A novel Ca2+ entry mechanism is turned on during growth arrest induced by Ca2+ pool depletion.

Ca2+ pool depletion with Ca2+ pump blockers induces growth arrest of rapidly dividing DDT1MF-2 smooth muscle cells and causes cells to enter a stable, quiescent G0-like growth state (Short, A. D., Bian, J., Ghosh, T. K., Waldron, R. T., Rybak, S. L., and Gill, D. L.(1993) Proc. Natl. Acad. Sci. U.S.A. 90, 4986-4990). Here we reveal that induction of this quiescent growth state with the Ca2+ pump blocker, thapsigargin, is correlated with the appearance of a novel caffeine-activated Ca2+ influx mechanism. Ca2+ influx through this mechanism is clearly distinct from and additive with Ca2+ entry through store-operated channels (SOCs). Whereas SOC-mediated entry is activated seconds after Ca2+ pool release, caffeine-sensitive influx requires at least 30 min of pool emptying. Although activated in the 1-10 mM caffeine range, this mechanism has clearly distinct methylxanthine specificity from ryanodine receptors and is not modified by ryanodine. It is also unaffected by the Ca2+ channel blockers SKF96365 or verapamil and is independent of modifiers of cyclic nucleotide levels. Growth arrest by thapsigargin-induced Ca2+ pool depletion can be reversed by treatment with 20% serum (Waldron, R. T., Short, A. D., Meadows, J. J., Ghosh, T. K., and Gill, D. L.(1994) J. Biol. Chem. 269, 11927-11933). The serum-induced return of functional Ca2+ pools and reentry of cells into the cell cycle correlates exactly with the disappearance of the caffeine-sensitive Ca2+ influx mechanism. Therefore, appearance and function of this novel Ca2+ entry mechanism are closely tied to Ca2+ pool function and cell growth state and may provide an important means for modifying exit from or entry into the cell cycle.

Ca 2ϩ inside intracellular pools not only provides a source for cytosolic Ca 2ϩ signals (1) but controls diverse cellular and lumenal functions including the activation of external Ca 2ϩ influx (2), folding and processing of proteins in the endoplasmic reticulum (3), and cell proliferation and cell cycle progression (4 -7). Emptying of Ca 2ϩ pools in rapidly dividing DDT 1 MF-2 smooth muscle cells with the Ca 2ϩ pump blockers, thapsigargin or 2,5-di-tert-butylhydroquinone, induces entry of cells into a stable, quiescent G 0 -like growth state (5,6). Here we reveal that while in this Ca 2ϩ pool-depleted, growth-arrested state, cells turn on a novel methylxanthine-sensitive Ca 2ϩ influx mechanism. This influx process is clearly distinct from Ca 2ϩ entry through store-operated channels (SOCs), 1, 2 is independent of ryanodine receptor or InsP 3 receptor function, and is not related to changes in cyclic nucleotide levels. Induction of new pump synthesis and return of functional Ca 2ϩ pools by treatment of Ca 2ϩ pool-depleted cells with high serum or arachidonic acid causes the methylxanthine-sensitive Ca 2ϩ influx mechanism to be turned off, normal receptor-operated Ca 2ϩ signaling to resume, and re-entry of cells into the cell cycle. The function of this novel Ca 2ϩ influx mechanism and its precise turning on and off may be important events in the relationship between Ca 2ϩ signaling and the growth state of cells.

EXPERIMENTAL PROCEDURES
Growth of Cells and Thapsigargin Treatment-Culture of DDT 1 MF-2 smooth muscle cells was as described previously (8). Thapsigargintreated cells (attached to poly-L-lysine-coated glass coverslips in Dulbecco's modified Eagle's medium with 2.5% CalfPlus) had been exposed to 3 M thapsigargin in culture for 18 h, followed by washing and culture in thapsigargin-free medium for 48 h as described previously (5,6). Untreated cells were exposed to the same conditions except without thapsigargin.
