Nitric oxide action on growth factor-elicited signals. Phosphoinositide hydrolysis and [Ca2+]i responses are negatively modulated via a cGMP-dependent protein kinase I pathway.

The role of nitric oxide (NO) in the phosphatidylinositol 4,5-bisphosphate (PIP2) hydrolysis and intracellular Ca2+ release responses induced by epidermal, platelet-derived, and fibroblast growth factors was investigated in three cell lines, a clone of NIH-3T3 fibroblasts overexpressing epidermal growth factor receptors and the tumoral epithelial cells A431 and KB. In all three cell types, pretreatment with NO donors decreased growth factor-induced PIP2 and Ca2+ responses, whereas pretreatment with NO synthase inhibitors increased them. The Ca2+-dependent PIP2 hydroysis induced by micromolar concentrations of the Ca2+ ionophore, ionomycin, was also modulated negatively and positively by NO donors and synthase inhibitors, respectively. In contrast, the Ca2+ content of the intracellular stores was unaffected by the various pretreatments employed. NO donors and synthase inhibitors induced an increase and decrease, respectively, of the intracellular cGMP formation in all three cell lines investigated. All of the effects of the NO donors were mimicked by 8-bromo-cGMP administration and abolished by pretreatment with the specific blocker of the cGMP-dependent protein kinase I, KT5823, which by itself mimicked the effects of the synthase inhibitors. Together with previous observations on G protein-coupled receptors, the present results demonstrate that PIP2 hydrolysis and Ca2+ release occur under the feedback control of NO, independently of the phospholipase C (β, γ, or δ type) involved and of the mechanism of activation. Such a control, which appears to be effected by the cGMP-dependent protein kinase I acting at the level of the phospholipases C themselves, might ultimately contribute to the inhibitory role of NO on growth previously observed with various cell types.

Individual molecules of the signal transduction cascades turned on by receptor agonist binding can play important roles not only in the intracellular activation process but also in the fine feedback regulation of signaling itself. In this respect particular attention has been devoted to nitric oxide (NO). 1 In the cells competent for the Ca 2ϩ -dependent, constitutive forms of NO synthases (NOSs), this highly reactive radical gas, generated in response to appropriate increases of the cytosolic Ca 2ϩ concentration ([Ca 2ϩ ] i ), works as the controller of a number of enzymes including guanylyl cyclase (1). The ensuing increase of cGMP formation, with activation of cGMP-dependent protein kinase I (G kinase), yields responses that may be variable from cell to cell (2,3). In the case of receptors coupled to phosphatidylinositol 4,5-bisphosphate (PIP 2 ) hydrolysis via the activation of heterotrimeric G proteins of the Gq family (4), negative modulations by NO have been described, with decreased generation of the two second messengers, inositol 1,4,5-trisphosphate (IP 3 ) and diacylglycerol, and ensuing blunting of Ca 2ϩ release from intracellular stores (3). With these receptors the modulation was shown to depend upon G kinase activation, and the site of action was proposed at the G protein/phospholipase C (PLC) interface (5). PIP 2 hydrolysis is induced not only by G protein-coupled receptors but also by growth factor receptors, working, however, on different PLCs and by a different activation process, i.e. by direct tyrosine phosphorylation of the PLCs of the ␥ family rather than by G protein activation of those of the ␤ family (4,6). Because of these distinct molecular and functional differences, we thought it worth investigating whether NO and cGMP had any effect on the growth factor-induced PIP 2 hydrolysis and [Ca 2ϩ ] i responses.
The results reported here indicate in three types of cells that the above responses induced by growth factors are indeed inhibited by NO working through the cGMP/G kinase I pathway. Similar inhibition by NO was observed also when the same responses were induced by persistent [Ca 2ϩ ] i increases triggered by the Ca 2ϩ ionophore, ionomycin. Under the latter condition activation is not restricted to a single family of PLCs but affects them all: ␤, ␥, and ␦ families altogether (7)(8)(9)(10). We conclude therefore that the G kinase-sustained negative modulation is a [Ca 2ϩ ] i and NO-induced feedback regulation process occurring most probably at the level of PLC.
