Thrombin induces a calcium transient that mediates an activation of the Na+/H+ exchanger in human fibroblasts.

The calcium dependence of growth factor-induced cytoplasmic alkalinization was determined in serum-deprived human fibroblasts (WS-1 cells). Intracellular pH (pHi) and intracellular calcium (Ca2+i) were measured using the fluorescent dyes 2',7'-bis-(2-carboxyethyl)-5(6)-carboxyfluorescein and fura2, respectively. Thrombin (10 nM) induced an alkalinization (0.18 +/- 0.01 pH units, n = 23) that was Na+-dependent and amiloride-sensitive, suggesting that the alkalinization was mediated by the Na+/H+ exchanger. Thrombin treatment caused a transient increase in Ca2+i (325 +/- 39 nM, n = 12) that preceded the observed increase in pHi. The increases in Ca2+i and pHi were dependent on the concentration of thrombin. The thrombin-induced increase in Ca2+i occurred in the absence of external calcium indicating that thrombin released calcium from internal stores. Inhibition of the thrombin-induced increase in Ca2+i with 8-diethylaminooctyl 3,4,5-trimethoxybenzoate hydrochloride or bis-(o-aminophenoxy)ethane-N,N,N',N'- tetraacetic acid also inhibited the thrombin-stimulated increase in pHi. The calcium ionophore ionomycin was used to increase Ca2+i independent of growth factor stimulation. When Ca2+i was elevated with ionomycin, a concomitant increase in pHi was observed. The increase in pHi due to ionomycin was dependent on Na+ and sensitive to amiloride. The removal of external Ca2+i inhibited the ionomycin-induced elevation of both Ca2+i and pHi. The ionomycin-induced increases in Ca2+i and pHi were not inhibited by 8-diethylaminooctyl 3,4,5-trimethoxy-benzoate hydrochloride. The results suggest that thrombin treatment can activate the Na+/H+ exchanger, and this activation is mediated by an increase in Ca2+i.

The calcium dependence of growth factor-induced cytoplasmic alkalinization was determined in serumdeprived human fibroblasts (WS-1 cells). Intracellular pH (pHi) and intracellular calcium (Ca2*i) were measured using the fluorescent dyes 2',7'-bis-(2-carboxyethyl)-5(6)-carboxyfluorescein and fura2, respectively. Thrombin (10 nM) induced an alkalinization (0.18 f 0.01 pH units, n = 23) that wasNa+-dependent and amiloride-sensitive, suggesting that the alkalinization was mediated by the Na+/H+ exchanger. Thrombin treatment caused a transient increase in Ca2+i (325 f 39 nM, n = 12) that preceded the observed increase in pH;. The increases in Ca2+i and pHi were dependent on the concentration of thrombin. The thrombin-induced increase in Ca2+, occurred in the absence of external calcium indicating that thrombin released calcium from internal stores. Inhibition of the thrombininduced increase in Caz+i with 8-diethylaminooctyl 3,4,5-trimethoxybenzoate hydrochloride or bis-(o-aminophen0xy)ethane-N,N,N',N'-tetraacetic acid also inhibited the thrombin-stimulated increase in pHi. The calcium ionophore ionomycin was used to increase CaZf independent of growth factor stimulation. When Ca2+i was elevated with ionomycin, a concomitant increase in pHi was observed. The increase in pHi due to ionomycin was dependent on Na+ and sensitive to amiloride. The removal of external Caz+i inhibited the ionomycin-induced eIevation of both Ca2+i and pHi. The ionomycin-induced increases in Ca2+i and pHi were not inhibited by 8-diethylaminooctyl 3,4,5-trimethoxybenzoate hydrochloride.
