Abstract
Growth factor receptor tyrosine kinase (RTK)-activated signaling pathways are well established regulators of neuronal growth and development, but whether these signals provide mechanisms for acute modulation of neuronal activity is just beginning to be addressed. We show in pheochromocytoma (PC12) cells that acute application of ligands for both endogenous RTKs [trkA, basic FGF (bFGF) receptor, and epidermal growth factor (EGF) receptor] and ectopically expressed platelet-derived growth factor (PDGF) receptors rapidly inhibits whole-cell sodium channel currents, coincident with a hyperpolarizing shift in the voltage dependence of inactivation. Sodium channel inhibition by trkA and PDGF receptors is mutually occlusive, suggestive of a common signal transduction mechanism. Furthermore, specific inhibitors for trkA and PDGF RTK activities abrogate sodium channel inhibition in response to NGF and PDGF, respectively, showing that the intrinsic RTK activity of these receptors is necessary for sodium channel inhibition. Use of PDGF receptor mutants deficient for specific signaling activities demonstrated that this inhibition is dependent on RTK interaction with Src but not with other RTK-associated signaling molecules. Inhibition was also compromised in cells expressing dominant-negative Ras. These results suggest a possible mechanism for acute physiological actions of RTKs, and they indicate regulatory functions for Ras and Src that may complement the roles of these signaling proteins in long-term neuronal regulation.
Growth factors constitute a family of polypeptides essential for the establishment and maintenance of the mature nervous system. Expressed throughout the CNS and PNS, growth factors act by binding to their cognate receptor tyrosine kinases (RTKs), which autophosphorylate and subsequently activate a complex of signaling proteins including the well characterized Ras and mitogen-activated kinase (MAPK) pathway, phosphatidylinositol 3-kinase (PI3-K), phospholipase Cγ (PLCγ), and the Src family of nonreceptor tyrosine kinases (for review, see Panayotou and Waterfield, 1993;Johnson and Vaillancourt, 1994; Kaplan and Stephens, 1994;Szeberényi and Erhardt, 1994; van der Geer et al., 1994). The variety of neuronal responses to growth factors, including proliferation and migration, establishment and maintenance of synaptic connections, and survival of many mature neurons, ultimately depends on the activation of specific components comprising this signaling network (for review, see Korsching, 1993; Snider, 1994; van der Geer et al., 1994; Bothwell, 1995; Lewin and Barde, 1996).
In addition to the wealth of information on long-term cellular effects, there is increasing evidence that growth factors and their signaling pathways play pivotal roles in the acute regulation of the mature nervous system (for review, see Thoenen, 1995). In hippocampal neurons, depolarizing stimuli and stimulation protocols that induce long-term potentiation (LTP) increase neurotrophin mRNA levels (Lu et al., 1991;Patterson et al., 1992). In turn, growth factors can modulate synaptic activity in hippocampal neurons and developing Xenopusneuromuscular junctions (Terlau and Seifert, 1989; Abe et al., 1991;Kang and Schuman, 1995), as well as modulate short-term potentiation and LTP (Lohof et al., 1993; Stoop and Poo, 1996). These studies suggest that ion channel modulation may be an important component of acute growth factor actions, and they are complemented by work showing growth factors can acutely modulate the high-voltage-activated Ca2+ channel and Ca2+-activated K+ channels (Wildering et al., 1995; Holm et al., 1997). However, the signaling paths underlying these physiological events are undetermined, and although both NMDA receptors and the human Kv1.5 K+ channel associate with the nonreceptor tyrosine kinase Src (Holmes et al., 1996; Yu et al., 1997), it is not yet known whether this interaction underlies a growth factor-initiated action. Furthermore, studies suggesting that the activation of the MAPK cascade is essential for events associated with long-term facilitation in Aplysia neurons have not ascertained whether growth factors are vital in initiating this response (Bailey et al., 1997;Martin et al., 1997). Thus, there is much to be determined concerning how growth factors physiologically modulate the mature nervous system and which signaling pathways are essential to this modulation.
Past efforts to examine the signaling mechanisms that mediate the effects of chronic growth factor exposure in both the PNS and CNS have used the well characterized rat pheochromocytoma (PC12) model neuronal cell line (for review, see Chao, 1992; Keegan and Halegoua, 1993;Marshall, 1995). Application of NGF causes PC12 cells to differentiate into a neuronal-like phenotype characterized by numerous morphological and physiological changes, including the persistent activation of the Ras and MAPK pathway, cessation of cellular division, neurite extension, the induction of immediate early and neuronal-specific genes, and an induction of electrical excitability because of an increase in ion channel expression (Greene and Tischler, 1982). Via the analysis of PC12 cell lines expressing mutant receptors or intracellular signaling molecules, the signaling mechanisms that mediate these long-term changes have been well characterized (Rausch et al., 1990; Szeberényi et al., 1990; Ginty et al., 1992; Traverse et al., 1992; Fanger et al., 1993, 1997; Cavalié et al., 1994;Cowley et al., 1994; Loeb et al., 1994; Stephens et al., 1994;Vaillancourt et al., 1995; Pollock and Rane, 1996; Hilborn et al., 1997). For example, analyses of PC12 cell lines expressing mutated platelet-derived growth factor (PDGF) receptors deficient for activating specific signaling pathways have determined that both Src and PLCγ are required for growth factor-induced neurite extension (Vaillancourt et al., 1995). An analysis of these same cell lines, along with PC12 cells expressing the N17 Ras dominant-negative mutant, demonstrated that although growth factor-induced upregulation of brain type II/IIA Na+ channels is independent of Ras activation, the activation of Src is important for establishing the full functional expression of these channels (Fanger et al., 1993,1997; Pollock and Rane, 1996).