Measurement of Intracellular Ca 2ϩ -Intracellular free Ca 2ϩ measurements were as described previously (6,9). Briefly, coverslip-attached cells were transferred into HKM (107 mM NaCl, 6 mM KCl, 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 1 mM CaCl 2 , 11.5 mM glucose, 0.1% bovine serum albumin, 20 mM Hepes-KOH, pH 7.3) and loaded with 2 M fura-2 acetoxymethyl ester for 10 min at 20°C in the dark. After deesterification in fresh loading medium for 15 min at 20°C, coverslips were inserted into a Dvorak-Stotler chamber. Groups of 5-10 fura-2loaded cells were viewed through a Nikon 40ϫ UV-fluor objective. Excitation at 340, 358, or 380 nm was generated using a PTI D103 light source, and fluorescence emission at 505 nm was monitored at 24°C using a PTI D104 photometer. Free intracellular Ca 2ϩ concentrations were calculated from either 340/380 or 358/380 ratios of fluorescence intensities using the method of Grynkiewicz et al. (10)

RESULTS AND DISCUSSION
The Ca 2ϩ pump blocker, thapsigargin, empties intracellular Ca 2ϩ pools (11) and rapidly activates Ca 2ϩ entry through SOCs (2). In many cells including DDT 1 MF-2 cells this entry becomes efficiently deactivated a few minutes after pool emptying (7). SOCs can be transiently reactivated by brief removal of Ca 2ϩ o (12). As shown in Fig. 1A, in DDT 1 MF-2 cells, even 24 h after thapsigargin-induced Ca 2ϩ pool depletion and the establishment of growth arrest, Ca 2ϩ o removal results in an immediate decrease in resting cytosolic Ca 2ϩ , presumably reflecting the contribution of some residual non-deactivated SOC activity. During the absence of Ca 2ϩ o , SOC activity becomes reactivated as reflected by the large "overshoot" in cytosolic Ca 2ϩ observed upon readdition of Ca 2ϩ o ; the transient nature of the overshoot reflects a temporary activation of SOCs, which again undergo rapid deactivation. Normal DDT 1 MF-2 cells with filled Ca 2ϩ pools and presumably without any activated SOCs exhibit little change in cytosolic Ca 2ϩ as a result of transient Ca 2ϩ o removal. Interestingly, addition of 10 mM caffeine to the pool-depleted cells induces a further rapid and substantial increase in Ca 2ϩ i , which also appears to undergo deactivation (Fig. 1A). This caffeine effect is not dependent on brief Ca 2ϩ o removal since, added directly to thapsigargin-arrested cells, caffeine induces the same large transient Ca 2ϩ i increase (Fig. 1B). Cells treated with the alternative Ca 2ϩ pump blocker, 2,5-di-tert-butylhydroquinone, which also induces pool emptying and growth arrest (6), develop the same caffeine-induced Ca 2ϩ response (data not shown). Importantly, normal cells (that is untreated with Ca 2ϩ pump blockers) are devoid of any caffeine response ( Fig.  1, A and B).
The effect of caffeine is clearly on influx of Ca 2ϩ since it induces no change in the absence of Ca 2ϩ o (Fig. 1C). When Ca 2ϩ is added back in the presence of caffeine, a rapid and even larger transient of Ca 2ϩ is observed (Fig. 1C). The almost exact doubling in the size of the transient after readdition of Ca 2ϩ reflects clear additivity of SOC-and caffeine-mediated Ca 2ϩ entry, providing further evidence that the two mechanisms are independent. The Ca 2ϩ entry blocker SKF96365 at 50 M completely blocks SOC-mediated Ca 2ϩ entry but even as high as 2 mM has no effect on caffeine-mediated Ca 2ϩ entry (data not shown). These data indicate that caffeine induces an influx of Ca 2ϩ distinct from that mediated by SOCs but only in Ca 2ϩ pool-depleted growth-arrested cells.
Even though both depend on Ca 2ϩ pool emptying, an interesting divergence between the two influx mechanisms is their time dependence of activation following pool depletion. SOCmediated Ca 2ϩ influx becomes activated rapidly after thapsigargin-induced pool depletion but within 5 min has become almost completely deactivated ( Fig. 2A). Indeed, the efficient turn-off of SOC-mediated entry prevents long-term increased cytosolic Ca 2ϩ levels following pool depletion and may be im-portant in the survival of DDT 1 MF-2 cells, albeit in a growtharrested state, following pool emptying (13,14); cells in which SOC-mediated entry is not deactivated can enter an irreversible apoptotic cycle (13,14). Reactivation of SOCs can occur by transient Ca 2ϩ removal as early as a few minutes after pool emptying ( Fig. 2A). At this time addition of caffeine has no effect (Fig. 2B), further supporting an SOC-independent action of caffeine. The time dependence of pool emptying and appearance of caffeine-induced influx is shown in Fig. 2, C-E. After 10 min of thapsigargin-induced pool emptying no caffeine effect can be detected. The shortest period of thapsigargin treatment after which a significant caffeine-induced Ca 2ϩ influx can occur is 30 min. However, immediately following this minimally effective period of