The results suggest that thrombin treatment can activate the Na+/H+ exchanger, and this activation is mediated by an increase in Caz+i. lar calcium, and an intracellular alkalinization. In order to study the mechanisms of mitogen-induced signal transduction, we have utilized human fibroblasts of fetal skin origin (WS-1 cells) as a model system. Unlike many of the immortal rodent cell lines used for similar studies, human diploid fibroblasts senesce at high passage number (1). Previous work has demonstrated that WS-1 cells can be growth-arrested by incubation for 2 days in serum-free medium (2). Adding 10 nM thrombin causes the turnover of phosphoinositides, a cytoplasmic alkalinization, and a mitogenic response as measured by increased DNA synthesis (2). This report examines the effect of thrombin stimulation on intracellular calcium (Ca2+i) and the relationship between changes in Ca2+; and intracellular pH (pHi), A variety of mitogenic agents (serum, platelet-derived growth factor, bombesin, and others) transiently elevate Ca2+; (3)(4)(5)(6). Growth factor stimulation of PI turnover is thought to mediate this release of calcium from internal stores via the production of inositol 1,4,5-trisphosphate (3, 7). An alternative pathway for increasing Caz+i is suggested by studies using EGF. Stimulation with EGF results in the uptake of external calcium in some cell types (8,9).
An intracellular alkalinization mediated by the Na+/H+ exchanger occurs in a variety of cells following growth factor stimulation (10)(11)(12). In the absence of bicarbonate, the activation of the Na+/H+ exchanger by growth factors raises the pHi to a level permissive for DNA synthesis (13,14).
The transient Ca2+; rise induced by a number of growth factors precedes the increase in pH; (15). The relationship between the two responses is not well understood, but three hypotheses have been proposed I) the observed rise in CaZfi is necessary for the activation of the Na+/H+ exchanger (11,16); 2) the increase in pHi is independent of the rise in Ca2+; (5,12,17); or 3) a certain level of intracellular free calcium is necessary but not sufficient to induce the pHi increase (18, 19). The discrepancies concerning the role of calcium in the activation of the Na+/H+ exchanger suggest the existence of several pathways for the activation of the Na+/H' exchanger within a single cell line (8, 20). Furthermore, the expression of different pathways may vary in closely related cell types (21).
In order to investigate the relationship between mitogenstimulated increases in Ca2+; and pH;, thrombin was used in growth-arrested WS-1 cells to demonstrate the following: 1) thrombin increases both Ca2'; and pH; in a concentrationdependent fashion; 2) ionomycin-induced elevation of Ca2+; results in cellular alkalinization; 3) both thrombin and ionomycin-induced alkalinization are due to the activation of the Na+/H+ exchanger; and 4) inhibition of thrombin or ionomycin-induced increase of Ca2+; inhibits the increase in pH;. These data represent the first demonstration of a thrombininduced release of calcium from internal stores in human fibroblasts and suggest that this rise in Ca2+i can mediate the activation of the Na+/H' exchanger in WS-1 human fibroblasts.

EXPERIMENTAL PROCEDURES
Materials-The pentaacetoxymethyl ester of BAPTA (BAPTA-AM), the tetraacetoxymethyl ester of BCECF (BCECF-AM), and TMB8 were purchased from Molecular Probes (Eugene, OR). Fura 2 pentaacetoxymethyl ester (fura2-AM) and ionomycin were obtained from Calbiochem. Human thrombin (3000 NIH units per mg of protein) was obtained from Sigma. Amiloride was a gift from Merck, Sharp and Dohme.
CeE Growth and Arrest-WS-1 cells are human diploid fibroblasts of fetal skin origin (1). Stock cultures were grown on 100-mm dishes in HEPES-buffered minimal essential media supplemented with 10% fetal bovine serum. Only rapidly growing cells of passage numbers 24 through 29 were used. Experimental cultures were plated at high density (6 X 10' cm") on 10 X 17-mm glass chips in 35-mm dishes with 3 ml of minimal essential media and 10% fetal bovine serum. The cells were growth-arrested 24 h after plating by rinsing twice with NHB (140 mM NaCl, 5 mM KC1,1.25 mM CaCl,, 1 mM MgCl,, 10 mM glucose, 10 mM HEPES, pH 7.4) followed by the addition of serum-free minimal essential media.
