenteffects on neuronal transmission such as excitability and LTP.Rapid signaling through integrin and NMDA receptors uponpresentation of an integrin-binding ligand may therefore representa novel pathway that triggers more persistent integrin-dependenteffects. Although GRGDSP activated NMDA receptors andincreased intracellular calcium, which are both key upstreammediators of excitotoxicity (Arundine and Tymianski, 2003;Nicholls and Budd, 2000), the peptide was not toxic, nor did italter neuronal adhesion or morphology (data not shown), and it didnot modulate NMDA receptor-mediated cell death indicating aselective action on NMDA-mediated transmission.

These results suggest that integrin–ligand binding may veryrapidly activate several intracellular pathways in neurons. NMDAreceptor activation leads to the rapid activation of MAPK and Srcsignaling, with the potential to influence neuronal gene expressionthrough the nuclear translocation of phosphorylated ERK1/2. Our

Fig. 5. GRGDSP peptides do not alter neuronal cell death in response to chronic or acute excitotoxic insults. Cortical neurons were treated with: 2.5 mMGRGDSP, GRADSP or 500 μM glutamate alone for 24 h (a), 50 μMNMDA or 10 μM glutamate±2.5 mM GRGDSP or GRADSP for 24 h (b), pre-treated with2.5 mM GRGDSP for 0, 2, 4, 6 or 24 h prior to treatment with 10 μM glutamate for 24 h (c) or exposed to 200 μM glutamate for 1 h in the presence of 2.5 mMGRGDSP or GRADSP (d). Peptides and excitotoxins were added directly to the culture medium and cell death was quantified by lactate dehydrogenase assay.Cell death is expressed as mean increase in LDH release normalised to basal levels of LDH present in culture medium in each separate experiment 24 h aftertreatments. Data are presented as mean±SEM. **P<0.01 versus control, *P<0.05 versus control.

151P.M.D. Watson et al. / Mol. Cell. Neurosci. 34 (2007) 147–154

Autho

r's

pers

onal

co

py

data are consistent with previous reports indicating roles forNMDA receptors and Src in integrin signaling, such as GRGDSP-induced potentiation of AMPA receptor currents, phosphorylationof CaMKII and GluR1 and endocytosis of β-amyloid peptide, butsuggest that these events are initiated much faster than previouslythought (Bernard-Trifilo et al., 2005; Lin et al., 2003, Kramar et al.,2003; Bi et al., 2002). In hippocampal slices, GRGDSP potentiatesSrc-dependent NMDA receptor-mediated responses within 15–20 min (Bernard-Trifilo et al., 2005; Lin et al., 2003; Kramar et al.,2003); this is consistent with the time scale over which GRGDSPand integrin function-blocking antibodies disrupt stabilization ofLTP (Kramar et al., 2002; Chun et al., 2001; Staubli et al., 1998;Bahr et al., 1997). Our results however indicate that GRGDSPactivates NMDA receptors much more rapidly in cortical neurons,with intracellular calcium increases detected within seconds ofpeptide application (Fig. 1). This rapid activation of NMDAreceptors may precede and modulate subsequent synapticresponses involving phosphorylation of receptors, cytoskeletalrearrangements or proteolysis by enzymes such as calpain (Bahr,2000). The rapid response to integrin-ligand binding is consistentwith other reports of disruption of LTP stabilization by GRGDSPon a shorter time scale (LeBaron et al., 2003).

