INTRODUCTION

Platelet-derived growth factor (PDGF),¹ like other growth factors and hormones, produces biphasic changes in intracellular Ca²⁺ concentration in many cell types (1-4). The initial Ca²⁺ transient results from InsP₃-mediated release of intracellular Ca²⁺ stores (5). This Ca²⁺ transient is usually followed by sustained influx of Ca²⁺ from the medium. The signaling mechanism for this extracellular Ca²⁺ entry process is not well defined. The most widely accepted model for regulation of Ca²⁺ entry, termed the "capacitative" model, argues that entry is activated by prior depletion of the InsP₃-mediated intracellular Ca²⁺ stores (6). A weakness of this model is that the interaction between the intracellular stores and the plasma membrane Ca²⁺ channel has never been characterized.

Receptor-mediated Ca²⁺ entry may involve pathways other than the proposed capacitative entry system. In hepatocytes (7-9), Jurkat T cells (10), neurosecretory PC12 cells (11), adrenal chromaffin cells (12), fibroblasts (13), and pancreatic acinar cells (14), receptor-mediated Ca²⁺ entry appears to occur in the absence of prior intracellular Ca²⁺ mobilization. This laboratory (1) and others (3, 4) demonstrated that PDGF-induced Ca²⁺ entry can occur in the absence of intracellular Ca²⁺ mobilization in several cell types. Taken together, these data suggest that some receptors may signal intracellular Ca²⁺ release and Ca²⁺ entry separately, such that Ca²⁺ entry would not require prior release of intracellular stores.

One candidate for a separate signal for Ca²⁺ entry is direct G-protein regulated opening of the membrane Ca²⁺ channels (15). Alternatively, second messengers such as inositol phosphates (15, 16) or diacylglycerol (17, 18) might be involved in regulation of plasma membrane Ca²⁺ channels. In support of the latter idea, Huang et al. (1) showed that antibodies to phosphatidylinositol 4,5-bisphosphate (PIP₂) block PDGF-mediated Ca²⁺ entry, suggesting that the mechanism involves a metabolite of PIP₂.

In this study, we used PDGF receptor mutations, microinjected heparin, and caged InsP₃ to examine PDGF-induced Ca²⁺ entry. We show that PDGF-induced Ca²⁺ entry does not require InsP₃-mediated release of intracellular Ca²⁺ stores and that PDGF causes substantially more Ca²⁺ entry than InsP₃ alone. We propose that in this system PDGF does not induce Ca²⁺ entry predominantly via the capacitative mechanism, but rather via a direct pathway involving a metabolite of PIP₂.

EXPERIMENTAL PROCEDURES

Unless otherwise specified, all chemicals were purchased from Sigma. Low molecular weight heparin was Sigma H-5271. Fura-2-AM and Ca²⁺-green-1 pentapotassium salt were purchased from Molecular Probes (Eugene, OR). Bovine serum albumin was fatty acid-poor fraction V from Miles (Kankakee, IL). PDGF-BB was purchased from Boehringer Mannheim. Caged-InsP₃ was purchased from Calbiochem (San Diego, CA).

Chinese hamster ovary (CHO) cells, which normally lack PDGF receptors (19), were stably transfected with wild-type PDGF-BB receptor cDNA (CHO-PDGF) (19). PDGF receptor mutants were constructed by site-directed mutagenesis (20). Plasmids containing the mutated receptor cDNA were also used to stably transfect CHO cells (19). Mutation of Tyr⁷³⁹ to Phe (Y739F) prevents binding of GTPase-activating protein; Y708F/Y719F prevents the association of phosphatidylinositol 3-kinase (19), and Y977F/Y989F interfere with bindings of PLC (based on homology with Tyr¹⁰⁰⁹ and Tyr¹⁰²¹ of the human receptor (21)).

