in the accompanying study in human skin fibroblasts (31), we confirmed that maitotoxin (MTX), one of the most potent marine toxins known, produces a dramatic increase in cytosolic free Ca²⁺ concentration ([Ca²⁺]_i). This rise in [Ca²⁺]_iis unrelated to mobilization of Ca²⁺ from internal stores but rather results from a MTX-induced activation of nonselective cation channels present in the plasmalemma. The activation of these channels by MTX is followed closely in time by the formation of large pores that allow the flux of vital dyes, such as ethidium and YO-PRO-1, into the cell and the release of cell-associated fura 2. At much later times, MTX causes the release of lactate dehydrogenase (LDH), an event that presumably signals oncotic cell death. This sequence of permeability changes is reminiscent of the changes observed in many cells following stimulation of P2Z/P2X₇ receptors (10, 12, 26,43). The P2Z/P2X₇ receptor is the seventh member of the P2X receptor family (35). P2X receptors are thought to form Ca²⁺-permeable, nonselective cation channels activated by extracellular ATP. The P2Z/P2X₇ receptor has the novel characteristic of inducing the formation of large pores, which, like MTX, allow the entry of large organic vital dyes into the cell and ultimately causes the release of LDH and cell death by oncosis.

Although it is clear that heterologous expression of the P2Z/P2X₇ receptor is associated with expression of both functional channels and pores, the actual structure of the pore-forming subunit remains unknown. Likewise, the actual biophysical properties of the P2Z/P2X₇ channel that allow formation of pores have not been defined. It has been suggested that individual P2Z/P2X₇ subunits may aggregate to form pores of increasing size and hence allow larger and larger molecules to cross the plasmalemma (26, 36). However, there is little evidence to date that the pore is actually composed of P2Z/P2X₇ protein subunits, and it remains possible that an endogenous pore structure present in the cells employed for heterologous expression is activated following stimulation of P2Z/P2X₇ channels by purinergic agonists.

The finding that MTX induces formation of large pores with characteristics similar to P2Z/P2X₇ receptor-induced pores suggests two possibilities: 1) MTX is a high-affinity ligand for the P2Z/P2X₇ receptor or2) MTX and P2Z/P2X₇ receptor stimulation activate distinct channels but a common cytolytic pore. To distinguish between these two possibilities, the effect of MTX was examined in cells in which the expression of functional P2Z/P2X₇ channels and pores could be varied either by alterations in the growth conditions or by heterologous expression. In the first set of experiments, the effect of MTX was examined in THP-1 monocytes. These cells express low levels of P2Z/P2X₇ receptor under normal growth conditions, but pretreatment of these cells with bacterial lipopolysaccharide (LPS) and interferon-γ (IFN-γ) dramatically upregulates P2Z/P2X₇ message and function (18). In the second set of experiments, the effect of MTX was examined in control HEK cells and in HEK cells stably expressing the P2Z/P2X₇ receptor. Last, the effect of MTX on BW5147.3 cells, a cell line that expresses P2Z/P2X₇ receptors that are poorly coupled to pore formation (20), was examined. The results suggest that MTX activates channels that are distinct from those activated by stimulation of P2Z/P2X₇ receptors; however, the characteristics of the MTX-induced pores are indistinguishable from the pores activated by P2Z/P2X₇ receptor stimulation. A physical model is proposed in which MTX-activated channels and P2Z/P2X₇ receptor-activated channels compete for a common cytolytic pore.

MATERIALS AND METHODS

Solutions and reagents.

Unless otherwise indicated, HEPES-buffered saline (HBS) contained 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 10 mM d-glucose, 1.8 mM CaCl₂, 15 mM HEPES, and 0.1% BSA, pH adjusted to 7.40 at 37°C with NaOH. Ca²⁺-free HBS contained 0.3 mM EGTA and the same salts as HBS without added CaCl₂. Fura 2-AM, ethidium bromide, YO-PRO-1, and POPO-3 were obtained from Molecular Probes (Eugene, OR). MTX was obtained from Calbiochem (San Diego, CA) or from LC Laboratories (Woburn, MA) and was stored in amber-colored glass or plastic vials at −20°C as an aqueous stock solution (>1 μM) with 0.1% BSA. 2′ and 3′-O-(4-benzoylbenzoyl)-ATP (Bz-ATP) was obtained from Sigma Chemical (St. Louis, MO). All other salts and chemicals were of reagent grade.