In this report, we use this system to show that growth factors can acutely inhibit voltage-gated Na+ currents in undifferentiated and differentiated PC12 cells and that this inhibition is dependent on RTK coupling to Src and Ras. This inhibition does not share the mechanistic properties of Na+ current inhibition induced by signaling molecules such as the cAMP-dependent protein kinase (PKA) or protein kinase C (PKC), further suggesting that this modulation relies on intracellular mechanisms distinguishable from pathways such as those driven by receptors of the seven transmembrane spanning region superfamily. Thus we show that signals mediated by growth factors and their RTKs are capable of regulating an ion channel fundamental to neuronal excitation and that this regulation occurs through Src and Ras. Although Src is extensively expressed in the mature PNS and CNS and contributes to long-term growth factor effects, its physiological function has remained elusive. Our results suggest that acute regulation of neuronal Na+ channels may be one such function.
MATERIALS AND METHODS
Tissue culture. Stock PC12 cell cultures were grown on rat tail collagen-coated Petri dishes and maintained in DMEM. For cells ectopically expressing N17 Ras (for details of cell line generation, see Szeberényi et al., 1990), the media were supplemented with 5% fetal calf serum and 10% horse serum. For PC12 cells expressing either the human form of the βPDGF [wild-type PDGF (wtPDGF)] receptor or mutant PDGF receptor (for details of cell line generation, see Vaillancourt et al., 1995), the media were supplemented with 2.5% fetal calf serum (Sigma, St. Louis, MO) and PDGF-deficient 12.5% plasma-derived horse serum (Sigma) to obviate differentiation. All cell lines were kept under G418 selection and maintained in 25 U/ml penicillin and 25 μg/ml streptomycin. In preparation for electrophysiology, cells were seeded onto 35 mm collagen-coated dishes and allowed to grow for 1 d in standard culture medium. Then, fresh medium with either NGF (100 ng/ml), PDGF (20 ng/ml), or basic FGF (bFGF) (50 ng/ml) was applied. Medium and growth factor were replaced every other day.
Electrophysiology. Conventional patch-clamp techniques and computer-assisted data acquisition and analysis (Pulse and PulseFit software; Instrutech Corporation, Great Neck, NY) (Pollock and Rane, 1996) were used to study whole-cell Na+ and Ca2+ channel currents. All current records were leak subtracted with a standard P/N procedure and filtered at 5 kHz before storage on disk for off-line analysis. For recording Na+ currents, the bath solution was 138 mm NaCl, 9 mm KCl, 1 mmCaCl2, 1 mm MgCl2, 10 mm TEA-Cl, 200 μm CdCl2, and 10 mm HEPES, pH 7.3. For recording voltage-gated Ca2+ channel currents, the bath solution consisted of 135 mm TEA-Cl, 4 mm KCl, 10 mmBaCl2, 1 mm MgCl2, 5 mm glucose, 10 mm HEPES, and 1 μmtetrodotoxin, pH 7.3. In all cases, the patch pipette solution was 150 mm CsCl, 2 mm MgATP, 0.5 mm GTP, 2 mm BAPTA, and 10 mm HEPES, pH 7.3. Whole-cell capacitive transients, elicited by 20 mV depolarizing steps, were compensated with the analog compensation circuitry of the patch-clamp amplifier. Whole-cell capacitance, a measure of membrane area, was read directly from the compensation control dial of the amplifier.
We waited from 2 to 5 min after establishing whole-cell patch-clamp access before beginning experimental data collection. During this time, the Na+ I–V relationship was assessed by holding the cell at −90 mV and by evoking command steps from −60 to 60 mV in 10 mV increments at 1.5–3 sec intervals. Two changes were observed when this protocol was repetitively administered. First, as has been noted in a number of studies on voltage-gated currents, the Na+ I–V relationship shifted to the left by 5–10 mV during the first 30–90 sec of recording. We also observed that in addition to this shift, current amplitudes increased at any given voltage during the first 1–3 min of recording, and then they stabilized. Growth factor inhibition of Na+ current was assessed at the command voltage giving maximal current (as shown by the I–V protocol) only after these spontaneous changes in Na+ current were complete. Any cell that showed continuous variations in current of >5% at this voltage was rejected. To test for significant differences in growth factor inhibition of Na+ current between experimental groups, we used a two-tailed, nonpaired, t test (at p < 0.05). All experiments were performed at 22–25°C.
Growth factor and drug application. Growth factors dissolved in the bath solution were acutely applied to cells via pressure ejection from blunt-tipped, fire-polished micropipettes positioned ∼5–10 μm from the cell. For RTK and Src inhibitor studies, tyrphostin AG9, tyrphostin AG879, and PP1 were dissolved in 70% ethanol, 70% methanol, and DMSO, respectively, before dilution to their appropriate concentrations in intracellular solution (for details of inhibitor specificity, see Bilder et al., 1991; Levitzki and Gilon, 1991; Ohmichi et al., 1993; Levitzki and Gazit, 1995; Hanke et al., 1996). After whole-cell access, the intracellular solution containing the inhibitor was allowed to diffuse into the cell for ∼3–5 min before recording. For all experiments, inclusion of inhibitors did not affect basal Na+ current densities. Furthermore, recordings with intracellular solutions containing only ethanol, methanol, or DMSO did not alter basal Na+ channel current densities or acute responses to growth factors.