Cell Culture and Preparation-The NIH 3T3 cell clone EGFR-T17, overexpressing the human EGF receptor, A431, and KB carcinoma cell lines were routinely grown as previously described (11) and used before the 10th week from thawing. The day of the experiment, cell monolayers were detached from Petri dish by gentle trypsinization and resuspended in Krebs-Ringer-Hepes (KRH) medium containing 125 mM NaCl, 5 mM KCl, 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 2 mM CaCl 2 , 6 mM glucose, 0.8 mM L-arginine, 25 mM Hepes-NaOH (pH 7.4). Cells were then washed three times by centrifugation and resuspended as required by the various experimental procedures. Cell viability in the presence or absence of the various drugs employed was more than 95%, as assessed by the trypan blue exclusion test.
[Ca 2ϩ ] i Measurements-Cell suspensions were loaded with the Ca 2ϩsensitive dye fura-2/AM (3 M final concentration) for 30 min at 25°C in KRH medium and kept at 37°C until use. Cell aliquots (4 ϫ 10 6 cells) were then transferred to a thermostatted cuvette in a Perkin-Elmer LS-5B fluorimeter and maintained at 37°C under continuous stirring. Preincubations with the various drugs interfering with the L-arginine/NO pathway were carried out for 15 min at 37°C, a condition selected in preliminary experiments to be long enough for maximal drug effects in the time range effective in other cell types (12)(13)(14)(15). One min before addition of the Ca 2ϩ -releasing agents (EGF, PDGF, FGF, UTP, or thapsigargin), the samples were supplemented with excess EGTA (2 mM; estimated extracellular [Ca 2ϩ ], Ͻ10 Ϫ8 M). Under this condition only Ca 2ϩ release from intracellular stores, rather than Ca 2ϩ influx, can be detected. Traces were recorded and analyzed as described elsewhere (16). Results shown are averages of 8 -10 separate experiments Ϯ S.D.

45
Ca 2ϩ Measurements-During the last 72 h of cell growth the incubation medium was supplemented with 45 Ca 2ϩ (4 Ci/ml). Labeled cells were extensively washed and resuspended in KRH medium. A pellet obtained from the suspension was used for the measurement of total cell 45 Ca 2ϩ content. The rest was incubated at 37°C in EGTA-containing Ca 2ϩ -free KRH medium. At different times, aliquots of 4 ϫ 10 6 cells were centrifuged, and the 45 Ca 2ϩ recovered into the medium was assayed in a Beckman ␤-counter (for further details, see Fasolato et al. (17)). Results shown are means of four separate experiments Ϯ S.D.
Inositol Phosphate Measurements-Cells labeled for 24 h with 3 Ci/ml myo- [2-3 H]inositol in basal Eagle's diploid modified medium containing 8% inositol-free fetal calf serum were detached, washed, and resuspended in KRH medium. After preincubation (15 min at 37°C) with the various drugs interfering with the L-arginine/NO pathway, in the presence (total inositol phosphates (IPs) measurements) or absence (IP 3 measurements) of 20 mM LiCl, aliquots of 4 ϫ 10 6 cells were challenged with either growth factors or ionomycin and the reaction stopped by the addition of ice-cold formic acid (20 mM final concentration). Samples were kept on ice for 30 min, then centrifuged, and the supernantants were loaded onto anion-exchange columns. Radioactive IPs or IP 3 were separated by stepwise elution as previously described (18), and radioactivity was counted in a Beckman ␤-counter. Results shown are means of four to six independent experiments Ϯ S.D. Control experiments in which production of IPs and IP 3 was stimulated in Ca 2ϩ -free KRH medium gave similar results.
