Reinforcement of signal generation at B2 bradykinin receptors by insulin, epidermal growth factors, and other growth factors.

Insulin and various growth factors (epidermal growth factor (EGF), insulin-like growth factor, fibroblast growth factor, and transforming growth factor alpha), which fail to modify the resting [Ca2+]i in PC12 rat pheochromocytoma and SKNBE human neuroblastoma cells when administered alone, became capable of inducing [Ca2+]i increases when administered a few (4-20) min after another agent, bradykinin. The latter peptide, working through a B2 receptor, caused hydrolysis of polyphosphoinositides and a large, biphasic [Ca2+]i transient (an initial (1-2 min) spike, originated primarily from intracellular stores, followed by a steady-state elevation dependent on Ca2+ influx). Priming by bradykinin of the growth factor effects was quickly dissipated by the addition of a B2 blocker. Activation of other receptors coupled to polyphosphoinositide hydrolysis: muscarinic and purinergic (in PC12 and SKNBE cells); bombesin and vasopressin receptors (in Swiss 3T3 cells), was without effect in priming. Bradykinin-primed, growth factor-induced [Ca2+]i rises in PC12 cells appeared after a 20-30-s delay; they were relatively small, but persistent; their concentration dependence was similar to that of other effects of the factors; and they included both release of Ca2+ from intracellular stores and stimulation of Ca2+ influx, preceded (in PC12 cells) by a transient increase of polyphosphoinositide hydrolysis. Thus the effect of growth factors (possibly dependent on the tyrosine kinase activity of their receptors) consisted in the reinforcement of the transmembrane signaling at B2 receptors. This is the first direct demonstration of a [Ca2+]i rise induced by insulin and insulin-like growth factor-I, and of such an effect of EGF in cell types endowed with a small number of specific EGF receptors.

Most of our knowledge on transmembrane signaling at GF receptors comes from studies carried out by treating in vitro serum-starved cells with one GF administered alone. In a multicellular living organism, however, cells are simultaneously exposed to a variety of stimuli, whose intracellular effects might interfere extensively with one another. A host of examples have already been reported in the literature demonstrating the synergism of various GF, and even of a GF with other, apparently unrelated, agents, with respect to cell growth (5,18,19). In contrast, the interaction of these various agents at the level of signal generation, in particular of [Ca2+]i control, has been investigated to a smaller extent (5,16,18). In the present work, a study along these lines has been carried out in various cell lines by the use of insulin, EGF, and other GF. GF unable to raise [Ca2+]; when administered alone were found to cause a positive modulation (reinforcement) of the PPI hydrolysis and [Ca2+]i responses induced by the previous application of an unrelated peptide agent, bradykinin (BK). The [Ca2+]i effects of GF were specific for the transduction machinery of the BK receptor of the B2 type. In fact, they were blocked by a BZ antagonist and failed to appear when the cells were pretreated with agents other than BK, working either via the activation of PPI hydrolysis or by other mechanisms.

EXPERIMENTAL PROCEDURES
Cell Culture"PC12 cells (20) were cultured in RPMI 1640 medium supplemented with 10% fetal calf serum, 5% horse serum, and antibiotics (penicillin, 100 units/ml; streptomycin, 0.1 mg/ml). They were subcultured weekly in a split ratio of 1:4. SKNBE human neuroblastoma cell monolayers were cultured as PC12 but horse serum was omitted from the media. The cells were split 1:6 weekly by treatment with trypsin as previously described (8,21). Mouse swiss 3T3 cells were cultured in Dulbecco's modified Eagle's medium, with antibiotics and 10% fetal calf serum. The cells were split 1:5 weekly and main-  Hepes,25 (pH 7.4). Final cell concentration was 107/ml. An aliquot of lo6 unloaded cells was used for autofluorescence measurements.