Cell culture.

The THP-1 monocytic cell line was obtained from American Type Culture Collection (ATCC) and grown in RPMI 1640 containing 10% heat-inactivated fetal bovine serum (FBS), 1% penicillin-streptomycin-neomycin (PSN) antibiotic mixture, and 2 mM l-glutamine in spinner flasks as previously described (18). In some experiments, the P2Z/P2X₇ receptor was upregulated by pretreatment of THP-1 cells for 48 h in complete RPMI 1640 medium containing 1 μg/ml bacterial LPS and 1,000 U/ml IFN-γ. Wild-type HEK cells were obtained from ATCC and grown in MEM supplemented with 10% FBS, 1% PSN antibiotic mixture, and 2 mMl-glutamine. For passage, cells were dispersed by trypsin treatment and seeded at a density of ∼3 × 10³cells/cm³. The medium was changed every 2–3 days after seeding. HEK cells were harvested for experimentation when the cells had reached confluence. HEK cells stably expressing the human P2Z/P2X₇receptor were generated as previously described (19) and cultured in a fashion identical to wild-type HEK. BW5147.3, a murine lymphoma cell line, was obtained from ATCC and grown in spinner flasks in DMEM supplemented with 10% FBS, 1% PSN antibiotic mixture, and 2 mM l-glutamine.

Uptake of vital dyes.

The uptake of ethidium, YO-PRO-1, and POPO-3 was determined in the absence and presence of MTX or Bz-ATP as previously described (31).

Measurement of the apparent [Ca²⁺]_i.

[Ca²⁺]_iwas measured using the fluorescent indicator fura 2 as previously described (30, 31). Unless otherwise indicated, all measurements were performed at 37°C. Calibration of the fura 2 associated with the cells was accomplished using Triton X-100 lysis in the presence of saturating divalent cation concentrations followed by addition of EGTA (pH 8.5). [Ca²⁺]_iwas calculated by the equation of Grynkiewicz et al. (15) using a dissociation constant value for Ca²⁺ binding to fura 2 of 224 nM. As reported in the accompanying paper (31), MTX causes fura 2 efflux from cells. Therefore, where indicated, [Ca²⁺]_ivalues are reported as “apparent” changes in [Ca²⁺]_i.

Data presentation and statistical treatment.

Figures 1-12 show representative traces from experiments performed at least three times. For mean values, traces recorded under identical conditions on one day were pooled yielding ann of 1. Where indicated,n equals the number of independent experiments and statistical differences between means were determined by Student’s t-test, withP < 0.05 considered significant. To control for cell-to-cell and day-to-day variation, all traces shown within each panel of Figs. 1-12 were obtained on the same day and on the same batch of cells. Where indicated, traces in multiple panels of a figure were also paired. For example, the traces shown in bothA andB of Fig. 3 were obtained on the same day on the same batch of THP-1 cells, divided into control and treated groups.

RESULTS

Effect of MTX on THP-1 monocytes.

The results of our accompanying study (31) on fibroblasts suggested two possibilities: 1) MTX activates P2Z/P2X₇ or2) MTX and the P2Z/P2X₇ receptor activate distinct channels but a common cytolytic pore. If MTX activates the P2Z/P2X₇ receptor, then the response to MTX should be proportional to the level of expression of the P2Z/P2X₇ receptor. Previous studies have shown that P2Z/P2X₇expression is low in control THP-1 monocytes but can be dramatically upregulated by pretreatment of the cells with LPS and IFN-γ for 24–48 h (18). The effect of MTX on [Ca²⁺]_iin control THP-1 cells is shown in Fig. 1. MTX produced a concentration-dependent, biphasic change in [Ca²⁺]_i. At the lowest concentration tested (2 pM), MTX initially produced a significant 1.8-fold increase in [Ca²⁺]_ifrom a resting level of 71.8 ± 6.7 nM to a sustained level of 129.6 ± 20 nM (means ± SE, n = 7,P < 0.01). After a delay of ∼5 min, [Ca²⁺]_ibegan to increase again with time. Increasing the concentration of MTX produced a decrease in the delay time between the first and second phase with little change in the subsequent rate of increase of [Ca²⁺]_i(Fig. 1 A). Thus the first phase appears to be concentration dependent, whereas the second phase appears to be all or nothing. As the concentration of MTX increased from 25 pM to 2 nM, the first and second phase gradually merged into a single response (Fig. 1 B). Addition of MTX in the absence of extracellular Ca²⁺ had no effect, but a large increase in [Ca²⁺]_iwas observed on subsequent addition of the Ca²⁺ to the bath solution (data not shown). Thus, as previously shown, MTX does not release Ca²⁺ from internal stores but initially stimulates Ca²⁺ influx from the extracellular space. Note that there is a “crossover” phenomenon seen in this experiment at the three highest concentrations of MTX tested. This was a consistent finding (see Figs.3 A and8 A); the maximum level of apparent [Ca²⁺]_iwas generally greater for the lower concentrations vs. the higher concentrations of MTX, at least over this time frame.