RESULTS
Initial characterization of growth factor-induced Na+ channel current inhibition in PC12 cells expressing wild-type PDGF receptors
In PC12 cells, cellular differentiation is mediated by NGF binding to its cognate RTK. The wtPDGF receptor, when expressed in PC12 cells, mediates neuronal differentiation in a manner characteristic of NGF; both growth factors induce the persistent activation of the Ras and MAPK pathway and of PLCγ, induce the expression of the early gene c-fos, and upregulate Na+ channels and other neuronal-specific mRNAs (Heasley and Johnson, 1992; Fanger et al., 1993, 1997). Furthermore, observations with mutant PDGF receptors demonstrate that both Src and PLCγ are important in regulating morphological differentiation (Vaillancourt et al., 1995). Thus, wtPDGF and mutant PDGF receptors serve as reasonable models for neurotrophin receptors and the signaling pathways underlying their actions.
We first confirmed the robustness of signaling in PC12 cells ectopically expressing the wtPDGF receptor by showing that chronic application of PDGF (20 ng/ml) could induce both morphological and physiological differentiation. In agreement with previous observations (Heasley and Johnson, 1992; Vaillancourt et al., 1995), the morphological differentiation elicited by chronic application of PDGF for a week, specifically neurite extension, was comparable with that in cells treated with NGF (100 ng/ml) for the same period of time (data not show). To determine whether morphological differentiation was accompanied by ion channel upregulation, we used whole-cell patch-clamp electrophysiology to examine whether chronic application of PDGF to PC12 cells expressing wtPDGF receptors could increase functional Na+ channel expression. Peak Na+currents of undifferentiated and differentiated PC12 cells were normalized to whole-cell capacitance (as an indirect measure of cell size) and used as an estimation of the Na+ channel current density in individual cells. In accordance with previous observations (Fanger et al., 1995, 1997), PDGF elicited an approximately threefold increase in the functional expression of Na+ channels, elevating Na+current densities from 32.7 ± 5.9 pA/pF (n = 14) in undifferentiated cells to 90.3 ± 8.4 pA/pF (n= 17) in differentiated cells, which paralleled an increase in Na+ current densities observed in cells that had been differentiated with NGF (85.9 ± 12.5 pA/pF;n = 22). As an additional control, we compared Ca2+ current densities in both undifferentiated cells and cells differentiated with PDGF and found that PDGF elicited an increase in Ca2+ current densities comparable with increases seen in positive controls treated with NGF (data not shown). These results, combined with previous observations from other laboratories (Heasley and Johnson, 1992; Fanger et al., 1995, 1997;Vaillancourt et al., 1995), demonstrated that PC12 cells differentiated with PDGF are both morphologically and physiologically comparable with PC12 cells differentiated with NGF and thus serve as an excellent model for examining the signaling mechanisms underlying the physiology of fully differentiated neurons. Furthermore, the similarities in chronic effects of wtPDGF and NGF receptor activation suggest similarities in their signaling mechanisms, making the wtPDGF receptor a valuable surrogate for analyzing neurotrophin receptor signaling.
To analyze the possibility that NGF could acutely regulate Na+ channels, we differentiated PC12 cells with PDGF (20 ng/ml) over a period of 4–6 d during which Na+current densities are upregulated. Cells differentiated with PDGF were stepped to voltages that induced peak Na+ current amplitudes (typically between −10 and 0 mV) and then were subjected to acute application of NGF (100 ng/ml). Figure1, A and C, shows that NGF caused a rapid decrease in Na+ current amplitude (36.5 ± 3.0%; n = 13) that was partially recoverable to 80.0 ± 4.0% (n = 6) of the total current observed before NGF application. NGF-induced inhibition was not associated with any change in the voltage dependence or time course of the currents. Because future experiments would entail analyzing Na+ channel currents in PC12 cells expressing mutant receptors, it was essential to determine whether PDGF could acutely modulate Na+ currents in a manner similar to NGF. Figure 1, B and C, shows that application of PDGF (20 ng/ml) to cells differentiated with NGF (100 ng/ml) also caused an acute decrease in Na+ current amplitudes (54.0 ± 3.8%; n = 21) that was partially reversible to 71.5 ± 7.3% (n = 12) of the total current observed before PDGF application. Again, this was not the result of a shift in the voltage dependence of activation or current kinetics. A comparison of cells acutely treated with NGF and PDGF showed that, in general, the time to maximal PDGF-induced Na+ current inhibition (160 ± 15 sec;n = 21) was slow relative to the NGF time course (82 ± 11 sec; n = 12). Furthermore, the magnitude of inhibition observed in cells acutely treated with PDGF was generally greater than was the inhibition caused by NGF.
Growth factors inhibit Na+currents. A, Representative time course of NGF-induced inhibition of whole-cell Na+ currents in PC12 cells expressing wtPDGF receptors. PC12 cells were differentiated with 20 ng/ml PDGF for a period of 4–6 d. Differentiated cells were then held at −90 mV and given sequential 10 mV steps from −60 to 60 mV in 1.5 sec intervals to determine the voltage of peak current amplitude. Command steps to induce peak current amplitudes (open diamonds) were then given every 5 sec before, during, and after acute application of 100 ng/ml NGF (horizontal bar); *selected current traces shown in theinset. B, Representative time course of 20 ng/ml PDGF-induced inhibition of whole-cell Na+currents in PC12 cells differentiated by chronic treatment with 100 ng/ml NGF. Electrophysiological protocols are identical to those described in A; * indicates selected currenttraces that are shown in the inset.C, Cumulative data for the inhibition and recovery of Na+ currents in response to acute application of either 100 ng/ml NGF or 20 ng/ml PDGF. Whole-cell Na+ currents recorded during or after growth factor treatment were normalized to whole-cell currents recorded before growth factor applications. Error bars represent the mean ± SEM of normalized Na+ currents for the indicated growth factor; †The current after recovery from growth factor inhibition is statistically significant from the fully inhibited current. The values of n for experimental groups are, from left to right, 13, 6, 21, and 12.