Loading of cells with fura-2 (21, 22) was carried out by a 30-min incubation with 5 I.~M fura-2/acetoxymethylester (fura-B/AM) at 37 "C in a water bath with occasional shaking. Suspensions were then diluted 1:5 with warm KRH, incubated for a further 15 min at 37 "C and washed. Aliquots (lo6 cells/measurement) in Eppendorf tubes were stored on ice until the begining of the assays. Before each assay the cells were pelleted by centrifugation (2 s, lo4 X g) and resuspended in 1.5 ml of warm KRH. The suspension was transferred to a fluorimeter cuvette housed in a thermostatted holder. Fluorescence readings were taken in a Perkin-Elmer LS-5B fluorimeter, at excitation and emission wavelengths of 345 and 490 nm, respectively, with slits of 5 nm. At the end of each measurement, fura-2 fluorescence changes were calibrated in terms of [Ca2+]i as described by a single wavelength protocol similar to that reported by Arslan et al. (23) for quin-2. The K d of the fura-2/Ca2+ interaction was taken to be 225 nM (22). [Ca2+Ii measurements were routinely corrected for cell autofluorescence and dye leakage. This same protocol of [Ca2+Ii measurement was used with both 3T3 fibroblasts and SKNBE neuroblastoma cells after detachment from the dish by trypsin treatment as described in Ref. 8. 45Ca2+ Influx-For the measurement of unidirectional ,Ta2+ influx, PC12 cells were harvested, washed twice with warm KRH, and resuspended in the same medium (5 X lo6 cells/ml). The cells were preincubated at 37 "C for 3 min before addition of either BK (100 nM) or vehicle. After a further 6 min at 37 "C. 45Ca2+ (10 &i/ml) was added together with [3H]sucrose (1 pCi/ml, as a marker for extracellular space) with or without GF. Aliquots (200 rl) of the incubated samples were withdrawn at the indicated time points, mixed with EGTA and ruthenium red to instantaneously block influx, and rapidly spun through a mineral oil layer (24).
PPI Hydrolysis-PC12 cell monolayers were incubated for 24 h with 3-5 pCi/ml of my~-[~H]inositol in basal Eagle's medium (inositol-free) supplemented with dialyzed fetal calf serum (10%) and horse serum (5%). Before use, cells were detached from the dish, washed twice with warm KRH, and preincubated at 37 "C in KRH with or without 10 mM LiCl for 10 min, after which the various agents were added. Two different experimental protocols were investigated. In the A IGF-I first, insulin or EGF were added to the cells, and the incubations terminated 120 s later. In the second set of experiments BK was added first, followed 6 min later by 30-120-s treatments with insulin, IGF-I, or EGF. In all cases, the experiments were terminated by the addition of trichloroacetic acid (7.5%, final concentration) and the preparations were stored on ice for 30 min. Total inositol phosphates (Ins-P) generated during cell incubation were recovered by anion exchange chromatography (25). For the separation of the various Ins-P, the HPLC procedure described by Batty et al. (26) was used in ether-washed trichloroacetic acid supernatants supplemented with a phytic acid hydrolysate to improve Ins-P, recovery (27). Radioactivity of total and separated Ins-P was counted in a Beckman LS 7500 spectrometer. Trichloroacetic acid-insoluble material was solubilized in 500 ~1 of 0.2 N NaOH, and protein assayed (28) using bovine serum albumin as the standard. Experiments were carried out in quadruplicate.