Previous studies in fibroblasts showed that the [Ca²⁺]_iresponse to MTX was composed of two components; the first phase reflects the influx of Ca²⁺ from the extracellular space, whereas the second phase reflects pore formation and the concomitant efflux of fura 2. To test the hypothesis that the second phase of the [Ca²⁺]_iresponse to MTX in control THP-1 cells reflects pore formation, the effect of MTX on ethidium uptake was examined (Fig.2). MTX produced an increase in ethidium uptake following a time delay that decreased as the concentration of MTX increased; the uptake of ethidium correlated closely in time with the second phase of the [Ca²⁺]_iresponse. Thus, as was seen in fibroblasts (31), the fura 2 response shown in Fig. 1 at early times reflects Ca²⁺ influx via MTX-induced channels but at later times probably reflects fura 2 efflux following pore formation. At intermediate concentrations of MTX, the maximum level of ethidium uptake appeared to be attenuated as evidenced by the crossover of the lower concentrations (Fig.2 A, traces c and d). At higher MTX concentrations (>50 pM), there was a progressive decrease in the time delay and an increase in the maximum level of ethidium uptake (Fig2 B). The crossover phenomenon observed in the ethidium uptake experiments was a consistent finding (see Figs. 4 A and9 A).

To more closely examine the extent of the crossover phenomenon, longer time courses were performed. As seen in Fig.2 C, ethidium uptake at 20 and 50 pM produced almost maximum ethidium uptake in this batch of THP-1 cells. Ethidium uptake occurred with a significant time delay that decreased as the concentration of MTX increased. Once activated, however, ethidium uptake proceeded in an all-or-nothing fashion, reaching >80% within 8 min of activation. In contrast, ethidium uptake induced by 300 pM MTX occurred with shorter delay but was clearly attenuated, giving rise to a prominent crossover phenomenon. This attenuation was sustained, but ethidium uptake continued to increase slowly with time, reaching ∼60% over this time frame.

Effect of P2Z/P2X₇ receptor upregulation on the response to MTX.

THP-1 monocytes were incubated with LPS and IFN-γ for 48 h to upregulate P2Z/P2X₇ as previously described (18). Increased expression of P2Z/P2X₇ was confirmed in the present study by primer-specific amplification using RT-PCR (data not shown). Compared with matched controls, LPS and IFN-γ pretreatment increased basal resting [Ca²⁺]_ifrom 65.2 ± 12 to 115 ± 27 nM (means ± SE,n = 3,P = 0.085) and produced an attenuation the response of the cells to MTX (Fig. 3). The primary effect of LPS and IFN-γ pretreatment was on the second phase of the response to MTX (e.g., compare the [Ca²⁺]_iresponse to 20 pM MTX in control and LPS- and IFN-γ-treated cells; Fig. 3, trace c inA andB). This result suggests that upregulation of P2Z/P2X₇ may reduce MTX-induced pore formation. Consistent with this hypothesis, LPS and IFN-γ pretreatment slightly inhibited MTX-induced ethidium uptake at concentrations <0.5 nM (Fig. 4). At 0.5 nM MTX, ethidium uptake in LPS- and IFN-γ-treated cells was actually greater than in controls and the crossover phenomenon was not observed over this concentration range. This suggests that the crossover may be related to [Ca²⁺]_i, which is clearly attenuated in the cells expressing the P2Z/P2X₇ receptor. Importantly, control THP-1 cells, which essentially lack the P2Z/P2X₇ receptor, and pretreated cells, which have a high level of receptor expression, both respond to MTX. This result supports the hypothesis that the channels activated by MTX are distinct from P2Z/P2X₇.

