Platelet-derived Growth Factor Receptor-induced Feed-forward Inhibition of Excitatory Transmission between Hippocampal Pyramidal Neurons*

Growth factor receptors provide a major mechanism for the activation of the nonreceptor tyrosine kinase c-Src, and this kinase in turn up-regulates the activity ofN-methyl-d-aspartate (NMDA) receptors in CA1 hippocampal neurons (1). Unexpectedly, applications of platelet-derived growth factor (PDGF)-BB to cultured and isolated CA1 hippocampal neurons depressed NMDA-evoked currents. The PDGF-induced depression was blocked by a PDGF-selective tyrosine kinase inhibitor, by a selective inhibitor of phospholipase C-γ, and by blocking the intracellular release of Ca2+. Inhibitors of cAMP-dependent protein kinase (PKA) also eliminated the PDGF-induced depression, whereas a phosphodiesterase inhibitor enhanced it. The NMDA receptor-mediated component of excitatory synaptic currents was also inhibited by PDGF, and this inhibition was prevented by co-application of a PKA inhibitor. Src inhibitors also prevented this depression. In recordings from inside-out patches, the catalytic fragment of PKA did not itself alter NMDA single channel activity, but it blocked the up-regulation of these channels by a Src activator peptide. Thus, PDGF receptors depress NMDA channels through a Ca2+- and PKA-dependent inhibition of their modulation by c-Src.

In the central nervous system platelet-derived growth factor receptors (PDGF-R) 1 play a role in the development and survival of neurons, in the regulation of synapses, and in adaptive responses following neuronal injury. These receptors are expressed in the hippocampus, and their activation regulates both excitatory and inhibitory synaptic transmission mediated by ␥-aminobutyric acid (2) and NMDA receptors (3), respec-tively (4). PDGF itself forms dimers (PDGF-AA, -BB, and -AB) that bind to two classes of PDGF-R (␣ and ␤) (5), and ligand binding induces receptor dimerization and tyrosine autophosphorylation. Phosphotyrosyl residues of the receptor can then bind a variety of signal transduction molecules via their Src homology 2 domains. These include phospholipase C-␥ (PLC-␥), members of the Src family of protein tyrosine kinases including c-Src, the protein tyrosine phosphatase SYP-2, Ras-GTPaseactivating protein, and phosphatidylinositol 3-kinase, as well as the adaptor molecules Grb2, Nck, and Shc (5)(6)(7). Activation of PDGF-R can therefore signal through various parallel pathways to cell growth and motility. However, recent studies have suggested that these parallel pathways can signal both stimulatory and inhibitory signals (8).
Of the growth factors, there is considerable evidence that PDGF-R receptor activation of Src or PLC-␥ leads to stimulation of the mitogen-activated protein kinase (MAPK) cascade and cell mitogenesis (5). Similar activation of the MAPK cascade and mitogenesis has also been reported for G-proteincoupled receptors (9), and we recently reported that several G-protein-coupled receptors act through a sequential activation of PKC and c-Src to up-regulate the activity of N-methyl-Daspartate (NMDA) receptors (1). This subtype of glutamate receptor is an ion channel that contributes to excitatory synaptic transmission, to plasticity, and to cell growth of central neurons (10). Paradoxically, we also demonstrated that PDGF induces a long term depression of synaptically evoked NMDA receptor-mediated currents at CA1 synapses as well as of NMDA-evoked currents recorded in cultured hippocampal neurons (3).
The PDGF-R-mediated depression of NMDA channels requires activation of the PLC-␥ pathway as a tyrosine site (Tyr 1021 ) add-back mutant of the PDGF-␤ receptor restores this function when receptors are co-expressed in Xenopus oocytes (3). PLC-␥ catalyzes the hydrolysis of phosphatidylinositol-4,5bisphosphate to two potent second messengers: diacylglycerol, which activates various isoforms of protein kinase C (PKC), and inositol triphosphate, which releases intracellular Ca 2ϩ (11). Therefore PDGF-R might be anticipated to activate the PKC/Src signaling cascade in CA1 pyramidal hippocampal neurons responsible for up-regulation of NMDA channel activity. Instead, we show here that the PDGF-R signaling is mediated through a second but functionally opposing pathway. This pathway signals a Ca 2ϩ -and PKA-dependent inhibition of the Src regulation of NMDA channels.