CHO cells transfected with PDGF wild-type and mutant receptors were grown in Ham's F-12 medium in a humidified atmosphere of 5% CO₂, 95% air at 37 °C. Media contained 10% (v/v) fetal calf serum, penicillin (50 units/ml), and streptomycin (50 units/ml). Medium for stable transfectants was supplemented with 400 µg/ml G418. Culture medium was changed every 2-3 days until cells were confluent. Cells were made quiescent by replacement of serum with bovine serum albumin (0.5 mg/ml) and transferrin (5 µg/ml).

For measurement of Ca²⁺_i, cells were plated at 500 cells/ml on a microscope cover glass (Fisher, 25-mm circle) and grown in serum-containing medium until 70-80% confluence was achieved. Serum was removed 24 h prior to Ca²⁺_i measurements to eliminate any residual PDGF activity in the serum. Cells grown on coverslips were incubated with 5 mM fura-2-AM in assay medium at room temperature for 1 h. The coverslips were mounted on a temperature-controlled chamber at 31 °C in assay medium (140 mM NaCl, 5 mM KCl, 1 mM Na₂HPO₄, 25 mM glucose, 25 mM Hepes/NaOH (pH 7.2), and 0.5 mg/ml bovine serum albumin with or without 2 mM CaCl₂ (22). Solution without added CaCl₂ was termed "0" Ca²⁺ medium. Fura-2 fluorescence was measured using a Nikon epifluorescence inverted microscope fitted with a rotating holder for excitation filters (340 and 380 nm) as described previously (1). Signals were digitized using a Labmaster interface board (Scientific Solutions, Solon, OH) and were recorded in an IBM style computer using the UMANS software package (Chester Regen, Bio-Rad). To calibrate the fluorescence signals, the ratio of fluorescence at 340/380 nm was compared with ratios obtained at maximal Ca²⁺_i (achieved by the addition of 10 µm 4BR-A23187) and 0 Ca²⁺_i (achieved by the addition of 20 mM EGTA). Ca²⁺_i was calculated as described previously (23).

For Ca²⁺_i measurements with caged 1,4,5-InsP₃, single cells grown on coverslips were microinjected with Ca²⁺ green-1-pentapotassium salt (0.1 mM) with caged 1,4,5-InsP₃ in the absence or presence of heparin. The fluorescence signal for the photorelease experiments was measured with excitation light passed through a narrow-band interference filter centered at 490 nM to excite Ca²⁺ green-1 fluorescence. Increases in fluorescence emission collected at 515 nM correspond to increases in the free Ca²⁺_i concentration (24).

Following microinjection of caged 1,4,5-InsP₃, photolysis of caged 1,4,5-InsP₃ was achieved using a modification of the method of Bird et al. (24). Light from a continuously burning xenon arc flashlamp was passed through a broad 120-nm bandwidth filter centered at 350 nm and directed to the specimen for times ranging from 0.5 to 4.0 s. The timing and duration of the UV flashes were controlled by the computer. Immediately following these computer-controlled flashes, Ca²⁺_i measurements were started.

Cells were microinjected using glass capillary needles held in a Narishige micromanipulator. Pipettes were made from borosilicate glass tubes (0.9 mm) using a Flaming/Brown Micropipette Puller from Sutter Instrument Co. (Novato, CA). Volumes of approximately 5 × 10¹⁴ liters are microinjected by this method (25). Various concentrations of heparin in buffer (27 mM K₂HPO₄, 8 mM NaH₂PO₄, 26 mM KH₂PO₄, pH 7.3) were microinjected with 0.1 mM Ca²⁺ green-1-pentapotassium salt with or without caged 1,4,5-InsP₃ to achieve an estimated intracellular concentration that was 10-fold diluted.

For whole cell patch-clamp measurements, cells were plated on glass coverslips and cultured in serum-containing medium. Coverslips were transferred from the culture dish onto the bottom of the recording chamber, which was then perfused with Ringer's solution (155 mM NaCl, 4.5 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 10 mM glucose, and 10 mM Hepes, pH 7.3, adjusted with NaOH) for 2-3 min to remove the cell culture medium. The internal solution loaded in the recording pipettes contained 140 mM cesium aspartate, 10 mM CsCl, 1 mM CaCl₂, 10 mM EGTA, 2 mM MgATP, and 10 mM Hepes (pH 7.3, adjusted with CsOH). 10 µl of PDGF was added into the recording chamber with a scaled pipette. Heparin was mixed with the internal solution and loaded in the recording electrode.