Effect of MTX on heterologously expressed P2Z/P2X₇ receptor.

To further test the hypothesis that MTX activates the P2Z/P2X₇ receptor, the effect of MTX on [Ca²⁺]_iand ethidium uptake was examined in wild-type HEK cells and in HEK cells stably expressing the P2Z/P2X₇ receptor. Basal, nonstimulated [Ca²⁺]_iwas 56.4 ± 6.9 and 50.2 ± 3.8 nM (means ± SE;n = 3) in wild-type HEK and P2Z/P2X₇-expressing cells (Fig. 5). The wild-type HEK cells were essentially unresponsive to Bz-ATP (up to 200 μM), a specific P2Z/P2X₇ receptor agonist. In sharp contrast, addition of Bz-ATP to HEK cells expressing the P2Z/P2X₇ receptor produced a concentration-dependent increase in both [Ca²⁺]_i(Fig. 5 A) and ethidium uptake (Fig.5 B), consistent with both channel and pore formation. The responses of these two cell types to MTX are shown in Figs. 6 and7. MTX produced a time- and concentration-dependent increase in both apparent [Ca²⁺]_iand ethidium uptake. The response was virtually identical for wild-type HEK cells and the P2Z/P2X₇-expressing cells. Together with the results obtained in the THP-1 monocytes, these results strongly suggest that expression of the P2Z/P2X₇ receptor has little effect on MTX-induced channel activation, a result that is inconsistent with MTX acting as a high-affinity ligand for the P2Z/P2X₇ receptor.

Effect of temperature on MTX-induced pore formation.

The above results provide evidence that MTX activates a Ca²⁺-permeable channel that is distinct from the P2Z/P2X₇receptor channels. However, the experiments do not eliminate the possibility that activation of the MTX-sensitive channel and the P2Z/P2X₇ channel induce the formation of a common cytolytic pore. To test this hypothesis, we have compared the properties of the P2Z/P2X₇-induced pores to those activated by MTX. Previous studies have shown that a decrease in temperature has little effect on P2Z/P2X₇-induced channel activity (i.e., the change in [Ca²⁺]_i) but substantially attenuates pore formation (i.e., dye uptake) (26,33). To determine if the MTX-induced pores have the same characteristic, we compared the MTX response of the control THP-1 monocytes at 37 and 22°C (Figs. 8 and9). Although there is a shift in the apparent ED₅₀ for MTX-induced change in [Ca²⁺]_i, large increases in [Ca²⁺]_iwere clearly observed at 0.2 and 0.5 nM MTX at 22°C (Fig.8 B). The response, however, lacked the biphasic characteristic seen at 37°C, suggesting that the second phase of the response was attenuated. Consistent with this hypothesis, ethidium uptake was dramatically inhibited at all concentrations of MTX when examined at 22°C (Fig. 9). Thus MTX-induced pore formation has a temperature sensitivity similar to that previously reported for the P2Z/P2X₇ pore.

Effect of MTX on [Ca²⁺]_iand ethidium uptake in BW5147.3 lymphoma cells.

Previous studies suggest that BW5147.3 lymphoma cells express functional P2Z/P2X₇ channels that are poorly coupled to pore formation (20). The effect of MTX in this cell line is shown in Fig. 10. Basal, nonstimulated [Ca²⁺]_iin BW5147.3 cells was 197 ± 16 nM (means ± SE;n = 4), a value that is approximately threefold greater than that measured in either control THP-1 monocytes (P < 0.002) or in wild-type HEK cells (P < 0.001). MTX (0.05 and 0.2 nM) produced a large, concentration-dependent increase in [Ca²⁺]_iin BW5147.3 lymphoma cells that increased gradually with time over 6 min. In contrast, uptake of ethidium was unaffected by MTX at these concentrations, even when examined for extended periods of time (13 min; see Fig. 10, inset). Ethidium uptake was, however, observed in BW5147.3 cells in response to supermaximal concentrations (2–20 nM) of MTX (not shown). Thus BW5147.3 cells apparently have MTX-activated channels that are poorly coupled to pore formation, as previously noted for the P2Z/P2X₇ response in this cell line.

Comparison of MTX- and Bz-ATP-induced pore size.