DNA Synthesis-Swiss 3T3 fibroblasts were plated in 24-well plates at 50,000 cells/well. Two days thereafter, when cells were confluent, the media was removed and substituted with serum-free Dulbecco's modified E8.gle's medium. Thirty h later, BK and/or GF were added to cdiiures. A [3H]thymidine (3 pCi/well) pulse was administered during the last 6 h of a 24-h incubation with the different agents. Monolayers were then washed twice with cold phosphatebuffered saline, and acid-insoluble radioactivity extracted and counted as described (29

Ka2+Ii in PC12 Celk
Effects of GF with and without BK-Figs. 1-3 illustrate the [Ca"']i effects induced in PC12 cell suspensions loaded with fura-2 by various GF, administered either alone or in combination with the natural nonapeptide BK. As can be was observed (see also Ref. 13). Likewise, no change of [Ca2+], was observed with either TGFa and TGFP (10 and 0.8 nM, data not shown and Fig. 3E, respectively) or basic FGF (not shown). A completely different situation was observed, however, when GF were administered to PC12 cells previously challenged with BK. The [Ca2+]i effect of this peptide is due lCoz'l,, nM [Caa+]i effects of GF administered to BK-primed PC12 cell suspensions in Ca2+-containing and Ca2+-free media. In A and B, excess (3 mM) EGTA was added 1 min before BK (100 nM). The GF used in A was EGF, but identical results were obtained with insulin and IGF-I. In B, CaC12 (3 mM) was reintroduced into the medium shortly before EGF addition. In C, excess EGTA was administered after BK, shortly before EGF administration. Concentration of cells and GF were described in the legend to Fig. 1.

FIG. 3. Unidirectional 4aCa2+ influx induced in PC12 cell suspen-
sions by BK f GF. PC12 cells (5 X lo6/ ml) were incubated with BK (100 nM for 6 min) after which '%az+ (10 pCi/ml) and [3H]sucrose (1 pCi/ml) were added with or without EGF (40 nM) or IGF-I (20 nM). At the indicated time points, samples containing IO6 cells were withdrawn, mixed with a drop of EGTA and ruthenium red (to block Ca2+ influx), and rapidly centrifuged through a mineral oil layer. Results are averages f S.E. of quadruplicates of an experiment that was repeated twice. *, P < 0.05 by unpaired t test.

Lent of BB Receptor Signals
to a dual mechanism: release of Ca2+ from intracellular store(s) mediated by Ins-1,4,5-P3, and stimulation of Ca2+ influx across the plasma membrane (27,30,31). The first such mechanism is responsible for a large part of the [Ca2+Ii spike during the initial 60-120 s, whereas the second mechanism accounts for the subsequent, prolonged steady-state [Ca2+Ii elevation . Insulin, IGF-I, EGF, FGF, and TGFa (but not TGFP, not shown), administered during the [Ca2+Ii steady-state induced by BK, were found able to trigger further increases of [Ca"],. Such increases became appreciable after a lag of 20-40 s from the GF application, and remained visible for several minutes thereafter (Fig. 1, B-F, and see below).
In order to clarify the source(s) of the [Ca2+Ii increases induced by GF, experiments were carried out in which the Ca2+ concentration in the medium, [Ca2+],, was switched from millimolar to very low levels by the use of EGTA (Fig. 2, A-C). When excess EGTA was added to the cells before BK, the initial [Ca2+]; peak induced by the peptide was largely maintained, whereas the subsequent steady-state plateau and the GF-induced transients were lost (Fig. 2 A ) . In contrast, the GF-induced [Ca"']i transients were still observed when either Ca2+ was reintroduced into the medium before GF addition ( Fig. 2B) or EGTA was administered after BK, shortly before GF, (Fig. X ) . The effect of GF on CaZ+ homeostasis in BKprimed cells was also investigated by unidirectional 45Ca2+ flux experiments. 