All electrophysiological recordings were performed at room temperature (22-24 °C). Patch-clamp recordings were performed using the whole cell recording configuration of the patch-clamp technique (26) at the holding potential of 60 mV with an Axopatch-1D amplifier (Axon Instruments, Foster City, CA). Electrodes pulled from KG-33 borosilicate glass capillaries (Garner Glass, Inc., Claremont, CA) with an outer diameter of 1.5 µm and inner diameter of 1.0 µm were filled with the internal solution. The current responses were evoked by adding PDGF (final concentration 30 ng/ml).

The generation of voltage-clamp protocols and the acquisition of data were controlled by an IBM/AT-compatible computer through pClamp software (Axon Instruments) via a 12-bit digital/analog and analog/digital converter (TL-1 DMA Labmaster Interface, Axon Instruments). The signal was filtered at 5 kHz (3 dB, 8 pole, low pass Bessel filter; Frequency Devices, Haverhill, MA) and digitized on-line at a sampling interval of 0.25 ms, and directly stored on the computer hard disk for off-line analysis. Specific patch-clamp protocols are described in the legend of Fig. 3.

[View Larger Version of this Image (27K GIF file)]

Under a microscope (× 600 magnification), CHO-PDGF cells were visually identified. Each cell was held at 60 mV and data was generated by a voltage step-ramp-step protocol during each experiment. The protocol consisted of the following pulses: holding potential (60 mV, 37.5 ms); step (90 mV, 18.75 ms); ramp (90 to 50 mV, 400 ms); step (50 mV, 18.75 ms), and holding potential (60 mV, 37 ms). This protocol was repeated continuously every 4 s until the experiment was stopped. During whole cell recordings, the access resistance was usually 4-10 M, 80% of which was canceled by electronic compensation. The currents studied in this paper were typically less than 200 pA, so the maximum error in voltage due to uncompensated access resistance was less than 2 mV. Only those recordings with stable and low access resistance were included in the final analysis. Our patch-clamp results are derived from a total population of 34 CHO-PDGF cells.

The PDGF evoked responses were recorded with pClamp 5.6 and analyzed with pClamp 6.1 Data plotting and curve fittings were performed with Origin (MicroCal Software, Inc., Northampton, MA). The data are presented as the mean ± S.E. Statistical analysis was performed by the paired or unpaired Student's t test. p < 0.05 was considered significant.

RESULTS

Ca²⁺ transients were measured in CHO cells using either fura-2 or Ca²⁺ green. For the purposes of this study, the first phase measured in nominally 0 Ca²⁺ extracellular medium was termed "Ca²⁺ release" and the second, measured upon re-addition of Ca²⁺ to the medium, was termed "Ca²⁺ entry." Previous work from this laboratory showed that PDGF-induced Ca²⁺ entry can occur in the absence of intracellular Ca²⁺ release in vascular smooth muscle cells (1). Notably, microinjection of anti-PIP₂ antibodies blocked both PDGF-mediated Ca²⁺ release and Ca²⁺ entry (1), suggesting that metabolites of PIP₂, possibly acting through two separate pathways, are responsible for both intracellular Ca²⁺ release and extracellular Ca²⁺ entry. We sought to confirm these findings using various signaling mutations in the PDGF receptor.

In wild-type CHO cells, which do not express native PDGF receptors, PDGF did not produce changes in Ca²⁺_i (Fig. 1A). Solution changes from 2 mM to 0 Ca²⁺ produced small intracellular Ca²⁺ transients in the direction of the solution change. In CHO cells expressing the wild-type PDGF receptor, PDGF (5 or 25 ng/ml) raised Ca²⁺_i in a biphasic fashion (Fig. 1B). In 0 Ca²⁺ medium, PDGF produced a transient elevation of intracellular Ca²⁺ to 625 ± 32 nM (Ca²⁺ release). Subsequent addition of 2 mM Ca²⁺ to the medium produced a greater and more prolonged Ca²⁺ transient to 845 ± 48 nM, representing Ca²⁺ entry.