The results reported here implicate a specific subtype of NMDAreceptor, containing the NR2B subunit, in GRGDSP-inducedactivation of ERK1/2 MAP kinase. NR2B-containing NMDAreceptors have been implicated in ERK1/2 signaling in neurons(Kim et al., 2005; Krapivinsky et al., 2003) and colocalise withERK and its upstream signaling partners (Husi et al., 2000).Integrins have also been implicated in activity-dependent deve-lopmental switches in the composition of NR2B-containing NMDAreceptors (Sinagra et al., 2005; Chavis and Westbrook, 2001).GRGDSP-induced increases in intracellular calcium are howevernot solely dependent upon this subtype of NMDA receptor sinceifenprodil did not reduce the calcium increases as much as MK-801.It is likely therefore that the calcium increase induced by RGDpeptide has a complex nature: although triggered by activation ofNMDA receptors with calcium influx through them, there may be aparallel depolarization and calcium influx through voltage-gatedchannels and possibly release from intracellular stores. Ifenprodilwould be predicted only to affect the NR2B component.

In hippocampal slices and synaptosomes, GRGDSP inducesphosphorylation of CaMKII, GluR1 (Kramar et al., 2003), Src,Pyk2 and the NMDA receptor subunits NR2A and NR2B(Bernard-Trifilo et al., 2005) over similar time scales to thosedescribed here for ERK1/2. We observed no increase in FAKphosphorylation however, and no effect of an inhibitor of CaMKIIon ERK1/2, perhaps reflecting intrinsic differences in integrin orNMDA receptor signaling pathways in cortical versus hippocampalneurons.

Intracellular calcium is a critical second messenger andregulator of homeostasis in neurons (Verkhratsky, 2005) thatmodulates synaptic plasticity and memory formation (Sweatt,2004; Thomas and Huganir, 2004; Agell et al., 2002), processesalso regulated by integrins (Gall and Lynch, 2004). The very rapidintegrin-mediated MAPK and calcium signaling may thereforecontribute to the various neurochemical changes that underlie theearly stages of memory and learning. Longer-term changes inmemory and LTP also require new gene expression. Activation andnuclear translocation of ERK1/2 in response to GRGDSP couldmodulate gene expression within cortical neurons. Indeed, ERKand intracellular calcium signaling can stimulate gene expression

(Xia et al., 1996; Bading et al., 1993, 1995), and in hippocampalneurons, GRGDSP induces L-type calcium channel-dependentexpression of neurotrophin and neurotrophin receptor genes (Gallet al., 2003).

Integrins may modulate NMDA receptor activity indirectly viaintracellular signaling and/or directly via physical association. Srcis a good candidate for signaling between integrins and NMDAreceptors, since Src inhibitors block the GRGDSP effects onNMDA and AMPA receptor transmission (Bernard-Trifilo et al.,2005; Lin et al., 2003; Kramar et al., 2003), and the NMDAreceptor-dependent ERK1/2 phosphorylation reported here. In thepost-synaptic membrane, NMDA receptors reside in large com-plexes, containing cytoskeletal and signaling molecules (Husi etal., 2000), many of which are also found in integrin-associatedcomplexes. Integrins themselves however have not been detectedin NMDA receptor complexes (Husi et al., 2000), although thepossibility remains that integrins may be transiently present orrecruited upon synaptic stimulation or strengthening in paradigmssuch as LTP. AMPA receptor stimulation, for example, increasessurface expression of integrin α5β1 (Lin et al., 2005). Interestingly,some integrin ligands such as the molecule L1 are present inNMDA receptor complexes (Husi et al., 2000). The precise natureof the integrin heterodimers that regulate these rapid signalingresponses when binding GRGDSP also remains to be definitivelydetermined, although, using RT-PCR and immunoblotting, wehave found that cultured rat cortical neurons express α3, α4, α5, α6,α7, α8, α9, α10 α11 αV and β1, β3, β6, β8 subunits (data notshown), implying that RGD-binding heterodimers α5β1, α8β1,αVβ1, αVβ3, αVβ6 or αVβ8 may play a role. It is plausible thatdifferent integrin heterodimers might differentially affect NMDAreceptor signaling properties, depending upon the subunitcomposition of the receptors.