Measurement of Intracellular pH-After 2 days of growth arrest, the cells were loaded with the pH-sensitive dye BCECF by placing the chip in 3 ml of 1 p~ BCECF-AM for 35 min. After washing with NHB, the chip was placed in a holder designed to fit in a standard polystyrene cuvette (22) which was inserted into a water-jacketed cuvette holder in an Aminco SLM 8000C Spectrofluorometer (SLM Instruments, Urbana, IL). A syringe connected to the chip holder with Tygon tubing was used to inject solutions without any interruption of the fluorescence measurements. The chip holder was designed to allow the perfusion of a solution from the bottom of the cuvette and the removal by a suction line at the top of the cuvette. All buffer solutions were maintained in a water bath at 37 "C. Growth factors and/or inhibitors were added to 10 ml of NHB immediately prior to injection. Ten ml of a solution could be injected within 10 s and allowed for the full exchange of the solution in cuvette.
The ratio of fluorescence (Rn = fl504/fl440) at the pH-sensitive excitation wavelength of 504 nm versus the pH-insensitive wavelength 440 nm was measured at an emission wavelength of 530 nm. The emissions at each excitation wavelength were averaged over a 1s interval and ratios were constructed at 2-5 intervals. The fluorescence was not corrected for light scattering and autofluorescence as these values amounted to less than 5% of the value of dye-loaded cells. At the end of an experiment, Rn was calibrated using the high potassium and nigericin procedure (145 mM KC1 and 5 pg/ml nigericin) described by Thomas et al. (23). After each experiment, Rn was measured at two pH values ranging from 6.8 to 7.5; all pH values were determined at 37 "C. Over this range, Rn and pH were linearly related according to the least squares line: pHi = 5.84 + 0.20 Rfl (R = 0.99). This equation was used to convert Rn to pH;. A similar calibration curve was obtained in the presence of 1 mM amiloride, suggesting that no correction was necessary when using amiloride-containing solutions.
Measurement of Intracellular Calcium-The cells were grown on glass chips and growth-arrested as in the pH experiments. Following growth arrest, cells were loaded with 1 p M fura2-AM in NHB for 35 min. Loading was performed at 21 "C to minimize endocytosis of the dye and the preferential labeling of endosomal compartments (24). After loading, the cells were rinsed for 15 min in NHB at 37 "C and placed in the spectrofluorometer as in the pH experiments described above.
Fluorescence was measured at a constant emission wavelength of 510 nm. Excitation wavelengths alternated between 340 nm (the peak for calcium-bound fura2) and 380 nm (the peak for free fura2). The emission for each excitation wavelength was averaged over a [1][2][3][4][5] period. Autofluorescence at both wavelengths was measured at the start of each experiment using unlabeled cells on a glass chip. These values were then used to correct the 340-and 380-nm readings prior to the calculation of the 340/380 nm fluorescence ratio (Rn). Rn values were recorded at 2-5 intervals. Following the experimental treatment, each chip was calibrated using the calcium ionophore ionomycin in the absence of calcium (NHB with no added calcium and 10 mM EGTA) to yield a minimum 340/380 ratio (Rfl min) and a maximum 380 nM (380 max) response. Subsequent addition of NHB buffer with 10 p~ ionomycin yielded a maximum 340/380 ratio (Rfl max) and a minimum 380 nm (380 min) response. Intracellular calcium levels were then estimated (25) using the following formula: where K = Kd (380 max/380 min) and Kd for fura2 = 224 nM (26).