[View Larger Version of this Image (22K GIF file)]

In CHO cells expressing the Y977F/Y989F mutation (which disrupts the binding of PLC to the activated PDGF receptor) (21), PDGF (5 ng/ml, Fig. 1C) failed to elicit any normal Ca²⁺ transients. Similar results were obtained when PDGF concentration was increased to 25 ng/ml (data not shown). Following addition of PDGF in 0 Ca²⁺ medium, re-addition of 2 mM Ca²⁺ resulted in only a small Ca²⁺ transient, no different from that observed in wild-type CHO cells in Fig. 1A. Thus, PLC appears critical to the development of all Ca²⁺ signals produced by PDGF.

Other signaling mutations in the PDGF receptor (20) did not significantly affect Ca²⁺ responses. Activation of the PDGF receptor in CHO cells expressing PDGF receptors mutated at either the PI 3-kinase site (Y708F/Y719F) (Fig. 1D) or GTPase-activating protein-binding site (Y739F) (Fig. 1E) induced both Ca²⁺ release (525 ± 38 and 575 ± 27 nM) and Ca²⁺ entry (725 ± 43 and 630 ± 51 nM). Ca²⁺ responses to PDGF were somewhat variable for Y708F/Y719F and Y739F. However, for neither mutant was the Ca²⁺ response eradicated as it was for the PLC mutant. These findings are identical to results obtained with normal murine mammary gland epithelial cells expressing the same PDGF receptor mutations (data not shown).

To determine whether PDGF-mediated Ca²⁺ entry depends on InsP₃-dependent release of intracellular Ca²⁺ stores, cells were microinjected with heparin to block the InsP₃ receptor. To first demonstrate that heparin blocks the InsP₃ receptor, Ca²⁺ green (0.1 mM) and caged 1,4,5-InsP₃ (0.375 mM) were microinjected into CHO-PDGF cells in the absence or presence of low molecular weight heparin (5-50 mg/ml). Photorelease of InsP₃ was achieved with flashes of UV light (see "Experimental Procedures"). Cells in 0 Ca²⁺ medium were exposed to multiple flashes (~8) of 0.5-2 s duration (Fig. 2A). Each flash resulted in a single transient Ca²⁺ spike. In the absence of caged 1,4,5-InsP₃, multiple flashes of UV light (0.5 to 4 s duration) had no effect on intracellular Ca²⁺ levels (data not shown). Heparin, a competitive inhibitor of the InsP₃ receptor (27), when injected at 20 mg/ml, blocked intracellular Ca²⁺ release in 0 Ca²⁺ medium both following flash photolysis of caged 1,4,5-InsP₃ and after subsequent addition of PDGF (Fig. 2B). However, addition of 2 mM Ca²⁺ to the medium in the ongoing presence of PDGF resulted in a substantial increase in intracellular Ca²⁺. These data suggest that PDGF-mediated Ca²⁺ entry can occur despite blockade of InsP₃-dependent intracellular Ca²⁺ release by heparin.

[View Larger Version of this Image (14K GIF file)]

It has previously been reported in fibroblasts (28) and rat kidney mesangial cells (4, 29) that activation of the PDGF receptor results in opening of a nonselective cation channel. Using a whole cell patch-clamp, we measured cellular currents in CHO-PDGF cells following the addition of PDGF to the extracellular medium. To reduce noise and membrane breakdown, these studies were performed at 22-24 °C. PDGF (30 ng/ml) caused an inward current which slowly inactivated (Fig. 3A). The peak amplitude was 124 ± 40 pA at 60 mV. There was a latency from the time of PDGF addition to the peak current of 110 ± 28 s (n = 5). The ramp plot shows the current-voltage relationship for this current (Fig. 3B). The reversal potential was 5.6 ± 6.5 mV (n = 5). In most recordings, there was a noticeable inward sodium current on the ramp plot (> 20 mV). In 18 cells studied, 5 cells responded to the addition of PDGF (response rate ~27%). This response rate following PDGF was nearly identical to the response rate (28%) when Ca²⁺_i was measured by fura-2 at 23 °C. In 8 of 8 cells, there was no comparable current observed in response to thapsigargin.