Another characteristic of the P2Z/P2X₇-induced pore is that the apparent size exclusion varies with cell type (10). We, therefore, compared MTX- and Bz-ATP-induced ethidium (314 Da), YO-PRO (375 Da), and POPO-3 (715 Da) uptake in HEK cells and THP-1 monocytes expressing the P2Z/P2X₇ receptor. In HEK cells, both Bz-ATP (200 μM) and MTX (1 nM) produced a time-dependent increase in ethidium uptake (Fig. 11); however, uptake of YO-PRO was much slower, and uptake of POPO-3 was essentially zero over this time frame. Thus MTX- and Bz-ATP-induced pores activate with a similar time course and have the same relative permeability in HEK cells, i.e., ethidium > YO-PRO ≫ POPO-3. In sharp contrast to the HEK cells, MTX (50 pM) and Bz-ATP (200 μM) increased the uptake of all three dyes into THP-1 monocytes (Fig.12). Thus the pores formed in THP-1 monocytes appear to be larger than those seen in HEK cells and, as suggested above, seem to activate in an all-or-nothing fashion. One note of caution, release of LDH in response to Bz-ATP occurs very soon after ethidium uptake in THP-1 cells pretreated with LPS and IFN-γ (18). This could explain, in part, the differences between HEK and THP-1 cells with respect to apparent size exclusion. Importantly, however, the profiles of permeability change induced by MTX and Bz-ATP are essentially identical in the two cell types with respect to kinetics of activation and apparent size exclusion.

DISCUSSION

MTX is one of the most potent marine toxins isolated to date. This toxin produces increases in [Ca²⁺]_iin all cell tested via the activation of Ca²⁺-permeable, nonselective cation channels. The results of the present study demonstrate that THP-1 monocytes are extremely sensitive to MTX with apparent ED₅₀ of ∼10 pM, estimated from the initial change in [Ca²⁺]_i. In our accompanying study using human skin fibroblasts (31), we showed for the first time that MTX also activates large pores that allow molecules with molecular masses up to 800 Da to enter or exit the cell. In fibroblasts, this pore formation occurred with a short delay following channel activation, and a similar result was obtained in the present study in HEK cells. However, in THP-1 monocytes, a considerable lag was observed between the initial activation of the channel and subsequent pore formation. Once pore formation was initiated in the THP-1 monocytes, it occurred in an all-or-nothing fashion, causing large increases in ethidium, YO-PRO-1, and POPO-3 uptake into the cells. The lag phase would suggest that there may be several biochemical steps between channel activation and pore formation or that the concentration of a critical messenger or factor may need to exceed a threshold value before activation of the pore is observed. An enzymatic step(s) between channel activation and pore formation is also supported by the sensitivity of pore formation to temperature. As shown in our previous studies on fibroblasts, reducing the temperature substantially increased the lag time between channel activation and pore formation, and the results of the present study on THP-1 monocytes clearly shows that, at concentrations of MTX that substantially increase [Ca²⁺]_i, pore formation is essentially reduced to zero at 22°C. The biochemical events responsible for coupling of channel activation to pore formation remain unknown. In this regard, the crossover phenomenon seen in the THP-1 monocytes may provide a clue. At low concentrations of MTX, pore formation appears to occur in an all-or-nothing fashion after a considerable lag time. However, at the higher concentrations of MTX associated with the crossover phenomenon, (i.e., inhibition of pore formation), a rapid increase in [Ca²⁺]_iwas observed. Thus an increase in [Ca²⁺]_imay produce feedback inhibition. Interestingly, the resting [Ca²⁺]_iin the BW5147.3 lymphoma cells is approximately threefold higher than in THP-1 cells and pore formation induced by either Bz-ATP or MTX is substantially attenuated, supporting this hypothesis. Understanding this regulation will require greater knowledge of the molecular/biochemical link between channel activation and pore formation.