The cells were loaded with either BCECF-AM or fura2-AM as described above. Following the placement of a chip into the spectrofluorometer, a base line was obtained. The base line was the average of a stable resting pH, or Ca2+, level maintained for 60 to 100 s. The difference in the rapidity of the pH and calcium responses necessitated a difference in the time frames for measuring the maximal calcium and pH responses. In the case of pH,, an average maximal reading maintained for at least 120 s was determined during an observation period that was limited to the first 10 min following thrombin or ionomycin stimulation. The calcium response was much more rapid than the pH, response. The maximal Ca2+, response was obtained by averaging the maximal Ca2+i occurring over a 10-5 period within 1 min of stimulation. The difference between the average maximal response and the average base line was calculated and used as the maximal stimulated change for that chip of cells.
The pretreatment of the chips with TMB8 or with EGTA was accomplished by the perfusion of the appropriate solution through the cuvette containing the chip while in the spectrofluorometer. In all cases, a stable base line was obtained prior to stimulation with either thrombin or ionomycin. In the case of the calcium chelator BAPTA, the cells were preloaded with 30 p~ BAPTA-AM for 1 h in a small Petri dish. During the final 35 min of BAPTA-AM loading, the cells were also loaded with either 2 p~ fura2-AM or 2 p~ BCECF-AM. The cells were rinsed in NHB for a 45-min period prior to the start of the experiment.

RESULTS
Thrombin Effects on Ca2+i and pH,-Exposure of growtharrested WS-1 cells to 10 nM thrombin generated a transient increase in Ca2+; (Fig. U). The thrombin-stimulated increase in Ca2+i was transitory with peak values generated within 20 s from the start of thrombin addition. The peak Ca2+, values lasted for 6-10 s and were followed by a decrease in Caz+i to below base-line levels. Ca2+; returned to the initial base line after 5-7 min. Thrombin addition also caused a change in pH; that was detectable within 1 min and complete within about 10 min ( Fig. 2A).
The increase in Ca2+i and pHi was dependent on the concentration of thrombin. Table I shows the increase in Ca2+i and pH; above a base-line value established at the start of each experiment. The average base line for Ca2+i was 184 & 12 nM ( n = 48) while the average base line for pH, was 7.06 & 0.02 pH units (n = 111). At 0.1 nM thrombin, there was a small pH change and no detectable increase in Ca2+;. As the concentration of thrombin was increased from 0.1 to 10 nM, there was a concomitant increase in both the pHi and Ca2+i responses. The half-maximal concentration of thrombin required for the increased pHi and Ca2+; was similar to the concentration required for DNA synthesis and inositol phosphate accumulation (2).
The peptide bombesin caused the accumulation of inositol phosphates in WS-1 cells but was ineffective in stimulating DNA synthesis (2). Even though bombesin stimulated the accumulation of inositol phosphates (2), it did not produce a detectable increase of Ca2+, (Fig. 1B) and no change in pHi was observed (Table I).
Thrombin Actiuation of Na+/H+ Exchange-The involvement of the Na+/H+ exchanger in the thrombin-induced alkalinization was tested by measuring the sensitivity of the pH response to amiloride or Na+-free buffer (NHB with Na+ isoosmotically replaced by N-methyl-D-glucammonium). Exposure to Na+-free solution caused an acidification of 0.08 & 0.02 pH units in 100 s (n = 4) (Fig. 2B). A slower acidification was observed with 1 mM amiloride, 0.03 5 0.01 pH units in 100 s (n = 4) (Fig. 2C). The addition of 10 nM thrombin did A . not cause an alkalinization under either of these conditions (Fig. 2, B and C). In the presence of thrombin, Na+-free solutions caused a rapid acidification of 0.16 f 0.02 pH units over 100 s (n = 5). In contrast, thrombin and 1 mM amiloride resulted in an acidification similar to that observed in the absence of thrombin, 0.04 k 0.02 pH units over 100 s (n = 4).
The amiloride and Na+ sensitivity of the pH response indicate that the thrombin-induced change in pH, occurred by activation of the Na+/H' exchanger.