We next determined the effect of intrapipette application of heparin on PDGF-activated cation currents. In a CHO-PDGF cell in which the patch-clamp was loaded with heparin (1 mg/ml), PDGF (30 ng/ml) evoked an ion current in 25% (2 of 8) of cells studied (Fig. 3C). This current was similar to the PDGF-activated cation current observed without heparin. The peak current amplitude was 255 pA at 60 mV with a latency of 122 s. The reversal potential (Fig. 3D) of the PDGF-activated current in the presence of heparin was also near 0 (2.6 mV). The similar percentage of cells activated by PDGF and the similar characteristics of the channel following heparin exposure suggest that heparin did not alter the ability of PDGF to activate the cation channel.

To determine whether photolysis of caged 1,4,5-InsP₃ could (through release of Ca²⁺ from intracellular stores) activate the Ca²⁺ entry pathway, Ca²⁺ green (0.1 mM) and caged 1,4,5-InsP₃ (0.375 mM) were microinjected into CHO-PDGF cells. Cells exposed to sequential changes in medium Ca²⁺ concentration without flash photolysis or PDGF demonstrated a small, short-lived Ca²⁺ transient, in the same direction as the medium change (Fig. 4, left). The same cell containing caged-InsP₃ and incubated in 0 Ca²⁺ medium was then exposed to multiple flashes (~11) of 0.5 to 4 s duration. Each flash caused a single spike in intracellular Ca²⁺. Subsequent re-addition of 2 mM Ca²⁺ to the medium caused a Ca²⁺ transient that was similar in magnitude from that observed in the same cell when medium Ca²⁺ concentration was increased prior to flash photolysis. When PDGF (25 ng/ml) was added to the cells, it produced a single intracellular Ca²⁺ spike that was similar in magnitude and duration to the spikes caused by each flash photolysis of caged InsP₃. However, the increase in Ca²⁺_i following re-addition of Ca²⁺ to the medium was much larger and longer lasting than was seen in either control cells or in cells exposed to photolysis of caged-InsP₃. In 11 experiments, re-addition of Ca²⁺ after caged-InsP₃ increased fluorescence by 3.2 ± 0.2 arbitrary fluorescence units. The identical procedure following PDGF increased fluorescence by 4.4 ± 0.1 units (p = 0.0002). Thus, PDGF appears to produce a signal for Ca²⁺ entry that differs from that produced by photolysis of caged InsP₃. This additional signal results in a larger Ca²⁺ response than that caused by InsP₃ alone.

[View Larger Version of this Image (15K GIF file)]

One explanation for the finding of Fig. 4 that would fit with the capacitative entry model is that PDGF results in more complete emptying of intracellular stores despite the similar appearance of Ca²⁺ spikes generated by PDGF and photorelease of caged-InsP₃. To address this possibility, CHO-PDGF cells containing microinjected Ca²⁺ green (0.1 mM) and caged 1,4,5-InsP₃ (0.375 mM) were exposed to PDGF and then to multiple UV flashes in 0 Ca²⁺ medium. PDGF (25 ng/ml) produced a single Ca²⁺ transient, and subsequent photolysis of caged 1,4,5-InsP₃ generated multiple Ca²⁺ transients (Fig. 5), showing that PDGF alone did not completely empty intracellular Ca²⁺ stores. Addition of 2 mM Ca²⁺ to the medium then resulted in Ca²⁺ entry that was comparable to that observed with PDGF alone (Fig. 1B). These findings suggest that complete emptying of InsP₃-dependent intracellular Ca²⁺ stores is not necessary for activation of the Ca²⁺ entry pathway by PDGF.