The only documented example of channels linked to activation of large pores are members of the P2X ionotropic receptor family (21, 40), of which P2Z/P2X₇ is the best-studied example. The P2Z/P2X₇ receptor is a protein of 595 amino acids in length possessing two putative transmembrane segments with cytoplasmic NH₂- and COOH-terminal domains (35). It has a structure similar to the inwardly rectifying K⁺ channels, amiloride-sensitive Na⁺ channels and the mechanosensitive Mec channels of Caenorhabditis elegans (2). Each of these channel families has a cytoplasmic NH₂ and COOH terminus, two membrane spanning segments, and a permeation pathway formed by a reentrance loop immediately preceding the second transmembrane segment. These channels are thought to form homo- and heteromeric channel structures. On the basis of structural similarity, the P2X purinergic receptor family may form channels in a similar fashion, i.e., by homo- or heteromeric assembly (39), and it has been suggested that P2Z/P2X₇-induced pore formation reflects the progressive association/aggregation of channel-forming subunits into a large pore structure (26, 36), although evidence to support such a biophysical model is scant. It became clear early in the present study that MTX-induced pores share many characteristics with the pores formed following activation of P2Z/P2X₇ receptors/channels. Therefore, we initially focused on the intriguing possibility that MTX was a high-affinity ligand for P2Z/P2X₇. The results of the present study however, demonstrate that this is not the case. First, untreated THP-1 cells and wild-type HEK-293 cells show little response to Bz-ATP, suggesting that functional P2Z/P2X₇ receptors are not expressed or are expressed at very low levels, yet both cell types are sensitive to MTX, exhibiting both large increases in [Ca²⁺]_iand fluorescent dye uptake. Thus the MTX-activated channel and pore are functional in the apparent absence of P2Z/P2X₇ receptor. Second, upregulation of the P2Z/P2X₇ in THP-1 cells by pretreatment with LPS and IFN-γ greatly enhances both the rise in [Ca²⁺]_iand the uptake of fluorescent dyes in response to Bz-ATP (18) but slightly attenuates the response of the cells to MTX. Third, heterologous expression of P2Z/P2X₇ in HEK-293 cells dramatically increases the response of the cells to Bz-ATP but has little or no effect on the response to MTX. Together, these results demonstrate that expression of P2Z/P2X₇ does not correlate with MTX responsiveness, and provide strong support for the conclusion that P2Z/P2X₇ is not the MTX-activated channel.

Although the channels appear to be different, the MTX- and Bz-ATP-induced pores are indistinguishable, leading to the hypothesis that MTX and P2Z/P2X₇ activate a common cytolytic pore. This conclusion is based on the following results. First, all cells examined in the present study that exhibit Bz-ATP-induced pores also possess the MTX-induced pore. Second, the pores for both agonists activate with a similar time course following stimulation of the channel activity and this time course of activation is cell specific. Thus, as discussed above, MTX- and Bz-ATP-induced channel activation is followed closely in time by the formation of pores in HEK cells, but this activation requires a considerable lag time in THP-1 cells. Third, the MTX and Bz-ATP-induced pores have similar size exclusions. In HEK, there is a clear difference in ethidium, YO-PRO-1, and POPO-3 uptake for both MTX and Bz-ATP, whereas, in THP-1 cells, uptake of each dye is observed for both agonists. Fourth, BW5147.3 lymphoma cells appear to have Bz-ATP- (20) and MTX-induced channels that are poorly coupled to pore formation. Finally, fifth, both MTX- and Bz-ATP-induced pores have similar temperature sensitivity. Overall, we could find no characteristic that would distinguish between these two pores. Although we cannot eliminate the possibility that each agonist activates different channels and different pores, it is difficult to reconcile the similarities with such a model; even cell-specific characteristics such as kinetics of activation, size exclusion, and the inefficient coupling between channel and pore seen in BW5147.3 cells are essentially identical for MTX- and Bz-ATP-induced pores. Thus the simplest explanation for these results is that MTX and Bz-ATP activate distinct channels but a common cytolytic pore.

Evidence in the literature demonstrating identity between P2Z/P2X₇ channels and pores is very limited and relies on heterologous expression. These reports show that P2X receptor expression is required for responsiveness to externally applied purinergic receptors agonists and that channels formed by different P2X subunits, either alone or in various combinations, possess unique biophysical properties (13, 14, 21, 24,26, 28, 34, 39, 40). Thus it is likely that the P2X protein forms at least part of both the agonist binding domain and the channel permeation pathway. The results of the present study support this hypothesis. However, it remains possible that heterologously expressed P2Z/P2X₇ protein does not form the pore but rather is coupling (either physically or biochemically) to endogenous pores in the cell types used for expression. The most common cell types used for heterologous expression of P2Z/P2X₇ receptors are the HEK cell and frog oocyte. In this regard, it is clear from the present study that HEK cells have an endogenous pore that is activated by MTX. Furthermore, it has recently been shown (1) that frog oocytes have an MTX-induced nonselective cation current that initially excludes the large cation N-methyl-d-glucamine (NMDG) (167 Da). Although these authors did not specifically test for pore formation, it is clear from their current recordings that a conductance to NMDG developed at later times following MTX addition to the external bath, consistent with pore formation. Furthermore, all cells examined to date respond to MTX with large increases in [Ca²⁺]_iand/or downstream Ca²⁺-dependent signaling events (1, 6, 8, 9, 16, 17, 22, 25, 29). Thus it seems likely that most, if not all cells, have an endogenous MTX-activated cytolytic pore, including cell types used for heterologous expression of P2Z/P2X₇ receptors.