EXPERIMENTAL PROCEDURES
Preparation of Acutely Isolated Hippocampal CA1 Neurons-CA1 hippocampal pyramidal neurons were acutely isolated using modified procedures of Wang and MacDonald (12). Briefly, Wistar rats 2-3 weeks old were decapitated via a guillotine after halothane anesthesia. Hippocampi were removed quickly and placed in a dish containing cold oxygenated external solution consisting of 140 mM NaCl, 1.0 mM CaCl 2 , 5.4 mM KCl, 25 mM Hepes, 33 mM glucose, 1 mM MgCl 2 , and 0.0003 mM tetrodotoxin, pH 7.4 (osmolarity, 320 -335 mosmol liter Ϫ1 ). The hippocampi were cut into 300 -500-m-thick slices by hand with a razor blade. The hippocampal slices were digested at room temperature (20 -22°C) in the external solution with 5 mg/ml papaya latex (Sigma). The incubation medium was stirred with 95% O 2 /5% CO 2 . After 30 min of enzymatic digestion, the slices were rinsed three times with the external solution. The slices were kept in the external solution bubbled with oxygen and used for 8 -10 h. The CA1 region was cut out with a scalpel under a phase contrast microscope and triturated with a fire-polished glass pipette.
Whole Cell Recordings-Whole cell recordings were performed with an Axopatch-1B amplifier (Axon Instruments) under voltage-clamp mode. Recording electrodes, with resistance of 3-5 M⍀ were constructed from thin walled borosilicate glass (1.5-mm-diameter; World Precision Instruments) using a two-stage puller (PP83, Narishige). Data were digitized, filtered (2 kHz), and acquired on-line using the programs pClamp (Axon Instruments). The standard internal solution for recording electrode consisted of the following 140 mM CsF, 35 mM CsOH, 10 mM Hepes, 2 mM MgCl 2 , 2 mM tetraethylammonium, 4 mM Na 2 ATP, pH 7.3 (osmolarity, 300 mosmol liter Ϫ1 ). A multibarrel fast perfusion system (13) was employed to achieve a rapid exchange of solution. NMDA (50 M) and glycine (3 M) in the external solution were applied to the neurons via one barrel for 2 s to evoke NMDA-mediated currents. Isolated hippocampal neurons were first patch clamped and then lifted into the outflow of the control barrel. The distance between the outlet of the barrel and the neuron was within 100 m. The holding potential was Ϫ60 mV. PDGF and other drugs were diluted in the external solution to the required concentrations and applied to the neurons via the control barrel unless otherwise stated. Some hydrophobic drugs were initially dissolved in dimethyl sulfoxide and then diluted to the desired concentration in the external solution. The final concentration of Me 2 SO applied to the cells was less than 0.1%, and this concentration of Me 2 SO did not affect NMDA currents.
Miniature Synaptic Currents in Cultured Hippocampal Neurons-Cultures of fetal hippocampal neurons were used for electrophysiological recordings 12-17 days after plating. Spontaneous miniature mEP-SCs in cultured hippocampal neurons were recorded after formation of the whole cell patch configuration. The bathing solution was supplemented with 0.5 M tetrodotoxin, 1 M strychnine, 10 M bicuculline methiodide, and 3 M glycine. Miniature EPSCs were filtered at 2 kHz and stored on tape prior to their off-line acquisition with an event detection program (SCAN, Strathclyde software; courtesy of Dr. J. Dempster). For detection, the trigger level was set approximately three times higher than the base-line noise. False events were eliminated by subsequent inspection of the raw data.