[View Larger Version of this Image (16K GIF file)]

DISCUSSION

The capacitative model of receptor-mediated Ca²⁺ entry proposes that Ca²⁺ entry occurs when intracellular Ca²⁺ stores become depleted (6). Activation of many receptors results in the release of InsP₃ into the cytoplasm which then binds to InsP₃ receptors on intracellular membranes, opening InsP₃-gated Ca²⁺ channels (5). Depletion of these InsP₃-dependent intracellular stores is proposed to generate a signal that opens Ca²⁺ channels on the plasma membrane, causing Ca²⁺ influx from outside the cell (6). Several models for such a signal have been proposed, including: (a) formation of a factor (CIF, Ca²⁺ influx factor) which could diffuse from the Ca²⁺ stores to the plasma membrane (30), or (b) direct physical coupling of intracellular Ca²⁺ stores to Ca²⁺ entry channels (31). A weakness of the capacitative model is that the signal necessary to couple the state of the intracellular stores to the Ca²⁺ entry channel has never been clearly defined.

A number of recent studies indicate that in some systems Ca²⁺ entry can occur by a non-capacitative mechanism. For example, angiotensin II in adrenal chromaffin cells (14) or the cholecystokinin analogue, CCK-JMV-180, in pancreatic acinar cells (12) activated Ca²⁺ entry in the absence of intracellular Ca²⁺ mobilization. We have found that blockade of intracellular Ca²⁺ release by microinjected heparin does not interfere with PDGF-mediated Ca²⁺ entry in single vascular smooth muscle cells (1) or in CHO cells (the current study). Blockade of intracellular Ca²⁺ release by anti-InsP₃ receptor antibodies also failed to prevent PDGF-mediated Ca²⁺ entry.² Taken together, these data indicate that in at least some systems, Ca²⁺ entry can occur without activation of the capacitative system.

Further evidence to support that release of Ca²⁺ from heparin-sensitive InsP₃-dependent intracellular Ca²⁺ stores is not necessary for opening of plasma membrane Ca²⁺ channels was provided by patch-clamp studies in which the response rate for activation of cation currents by PDGF was unaffected by intracellular heparin. These findings are consistent with the work of Ma et al. (4) who found in excised patches of mesangial cells that intrapipette application of heparin failed to block PDGF-evoked ion currents. These findings contrast with other patch-clamp studies in which heparin did block cholecystokinin Ca²⁺-dependent currents (32), vasopressin-mediated Ca²⁺ transients (33), acetylcholine-dependent Ca²⁺ oscillations (34), and angiotensin II-dependent channel activity (35). Since heparin is a competitive inhibitor of the InsP₃ receptor, one might speculate that the failure of heparin to block PDGF-evoked currents in our patch-clamp experiments may be due to a heparin concentration that was insufficient to completely block the InsP₃ receptor. However, the intrapipette heparin concentration used in our patch-clamp studies was 2-20-fold greater than that used in experiments in other systems in which heparin did block receptor-mediated ion currents (32-35). Furthermore, the intrapipette heparin concentration we used was probably substantially higher than the intracellular heparin concentrations achieved in previous microinjection studies in which we demonstrated blockade of PDGF-mediated intracellular Ca²⁺ release.²

Currents that flow-through capacitative Ca²⁺ entry channels have been referred to as Ca²⁺ release-activated Ca²⁺ currents (I_crac) (36, 37). Several findings in the current study and earlier work by others (4, 38) suggest that the PDGF-activated current may differ from I_crac. I_crac has been reported to have a low conductance, approximating 24 fS (39). Although we were not able to directly assess unitary conductance of the PDGF-activated ion channel, Frace and Gargus (28) demonstrated in fibroblasts that PDGF activated an ion channel with a much larger conductance of 28 pS. The PDGF-activated cation current we observed was not inwardly rectified, in comparison to I_crac which is typically inwardly rectified (40). I_crac is generally described as being Ca²⁺ selective (40), while Gargus et al. (38) and Ma et al. (4) found that the PDGF-activated ion current occurs through a nonselective cation channel. The reversal potential near 0 which we observed in the present work also suggests that the current is not Ca²⁺ selective. Finally, 8 of 8 cells exposed to thapsigargin in our study failed to evoke any cation currents, suggesting that I_crac is too small to detect under the conditions we used. Taken together, these findings all suggest that ion currents produced by activation of the PDGF receptor are not I_crac. The fact that PDGF can lead to opening of a channel other than I_crac is consistent with the notion that PDGF can cause Ca²⁺ entry via a pathway that bypasses the capacitative Ca²⁺ entry system.