45Ca2+ was administered to the cells (with or without a GF) 6 min after priming with BK, i.e. during the [Ca2+Ii steady-state phase induced by the peptide. As can be seen in Fig. 3, at 60-120 s the tracer accumulation was found to be significantly greater in the cells treated with either EGF or IGF-I with respect to controls. Taken together, the results of Figs. 2 and 3 demonstrate that, similar to the BK-induced [Ca2+li increases, also the GF-induced transients include two components, one originated from intracellular Ca2+ stores, which is maintained even after [Ca2+j0 chelation (Fig. 2C), the other from extracellular Ca2+, which is responsible for both the [Ca2+], increase after Ca2' reintroduction into the medium (Fig. 2B) and the increased 45Ca2+ accumulation (Fig.  3). The intracellular store activated by the GF was apparently depleted (or otherwise inactivated) within minutes from the application of BK in the EGTA-containing medium, and required therefore the chelator and GF to be administered in rapid sequence in order to be revealed (compare Fig. 2 [Ca2+]i effects of two successive administrations of some or different GF, given to PC12 cell suspensions before or after BK (100 nM). Concentrations of cells and GF as described in the legend to Fig. 1, except for TGFp, that was given at 0.8 nM. increase expected from a second, post-BK administration of the same factor ( Fig. 44 and data not shown). In contrast, the effect of the post-BK administration was maintained, although only in part, when the pre-BK treatment was made with a different GF (compare Fig. 4, A, C, and D ) . Likewise, when saturating concentrations of different GF were administered in sequence after BK, each of them induced measurable responses (Fig. 4B), but their sum was smaller than the sum of the responses elicited by the same concentration of GF in separate cell aliquots. In contrast, when the same GF was administered twice to the same aliquot of BK-pretreated cells, the second administration was ineffective (not shown in Figures). Among the GF used, only TGFP was unable to induce the post-BK [Ca2+]i transient in PC12 cells. In this respect, however, it is worth mentioning that the complement of receptors for this GF has been recently shown to be very low in this cell line (32).

Concentration and Time Dependence of the [Ca2+]
i Effects of GF-EGF proved to be the most potent (apparent ECb,,: 1 nM with respect to 6 and 10 nM for IGF-I and insulin, respectively) and efficaceous (approximately 90 and 40% rise over the BK-induced steady-state plateau with maximal concentrations of EGF uersus insulin and IGF-I, respectively) stimulator of post-BK [Ca2+]i transients (Fig. 5 ) . The transients were also found to differ in their kinetics depending on the nature and concentration of the factors employed (Fig. 6). With EGF the lag phase was shorter and the [Ca2+], rise steeper with respect to insulin and IGF-I (Fig. 6, A and B ) .
[Ca2+], transients by EGF, on the other hand, were distinctly shorter lived (on the average 5.0 k 0.3 min, n = 12) with respect to those induced by insulin and IGF-I (7.4 f 0.4 and 7.1 f 0.7 min; n = 29 and 12, respectively) (Fig. 6B)  [CaZ+], transients from both the time after BK addition and the concentration of the peptide was also investigated. The priming effect of BK became appreciable within 2 min, and persisted for a t least 20 min thereafter (not shown). Fig. 7 compares the effects of the same dose of insulin (200 nM) administered 5 min after 20 or 0.5 nM BK. The time course of the insulin-induced transients was much faster with the higher [BK], but their final size was similar (Fig. 7, inset). When, however, insulin was administered after an even lower [BK] (0.1 nM) (Fig. 7C), which by itself still induced a small [Caz+]i increase, no response could be seen. Studies with specific BK inhibitors were carried out to identify the receptor responsible for cell priming. The BI antagonist, des-Argg[Leu8]BK was ineffective. In contrast, the post-BK, GF-induced [Ca2+]i transients were completely inhibited by the specific Bz blocker Arg"-[Hyp3,Thi5>',~-Phe7] BK (33), administered before or even immediately after the GF, i.e. during the lag phase preceding the GF-induced [Ca2+], increase (Fig. 7, D and E ) . When  increases, on the one hand, demonstrates the strict dependence of BK priming from Bz receptors; on the other hand, suggests that priming might not be due to a second messenger, whose dissipation after receptor blockade is expected to occur in a longer time.