We also examined whether the acute responses to NGF and PDGF were the result of signaling mechanisms specific to the differentiated phenotype. Undifferentiated cells expressing wtPDGF receptors were exposed to acute application of NGF (100 ng/ml) or PDGF (20 ng/ml). As seen in differentiated cells, NGF elicited a decrease in Na+ current levels (40.6 ± 4.3%;n = 18) with an onset of 96 ± 13 sec (n = 15) and a rapid partial recovery to 90.0 ± 4.8% (n = 7) of the total current observed before NGF application. Likewise, PDGF induced a decrease in Na+ current (42.4 ± 3.4%; n = 12) with an onset of 154 ± 17 sec (n = 12) which was partially reversible to 80.4 ± 6.1% (n = 4) of the total current observed before PDGF application. These results suggest that the signaling pathways required for growth factor-induced Na+ current inhibition were present in both undifferentiated and differentiated cells expressing wtPDGF receptors and thus are not dependent on physiological events necessary for neuronal differentiation. In subsequent studies, however, we limited our experiments to the acute regulation involved in differentiated PC12 cells because the presence of several markers specific for neuronal differentiation was necessary to ensure that the effects of mutant PDGF receptors were specific to the inhibitory response (see below).
To test whether the acute effects of growth factors were dependent on the activation of their respective RTKs, we took advantage of a family of protein tyrosine kinase inhibitors, the tyrphostins (for review, seeLevitzki and Gilon, 1991; Levitzki and Gazit, 1995), specifically tyrphostin AG9, which inhibits PDGF receptor autophosphorylation and the ability of the receptor to phosphorylate intracellular signaling molecules (Bilder et al., 1991), and AG879, which prevents the activation of the NGF receptor and neurite extension in PC12 cells (Ohmichi et al., 1993). For our experiments, individual cells were loaded with the appropriate inhibitor by including either AG9 or AG879 in the intracellular recording solution and by allowing the solution to diffuse into the cell for 3–5 min before recording Na+ current responses to growth factor application (Fig. 2). We found that inclusion of AG9 (10 μm) in the intracellular solution dramatically reduced PDGF inhibition of Na+ currents (10.4 ± 4.8%; n = 13), whereas the acute effects of NGF remained unaffected (38.9 ± 5.9%; n = 9). The converse was true in differentiated cells loaded with AG897 (25 μm). The inhibition of Na+ current caused by NGF (13.6 ± 4.6%; n = 11) was significantly reduced, whereas the inhibition caused by PDGF (29.2 ± 2.7%; n = 11) was only slightly affected when compared with the inhibition observed in control cells. These results strongly suggest that growth factors inhibit Na+channels via the activation of their cognate RTKs.
Inhibition of Na+currents by NGF (100 ng/ml) or PDGF (20 ng/ml) is reduced by inhibitors specific for the wtPDGF and trkA receptors. A, Representative time courses of NGF- and PDGF-induced inhibition of whole-cell Na+ currents in differentiated cells loaded with either 25 μm AG879 (open diamonds) or 10 μm AG9 (closed diamonds). Differentiation protocols and electrophysiological protocols for determining voltages of peak current amplitudes are identical to those described in Figure 1. After achieving whole-cell access, the intracellular solution containing the indicated inhibitor was allowed to diffuse into the cell for a period of 3–5 min. Peak Na+ current amplitudes were subsequently recorded before and during acute application of the indicated growth factor (horizontal bar). For comparison, currents were normalized to the total whole-cell current recorded before growth factor application. B, Representative currenttraces selected from A.Traces depict total whole-cell current and the total current remaining after the indicated growth factor in the presence of the indicated inhibitor reached its maximum effect. C, Cumulative data from A. Error bars represent mean ± SEM of peak Na+ currents normalized to the total current recorded before indicated growth factor application; † indicates a statistically significant difference from cells that were not subjected to RTK inhibitors. The values of n for experimental groups are, from left toright, 13, 11, 9, 10, 11, and 13.
Finally, because there is significant conservation of signaling mechanisms among the growth factor RTKs, especially in early events (Chao, 1992; Marshall, 1995), we asked whether other growth factors were capable of acutely modulating Na+ channels (Fig. 3). In addition to the NGF RTK trkA, PC12 cells express RTKs for epidermal growth factor (EGF) and bFGF. Therefore, either EGF (100 ng/ml) or bFGF (50 ng/ml) was acutely applied to cells that had been differentiated with NGF. Like PDGF and NGF, acute application of EGF inhibited Na+ currents by 44.4 ± 3.6% (n = 7) with a time to maximum inhibition of ∼50–75 sec and a recovery to 80.2 ± 6.5% (n = 6) of the total current observed before growth factor application. Likewise, bFGF initiated a 38.7 ± 3.1% (n = 8) decrease in Na+ current with a time to maximum inhibition of ∼15–30 sec and a recovery to 86.5 ± 3.9% of the total current observed before growth factor application. Combined, these results demonstrate that the acute effects of NGF and PDGF on Na+ channels are shared by other growth factors and that the magnitude of current inhibition and recovery is remarkably preserved throughout the growth factor receptor family, although the time to reach maximum inhibition was dependent on the individual growth factor. The results again suggest that Na+ current inhibition is most likely mediated via a signaling pathway that is conserved among growth factor RTKs.
Activation of either EGF or bFGF receptors inhibits Na+ current. A,B, Results are shown from two PDGF receptor-expressing PC12 cells that had been differentiated by chronic NGF treatment. Peak Na+ current amplitudes (open diamonds) are plotted as a function of time, showing inhibition in response to EGF (A) or bFGF (B) application (horizontal bars). Currents were evoked every 5 sec; *Selection of raw datatraces shown in the insets.