Measurements of cGMP Levels-Cell suspensions were incubated for 15 min at 37°C in KRH medium supplemented with 0.6 mM 3-isobutyl-1-methylxanthine, with or without L-NIO (200 M). NOS activity was stimulated by cell exposure to EGF, PDGF, or FGF for 5 min at 37°C. As a control, cGMP formed upon stimulation for 5 min with SNP (30 M) was also measured. Reactions were terminated by addition of ice-cold trichloroacetic acid (final concentration, 7.5%). After ether extraction, cGMP levels were measured by radioimmunoassay kit (DuPont) and normalized on cellular proteins, determined using the bicinchoninic acid assay (BCA protein assay reagent; Pierce). Results shown are means Ϯ S.D. of three separate experiments.

RESULTS
All the results shown are from experiments in which L-NIO was used as a NOS inhibitor, and SNP as a NO donor. Qualitatively similar results were obtained using the NOS inhibitor, N -nitro-L-arginine methylester, and the NO donor, S-nitroso-N-acetylpenicillamine. With the less active enantiomer of Nnitro-L-arginine methylester, N -nitro-D-arginine methylester, the results did not differ significantly from those obtained with untreated, control cells.
Effects of NO on Growth Factor-induced Ca 2ϩ Release-Suspensions of fura-2 loaded EGFR-T17, A431, and KB cells were incubated for 15 min at 37°C in KRH medium, with or without the NOS inhibitor, L-NIO (200 M), the NO donor, SNP (30 M), or the membrane permeant, stable analogues of cGMP or cAMP, 8-Br-cGMP (200 M) and 8-Br-cAMP (200 M). They were then supplemented with excess EGTA and challenged with increasing concentrations of EGF. Under these experimental conditions any [Ca 2ϩ ] i changes can be safely attributed to Ca 2ϩ release from intracellular stores (4). In all three cell lines investigated, pretreatment with L-NIO potentiated Ca 2ϩ release induced by EGF, while SNP exerted an inhibitory effect with respect to untreated, control cells (Fig. 1). With 8-Br-cGMP, the effect observed on Ca 2ϩ release was analogous to that of SNP, while 8-Br-cAMP was without any appreciable effect (Fig. 1).
EGFR-T17 cells are known to exhibit [Ca 2ϩ ] i responses not only to EGF but to other growth factors, i.e. PDGF and FGF (19). The effects of L-NIO, SNP, 8-Br-cGMP, and 8-Br-cAMP on Ca 2ϩ release induced by the latter agonists were therefore investigated, with results qualitatively similar to those induced by EGF. Ca 2ϩ release elicited by PDGF and FGF was potentiated when cells were preincubated with L-NIO, inhibited after treatment with SNP or 8-Br-cGMP, and almost unaffected by 8-Br-cAMP (Fig. 2, A and B). The effects of NO-modulating drugs on Ca 2ϩ release induced by growth factors were then compared to those exerted on the responses elicited by activation of a G protein-coupled receptor. Fig. 2C shows the results obtained with UTP, a receptor agonist specific for the puriner- gic P 2U receptor (20). As already demonstrated in PC12 cells (15), preincubation with L-NIO increased, that with SNP or 8-Br-cGMP decreased, the UTP-induced Ca 2ϩ release to extents similar to those observed with growth factor-induced [Ca 2ϩ ] i responses.
Effects of NO on Ca 2ϩ Storage-In all cell types exchangeable Ca 2ϩ is known to be distributed into at least three major pools: (i) the IP 3 -sensitive Ca 2ϩ store, a subcompartment of the endoplasmic reticulum endowed with SERCA ATPases; (ii) an additional endoplasmic reticulum-located Ca 2ϩ pool also accumulated by SERCAs but insensitive to IP 3 ; (iii) a large, still poorly characterized pool, insensitive to IP 3 and devoid of SER-CAs, that can be discharged by Ca 2ϩ -specific ionophores (17,21). In order to investigate the effects of NO on these pools, cell monolayers were loaded to equilibrium (72 h) with 45 Ca 2ϩ , pretreated with L-NIO or SNP for 15 min at 37°C, and then challenged in sequence with EGF (to release Ca 2ϩ stored in the IP 3 -sensitive Ca 2ϩ store) followed by thapsigargin (an irreversible SERCA blocker that induces within a few minutes the leakage and emptying of Ca 2ϩ from the entire endoplasmic reticulum) (21), and finally by ionomycin (an electroneutral Ca 2ϩ ionophore) (15,17,21). Fig. 3A summarizes the results obtained with EGFR-T17 cells; qualitatively similar observations were made with A431 and KB cells. Pretreatment with L-NIO increased, that with SNP decreased the amount of 45 Ca 2ϩ released by EGF (100 nM) (Fig. 3A). In contrast, release by the subsequent administration first of thapsigargin (100 nM) and then of ionomycin (1 M) was unchanged by both pretreatments. Moreover, neither L-NIO nor SNP appeared to affect the basal leak of Ca 2ϩ from the resting cells. To further confirm these findings, thapsigargin-induced Ca 2ϩ release was analyzed also by the fura-2 technique. Again, no significant differences in Ca 2ϩ release induced by the SERCA blocker were revealed in cells pretreated with L-NIO or SNP in comparison with controls (Fig. 3B).