Mechanism of the GF-induced [Ca2+], Increases
High Ka2+Ii and PPI Hydrolysis Are Not Sufficient to Prime PC12 Celk-The possibility that the priming effect of BK was the mere consequence of the [CaZ+li increase induced by the peptide was excluded by experiments in which [Ca2+It was increased by a different mechanism, i.e. high K+ depolarization (34). Treatment with 50 mM KC1 caused [Ca2+]; to increase approximately to the steady-state level achieved after 50 nM BK, yet insulin administered 4 min after K+ was without effect on (Ca2+]i (Fig. 8A). This latter result was not due to a voltage-dependence of the GF effect because administration of 50 mM K' after BK did not inhibit the [Ca2+]; rise induced by a subsequent administration of insulin (Fig.  8B). The negative results with high K+ exclude the involvement in the GF effect of voltage-gated Ca2+ channels, i.e. the channels responsible for the [Ca2+Ji effects of depolarizing agents (35,36). This latter conclusion is supported also by experiments with blockers of voltage-gated Ca2+ channels, such as verapamil. At a concentration (10 p~) known to block the K+-induced [Ca2+]i increase (34,36), this drug failed to modify appreciably both the BK-induced and the subsequent GF-induced [Ca2+Ii transients (not shown in Figures).
Activation of muscarinic receptors by carbachol is known to induce in PC12 cells [Ca2+Ii transients due to mechanisms similar to those triggered by BK, i.e. PPI hydrolysis with Ca2+ release from internal stores accompanied by stimulated Ca2+ influx (37,38). Carbachol (100 p~) , however, failed to mimic the priming effect of BK on the insulin and EGF-induced [Ca2+]; effects (Fig. 8C). Negative results similar to those with carbachol were obtained with NGF (Fig. 8 0 ) and ATP (not shown). Both these agents are able to induce [Ca2+Ii increases in PC12 cells within seconds from their administration (Ref.

13).'
Prolonged (1-2 weeks) treatment with NGF causes PC12 cells to modify their phenotype by sprouting long neurites and acquisition of various receptors and channels (20). Other receptors, however, such as those for EGF, have been found recently to be decreased by the NGF-induced differentiation (39). Accordingly, the [CaZ+li effects of the various agents were differently modified by NGF-induced differentiation. The response to BK was unchanged, but the subsequent responses to both insulin and EGF were distinctly smaller (24 -+ 11% and 39 k 17% of controls, mean f S.D., n = 3, for insulin and EGF, respectively) than in NGF-untreated cells (compare Figs. 1B and 8E).
Another treatment that modified the GF-induced response was that with phorbol esters. Phorbol12-myristate 13-acetate (100 nM), administered before or after BK, caused the [Ca2+]; transients induced by insulin and IGF-I to be partially inhibited (72 k 19%, n = 4, and 69 _+ 12%, n = 3, respectively) and those by EGF to be completely blocked (Fig. 8, F and G ) .  for 6 min and then with or without GF (EGF or IGF-I, 20 nM) for additional periods of 30, 80, and 120 s. The reactions were quenched with ice-cold trichloroacetic acid, and neutralized extracts analyzed by anion exchange HPLC as in Ref. 26. The results are expressed as percent of BK-stimulated Ins-P production (mean e S.D. of duplicates). By themselves, EGF and IGF-I did not induce any detectable increase of the analyzed Ins-P. T h e radioactivity of BK-stimulated PC12 cells was 261, 217, and 251 cpm/mg of protein for Ins-1,3,4-P3, Ins-1,4,5-P3, and IP,, respectively. . Treatment with this toxin is known to cause ADP-ribosylation of the G proteins of the Go and Gi class (40). PPI Hydrolysis-The BK-induced [Ca2+Ii changes found to be reinforced by GF are known to be triggered via the activation of B2 receptors and the stimulation of PPI turnover (27). Experiments were therefore carried out to establish whether the administration of GF after BK results in a further increase in the generation of Ins-P, in particular of Ins-1,4,5-P3, the messenger responsible for Caz+ release from intracellular stores; and of Ins-P4, which has been implicated in the control of Caz+ influx (10, 11). When total Ins-P were measured together in cells primed with BK during incubation with LiCl (to block Ins-1-P phosphatase), no appreciable effect of either EGF or insulin (administered for 2 min, 6 min after BK) with respect to BK alone could be detected (data not shown). In contrast, without LiC1, various Ins-P, separated by HPLC, were found to be increased after GF addition (Table  I). At 30 s (i.e. the begining of the GF-induced [Ca2+]i increase), Ins-1,4,5-P3 and Ins-P4 were appreciably (about 20%) higher in IGF-I and EGF-treated samples with respect to those treated with BK only; a t 80 s (i.e. the time of the GFinduced [CaZ+li peak), Ins-l,4,5-P3 was back to control values, but Ins-P4 was still distinctly higher (Table I).