In conclusion, integrin binding to the peptide ligand GRGDSPinduces rapid increases of intracellular calcium, and phosphoryla-tion and nuclear translocation of ERK1/2 in primary corticalneurons, and these effects require the activity of NMDA receptors.Our results provide further evidence for the importance ofintegrins in neuronal biology and support for the hypothesis thatcell adhesion in neurons may play a critical role in complexneurological processes such as learning, memory and potentiallyalso in neurodegenerative disease.

Experimental methods

Reagents

All general chemicals were purchased from Sigma-Aldrich (UK), andcell culture reagents were purchased from Gibco (Invitrogen, UK) unlessstated otherwise. Integrin-binding peptides were obtained from Calbiochem(UK), NMDA receptor antagonists and the Src inhibitor PP2 were fromTocris-Cookson (UK) and MEK and PI3K inhibitors were from CellSignaling Technology (UK).

Primary culture of embryonic cortical neurons

Primary cultures of cerebrocortical neurons were prepared as describedpreviously (Moore et al., 2002), with minor modifications. Briefly,dissected cortices taken from E18 rats were mechanically dissociated inNeurobasal medium (containing 2% B27 supplement with antioxidants, 5%fetal bovine serum, 2 mM L-glutamine, 100 IU/ml penicillin and 100 μg/mlstreptomycin), and the cell suspension was passed through a 235 μm nylonmesh (Nytex nylon, John Staniar, UK). The cells were pelleted and re-

152 P.M.D. Watson et al. / Mol. Cell. Neurosci. 34 (2007) 147–154

Autho

r's

pers

onal

co

py

suspended in Neurobasal. Neurons were plated onto poly-D-lysine (PDL) ata concentration of 75,000 cells/cm2 and maintained in a humidifiedincubator at 37°C, 5% CO2. After 2 days in vitro (DIV), 10 μM cytosine-D-arabinofuranoside was added to inhibit the proliferation of non-neuronalcells. Media were changed after 5 DIV to serum-free Neurobasal. Furthermedia changes replacing half the volume of media were carried out after 7and 9 DIV. Neuronal cultures were 97% pure as determined byimmunostaining with cell type-specific antibodies (data not shown). Cellswere used for experiments after 14–16 DIV.

Preparation of protein samples and Western blotting

Following cellular treatments, proteins were extracted on ice usingNP40 protein lysis buffer (1% NP40, 25 mM HEPES, 0.4 M KCl, 5 mMEDTA) containing protease and phosphatase inhibitors. The concentrationof protein in the cell lysates was determined using the BCA assay (Pierce,UK). Proteins were separated and analysed by immunoblotting for ERK,FAK and AKT as described previously (Gibson et al., 2005).

Lactate dehydrogenase cell death assay

Neuronal cell death was assessed by measuring the release of thecytosolic enzyme lactate dehydrogenase (LDH) into the culture mediumusing the CytoTox 96 Non-Radioactive Cytotoxicity Assay (Promega, UK)according to the manufacturer’s guidelines. Sample absorbance values wereread at 490 nm using a spectrophotometric plate reader (Dynatech MR4000,UK).

Immunocytochemistry on primary neurons

Neurons were plated on glass coverslips, pre-coated with PDL. Cellswere washed in ice cold PBS and fixed in cold 4% (w/v) paraformaldehyde(PFA) for 15 min before permeabilisation with ice-cold methanol for15 min. Non-specific antibody binding was prevented by incubation with3% BSA (w/v) in PBS for 30 min at room temperature (RT). Antibodyraised against phosphorylated ERK1/2 protein (1:1000; Cell SignalingTechnology Inc, UK) was added to the cells for 1 h (RT). Cells were thenwashed 3 times in PBS before addition of secondary antibody (anti-rabbitIgG conjugated to Cy3, 1:1000, DakoCytomation, UK) for 1 h (RT) in thedark. Cells were washed again, and glass coverslips were mounted usingVectorshield mounting medium containing DAPI (Vector Laboratories,UK). Neurons were examined using a fluorescence microscope (LeicaDMR, UK).