Growth factors affect steady-state inactivation but not steady-state activation of Na+ channels
Similar to the acute effects observed with NGF and PDGF on PC12 Na+ current, phosphorylation of rat brain Na+ channels by either PKA or PKC results in a reduction of Na+ current amplitude; however, changes in kinetic behavior differ among individual inhibitory agents, suggesting different mechanisms of action. For example, application of purified PKA to excised membrane patches reduces peak Na+ current amplitudes without any effect on time course or the voltage dependence of activation or inactivation (Li et al., 1992), whereas phosphorylation of Na+ channels by PKC slows inactivation in addition to reducing peak current amplitude (Numann et al., 1994).
Current–voltage relationships for morphologically differentiated PC12 cells acutely treated with either NGF or PDGF indicated that the growth factor-induced reduction in current amplitude was not the consequence of a shift in the voltage dependence of channel activation. For differentiated PC12 cells held at −90 mV, the voltage threshold for activation of both control and growth factor-inhibited whole-cell Na+ currents is approximately −50 mV, and the currents reach a maximum amplitude between −10 and 0 mV (Fig.4A). Analysis of the steady-state inactivation of control cells and cells acutely treated with growth factors, however, demonstrated that acute application of NGF and PDGF shifted the voltage dependence of inactivation by −10.0 ± 1.3 mV (n = 10) and −8.5 ± 0.8 mV (n = 6), respectively (Fig. 4B). This suggests that the acute inhibition caused by NGF and PDGF is a direct result of a negative shift in the voltage dependence of Na+ channel inactivation. The similarity in NGF and PDGF effects on the Na+ current steady-state inactivation relationship also suggests that both receptors share a common mode of signaling in modulating Na+current.
Current–voltage and steady-state inactivation plots for Na+ currents inhibited by growth factor receptor activation. A, Results are shown from two wtPDGF receptor-expressing cells that had been differentiated by chronic application of 20 ng/ml PDGF (left) or 100 ng/ml NGF (right). Cells were held at −90 mV, and command steps from −60 to 60 mV in 10 mV intervals were given every 5 sec before growth factor application (closed diamonds) and after the acute application of the indicated growth factor had reached its maximum effect (open diamonds). B, Results are shown from two wtPDGF receptor-expressing cells that had been differentiated by chronic PDGF (left) or NGF (right) treatment. Cells were sequentially held at each of the indicated potentials for 3 sec before a command step to −10 mV to evoke Na+ current. The protocol was then repeated in the presence of acutely applied NGF (100 ng/ml) or PDGF (20 ng/ml). Peak current amplitudes were normalized to the current obtained from the −110 mV hold. Fitted curves are of the form:I/Imax = 1/1 + exp(Vhold −V1/2/k), whereVhold is the holding voltage from which the command step was evoked, V1/2 is the voltage corresponding to half-inactivation of the current, and kis the slope constant. The mean ± SEM shift inV1/2 in response to NGF was 10.0 ± 1.3 mV (n = 10 cells). The mean ± SEM shift inV1/2 in response to PDGF was 8.5 ± 0.8 mV (n = 6 cells).
Analysis of acute effects of growth factors on PC12 cells expressing mutant PDGF receptors
The use of exogenously expressed PDGF receptors in PC12 cells assumes that the physiological responses induced by these recep- tors are mediated by signaling cascades shared by endogenous RTKs that mediate similar, physiological events. One method of testing this possibility is to determine whether the application of one growth factor can occlude, or “mask,” the response of the second growth factor. We used this approach to determine whether NGF and PDGF modulate Na+ channel currents via similar, if not identical, signaling pathways. NGF was acutely applied to cells (Fig.5A). As expected, NGF induced an inhibition of Na+ current (29.2 ± 2.4%;n = 13), but subsequent application of PDGF after NGF reached its maximum effect caused only minor inhibition of Na+ current (10.5 ± 3.0%; n = 13). Similarly, acute application of PDGF to differentiated cells (Fig.5B) caused an inhibition of Na+ current (30.4 ± 3.5%; n = 10), whereas the inhibition of Na+ current caused by subsequent NGF treatment (7.7 ± 1.6%; n = 10) was dramatically reduced relative to inhibition in response to NGF alone. In conjunction with previous observations that signals from PDGF receptors serve as models for NGF-induced signaling in wild-type PC12 cells, our results strongly favor a common signaling pathway for growth factor-mediated Na+ channel modulation.
NGF- and PDGF-induced Na+ current inhibition are both mutually occlusive, suggesting a common signaling pathway. A, Na+ currents (open diamonds) were recorded from differentiated PC12 cells expressing the wtPDGF receptor. During whole-cell recording sessions, NGF (100 ng/ml) was applied (upper bar). When the NGF response reached a maximum, PDGF (20 ng/ml) was then applied (lower bar).Inset, Representative Na+ currenttraces depicting total current, the current remaining after subsequent application of NGF, and the current remaining after application of PDGF. B, Na+ currents were recorded from differentiated PC12 cells expressing the wtPDGF receptor. During whole-cell recording sessions, PDGF (20 ng/ml) was applied (upper bar). When the PDGF response reached a maximum, NGF (100 ng/ml) was then applied (lower bar). Inset, Representative Na+ current traces depicting total current, the current remaining after subsequent application of PDGF, and the current remaining after the application of NGF.