Growth Factor Reinforcement of BP Receptor Signals
[Ca2+/, in Other Cell Types Effects of GF Alone, with BK, and with Other Stimulators of PPI Hydrolysis-The effects of GF on [Caz+]i were investigated in two additional cell lines unrelated to PC12: the mouse Swiss 3T3 fibroblasts and human neuroblastoma SKNBE lines, with results ( Fig. 9) highly consistent with those described in the pheochromocytoma cell line. Thus, EGF caused no reproducible [Ca2+Ii increase when administered alone (Fig. 9, A and E), but was effective when administered after BK (Fig. 9, B and F). Similar results were obtained with insulin and IGF-I (not shown). In SKNBE neither high K+ (not shown), nor carbachol and ATP (Fig. 9, G and H ) , induced any priming effect with respect to GF. AS in PC12, these treatments increase [Ca"li in SKNBE cells, one by opening voltage-gated Ca2+ channels, the others by activating muscarinic and purinergic receptors. Finally, the specificity of the BK priming could be further investigated in Swiss 3T3 fibroblasts, which express various other receptors coupled to PPI hydrolysis. As can be seen in Fig. 9, C and D, neither bombesin nor vasopressin were able to reproduce the BK priming. Effects of BK and GF on Swiss 3T3 Cell Growth-Since BK and the various GF were able to interact at the level of signal generation, the possibility was raised that they could also trigger interacting effects on cell growth. As can be seen in Fig. 10, incorporation of ["]thymidine into DNA was not significantly modified by either BK (at a concentration that yields maximal [Ca2+Ii effects) or EGF administered alone, whereas insulin triggered an approximately 3-fold increase over basal. On the other hand, the combination of BK with either one of the two GF caused a marked increase of the incorporation: over &fold for insulin, and almost 4-fold for EGF.

DISCUSSION
The results herewith described demonstrate an up to now unreported effect of insulin, EGF, and various other GF, i.e. the reinforcement of the PPI hydrolysis and [Ca2+li increase effects elicited by the nonapeptide BK through the activation of Bz receptors. This new type of cross-talk was studied in detail in PC12 pheochromocytoma cells, and then demonstrated to exist also in the other two unrelated cell types we have investigated, SKNBE and Swiss 3T3 cell lines. The effect is therefore widespread and might be general to the cells endowed with both BS and GF receptors. By the reinforcement of the B, receptor signaling the GF become capable of increasing [CaZ++li. This might be of physiological importance because of a role of [Ca2+]i in the transduction of mitogenic messages (1, 5). Up to now, however, of the typical GF only PDGF (4, 5 ) and, in a few cell lines, EGF (6-9, 29) had been found to induce large increases of [Ca*+]j. In the case of insulin, a rise of [Ca2+]i was predicted based on indirect evidence (41) but the prediction was never documented. In the present study we have found that the combination of BK with either insulin or EGF is able to trigger a marked mitogenic effect in Swiss 3T3 fibroblasts, although the peptide (and also EGF) was devoid of any effect on cell growth when given alone.