Fura-2AM calcium imaging in primary cortical neurons

Ratiometric calcium imaging was carried out as previously described(Brough et al., 2003). Neurons grown on PDL-coated glass coverslips werewashed with HEPES-buffered saline (HBS) and then incubated for 30 min inHBS containing 5 μM fura-2AM (Molecular Probes, UK) at RT. The cellswere washed 3 times in HBS and allowed to rest for 30 min (RT). Thecoverslips were then placed in the microscope chamber filled with HBS,mounted on the stage of an inverted microscope (Olympus IX70). Thespecimen was alternately illuminated at 340 nm and 380 nm by amonochromator (Polychrom IV; T.I.L.L. Photonics, Germany) at a cyclefrequency of 1–2 Hz. Fluorescence images (515±10 nm) were capturedusing a charge-coupled (CCD), cooled intensified camera (Pentamax GeneIV; Roper Scientific, U.K.). Data were analysed using MetaFluor/MetaMorph software (Universal Imaging, Downingtown, PA).

Statistical analysis

All data are presented as mean±standard error of the mean (SEM). Dataconsisting of more than two groups were analysed using one-way ANOVAwith Tukey–Kramer post-test that compares all treatment groups. Allstatistical analyses and graphs were generated using GraphPad Prism 4.0software (GraphPad Software Inc. USA).

Acknowledgments

This work was funded by the BBSRC and BiogenIdec(PMDW), The Wellcome Trust (MJH, AV) and the MedicalResearch Council (NJR, RMG).

References

Akiyama, S.K., Hasegawa, E., Hasegawa, T., Yamada, K.M., 1985. Theinteraction of fibronectin fragments with fibroblastic cells. J. Biol.Chem. 260, 13256–13260.

Agell, N., Bachs, O., Rocamora, N., Villalonga, P., 2002. Modulation of theRas/Raf/MEK/ERK pathway by Ca2+ and calmodulin. Cell Signal. 14,649–654.

Arundine, M., Tymianski, M., 2003. Molecular mechanisms of calcium-dependent neurodegeneration in excitotoxicity. Cell Calcium 34,325–337.

Bading, H., Ginty, D.D., Greenberg, M.E., 1993. Regulation of geneexpression in hippocampal neurons by distinct calcium signalingpathways. Science 260, 181–186.

Bading, H., Segal, M.M., Sucher, N.J., Dudek, H., Lipton, S.A., Greenberg,M.E., 1995. N-methyl-D-aspartate receptors are critical for mediating theeffects of glutamate on intracellular calcium concentration andimmediate early gene expression in cultured hippocampal neurons.Neuroscience. 64, 653–664.

Bahr, B.A., 2000. Integrin-type signaling has a distinct influence on NMDA-induced cytoskeletal disassembly. J. Neurosci. Res. 59, 827–832.

Bahr, B.A., Staubli, U., Xiao, P., Chun, D., Ji, Z.X., Esteban, E.T., Lynch,G., 1997. Arg-Gly-Asp-Ser-selective adhesion and the stabilization oflong-term potentiation: pharmacological studies and the characterizationof a candidate matrix receptor. J. Neurosci. 17, 1320–1329.

Bernard-Trifilo, J.A., Kramar, E.A., Torp, R., Lin, C.Y., Pineda, E.A.,Lynch, G., Gall, C.M., 2005. Integrin signaling cascades are operationalin adult hippocampal synapses and modulate NMDA receptorphysiology. J. Neurochem. 93, 834–849.

Bi, X., Lynch, G., Zhou, J., Gall, C.M., 2001. Polarized distribution of alpha5integrin in dendrites of hippocampal and cortical neurons. J. Comp.Neurol. 435, 184–193.