Single-channel Recordings-Inside-out single channel recordings were carried out on cultured mouse hippocampal neurons (14). After coating with Sylgard the electrodes were fire-polished and filled with the standard external solution containing NMDA (10 M) and glycine (3 M). The intracellular solution for bathing the cytoplasmic face of the patch contained 120 mM CsCl, 35 mM CsOH, 10 mM Hepes, 2 mM MgCl 2 , 11 mM EGTA, 2 mM tetraethylammonium, 4 mM Na 2 ATP, pH 7.3 (osmolarity, 300 mosmol liter Ϫ1 ). The intracellular solution was supplemented as required with EPQ(pY)EEIPIA, EPQYEEIPIA, or the catalytic subunit of PKA (cPKA) just before use. During recording, the holding potential of the pipette was ϩ60 mV. Single channel events were recorded on videotape using a digital data recorder (VR-10, Instrutech Corp., Mineola, NY) and later played back and acquired using the pClamp 6 program (Axon Instruments). Single channel currents were sampled at 5 kHz and filtered at 2 kHz.
Calcium Imaging-Cultured hippocampal neurons were incubated in 6 M fura-2-acetoxy-methyl ester dissolved in Me 2 SO with pluronic acid in minimal essential for 30 -40 min at room temperature. Coverslips with fura-2-loaded cells were transferred to a perfusion chamber on an inverted microscope. Cells were illuminated using a xenon lamp and were observed using a 40ϫ UV fluor oil immersion objective on an inverted microscope (Nikon). Video images were obtained using an intensified CCD camera (PTI IC-100). Digitized images were acquired by averaging four frames at video rates using an image processing board (Axon Imagine Lightening) in a computer controlled by Axon Imaging Workbench software (Axon Instruments). The shutter and filter wheels also were controlled by Axon Imaging Workbench software to allow timed illumination of cells at 340 and 380 nm excitation wavelengths. Ratio images were analyzed by averaging pixel ratio values in circumscribed regions of all responding cells in the field of view. The values were exported from Axon Imaging Workbench software to a spreadsheet program (Sigmaplot or Graphpad).
Statistics Analysis-Currents were expressed as the means Ϯ S.E. normalized to the control values before the application of drugs. Values in parentheses refer to the number of different cells used in the statistical analysis. Statistics analyses were performed using either Student's t test or two-way analysis of variance.

PDGF Receptor Tyrosine Kinase Activity Is Required for PDGF-mediated Depression of NMDA Receptor Function-In
acutely isolated hippocampal CA1 pyramidal neurons applications of PDGF (6 nM) substantially depressed peak NMDAevoked currents (I p ), whereas steady-state currents were much less effected. The maximum effect was reached within 10 -15 min of applying PDGF (Fig. 1). I p was inhibited by 29.2 Ϯ 3.7% (n ϭ 9, p Ͻ 0.01 by Student's t test) 20 min following the application of PDGF. Recordings from matched cells that did not received PDGF demonstrated stable peak responses to NMDA over the same period of recording (Fig. 1). In all subse- quent whole cell recordings we examined the effects of agents on I p .
We first examined whether the tyrosine kinase activity of PDGF receptors in CA1 pyramidal neurons was required for the depression of I p by employing the novel and specific PDGF receptor tyrosine kinase inhibitor Win41662. This inhibitor suppresses the PDGF-stimulated autophosphorylation of PDGF receptors and the intracellular Ca 2ϩ mobilization induced by PDGF, as well as the cell proliferation by PDGF in human vascular smooth muscle cells (15). It is also approximately 500-fold more potent in inhibiting PDGF receptor tyrosine kinase activity than that of other tyrosine kinases (15). Application of Win41662 did not itself alter NMDA-activated currents, but it completely blocked the PDGF-induced depression of I p ( Fig. 2A). Therefore, PDGF receptor tyrosine kinase activity is required for expression of the PDGF response.
Activation of PLC-␥ Is Required for the PDGF-induced Depression of NMDA-evoked Currents-In the PDGF␣ receptor tyrosine 1021 is the minimal site required for PLC-␥ activation. Previously we demonstrated that this is also the minimal site required for the PDGF-induced depression of NMDA responses in Xenopus oocytes. However, it is not known whether PLC-␥ is required for the PDGF-mediated depression of NMDA-evoked currents in hippocampal neurons. Therefore, we employed a specific PLC-␥ inhibitor U73122. An inactive analogue U73343 was used as a control. We pretreated the cells with U73122 (10 M) or U73343 (10 M) for 15 min and included the same concentration of the appropriate compound in the patch pipettes to prevent recovery of PLC-␥ activity during the recording period. Treatment of the cells with the inactive analogue U73343 had no effect on the PDGF-induced depression of NMDA-evoked currents. However, the effect of PDGF on I p was completely blocked by U73122 (p Ͻ 0.01, two-way analysis of variance; Fig. 2B). This result provides strong evidence that PLC-␥ is indeed required for the PDGF-mediated depression of peak NMDA-activated currents in these neurons.