Priming by BK of the GF-induced [CaP+li Increase-The priming effect of BK appears to consist in the activation of Bz receptors. In fact, blockade of these receptors with the specific competitive antagonist, Arga-[Hyp3,Thi5*8,~-Phe7]BK was found to rapidly dissipate the priming, and thus render GF ineffective on [Ca2+]i. Moreover, when GF were administered first, no appreciable potentiation of the [Ca"]; transients induced by the subsequent addition of BK was ever observed. A variety of studies excluded the involvement in the priming process of mediators (or events) that are known to be generated (or to occur) distally with respect to BP receptor activation (27): the generation of 1,2-sn-diacylglycerol, because the analog phorbol esters did not increase, but rather block the effect of GFs, especially of EGF; the increase of [Ca2+Ii, because a variety of other treatments that induce this effect to an extent similar or even higher than BK, were ineffective in priming; the changes of membrane potential and conductance (hyperpolarization; activation of voltagegated Ca2+ channels), because BK was able to prime PC12 cells also when the latters were depolarized by high K' or  and lo7 (SKNBE, traces E-H)/ml and loaded with fura-2 as described under "Experimental Procedures." BK was administered at 100 nM, carbachol (CCh) at 500 g~, ATP at 10 g~, bombesin (BOM) at 50 nM, vasopressin (VASOP) at 1 g~. In the traces shown, the GF used was EGF (20 nM). treated with the channel blocker verapamil; the generation of cyclooxyge.nase products, because inhibitors of this pathway failed to modify BK priming. Finally, the lack of a priming effect by carbachol and ATP (in PC12 and SKNBE cells), bombesin, and vasopressin (in Swiss 3T3 fibroblasts), ix. by agents addressed to various receptors coupled to PPI hydrolysis, excludes that priming of GF-induced [Ca"], transients is a direct consequence of the latter process per se.

Mechanismfs) of GF-induced Bz Receptor Reinforcement-
The BK-primed [Ca2+Ii rises induced by insulin, IGF-I, and EGF appear to be generated independently, with each GF working through the activation of its specific receptor. In fact, (i) the concentration dependence of the [Caz+li rises was found to be similar to those of other physiological effects of the GFs; (ii) the concentration dependence of the insulin and IGF-Iinduced responses were almost identical, while receptor sharing of these two GF implies different binding affinities (14,15); and (iii) at least part of the responses were maintained in cells exposed to another GF before the application of BK, compared to the disappearance observed when the same GF was used for both, pre-and post-BK administrations. On the other hand, the responses triggered by successive administrations of two G F after BK were less than additive. These results suggest that, although originated at different receptors, the signals triggered by the various GF converge intracellularly to activate partially overlapping mechanisms. The nature of these mechanisms has been elucidated only in part.
The PPI hydrolysis experiments revealed that GF effects can be triggered when the generation of Ins-1,4,5-P3 (and of its phosphorylation product, Ins-P4) is already stimulated by BK, and consists in an appreciable increase of the Ins-1,4,5-P3 generation process. The occurrence of reinforcement over a large spectrum of [BK] seems to exclude that the GF effect consists in an increase of Bz receptor affinity. Rather, the process could consist in the transient recoupling of Bz receptors to the PIPz-specific phospholipase C, with consequent increase of [Ca2+Ii via both release from intracellular stores and activation of plasma membrane conductance. Receptors for the GF competent for reinforcement, in particular insulin, EGF, and IGF-I, are known to possess an intrinsic tyrosine kinase activity, which is activated following receptor occupancy (1-3). Tyrosine phosphorylation of some cellular proteins is stimulated also by FGF (42). The apparent specificity for the B2 receptor of the reinforcement process suggests the possibility that this receptor is, directly or indirectly, modified by the tyrosine kinase activity of GF receptors. Alternatively, the target of the GF receptors could not be the Bz receptor but the G protein that couples this (and possibly others, but not all) receptor to phospholipase C (40,43,44). In conclusion, the present results demonstrate that stimulation of PPI hydrolysis and increased [Ca2+], are part of the array of intracellular signals that can be elicited by not only PDGF, but also other GF and even insulin, in various cell types, via a reinforcement of transmembrane signaling. Whether the rein-