Use of PC12 cells expressing mutant PDGF receptors proved advantageous in identifying the signaling molecules necessary for neurite extension and for the chronic upregulation of Na+ channels during PC12 cell neuronal differentiation (Vaillancourt et al., 1995;Fanger et al., 1997). The tyrosine phosphorylation sites of the receptor are well characterized, and point mutations of these residues result in the inability of the receptor to associate with and activate specific signaling molecules. PLCγ associates with residues 1009 and 1021, PI3-K associates with residues 740 and 751, the nonreceptor tyrosine phosphatase Syp associates with tyrosine 1009, and the GTPase-activating protein (GAP) associates with the tyrosine at position 771. The tyrosine at position 716 enhances binding with Grb2 and is thought to be involved in Ras activation, although phenylalanine substitutions of this particular tyrosine have been unable to inhibit Ras activation to any great extent (for review, see Claesson-Welsh, 1994; van der Geer et al., 1994). Therefore, to elucidate the signaling pathways underlying the PDGF-induced acute Na+current inhibition that we observed, we took advantage of the following paradigm: cells expressing exogenous mutant PDGF receptors were morphologically differentiated with NGF; then, during recording sessions, the ability of PDGF to acutely inhibit Na+currents was analyzed (Fig.6A,B).
Analysis of growth factor-induced inhibition of Na+ current in PC12 cells expressing mutant PDGF receptors or a dominant-negative Ras mutant. A, PC12 cells expressing either F5, F579/581, F5/579, or F5/581 mutant PDGF receptors were differentiated by chronic treatment with 100 ng/ml NGF. Peak Na+ current amplitudes (open diamonds) are plotted as a function of time, showing inhibition in response to 20 ng/ml PDGF (horizontal bars). Recording protocols are described in Figure 1; *Raw datatraces selected for insets.B, Cumulative data for experiments described inA are shown. Whole-cell currents recorded during growth factor application were normalized to whole-cell currents recorded before growth factor application. Error bars represent the mean ± SEM of normalized Na+ current amplitudes;†Statistically significant difference from PC12 cells expressing the wtPDGF receptor. The values of n for each experimental group are, from left to right, 10, 20, 20, 15, and 14.C, Cumulative data for experiments with 17-2 PC12 cells, which express the dominant-negative mutant N17 Ras, are shown. Error bars represent the mean ± SEM of normalized whole-cell Na+ currents after acute application of 100 ng/ml NGF (solid columns) or 50 ng/ml bFGF (shaded columns) reached its maximum effect in either 17-2 cells or PC12 cells expressing the wtPDGF receptor; †Statistically significant difference from PC12 cells expressing the wtPDGF receptor. The values of n for each experimental group are, from left to right, 10, 8, 10, and 10.
The F5 mutant PDGF receptor is characterized by five tyrosine-to-phenylalanine substitutions at residues 740, 751, 771, 1009, and 1021. Biochemical analyses of this receptor in PC12 cells and other cell lines suggest that it is unable to associate with or activate PLCγ, Syp, PI3-K, or GAP (Valius and Kazlauskas et al., 1993; Vaillancourt et al., 1995). Despite the inability of this receptor to activate these signaling pathways, PC12 cells expressing the F5 receptor respond to PDGF with the extension of neurites and with the induction of mRNA for c-fos, Na+channels, and the neural-specific metalloprotease transin (Vaillancourt et al., 1995; Fanger et al., 1997). As expected, our cell lines, when chronically treated with NGF, upregulated the functional expression of Na+ channels, elevating current densities to 92.6 ± 8.8 pA/pF (n = 22). Furthermore, acute application of PDGF elicited a 39.8 ± 3.1% (n = 20) decrease in Na+ channel currents, suggesting that PLCγ, Syp, PI3-K, and GAP have little effect on growth factor-mediated Na+ current inhibition.
Because of the presence of tyrosine residues 579 and 581 located at the juxtamembrane region, the F5 PDGF receptor still retains its ability to activate members of the Src family nonreceptor tyrosine kinases. We next analyzed two PDGF mutants containing the same five-point substitution motif characteristic of the F5 receptor but possessing additional tyrosine-to-phenylalanine substitutions at either residue 579 (F5/579 receptor) or 581 (F5/581 receptor) and a third receptor with substitutions at residues 579 and 581 (F579/581 receptor) with the remainder of the intracellular region intact. Despite similarities in substitution motifs, the ability of each receptor to associate with Src and to induce Src phosphorylation differ, with PC12 cells expressing the F579/581 or F5/579 receptor showing no Src association or phosphorylation, whereas the F5/581 receptor shows modest Src association and phosphorylation (Vaillancourt et al., 1995). To test the ability of these mutant receptors to mediate PDGF-induced Na+ current inhibition, we differentiated cells expressing these receptors via chronic treatment with NGF. In agreement with previous observations (Fanger et al., 1997), NGF induced morphological differentiation and upregulated Na+channel currents, increasing current densities to 46.6 ± 5.7 pA/pF (n = 19), 78.3 ± 12.1 pA/pF (n = 14), and 44.6 ± 7.4 pA/pF (n= 13) in morphologically differentiated cells expressing, respectively, the F579/581, F5/579, and F5/581 mutant receptors. However, we found that the ability of PDGF to inhibit Na+ current was reduced in cells expressing the F579/581, F5/579, or F5/581 receptors. In these cells PDGF inhibited Na+ currents by only 20.2 ± 3.6% (n = 20), 17.0 ± 3.0% (n = 14), or 6.7 ± 3.0% (n = 15), respectively. This suggested that PDGF receptors with mutations in their Src binding domains were unable to mediate Na+inhibition in comparison with their wild-type counterparts, despite their ability to initiate morphological differentiation (with the exception of the F5/579 receptor), activate the Ras and MAPK pathway, and induce both c-fos and transin mRNA in a PDGF-dependent manner (Vaillancourt et al., 1995; Fanger et al., 1997). To confirm further that the decrease in PDGF-induced inhibition was the result of loss of Src activation, we loaded cells expressing the wtPDGF receptor with the Src kinase inhibitor PP1 (Hanke et al., 1996) by including the inhibitor in the intracellular recording solution and allowing the solution to diffuse into the cells for 3–5 min after establishment of whole-cell access and before PDGF application. Addition of PP1 (1 μm), which inhibits Src and other members of the Src family of kinases, including Fyn and Lyk, had little to no effect on Na+ current densities in cells that had been differentiated with either NGF [139.1 ± 13.1 pA/pF (n = 18) compared with 115.6 ± 11.4 pA/pF (n = 18) in cells without PP1] or PDGF [103.2 ± 17.5 pA/pF (n = 14) compared with 107.8 ± 14.7 pA/pF (n = 14) in cells without PP1]. However, PP1 effectively reduced Na+ current inhibition by both NGF (15.1 ± 3.4%; n = 14) and PDGF (17.2 ± 3.0%; n = 21), arguing further that the Src family of nonreceptor tyrosine kinases plays a role in growth factor-induced Na+ current inhibition.