PDGF Releases Intracellular Ca 2ϩ -Activation of PLC-␥ increases the production of IP 3 , which induces intracellular Ca 2ϩ release in many cells. Therefore, we measured relative changes in intracellular Ca 2ϩ in cultured hippocampal neurons using fura-2 ratiometric imaging during applications of PDGF. Absolute values of intracellular Ca 2ϩ were not measured. Bath applications of PDGF significantly increased the intracellular Ca 2ϩ concentrations within 30 s of applying this growth factor (Fig. 3).
The potential source of this Ca 2ϩ signal was then examined. Heparin selectively inhibits the IP 3 -induced release of Ca 2ϩ from intracellular stores. For example it completely blocked the Ca 2ϩ release induced following flash photolysis of caged IP 3 , and it blocked the intracellular Ca 2ϩ signal recorded after exposure to PDGF or fibroblast growth factor (16). We anticipated that the PDGF-induced depression of NMDA-evoked currents would be blocked by heparin if an IP 3 -dependent release of Ca 2ϩ released were responsible for this effect. Inclusion of heparin (400 M) in the patch pipette did not alter NMDAevoked currents on its own, but the inhibitory effect of PDGF was eliminated (Fig. 3C). Unexpectedly, in the presence of heparin applications of PDFG increased I p by 14.6 Ϯ 1.9% (n ϭ 17, p Ͻ 0.05, Student's t test). This suggests that PDGF-induced potentiation of peak NMDA-evoked currents is masked by the prominent Ca 2ϩ -dependent depression.
Activation of PKA Reverses PDGF-induced Depression of NMDA Currents-Next we considered the potential target of the elevated intracellular Ca 2ϩ . This lead us to consider the possibility of a Ca 2ϩ -dependent activation of adenylyl cyclase together with a subsequent activation of PKA as a basis of the depressant response to PDGF. We considered this pathway for a number of reasons. For example, several Ca 2ϩ -dependent adenylyl cyclases (I and VIII) are present in hippocampal neurons (17), and levels of cAMP can be increased in hippocampal neurons through a Ca 2ϩ -dependent activation of adenylate cyclase (18,19). In non-neuronal cells the activity of PKA is stimulated by PDGF (20), and PKA has been implicated in the regulation of PDGF related signal transduction pathways (21)(22)(23)(24). In addition, the NMDA receptor in hippocampal neurons can be directly phosphorylated by PKA (25,26), and PKA has been reported to modulate recovery of NMDA receptors from desensitization (27).
We considered the possibility that PKA might modify the depression of NMDA channels mediated by PDGF receptors. In this respect, inclusion of PKI (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24) , a relatively selective PKA inhibitor in the patch pipettes, reduced the PDGF-induced depression of NMDA-activated currents (Fig. 4). This inhibitor had no effect upon the amplitude of peak currents in the absence of PDGF (Fig. 4). These results suggest that activation of PKA is required for the PDGF-induced depression. To examine the potential role of PKA further, we examined the effects of the bath application of the selective PKA inhibitor Rp-cAMPS. This inhibitor also failed to alter I p , but it did completely block the PDGF-induced depression (Fig. 4). We then examined the effects of IBMX a potent membrane-permeable phosphodiesterase inhibitor on the PDGF-induced depression of I p . We anticipated that this phosphodiesterase inhibitor should enhance any potential PDGF-stimulated cAMP-mediated signal. Bath applications of IBMX significantly enhanced the PDGF-induced depression of NMDA currents (48.37 Ϯ 2.58%, n ϭ 9, for IBMX plus PDGF versus 29.2 Ϯ 3.66%, n ϭ 9, for PDGF alone; p Ͻ 0.01 by Student's t test; Fig. 4). These results strongly support the conclusion that the PDGF-induced depression requires activation of PKA but also demonstrate that PKA does not itself regulate the amplitude of peak NMDA-evoked currents, a result consistent with previous observations (28).