Bi, X., Gall, C.M., Zhou, J., Lynch, G., 2002. Uptake and pathogenic effectsof amyloid beta peptide 1–42 are enhanced by integrin antagonists andblocked by NMDA receptor antagonists. Neuroscience 112, 827–840.

Brough, D., Le Feuvre, R.A., Wheeler, R.D., Solovyova, N., Hilfiker, S.,Rothwell, N.J., Verkhratsky, A., 2003. Ca2+ stores and Ca2+ entrydifferentially contribute to the release of IL-1 beta and IL-1 alpha frommurine macrophages. J. Immunol. 170, 3029–3036.

Capaldi, D., Rosario, R., Esteban, E.T., Bahr, B.A., 1997. A 27-kDa matrixreceptor from rat brain synaptosomes: selective recognition of the Arg-Gly-Asp-Ser domain and unique resistance to calcium-dependentproteolysis. Neurosci. Res. 28, 275–279.

Chan, C.S., Weeber, E.J., Kurup, S., Sweatt, J.D., Davis, R.L., 2003.Integrin requirement for hippocampal synaptic plasticity and spatialmemory. J. Neurosci. 23, 7107–7116.

Chan, C.S., Weeber, E.J., Zong, L., Fuchs, E., Sweatt, J.D., Davis, R.L.,2006. Beta1-integrins are required for hippocampal AMPA receptor-dependent synaptic transmission, synaptic plasticity, and workingmemory. J. Neurosci. 26, 223–232.

Chavis, P., Westbrook, G., 2001. Integrins mediate functional pre- andpostsynapticmaturation at a hippocampal synapse. Nature 411, 317–321.

Chun, D., Gall, C.M., Bi, X., Lynch, G., 2001. Evidence that integrinscontribute to multiple stages in the consolidation of long termpotentiation in rat hippocampus. Neuroscience 105, 815–829.

Clegg, D.O., Wingerd, K.L., Hikita, S.T., Tolhurst, E.C., 2003. Integrins inthe development, function and dysfunction of the nervous system. Front.Biosci. 8, d723–d750.

Dityatev, A., Schachner, M., 2003. Extracellular matrix molecules andsynaptic plasticity. Nat. Rev., Neurosci. 4, 456–468.

153P.M.D. Watson et al. / Mol. Cell. Neurosci. 34 (2007) 147–154

Autho

r's

pers

onal

co

py

Einheber, S., Pierce, J.P., Chow, D., Znamensky, V., Schnapp, L.M., Milner,T.A., 2001. Dentate hilar mossy cells and somatostatin-containingneurons are immunoreactive for the alpha8 integrin subunit: character-ization in normal and kainic acid-treated rats. Neuroscience 105,619–638.

Gall, C.M., Lynch, G., 2004. Integrins, synaptic plasticity and epileptogen-esis. Adv. Exp. Med. Biol. 548, 12–33.

Gall, C.M., Pinkstaff, J.K., Lauterborn, J.C., Xie, Y., Lynch, G., 2003.Integrins regulate neuronal neurotrophin gene expression througheffects on voltage-sensitive calcium channels. Neuroscience 118,925–940.

Gary, D.S., Mattson, M.P., 2001. Integrin signaling via the PI3-kinase-Aktpathway increases neuronal resistance to glutamate-induced apoptosis.J. Neurochem. 76, 1485–1496.

Gary, D.S., Milhavet, O., Camandola, S., Mattson, M.P., 2003. Essential rolefor integrin linked kinase in Akt-mediated integrin survival signaling inhippocampal neurons. J. Neurochem. 84, 878–890.

Gibson, R.M., Craig, S.E., Heenan, L., Tournier, C., Humphries, M.J., 2005.Activation of integrin alpha5beta1 delays apoptosis of Ntera2 neuronalcells. Mol. Cell. Neurosci. 28, 588–598.

Grooms, S.Y., Terracio, L., Jones, L.S., 1993. Anatomical localization ofbeta 1 integrin-like immunoreactivity in rat brain. Exp. Neurol. 122,253–259.