Analysis of acute effects of growth factors in PC12 cells expressing a dominant-negative mutant form of Ras
In PC12 cells, activation of the Ras and MAPK pathway is essential for a variety of morphological and physiological events, including neurite extension (Szeberényi et al., 1990; Traverse et al., 1992; Cowley et al., 1994; Loeb et al., 1994; Stephens et al., 1994) and the upregulation of Ca2+ channels (Pollock and Rane, 1996). Furthermore, there is evidence that signals mediated by Src are required for the activation of Ras, suggesting that Ras might lie downstream of Src in certain RTK signaling cascades (Kremer et al., 1991; Rusanescu et al., 1995). The long-term upregulation of Na+ channels during PC12 cell differentiation, however, is completely independent of Ras (Fanger et al., 1993), although activation of Src seems to be important for establishing fully functional Na+ channel expression (Fanger et al., 1997).
To determine whether the signaling mechanisms underlying acute regulation paralleled those responsible for long-term Na+ expression, we examined the ability of NGF to acutely regulate Na+ channels in 17-2 cells, a PC12 cell line that overexpresses the N17 Ras dominant-negative mutant (Fig.6C). In these cells, NGF and bFGF fail to induce morphological differentiation and the early-response genes c-fos, c-jun, and zif268 (Szeberényi et al., 1990), and are incapable of upregulating Ca2+channels (Pollock and Rane, 1996). However, the upregulation of type II/IIA Na+ channel mRNA and functional Na+ channels is unaffected (Fanger et al., 1997).
We chronically treated 17-2 cells with either NGF or bFGF (50 ng/ml) for a period of 5–7 d. As expected, the cells failed to extend neurites but expressed elevated levels of Na+channels (data not shown). In cells that had been chronically treated with bFGF, acute NGF application reduced Na+ current by only 13.8 ± 1.0% (n = 10). Likewise, the ability of bFGF to inhibit Na+ currents in cells chronically treated with NGF was also impaired (15.9 ± 1.0%;n = 10) relative to our previous observations. Because we showed previously that growth factor-induced inhibition of Na+ currents does not depend on differentiation, it is unlikely that the decrease in inhibition is the result of the inability of Ras to induce differentiation. Therefore, these results indicate a role for Ras in the acute effects of growth factors on Na+ currents.
DISCUSSION
Our results are the first to ascertain that growth factors can acutely inhibit mammalian Na+ channels, which are fundamental for the regulation of electrical excitability in neuronal cells. Application of growth factors to differentiated PC12 cells, which express the PN-1 and brain type II/IIa Na+channels (Mandel et al., 1988; D’Arcangelo et al., 1993), resulted in a rapid, partially reversible inhibition of whole-cell Na+ current. The inhibition was dependent on the tyrosine kinase activity of the associated growth factor receptors, because RTK inhibitors reduced inhibition by at least two-thirds. These observations are consistent with an increasing amount of evidence that in addition to their ability to upregulate ion channel expression during neuronal development, growth factors and their signaling pathways may play another, fundamental role as acute regulators of ion channel activity in both developing and mature nervous systems. Indeed, in rat brain neurons, both neurotrophin-3 (NT-3) and NGF are capable of activating Ca2+-activated K+channels (Holm et al., 1997), which contribute to action potential repolarization and often underlie the modulation of action potential frequency. Additionally, both NMDA receptors and the human Kv1.5 K+ channel possess Src-specific SH3 binding domains, associate with Src, and are modulated by this association (Holmes et al., 1996; Yu et al., 1997), further arguing that signals activated by RTKs, particularly Src, are intimately involved with ion channel regulation. In agreement with this, our own observations with PDGF receptors deficient in activating the Src family of kinases demonstrate that Src is essential for growth factor-induced Na+channel inhibition. It should be noted, however, that neither the α subunits of the PN-1 and type II/IIa Na+ channels nor the β subunits possess the RPLPXXP SH3-binding domain motif indicative of direct Src association, suggesting that the influence of Src on Na+ channels is not direct and occurs via downstream effectors.