Role of PKC in PDGF-mediated Regulation of NMDA Currents-We next considered the possibility that the PDGF-induced depression masked a PKC-dependent potentiation of I p . If this were the case we would anticipate that blockers of PKC might appear to enhance the PDGF-induced depression, whereas activators of PKC would reduce the depression. We tested the first possibility by including a selective PKC inhibitor chelerythrine in the patch pipette. Applications of chelerythrine significantly enhanced the relative PDGF-mediated depression of NMDA-evoked currents (Fig. 5A). The second possibility was examined by pretreating the cells with 4␤-PMA, a potent PKC activator. This phorbol ester significantly decreased the relative depression of NMDA-activated currents induced by PDGF (Fig. 5B), whereas the inactive phorbol 4␣-PMA was without effect.
The PKC-induced potentiation of peak NMDA-evoked currents in these neurons is mediated via activation of the nonreceptor tyrosine kinase Src (1). We predicted that inhibition of Src kinases should block any potential PKC-dependent PDGFinduced potentiation of I p , leaving only the PDGF-induced depression. Therefore, we included in the pipette anti-cst1 (10 g/ml), an antibody that selectively inhibits the Src family of kinases (29,30). An IgG fraction (10 g/ml) was included in another series of recordings as a control (Fig. 5C). We also examined the effects of a Src (40 -58) , a peptide fragment that selectively inhibits Src function (29,30). A scrambled sequence of this peptide was used as the control for these experiments (Fig. 5D). Unexpectedly, both inhibitors of tyrosine kinase Src completely blocked the PDGF-induced depression (Fig. 5, C and  D). These results suggested that contrary to our initial hypothesis, the activity of Src is required for the PDGF-induced depression of I p .
PKA Disrupts Src-mediated Up-regulation of NMDA Currents-Tyrosine kinase Src and NMDA receptors co-immunoprecipitate suggesting a close physical proximity of these two proteins at the synapse. Furthermore, the activation of Src enhances NMDA channel activity (29) in inside-out patches and enhances NMDA-mediated synaptic currents in cultured hippocampal neurons (1). Because both PKA and Src activity appeared to be required for the PDGF-induced depression of NMDA-activated currents, we considered the possibility that PKA modulates the Src-induced enhancement of NMDA receptor function. To test this hypothesis we recorded NMDA receptor single-channel currents using inside-out patches. We employed a Src-activating peptide EPQ(pY)EEIPIA (31), which stimulates endogenous Src and enhances NMDA receptor function in such patches (1,29). The nonphosphorylated form of the peptide EPQYEEIPIA was applied as a control for the activator peptide. In addition, we used the cPKA to simulate activation of endogenous PKA and heat-denatured cPKA was used as a control for the catalytic subunit (1,29,31). The results are summarized in Table I. Applications of EPQ(pY)EEIPIA (1 mM) to the cytoplasmic side of the membrane significantly increased the open probability (P o ) of NMDA channels (Fig. 6, A, C, and  E). The mean open time (t o ) was also significantly increased by EPQ(pY)EEIPIA (Table I). The control peptide EPQYEEIPIA (1 mM) did not alter either of these parameters (Table I). In contrast to the active peptide, application of cPKA (100 units/ ml) had no significant effect on NMDA channel activity. However, cPKA blocked the effects of EPQ(pY)EEIPIA on NMDA single channel activity (Fig. 6, B, D, and E). Applications of heat-denatured cPKA failed to block the increase in P o and t o induced by the active peptide EPQ(pY)EEIPIA (Table I). These results demonstrate that activation of PKA can inhibit the stimulatory actions of endogenous Src upon NMDA channels in these patches.
PDGF-induced Depression of Miniature Synaptic Currents-We then examined the effects of PDGF receptor activation on the NMDA receptor-mEPSCs in cultured hippocampal neurons (Fig. 7). Applications of this growth factor depressed the NMDA receptor-mEPSCs, and the depression could be enhanced by including chelerythrine in the patch pipette, confirming that synaptically located NMDA receptors are likely simultaneously enhanced by PDGF through a PKC-dependent mechanism (Fig. 7, B and D). In contrast, the depression was eliminated by including the PKA inhibitory peptide in the patch pipette and replaced by a small PDGF-induced enhancement (Fig. 7, C and D).