Mobilization of Internal Calcium-The elevated Ca2+i induced by thrombin could be due to increased Ca2+ influx or to the release of Ca2+ from internal stores. In order to differentiate between these possibilities, the ability of thrombin to generate an increase in Ca2+; in the absence of external calcium was tested. The cells were pretreated with calciumfree buffer (NHB minus Ca2+ and with 10 FM EGTA) for 60-100 s. Pretreatment of WS-1 cells with calcium-free buffer resulted in a decrease in Ca2+; (Table 11). Stimulation with 10 nM thrombin in the absence of external calcium resulted in a calcium transient (Fig. 3A and Table 11) suggesting that thrombin releases calcium from internal stores. In order to confirm further the source of the elevated calcium, the drug TMB8 was used. TMB8 has been shown to block the release of internal calcium in human fibroblasts (27). The addition of 200 p~ TMB8 reduced the basal Ca2+; levels and also reduced the thrombin-stimulated calcium transient ( Fig. 3B and Table 11). Taken together, these results indicate that thrombin stimulation of WS-1 cells results primarily in the release of calcium from internal stores.  The BCECF and fura2 relative fluorescence values were converted to pH, and Ca2+,, respectively, as described under "Experimental Procedures." After pretreatment, a stable base line was established prior to the administration of 10 nM thrombin. The ApHi values were calculated using the base line established after pretreatment and represent the mean thrombin-induced change in pH;. In the absence of pretreatment, the resting Ca2+; was 184 f 12 nM (n = 48). following the addition of thrombin was measured under conditions in which the rise in Ca2+i was blocked by TMB8. The administration of 200 WM TMB8 reduced the base-line Ca2+L and the thrombin-stimulated Ca2+, increase (Fig. 3B). The reduction in the Ca", transient correlated with a reduction in the pH, response (Fig. 4B). Smaller reductions in Ca2+, and pH, were observed with lower doses of TMB8 (Table 11). In order to confirm further the importance of Ca2+; in mediating the thrombin-induced increase in pHi, the cell-permeable calcium chelator BAPTA-AM was used to reduce the calcium transient. Preincubation of the cells with 30 PM BAPTA-AM reduced the thrombin-stimulated increase in Ca2+L (Table I1 and Fig. 3C). BAPTA did not affect the resting pH, and attenuated the thrombin response (Fig. 4C). In contrast, the presence of 10 FM EGTA in the calcium-free external medium reduced the resting Ca2+; but still allowed a large increase in Ca2+, and an increase in pHi (Table I1 and Fig. 4A). Therefore, the thrombin-induced pH change was primarily dependent upon elevated Ca2+i derived from internal stores. Ionomycin Effects on Ca2+; and pH,-In order to examine the effect of Ca2+; on pHi independent of thrombin stimulation, the ionophore ionomycin was used to elevate cytoplasmic calcium. Concentrations of 1 and 2 PM ionomycin elevated Ca2+; (Fig. 5A). In contrast to thrombin, the ionomycininduced changes in Ca2+i were dependent upon external calcium. In the absence of external calcium (10 PM EGTA), ionomycin failed to generate a large change in Ca2+, (Fig. 5B). However, the ionomycin-induced changes in calcium were not markedly affected by 200 WM TMB8 (Fig. 5C). These data are quantitated in Table 111.

Relationship between
Ionomycin increased the cytoplasmic pH of growth-arrested WS-1 cells (Fig. 6A). The increase in pH, caused by ionomycin correlated with the ionomycin-induced increase in Ca2+i (Table 111). Ionomycin treatment in calcium-free buffer reduced the pH, response (Fig. 6B). Pretreatment with TMB8 raised the resting pH, but had little effect on the ionomycininduced increase in Ca2+, and pHi ( Fig. 6C and Table 111).