It is interesting to note, then, that studies indicate that Src, Ras, and MAP kinases act in sequence to initiate the induction of early genes and morphological differentiation in PC12 cells (Kremer et al., 1991; D’Arcangelo and Halegoua, 1993; Rusanescu et al., 1995). Because PDGF receptors deficient in activating the Ras and MAPK pathway have yet to be successfully expressed, we turned to the 17-2 PC12 cell line, which overexpresses the dominant-negative mutant N17 Ras, to analyze the role of Ras in Na+ current inhibition. Although we could not definitively conclude that Src lies upstream of Ras, it is intriguing to note that the block of growth factor-induced inhibition was comparable between cells loaded with RTK inhibitors and cell lines deficient for either Src or Ras signaling, which would be expected if these molecules are activated in series. Still, it is possible that Ras and Src inhibit Na+ channels via parallel, redundant pathways; thus we cannot eliminate the possibility that activation of one signaling molecule can partially compensate for inactivity of the other. Furthermore, Na+ channel inhibition is not completely blocked in cell lines deficient in Src or Ras signaling, which suggests that pathways independent of Ras and Src could also be contributing to acute Na+ channel regulation. However, the inability to completely block inhibition is probably the result of a small amount of residual Src and Ras activity, because growth factors applied to these cell lines are still capable of inducing Src- and Ras-dependent responses to a minor degree (Szeberényi et al., 1990; Kremer et al., 1991; D’Arcangelo and Halegoua, 1993; Vaillancourt et al., 1995; Fanger et al., 1997).
Our results also support an increasing amount of evidence that activation of Src and/or Ras through RTKs is a critical component of ion channel regulation in both developing and mature nervous systems. For example, chronic treatment of 17-2 PC12 cells with NGF or bFGF fails to upregulate Ca2+ channels. Likewise, PC12 cells morphologically differentiated by oncogenic Ras fail to show an increase in Ca2+ channel expression, suggesting that Ras is necessary but not sufficient for this upregulation (Pollock and Rane, 1996). PC12 cells induced to undergo morphological differentiation by oncogenic Src, however, upregulate Ca2+ channels in a manner comparable with wild-type PC12 cells differentiated by NGF or bFGF (Rausch et al., 1990). Furthermore, our own observations with PC12 cells expressing PDGF receptors deficient in Src signaling suggest that Src may also be a necessary component in this upregulation (M. D. Hilborn and S. G. Rane, unpublished observations). Similar to the situation with Ca2+ channels, the upregulation of Na+ channel expression during growth factor-induced PC12 cell differentiation seems to be partially dependent on the activation of Src (Fanger et al., 1997). However, this upregulation is independent of Ras activation (Fanger et al., 1993), an intriguing difference from our own observations that Ras activation is necessary for acute regulation. This difference in signaling, in conjunction with signals generated by sustained activation of RTKs, could account for the ability of RTKs to regulate Na+ and other ion channels both acutely and chronically. Further analysis of the role of Src, Ras, and their downstream effectors in acute ion channel regulation will perhaps resolve these issues.
Little is known about the receptor-mediated modulation of Na+ channels. Activation of G-protein-coupled pathways enhanced brain Na+ channel activity in Chinese hamster ovary (CHO) cells (Ma et al., 1994), and inhibition of cardiac Na+ channels by β-adrenergic agonists can also be mediated via G-proteins (Schubert et al., 1989). Stimulation of β2-adrenergic receptors expressed in Xenopus oocytes can either enhance or attenuate Na+ currents, depending on the levels of receptor expression (Smith and Goldin, 1992, 1996), and both PKA and PKC have been shown to modulate Na+channel currents by directly phosphorylating key residues on the α-subunit of rat brain channels (Rossie and Catterall, 1987, 1989;Rossie et al., 1987; West et al., 1991; Murphy and Catterall, 1992; for review, see Cukierman, 1996). Because NGF activates PLCγ and PKA during differentiation and because activation of the PKA pathway has been implicated in post-translational effects required for full, functional Na+ channel expression during PC12 cell differentiation (Kalman et al., 1990; Ginty et al., 1992; D’Arcangelo et al., 1993), both PKA and PKC might be considered candidates for mediating acute growth factor-induced Na+ channel inhibition. However, attenuation of Na+ currents by PKA phosphorylation occurs without a change in steady-state activation or inactivation properties of the channel (Gershon et al., 1992; Li et al., 1992, 1993; Smith and Goldin, 1996), an observation that differs from our own observation that growth factor-induced Na+ current inhibition is associated with a negative shift in steady-state inactivation. Additionally, the F5 mutant PDGF receptor, which is deficient in activating PLCγ, inhibited Na+ current to a level comparable with that observed with wild-type receptor, suggesting that downstream effectors of PLCγ, including PKC, are not required for this response. Furthermore, PKC inhibition of rat brain Na+ channels expressed in CHO cells is accompanied by a slowing of inactivation (Numann et al., 1994), and the attenuation of whole-cell Na+currents is the consequence of a depolarizing shift in the voltage dependence of channel activation (Dascal and Lotan, 1991; Schreibmayer et al., 1991), both of which differ from the shift in steady-state inactivation that we observed. And although a negative shift in the inactivation curve of Na+ channels in a mouse neuroblastoma cell line was observed in response to some PKC activators (Godoy and Cukierman, 1994a,b), a follow-up study demonstrated that the channel attenuation was a direct effect of the PKC activators (Renganathan et al., 1995).
Overall, our studies complement an increasing amount of evidence that growth factors, particularly neurotrophins, are involved in acute neuromodulatory processes in addition to their traditional roles in neuronal development (for review, see Thoenen, 1995). The ability of growth factors to modulate channels fundamental for regulating electrical excitability provides a mechanistic basis for the growth factor-dependent neural plasticity that has been observed in the mammalian visual cortex and hippocampus. Our results establish roles for Ras and Src as mediators of growth factor receptor-activated acute ion channel regulation, in addition to their recognized contribution to the long-term developmental actions of the growth factor receptor family.
Footnotes
This work was supported by a grant from the Whitehall Foundation (S.G.R.), Grant DK 08897 (R.R.V.), and a Predoctoral Fellowship from the American Heart Association, Indiana Affiliate (M.D.H.). We are grateful to Dr. S. Rossie for critical review of this manuscript.
Correspondence should be addressed to Dr. Stanley G. Rane, Department of Biological Sciences, Lilly Hall, West Lafayette, IN 47907.