The fact the PDGF may bypass the capacitative entry system does not rule out the possibility that this system still plays an important role in cellular Ca²⁺ metabolism under certain conditions. The data of Fig. 4 suggest that caged InsP₃ does not cause a Ca²⁺ entry signal any larger than that seen in control cells. However, caged InsP₃ under these conditions did not completely empty intracellular Ca²⁺ stores (data not shown). These data therefore cannot eliminate capacitative entry as a mechanism which participates in Ca²⁺ regulation. Fig. 4 does, however, show that PDGF produces a significantly larger entry signal than InsP₃ alone, suggesting that PDGF produces a signal in addition to the capacitative signal for the generation of Ca²⁺ entry.

Previous work from this laboratory showed that when intracellular Ca²⁺ stores were held in the depleted state, removal of PDGF from its receptor nearly abolished Ca²⁺ entry (1). This earlier data also indicates that PDGF receptor occupancy, and not simply depletion of intracellular stores, is essential to the generation of Ca²⁺ entry in vascular smooth muscle cells. What signal is produced by PDGF receptor occupancy that leads to non-capacitative Ca²⁺ entry? Previous work with anti-PIP₂ antibodies (1) and the current work with PDGF receptor mutants (Fig. 1) indicates that this signal derives from PLC-mediated breakdown of PIP₂. The failure of heparin to block Ca²⁺ entry and the finding that PDGF causes much larger Ca²⁺ entry than 1,4,5-InsP₃ alone suggests that this signal is not simply 1,4,5-InsP₃. Alternatives to InsP₃ include other inositol phosphates such as InsP₄, or lipid products of PIP₂ breakdown. The possibility of a role for InsP₄ in extracellular Ca²⁺ entry has been studied extensively in other systems with conflicting results (31, 41, 42).

A role for diacylglycerol and protein kinase C in regulating Ca²⁺ entry has also been studied extensively and the results are also contradictory (17, 18, 43-45). In some work, PKC has been shown to promote or enhance Ca²⁺ entry (17, 18), while other work suggests that PKC may blunt Ca²⁺ entry (43-45). One explanation for the variability in these findings is the failure to consider the isoforms of PKC expressed in various cell lines. Xu and Ware (44) recently showed the PKC may be a specific negative regulator of thrombin-mediated Ca²⁺ entry. On the other hand, preliminary work in our laboratory suggests that the -isoform of protein kinase C may play a key positive role in Ca²⁺ entry in CHO cells. A specific peptide inhibitor of this isoform blocked Ca²⁺ entry, but not intracellular Ca²⁺ release, following stimulation of CHO cells by endothelin (46). This new data suggests that diacylglycerol, rather than an inositol phosphate, may be the critical second messenger for Ca²⁺ entry following hormonal stimulation of CHO cells.

A model of the various pathways potentially involved in Ca²⁺ signaling by PDGF, and the points at which we have interfered with it, is shown in Fig. 6. Our earlier work showed that PIP₂ (A) is essential for both Ca²⁺ release and extracellular Ca²⁺ entry. This was confirmed by showing that mutations in the PDGF receptor that prevent PLC activation (B) failed to produce Ca²⁺ transients. However, blockade of InsP₃-dependent intracellular Ca²⁺ release (C) did not prevent PDGF-mediated Ca²⁺ entry. Finally, we showed that InsP₃-mediated release of intracellular Ca²⁺ stores (D) alone is insufficient to cause maximal Ca²⁺ entry in CHO cells. This leaves InsP₄ and diacylglycerol as the most likely candidates for molecules which can signal Ca²⁺ entry by bypassing the capacitative system.

[View Larger Version of this Image (23K GIF file)]