T o determine that the ionomycin-induced alkalinization (Fig. 7A) was due to the activation of the Na+/H+ exchanger, the changes in pH, were tested for Na+ and amiloride sensi- tivity. As with thrombin-induced alkalinization, ionomycin did not stimulate an increase in pH in the absence of Na+ or the presence of 1 mM amiloride (Fig. 7 ) . Ionomycin caused an acidification of 0.22 f 0.1 pH units in 100 s ( n = 4) in Na+free solution (Fig. 7 B ) , greater than that observed in the Na+free buffer control (Fig. 2B). In amiloride-containing solutions, similar rates of acidification were observed with ionomycin (0.04 f 0.01 pH units in 100 s (n = 4), Fig. 7C) or without ionomycin (Fig. 2C). Therefore, ionomycin or thrombin induced an increase in pH, that was dependent on a rise in Ca2+i, required external sodium, and was inhibited by amiloride.
Effect of TMB8 on Na+/H+ Exchanger-In the presence of TMB8, the thrombin-induced increases in intracellular calcium (Fig. 3B) and pH, (Fig. 4B) were both inhibited. One explanation for this correlation is that TMB8 blocks calcium release from internal stores, thereby inhibiting the activation of Na+/H+ exchange by thrombin. However, given the variety of possible cellular effects of TMB8 (28), TMB8 could inhibit another signal pathway, such as protein kinase C or have a direct inhibitory effect on Na+/H+ exchange. TMB8 does not block phorbol ester activation of Na+/H+ exchange in WS-1 A.

C.
Ca2'-free Buffer fluorescence was used as a measure of pHi as described in Fig. 2. The protocol for the administration of pretreatments was identical to that employed in Fig. 3. cells.' To determine a possible direct inhibitory effect of TMB8, the exchanger was activated by cellular acidification (induced after a NH&1 prepulse), and rates of recovery were measured in the presence or absence of TMB8 (Fig. 8). The minimal pH attained after removal of NH4Cl, pH 6.6, was matched in the treated and control cells, and the slopes were calculated over the same pH range. Slopes of recovery were determined as the least squares regression line, fit to at least 15 points, as pHi recovered toward the initial pHi. In control cells, this rate of recovery was 0.003 f 0.0013 pH units/s (n = 6) and was completely inhibited by 1 mM amiloride (0.00002 f 0.0002 pH units/s, n = 4). In the presence of 200 I.IM TMB8, cells recovered from acidification to pHi of 6.6 at a rate of 0.0029 f 0.0006 pH units/s ( n = 9), similar to the rate in the absence of TMB8. This result is consistent with the ionomycin-induced activation of Na'/H' exchange in the presence of TMB8 (Fig. 6C, Table 111), further demonstrating the ability of the Na+/H' exchanger to function in the presence of TMB8.

DISCUSSION
Thrombin is a potent mitogen for a variety of fibroblasts (29)(30)(31). Treatment of cells with thrombin initiates a number of early changes that have been associated with mitogenesis, including stimulation of PI turnover (32), cellular alkaliniza-B. Hendey, and M. D. Mamrack, manuscript in preparation.  The BCECF and fura2 relative fluorescence values were converted to pH, and CaZfi, respectively, as described under "Experimental Procedures." When pretreatments were used (TMB8, EGTA), the ApHi values were calculated using the base line established after pretreatment. The base-line Ca2+i and stimulated Ca2+; values were calculated as described in Table 11. Mean f S.E.; numbers in parentheses indicate number of observations. tion, and an activation of protein kinases (10,14). We previously demonstrated that thrombin is mitogenic and stimulated PI turnover in WS-1 cells (2). In the present study, we show that thrombin causes a calcium transient and a cyto-A . 8.

7.4
Ca*'-free Buffer plasmic alkalinization with a similar concentration dependence.
Growth factor-induced intracellular alkalinization is mediated by an increase in the activity of the Na'/H+ exchanger in a variety of cell types (12, 13). Amiloride, an inhibitor of Na+/H+ exchange (33), prevents the thrombin-induced alkalinization of WS-1 cells (Fig. 2C). In addition, removal of external Na' converts thrombin-induced alkalinization into an acidification (Fig. 2B). Such inhibition of mitogen-induced alkalinization by amiloride and Na+ removal is characteristic of the involvement of Na+ /H+ exchange (12,13).