Hama, H., Hara, C., Yamaguchi, K., Miyawaki, A., 2004. PKC signalingmediates global enhancement of excitatory synaptogenesis in neuronstriggered by local contact with astrocytes. Neuron. 41, 405–415.

Huettner, J.E., Bean, B.P., 1998. Block of N-methyl-D-aspartate-activatedcurrent by the anticonvulsant MK-801: selective binding to openchannels. Proc. Natl. Acad Sci. U. S. A. 85, 1307–1311.

Humphries, M.J., 2000. Integrin cell adhesion receptors and the concept ofagonism. Trends Pharmacol. Sci. 21, 29–32.

Humphries, M.J., Olden, K., Yamada, K.M., 1986. A synthetic peptide fromfibronectin inhibits experimental metastasis of murine melanoma cells.Science 233, 467–470.

Husi, H., Ward, M.A., Choudhary, J.S., Blackstock, W.P., Grant, S.G., 2000.Proteomic analysis of NMDA receptor-adhesion protein signalingcomplexes. Nat. Neurosci. 3, 661–669.

Hynes, R.O., 1992. Integrins: versatility, modulation, and signaling in celladhesion. Cell 69, 11–25.

Hynes, R.O., 2002. Integrins: bidirectional, allosteric signaling machines.Cell 110, 673–687.

Karanian, D.A., Brown, Q.B., Makriyannis, A., Bahr, B.A., 2005. Blockingcannabinoid activation of FAK and ERK1/2 compromises synapticintegrity in hippocampus. Eur. J. Pharmacol. 508, 47–56.

Kim, M.J., Dunah, A.W., Wang, Y.T., Sheng, M., 2005. Differential Roles ofNR2A-and NR2B-Containing NMDA Receptors in Ras-ERK Signalingand AMPA Receptor Trafficking. Neuron 46, 745–760.

Kramar, E.A., Lynch, G., 2003. Developmental and regional differences inthe consolidation of long-term potentiation. Neuroscience 118,387–398.

Kramar, E.A., Bernard, J.A., Gall, C.M., Lynch, G., 2002. Alpha3 integrinreceptors contribute to the consolidation of long-term potentiation.Neuroscience 110, 29–39.

Kramar, E.A., Bernard, J.A., Gall, C.M., Lynch, G., 2003. Integrinsmodulate fast excitatory transmission at hippocampal synapses. J. Biol.Chem. 278, 10722–10730.

Krapivinsky, G., Krapivinsky, L., Manasian, Y., Ivanov, A., Tyzio, R.,Pellegrino, C., Ben-Ari, Y., Clapham, D.E., Medina, I., 2003. TheNMDA receptor is coupled to the ERK pathway by a direct interactionby NR2B and RasGRF1. Neuron 40, 775–784.

LeBaron, R.G., Hernandez, R.V., Orfila, J.E., Martinez Jr., J.L., 2003. Anintegrin binding peptide reduces rat CA1 hippocampal long-termpotentiation during the first few minutes following theta burststimulation. Neurosci. Lett. 339, 199–202.

Leclerc, J.R., 2002. Platelet glycoprotein IIb/IIIa antagonists: lessonslearned from clinical trials and future directions. Crit Care Med. 30,S332–S340.

Lin, B., Arai, A.C., Lynch, G., Gall, C.M., 2003. Integrins regulateNMDA receptor-mediated synaptic currents. Neurophysiology 89,2874–2878.

Lin, C.Y., Lynch, G., Gall, C.M., 2005. AMPA receptor stimulationincreases alpha5beta1 integrin surface expression, adhesive function andsignaling. J. Neurochem. 94, 531–546.

Milner, R., Campbell, I.L., 2002. The integrin family of cell adhesionmolecules has multiple functions within the CNS. J. Neurosci. Res. 69,286–291.