DISCUSSION
The primary effect of applications of low concentrations of PDGF on NMDA-evoked currents was to depress I p . This action was mediated via PDGF receptor autophosphorylation as the selective PDGF receptor kinase inhibitor Win41662 blocked it. Furthermore, this inhibition likely takes place through autophosphorylation of Tyr 1021 (3) and the subsequent activation of PLC-␥ because the depression was also blocked by the PLC-␥ inhibitor U73122. Stimulation of PLC-␥ through activation of PDGF receptors leads to the hydrolysis of phosphoinositols, the production of IP 3 and diacylglycerol (4). The production of IP 3 following activation of PDGF receptors is known to release intracellular Ca 2ϩ (32), and a similar effect was observed by us in cultured hippocampal neurons. Consistent with this observation the PDGF-induced depression of I p in isolated neurons was blocked by including heparin in the patch pipette. The rise in intracellular Ca 2ϩ could potentially activate PKA through the stimulation of Ca 2ϩ -dependent adenylate cyclases (33,34). Indeed, PDGF receptor activation is associated with activation of PKA (20,35). This inhibitory signaling pathway is summarized schematically in Fig. 8.
PDGF Receptor Activation and PKC-In a previous study we demonstrated that G-protein coupled receptors, presumably acting through PLC-␥, enhance I p through a sequential activation of PKC and c-Src (1). Thus, stimulation of PLC-␥ by PDGF-R would also be anticipated to potentiate I p via a parallel excitatory pathway (Fig. 8). Whether I p was depressed or enhanced would therefore depend upon the relative strength of activation of inhibitory and excitatory signaling pathways. Our observations that intracellular applications of the selective PKC inhibitor chelerythrine enhanced the PDGF-induced depression, whereas pretreatment of cells with the active phorbol ester ␤-PMA reduced this depression, are consistent with a block or potentiation, respectively, of the excitatory pathway.
We previously reported that an inhibitor of the phosphatases PP1/PP2A blocked the PDGF-induced inhibition of I p (3). These inhibitors also dramatically enhance the phorbol-ester induced (PKC-and Src-dependent) potentiation of I p in these neurons (1). Therefore, inhibition of these phosphatases may have simply favored the PKC-dependent and facilitatory pathway over the PKA-dependent inhibitory pathway (an apparent block of the PDGF-induced inhibition of peak NMDA currents).
Role of PKA in the Regulation of NMDA Receptors-NMDA receptors can be directly phosphorylated by PKA (25,26). The physiological significance of the PKA-mediated phosphorylation of NMDA channels is nevertheless poorly understood. For example, previous experiments found that intracellular applications of cAMP (36) had little effect on peak NMDA-evoked currents in cultured hippocampal neurons. Applications of cPKA were also reported to have no effect on NMDA single channel activity in outside-out patches taken from cultured neurons (28), and we detected no effect on NMDA channel activity in inside-out patches in the present study. However, the PDGF-induced depression of I p we observed was blocked by the inhibitory PKA peptide PKI and by the potent and selective PKA inhibitor Rp-cAMPS. Applications of the phosphodiesterase inhibitor IBMX enhanced this depression, also suggesting that activation of PKA is required for this effect even though PKA does not seem to directly modulate the activity of NMDA channels.  In the presence of 4␤-PMA but not 4␣-PMA, the inhibitory effect of PDGF on I p currents was diminished. C, inclusion of the Src antibody, anti-cst1 (10 g/ml) in the pipette completely blocked the depression of NMDA-evoked currents by PDGF (n ϭ 11, squares, p Ͻ 0.05, Student's t test); whist control Ig-G antibody (10 g/ml) had no effect on this depression (n ϭ 9, circles, p Ͻ 0.01, Student's t test). D, effects of PDGF on NMDA-activated currents were blocked by the inclusion of the unique domain peptide fragment Src (40 -58) (

Role of c-Src in the PDGF-induced Depression-Intracellular
applications of either the c-Src selective inhibitory peptide (Src40 -58) or the Scr family inhibitory antibody (anti-cst1) but not appropriate controls blocked the PDGF-induced depression of I p . These results were surprising because we anticipated that inhibitors of c-Src would block the excitatory pathway and shift the effects of PDFG toward a predominance of the inhibitory one (Fig. 8). One possible explanation is that the inhibitory pathway also requires activation of c-Src. For example, autophosphorylation of the PDGF-R at Tyr 579 (and Tyr 581 ) is known to stimulate binding of c-Src to the receptor through an interaction of this phosphotyrosine with the Src homology 2 domain of c-Src (7). This interaction leads to activation of c-Src, and in turn Src phosphorylates residue Tyr 935 of the PDGF receptor (37). This sequence of events is thought to activate the MAPK kinase cascade by mechanisms that are not well understood. However, we previously demonstrated that only Tyr 1021 is required for the PDGF-induced depression of NMDA-evoked currents (3) and that the depression is also blocked by the PLC-␥ inhibitor Win41662; therefore, a Src-dependent phosphorylation of the PDGF receptor is unlikely to be required.