The presence of Na'/H+ exchange in these cells is also suggested by the more pronounced acidification observed in Na+-free solutions than in the presence of amiloride. In the nominal absence of COz, the predominant pH regulating transport system in most cells is Na+/H+ exchange (34). At steady state, this transport system has a low level of activity and counterbalances the acid loading processes in the cell (metabolic acid production, H+ influx, etc.). A slow acid drift in the presence of amiloride is a measure of the acid load on a cell (34), which is quite modest in WS-1 cells (Fig. 2C). In contrast, exposure to Na+-free solutions results in a rapid acidification (Fig. 2B). The most likely explanation for this increased rate of acid loading is reversal of Na+/H+ exchange (with a concomitant H+ influx) due to the reversal of the Na' Bombesin, platelet-derived growth factor, vasopressin, and other peptide hormones elevate Ca2+i by stimulating the release of Ca2+ from internal stores (4-6,15,17). Mitogens such as EGF elevate Ca2+i by increasing the influx of Ca2+ from the external medium (8,9). Thrombin mobilizes calcium in platelets (36) and in vascular smooth muscle cells (20). In this paper, we provide the first demonstration of thrombin-stimulated calcium mobilization in human fibroblasts. The increase in Ca2+i is transient (Fig. L4) and dependent on the concentration of thrombin (Table I). The removal of external calcium does not block the thrombin-induced Ca2+ transient (Fig. 3A, Table 11), but this transient is inhibited in a concentration-dependent fashion by TMB8 (Fig. 3B, Table 11), which has been shown to block the release of Ca2+ from internal stores (27). This release of calcium upon thrombin exposure may be mediated by the rapid thrombin-induced increase in  8. Effect of TMBS on rate of recovery of pH after a NH&l prepulse. Intracellular pHi was measured using BCECF as described in Fig. 2. A, the pHi response to a prepulse of 20 mM NH&l for the indicated time; B, similar experiment as in A except in the presence of 200 PM TMB8. Time of NH&l treatment was varied to produce similar minimal pHi in control and TMB8-treated cells.
inositol trisphosphate (2), which is a potent Ca'+ releasing agent (37). The elevation of inositol trisphosphate and Ca2+i occurred within the first 1-2 min of exposure to thrombin and subsided within a few minutes (Figs. lA and 3A) (2).
An unusual feature of the thrombin-induced Ca'+i transient is that Ca'+i actually drops below base-line levels for several minutes before returning to initial values (Figs. lA and 3A).
The basis for this effect may be a mitogen activation of Ca2+ sequestering mechanisms as well as Ca'+ transport pathways. When the Ca'+i transient is blocked by TMB8, Ca2+i drops below base line for several minutes (Fig. 3B). In addition, a similar drop in base-line Ca2+i has been observed following treatment with low concentrations of thrombin (0.1 nM), which fail to generate an observable calcium transient ( Table  I). These results suggest that Ca2+ sequestering or transport mechanisms are activated independent of elevated intracellular calcium.
In contrast to the thrombin-induced changes, bombesin had no effect on either pHi (Table I) or Ca2+i (Fig. lB, Table I). The inability of bombesin to induce a detectable increase in ca'+i was unexpected given our previous observation of inositol phosphate generation in WS-1 cells following bombesin treatment. Bombesin induced a significantly greater accumulation of inositol phosphate within 30 min of stimulation when compared with thrombin. However, a comparison of thrombin and bombesin inositol phosphate production indicated that thrombin produced larger quantities of inositol bisphosphate and trisphosphate at early time points (2). The rapid and greater production of inositol trisphosphate may explain thrombin's proficiency in releasing internal calcium. The failure of bombesin to induce early ionic events correlates with its lack of mitogenic stimulation of WS-1 cells (2). These results contrast with the effects of bombesin on Swiss 3T3 cells where bombesin is mitogenic (38) and results in an increase in both Ca2+i and pHi (15,39).