The model shown in Fig. 13 proposes that the P2X purinergic ionotropic receptors and the channels activated by MTX interact with a common cytolytic/oncotic pore, or COP. Although the model suggests that the coupling between channel and COP is membrane delimited, cytosolic proteins or factors may play a role in activation of COP. As was seen in Fig. 4, however, overexpression of P2Z/P2X₇ in THP-1 cells causes an attenuation of MTX-induced pore formation, which could be overcome by increasing the concentration of MTX. Thus it appears that P2Z/P2X₇ may compete with the MTX-induced channel for activation of COP, suggesting a preexisting channel-pore complex. In this regard, attenuation of the MTX response by expression of P2Z/P2X₇ in THP-1 cells may reflect a greater affinity of P2Z/P2X₇ for COP relative to the MTX channel or alternatively may reflect the amount and/or number of P2Z/P2X₇receptors relative to MTX channels in these cells. Irrespective of the exact mechanism of coupling between channel and COP, MTX should provide a useful tool for understanding this interaction.

Although the experiments of the present study focused on P2Z/P2X₇-mediated and MTX-induced COP activation, two recent studies demonstrate that other members of the P2X family can, to various extents, initiate pore formation and that mutations at specific residues within the second membrane spanning segment can differentially alter transition from channel to pore (21,40). These experiments were performed in both frog oocytes and HEK cells heterologously expressing the various P2X receptor subtypes and mutants. Thus, although these investigators interpreted their results as reflecting a conformational change or dilation of the P2X permeation pathway, the findings are not inconsistent with differential activation of the endogenous COP in these cell types. Furthermore, P2X- and MTX-activated channels may not be the only channels linked to COP. Indeed, activation of a number of different channel types has been shown to produce oncotic cell death. These include the VR₁ vanilloid receptor, which is responsible for pain sensation associated with stimulation by capsaicin, the active compound in chili peppers (3), the NR₁-NR₂ N-methyl-d-aspartate glutamate receptor, which has been implicated in excitatory neurotoxicity (5, 7, 37, 38), and the degenerins family of ion channels, responsible for hereditary neurodegeneration inC. elegans (4). Interestingly, heterologous expression in HEK cells of the mammalian degenerin MDEG, with the same mutation as that found in C. elegans neurodegeneration phenotypes, gives rise to constitutively active channels that cause cell death by swelling (41). Thus it is possible that the activity of a number of different channels may be coupled to oncosis by activation of COP.

Last, activation of COP need not always proceed to cell death by oncosis. Substantial evidence supports the hypothesis that apoptosis and oncosis reflect different endpoints of a common cell death pathway (11, 23, 27, 37). In this regard, activation of COP by extracellular ATP is associated with both apoptosis and oncosis in mesangial cells (32) and thymocytes (44). Furthermore, Warny and Kelly (42) recently showed that oncosis in THP-1 monocytes is mediated by loss of cellular K⁺ and activation of caspase-like proteases (42), a hallmark of apoptotic cell death. Under normal ionic conditions, it seems likely that activation of COP will lead to pronounced loss of cellular K⁺ in exchange for extracellular Na⁺ and Ca²⁺, which may lead to either apoptosis or oncosis, depending possibly on the metabolic status of the cell. In any event, it will be important to define the role of COP in both types of cell death and to identify and characterize the COP protein at the molecular level.

FOOTNOTES

• This work was supported by National Institute of General Medical Sciences Grants GM-52019 (W. P. Schilling) and GM-36387 (G. R. Dubyak) and by a Postdoctoral Fellowship awarded to W. G. Sinkins by the American Heart Association, Northeast Ohio Affiliate.

• The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.