A more plausible explanation is that the PDGF-R receptor modulates the ability of c-Src to directly enhance NMDA channel activity. We suggest that blockers of Src likely decrease an already tonic potentiation mediated by Src (1) and occlude the effects of PDGF. This is strongly supported by our observation that cPKA inhibits the ability of the Src activator peptide to enhance NMDA channel activity and by our observation that Binding of PDGF to its receptor activates PLC-␥, which in turn generates two second messengers, diacylglycerol (DAG) and IP 3 . Diacylglycerol stimulates the activity of PKC, which up-regulates NMDA receptor (NMDAR) function via the activity of channelassociated protein, Src. IP 3 induces intracellular Ca 2ϩ release. The released Ca 2ϩ increases the activity of adenylyl cyclase (AC), which hydrolyzes ATP to cAMP. cAMP further activates PKA. PKA disrupts the Src-mediated tonic enhancement of NMDA channel function and plays a preponderate role. The overall effect of PDGF is therefore inhibitory.
inhibitors of c-Src block the inhibitory modulation of I p by PDGF. Therefore, the PDGF-induced depression of I p is exerted indirectly at least in part through activation of PKA and inhibition of the effects of c-Src on NMDA channels (Fig. 8).
One possible mechanism is that PKA-induced phosphorylation of the NMDA channel alters the configuration of the channel and renders it insensitive to the enhancement by c-Src. Alternatively there is evidence that c-Src is phosphorylated by PKA at serine 17 (38,39), and this phosphorylation is known to decrease the hydrophobicity of c-Src and enhance the release of c-Src from the cytoplasmic membrane to the cytosol (21,40,41). PKA may therefore increase translocation of c-Src from the cell membrane to the cytosol, thus decreasing its ability to phosphorylate a relevant substrate and modulate NMDA channels. This possibility is consistent with co-immunoprecipitation of c-Src and NMDA receptor proteins from the postsynaptic density of central synapses (29,42). This latter mechanism is also consistent with serine-threonine phosphorylation and altered subcellular distribution of the NR1 subunit (43) or perhaps with changes in the trafficking of the receptor. For example, insulin can act to alter the number of ␥-aminobutyric acid receptors expressed at the membrane surface of hippocampal neurons (44).
G-protein-coupled receptors up-regulate the activity of NMDA channels in CA1 hippocampal neurons (1) as well as stimulate MAPK signaling (9,(45)(46)(47). In contrast, PDFG-R-dependent activation of MAPK signaling is inhibited by the stimulation of PKA in several different types of smooth muscle cells (22)(23)(24)48). We have shown that an increase in intracellular calcium and subsequent activation of PKA also serves as a switch to limit c-Src-mediated enhancement of NMDA channels in CA1 neurons. Our results demonstrate that the PDGF-R can control the activity of NMDA receptors through two functionally opposing signaling pathways. These pathways apparently converge at the level of c-Src with the PKA providing a feed-forward inhibition of c-Src.