Our data suggest that elevated Ca'+i is important for the activation of Na+/H+ exchange. Prevention of the thrombininduced Ca'+ increase by TMB8 or BAPTA significantly reduced the cytoplasmic alkalinization ( Fig. 4 and Table 11). In the absence of mitogen, an increase in CaZci induced by ionomycin caused a cytoplasmic alkalinization that was insensitive to TMB8 but required extracellular Ca'+ ( Fig. 6 and Table 11). Therefore, WS-1 cells have a pathway of mitogeninduced activation of Na'/H+ exchange that depends on elevated intracellular Ca2+.
Regulation of Na+/H' exchange in other cell types includes pathways activated by cellular acidification (34), volume change (40), and protein kinase C (8,17). Other pathways for the activation of Na+/H+ exchange probably exist in WS-1 cells. With 0.1 nM thrombin, a small (less than 0.1 pH unit over 10 min) but reproducible change in pH was observed with no detectable increase in Ca2+i. In addition, treatment of cells with BAPTA or TMB8 did not completely eliminate a pH response, while significantly reducing the calcium transient. A diacylglycerol-activated protein kinase C pathway may also be present in these cells and contribute in part to the regulation of Na+/H' exchange. Treatment of WS-1 cells with phorbol esters does not stimulate DNA synthesis, but does stimulate Na+/H+ exchange.' Taken together, our results suggest that the cytoplasmic alkalinization caused by thrombin is predominantly due to a calcium-dependent pathway, but another pathway may be evident at low thrombin concentration. Other human fibroblast cell strains have either protein kinase C or calcium/calmodulin-mediated regulation of Na'/H+ exchange (41).
The calcium requirement for activation of Na+/H+ exchange varies with cell type. Elevation of Ca'+i by exposure to calcium ionophores has been shown to activate Na+/H' exchange in some cells, such as HSWP human fibroblasts (16), mouse thymocytes, and Swiss 3T3 cells (15). However, in vascular smooth muscle cells (20), human platelets (361, Swiss 3T3 fibroblasts (5), and HF human fibroblasts (12), the elevation of Ca'+i with ionophores did not activate Na+/H+ exchange. In the case of stimulation by polypeptide hormones and mitogens, the role of elevated Ca2+i in mediating cellular alkalinization may vary with the cell system. For example, the prevention of mitogen-induced Ca2+i elevation inhibited the activation of the Na+/H+ exchanger in vascular smooth muscle cells (20), Chinese hamster embryo fibroblasts (81, and HSWP human fibroblasts (27). In contrast, the activation of the Na+/H+ exchanger could be observed even when the calcium transient was prevented in human platelets (36) and Swiss 3T3 cells (5). In WS-1 cells, the elevation of Caz+, with ionomycin is sufficient to stimulate the Na+/H+ exchanger. In addition, the transient elevation of Ca2+i caused by thrombin is necessary for the thrombin-stimulated alkalinization in these cells.
Several possible mechanisms for the effect of Ca2+ on Na+/ H+ exchange exist. The temporal difference between the rapid calcium response and the slower change in pH suggests that calcium is initiating a signaling cascade. Ca2+i could activate Na+/H+ exchange in WS-1 cells by altering the activity of protein kinase C. An alternative mechanism would involve calmodulin. Calmodulin-mediated stimulation of Na+/H+ exchange in human fibroblasts has been reported (42). Calmodulin-dependent protein kinase activity can remain active after the initial calcium transient (43). Alternatively, Ca2+i could interact directly with the Na+/H+ exchanger or induce the fusion of intracellular vesicles with the plasma membrane, thereby increasing the number of exchangers at the cell surface. The mechanism of the Ca2+ effect on the Na+/H+ exchange in WS-1 cells is currently under investigation.