thapsigargin treatment, the caffeine response is small; during a further period of 30 -60 min after removal of thapsigargin the caffeine response becomes larger. In Fig. 2, C-E, after treatment with thapsigargin for the times shown, cells were incubated a further 60 min without thapsigargin. Also, with the shorter thapsigargin treatment times, onset of influx following caffeine addition is more variable in cells, some cells exhibiting a lag of 1-2 min prior to activation of influx (this is apparent in Fig. 2D in which the data are an average of [Ca 2ϩ ] i was measured in normal cells (Untreated) or cells treated with thapsigargin to empty Ca 2ϩ pools and induce growth arrest (TG-treated). A, untreated and thapsigargin-treated cells were treated with nominally Ca 2ϩ -free HKM for 6 min and then returned to normal HKM. 3 min later, 10 mM caffeine was added to both cell types. B, 10 mM caffeine was added to untreated and thapsigargintreated cells at 30 s in normal HKM. C, thapsigargin-treated cells were treated at 30 s with Ca 2ϩ -free HKM and at 90 s with 10 mM caffeine; after 5 min in Ca 2ϩ -free HKM cells were returned to normal HKM. The specificity among methylxanthines in activating Ca 2ϩ entry indicates function of a novel influx mechanism. At 10 mM, 3,7-dimethyl-1-propargylxanthine (3,7-DMPX), a ryanodine receptor agonist 4-fold more effective than caffeine (15), is completely ineffective in inducing Ca 2ϩ influx (Fig. 3A). In fact DDT 1 MF-2 cells have no measurable ryanodine receptor activity; caffeine or ryanodine have no effect on intracellular Ca 2ϩ release; nor is there any detectable ryanodine receptor protein in DDT 1 MF-2 cells as determined by Western analysis. Theophylline (1,3-dimethylxanthine) consistently induces a more rapid increase in cytosolic Ca 2ϩ than caffeine (1,3,7-trimethylxanthine) (Fig. 3B). The sensitivity to theophylline and caffeine is similar; 10 mM gives a maximal response with a sharp dose-response curve between 1 and 10 mM. This caffeine-sensitivity range is similar to that for ryanodine receptors (15,16). 1,7-Dimethylxanthine induces a slower but almost full effect (Fig. 3B) whereas 3,7-dimethylxanthine has no effect (not shown). This methylxanthine specificity is clearly distinct from that of ryanodine receptors on which the latter two dimethylxanthines are actually more effective than caffeine (16).
The methylxanthine-activated Ca 2ϩ entry is not related to changes in cyclic nucleotide levels. The potent phosphodiesterase inhibitor, 3-isobutyl-1-methylxanthine (IBMX), has no effect on Ca 2ϩ entry at 10 mM. Addition of dibutyryl cyclic AMP, 8-bromo-cyclic AMP, forskolin (alone or in combination with IBMX), or dibutyryl cyclic GMP all have no effect. Methylxanthines are also effective adenosine receptor antagonists. However, the ineffectiveness of 3,7-DMPX, 3,7-DMX, and IBMX, which are potent antagonists of both A 1 and A 2 adenosine receptors (17), militates strongly against Ca 2ϩ influx being adenosine receptor-mediated. DDT 1 MF-2 cells have no measurable voltage-operated Ca 2ϩ channel activity, and 50 M verapamil has no effect on the influx induced by caffeine. InsP 3 receptors are reported to be inhibited by caffeine (18,19). Although there was no logical explanation of our results based on InsP 3 receptor alteration, examination of the effects of 10 mM caffeine on InsP 3 -mediated Ca 2ϩ release from permeabilized DDT 1 MF-2 cells revealed a negligible effect. Also, in contrast to the specificity described above, theophylline is reported to be much less potent on InsP 3 receptors than caffeine (15).
The entry of Ca 2ϩ induced by caffeine is clearly transient and becomes deactivated within a few minutes. Readdition of caffeine without washing induces no further effect; however, a 5-min wash of caffeine-treated cells followed by caffeine reapplication results in return of the full Ca 2ϩ entry effect indicating that the entry mechanism becomes reactivated. More recent experiments 3 have assessed cation selectivity and specificity of the caffeine-activated entry mechanism. These studies reveal that the entry mechanism is readily permeable to Mn 2ϩ . By assessing quench of cytosolic fura-2 in pool-depleted DDT 1 MF-2 cells, rates of Mn 2ϩ entry activated by different methylxanthines have been shown to correlate well with the effects on cytosolic Ca 2ϩ shown in Fig. 3. Also, we have observed that caffeine-activated Ca 2ϩ entry is blocked by other cations including Gd 3ϩ , La 3ϩ , and Co 2ϩ but is not blocked by Mn 2ϩ . In contrast, while SOC activation induces a modest influx of Mn 2ϩ , the entry of Ca 2ϩ through SOCs is substantially blocked by Mn 2ϩ , in keeping with the observations of others (20,21). These results provide evidence of yet another distinction between the methylxanthine-and SOC-mediated Ca 2ϩ entry mechanisms.