Effects of NO on PIP 2 Hydrolysis-The effects of L-NIO, SNP, and 8-Br-cGMP on growth factor-induced PIP 2 hydrolysis were next investigated. Fig. 4A shows the results obtained with EGF in EGFR-T17 cells. Pretreatment with L-NIO potentiated total IP accumulation stimulated by increasing concentrations of the growth factor, administered for 20 min in the presence of 20 mM LiCl. Under parallel conditions both SNP and 8-Br-cGMP exerted an inhibitory effect. Similar results were observed when the time-course of IP 3 production, generated after administration of 30 nM EGF, was assayed (Fig. 4B). The results obtained with PDGF and FGF were consistent with those of EGF (not shown). In additional experiments, total IP accumulation was investigated in EGFR-T17 cells in which PLC activity was stimulated by the Ca 2ϩ ionophore, ionomycin, administered at concentrations (1 or 3 M) that in preliminary experiments gave rise to persistently high values of [Ca 2ϩ ] i (10 Ϫ6 -10 Ϫ5 M).
With both concentrations of ionomycin tested, preincubation with L-NIO yielded higher levels, and those with SNP and 8-Br-cGMP lower levels of total IP accumulation (Fig. 4C).
Role of G Kinase I as NO Effector-Many effects of NO are known to be indirect, mediated by increases in cytosolic cGMP levels and activation of G kinase I (3,12). The involvement of such a mechanism in our experiments was already suggested by the parallel effects of SNP and 8-Br-cGMP on both [Ca 2ϩ ] i and total IP accumulation. The lack of effect of 8-Br-cAMP excludes that the effects of 8-Br-cGMP are mediated via cross activation of protein kinase A, a mechanism known to take place in other cell systems (22,23). The drugs employed in the present study did indeed induce considerable changes of the endogenous cGMP production in all the cell lines investigated: decrease (Ϫ30 -40%) with L-NIO and increases (10 -16 fold; 3.5-5 fold) with SNP and EGF, the latter almost completely prevented by prior L-NIO treatment (Table I).
In further experiments SNP preincubation of EGFR-T17 cell suspensions was carried out with or without KT5823, a widely employed inhibitor of the G kinase I activity (15,24,25). Fig.  5A shows that in the presence of KT5823 (10 M) the inhibitory effect of SNP on EGF-induced Ca 2ϩ release was almost completely abolished. Similarly, SNP inhibition of total IP accumulation triggered by either EGF or ionomycin was largely prevented by the kinase blocker (Fig. 5, B and C). KT5823 was also administered alone or in combination with L-NIO. Fig. 6 shows the results obtained in EGFR-T17 cells challenged with EGF (A and B) or ionomycin (C). The effects of the kinase blocker on [Ca 2ϩ ] i variations and total IP accumulation induced by either stimulant resembled those induced by L-NIO; when KT5823 and L-NIO were administered together, no additive effect was measured. Similar results were found when the effects of KT5823 were investigated in A431 and KB cells (not shown).