Moore, J.D., Rothwell, N.J., Gibson, R.M., 2002. Involvement of caspasesand calpains in cerebrocortical neuronal cell death is stimulus-dependent. Brit. J. Pharm. 135, 1069–1077.

Mori, H., Mishina, M., 1995. Structure and function of the NMDA receptorchannel. Neuropharmacology 34, 1219–1237.

Nicholls, D.G., Budd, S.L., 2000. Mitochondria and neuronal survival.Physiol. Rev. 80, 315–360.

Nikonenko, I., Toni, N., Moosmayer, M., Shigeri, Y., Muller, D., SargentJones, L., 2003. Integrins are involved in synaptogenesis, cell spreading,and adhesion in the postnatal brain. Brain Res. Dev. Brain Res. 140,185–194.

Nishimura, S.L., Boylen, K.P., Einheber, S., Milner, T.A., Ramos, D.M.,Pytela, R., 1998. Synaptic and glial localization of the integrinalphavbeta8 in mouse and rat brain. Brain Res. 791, 271–282.

Parsons, J.T., 2003. Focal adhesion kinase: the first ten years. J. Cell Sci.116, 1409–1416.

Pinkstaff, J.K., Detterich, J., Lynch, G., Gall, C., 1999. Integrin subunit geneexpression is regionally differentiated in adult brain. J. Neurosci. 19,1541–1556.

Rodriguez, M.A., Pesold, C., Liu, W.S., Kriho, V., Guidotti, A., Pappas,G.D., Costa, E., 2000. Colocalization of integrin receptors and reelinin dendritic spine postsynaptic densities of adult nonhuman primatecortex. Proc. Natl. Acad Sci. U. S. A. 97, 3550–3555.

Salter, M.W., Kalia, L.V., 2004. Src kinases: a hub for NMDA receptorregulation. Nat. Rev., Neurosci. 5, 317–328.

Sinagra, M., Verrier, D., Frankova, D., Korwek, K.M., Blahos, J.,Weeber, E.J., Manzoni, O.J., Chavis, P., 2005. Reelin, very-low-density lipoprotein receptor, and apolipoprotein E receptor 2 controlsomatic NMDA receptor composition during hippocampal maturationin vitro. J. Neurosci. 25, 6127–6136.

Smart, F.M., Edelman, G.M., Vanderklish, P.W., 2004. BDNF inducestranslocation of initiation factor 4E to mRNA granules: evidence fora role of synaptic microfilaments and integrins. Proc. Natl. Acad Sci.U. S. A. 100, 14403–14408.

Staubli, U., Chun, D., Lynch, G., 1998. Time-dependent reversal of long-term potentiation by an integrin antagonist. J. Neurosci. 18,3460–3469.

Stupack, D.G., Cheresh, D.A., 2002. Get a ligand, get a life: integrins,signaling and cell survival. J. Cell Sci. 115, 3729–3738.

Sweatt, J.D., 2004. Mitogen-activated protein kinases in synaptic plasticityand memory. Curr. Opin. Neurobiol. 14, 311–317.

Thomas, G.M., Huganir, R.L., 2004. MAPK cascade signaling and synapticplasticity. Nat. Rev., Neurosci. 5, 173–183.

van der Flier, A., Sonnenberg, A., 2001. Function and interactions ofintegrins. Cell Tissue Res. 305, 285–298.

Verkhratsky, A., 2005. Physiology and pathophysiology of the calciumstore in the endoplasmic reticulum of neurons. Physiol. Rev. 85,201–279.

Xia, Z., Dudek, H., Miranti, C.K., Greenberg, M.E., 1996. Calcium influxvia the NMDA receptor induces immediate early gene transcription bya MAP kinase/ERK-dependent mechanism. J. Neurosci. 16,5425–5436.

154 P.M.D. Watson et al. / Mol. Cell. Neurosci. 34 (2007) 147–154