Further reinforcement of the correlation between Ca 2ϩ pool content, growth state, and operation of caffeine-induced Ca 2ϩ influx is derived from experiments in which thapsigargintreated, growth-arrested cells are induced to reenter the growth cycle. We previously revealed that a 40-min treatment of thapsigargin-arrested cells with 20% serum induces synthesis of new pump protein, return of functional Ca 2ϩ pools, and resumption of normal growth (7). Recent studies reveal that 10 M arachidonic acid induces an identical recovery of Ca 2ϩ pools and cell growth. 4 As shown in Fig. 4, the reappearance of a functional bradykinin-activated Ca 2ϩ pool is directly correlated with the disappearance of the caffeine-sensitive Ca 2ϩ entry mechanism. In thapsigargin-arrested cells, the absence of an InsP 3 -sensitive Ca 2ϩ pool is reflected by the lack of response to receptor agonists such as bradykinin (Fig. 4A); in these cells caffeine always activates a substantial influx of Ca 2ϩ (Fig. 4C). Thapsigargin-arrested cells treated with 20% serum recover from growth arrest and when examined 16 h later show normal bradykinin-induced Ca 2ϩ signals (Fig. 4B) and a complete absence of the caffeine-induced Ca 2ϩ response (Fig. 4D). Thus, cells have returned to the pregrowth-arrested state in which the caffeine-sensitive influx mechanism is turned off and normal agonist-sensitive Ca 2ϩ pools are functional. Significantly, further experiments have shown that the time of disappearance of caffeine-induced influx closely correlates with that for reappearance of Ca 2ϩ pools following high serum-induced recovery. 3 h after thapsigargin-arrested cells are induced to recover with high serum, a significant bradykinin-sensitive Ca 2ϩ signal is observable in groups of cells as well as a still measurable caffeine response (although as yet a single cell with both activities has not been observed). At 6 h after induction, the response to caffeine has completely disappeared and the bradykinin response is the same as a normal cell. We have shown that following serum induction of thapsigargin-arrested cells, new Ca 2ϩ pump protein appears as early as 1 h and pools become fully operational at 6 h (7); thereafter cells progress through G 1 and begin to enter G s 16 h later (6,7). This entire sequence of events, including cessation of the caffeine response, is identically activated by 10 M arachidonic acid as opposed to 20% serum. From these experiments it is clear that function of the caffeine-sensitive Ca 2ϩ influx mechanism is restricted only 3  to Ca 2ϩ pool-depleted, growth-arrested cells.
The studies described here demonstrate specific activation of a novel and distinct Ca 2ϩ entry mechanism in pool-depleted growth-arrested DDT 1 MF-2 cells that permits a transient but substantial entry of Ca 2ϩ . Indeed, the levels of intracellular Ca 2ϩ achieved by activation of this mechanism are comparable with those attained by complete Ca 2ϩ pool emptying or activation of SOC-mediated entry. A Ca 2ϩ -conducting caffeine-sensitive influx channel was recently reported in adult gastric smooth muscle cells (22); however, in this case Ca 2ϩ influx was not rapidly deactivated and methylxanthine specificity was not examined. Whereas functional ryanodine receptors in the same cells precluded absolute proof that Ca 2ϩ influx was independent of ryanodine receptor function (23), it is intriguing that methylxanthine-sensitive Ca 2ϩ entry channels might be expressed in nondividing smooth muscle cells. Our results are the first to provide evidence for a pharmacologically defined and apparently unique methylxanthine-sensitive Ca 2ϩ entry pathway. Since caffeine has been widely used as a means of probing the action of Ca 2ϩ release channels in many cell types (24), the present results are significant in providing awareness of the existence of a distinct caffeine-activated Ca 2ϩ influx pathway. It is possible that this mechanism retains certain structural and/or functional similarities with ryanodine receptors. At this stage we do not know whether protein synthesis is required for turning on the entry mechanism. Another intriguing area of investigation will be to determine the means by which deactivation occurs. It is likely that operation of the influx pathway is transient within the cell cycle and/or restricted to discrete cell growth states.
Most significantly, although the physiological activation of the caffeine-sensitive Ca 2ϩ entry mechanism has yet to be characterized, the appearance of this clearly defined and substantial Ca 2ϩ entry mechanism under such specific conditions of pool emptying and growth arrest indicates a potentially important role in mediating Ca 2ϩ signals during transition into and out of the cell cycle or during cell division when Ca 2ϩ pools undergo substantial reorganization. As such, pharmacological modification of this channel may provide an important means for controlling cell growth.