DISCUSSION
The results reported here demonstrate that NO has a role in the chain of intracellular events elicited by activation of growth factor receptors. While the Ca 2ϩ storage machinery is unaffected by cell treatment with NO, the gaseous messenger is shown to modulate negatively PIP 2 hydrolysis and the ensuing generation of IP 3 . An important consequence is the reduction of the growth factor-induced release of Ca 2ϩ from the intracellular stores. Although obtained not by direct application of NO but by a pharmacological approach, our results appear unambiguous because of the contrast between the effects of NO donors, which induce release of the gas within the cells, with those of NOS blockers, which preclude the synthesis of endogenous NO. The fact that NO-induced negative signal modulations were observed in all three cell types investigated, NIH-3T3, A431, and KB lines, strongly suggests that they have a general significance. Moreover, their appearance with all the growth factors we have employed, i.e. EGF, PDGF, and FGF, suggests these effects to be generated at the level of the common signal cascade activated after receptor binding rather than at a level of receptors themselves. Indeed, it has been recently reported that EGF binding to its receptor is unaffected by NO (26).
Until now, negative effects of NO on PIP 2 hydrolysis and Ca 2ϩ release had been reported only with receptors coupled to PLC via heterotrimeric G proteins. In the latter case, the site of NO action was proposed to occur at the G protein/PLC interface (5). The PLCs activated by G proteins, however, are a family of isoenzymes (defined as PLC␤) molecularly and functionally different from those activated by growth factors, the PLC␥, which are activated at the receptor level by direct tyrosine phosphorylation (4,6). Taken together, the present and previous results indicate therefore that inhibition by NO is a widespread regulatory process that involves many (possibly all) types of transductive PLCs. In fact, also PIP 2 hydrolysis induced by administration of a Ca 2ϩ ionophore, ionomycin, a mechanism effective with all types of the enzyme known so far (see Refs. 7,8,9, and 10, for PLC␤ 1, 2, and 3 , PLC␥ 1 , PLC␥ 2 , and PLC␦ 1 , respectively), was inhibited by NO. The inhibition by NO of PLC activity appears to be mediated by accumulation of cGMP and activation of G kinase I. Whether this kinase phosphorylates PLC(s) directly, or whether its effect is mediated through the phosphorylation of regulatory, yet unidentified protein(s), remains to be established. Also to be elucidated is the mechanism of inhibition of the PLC␥ activity we have now  NO-and cGMP-induced inhibitory modulation of growth factor receptor function may account for a number of cell growth effects that up to now had remained without an adequate explanation. Inhibition of mitogenesis by the gas and G kinase have been reported in various cell systems including vascular smooth muscle (3), rat hepatocytes (27), and retinal pigmented epithelial cells (28). Moreover, proliferation and development of bone marrow were reported to be inhibited (29), and neuronal and muscular differentiation to be stimulated by NO (30 -32). In the array of intracellular signals elicited by growth factor receptor activation, PIP 2 hydrolysis and [Ca 2ϩ ] i responses are now recognized to promote mitogenesis and differentiation in various cell systems (33,34), while impaired Ca 2ϩ homeostasis or altered Ca 2ϩ release exert an inhibitory effect on growth (see e.g. Refs. 20, 35, and 36). The possibility can therefore be considered that the effects of NO and cGMP on cell growth are mediated, at least in part, by their negative modulatory actions here described. Whether these actions are accompanied by others as yet still unknown, also induced by NO and cGMP, remains to be investigated.
In conclusion, our results expand the importance of the NO-cGMP-mediated modulation in transmembrane signaling, demonstrating that it covers the entire PIP 2 hydrolysis field, independent of the PLC isoform families involved and the mechanisms of their activation. In addition to the inhibition of cell growth discussed above, an important role of the NO modulation of PIP 2 hydrolysis could be in the control of Ca 2ϩ homeostasis, with prevention of excess Ca 2ϩ release. This possibility is supported by the recently recognized, NO-induced positive modulation of the surface Ca 2ϩ channels of the type activated by intracellular Ca 2ϩ release (13,14), which are believed to be responsible for the replenishment of discharged Ca 2ϩ stores.