Caspase Independent/Dependent Regulation of K+, Cell Shrinkage, and Mitochondrial Membrane Potential during Lymphocyte Apoptosis*

The loss of cell volume is a fundamental feature of apoptosis. We have previously shown that DNA degradation and caspase activity occur only in cells which have shrunken as a result of potassium and sodium efflux (Bortner, C. D., Hughes, F. M., Jr., and Cidlowski, J. A. (1997) J. Biol. Chem. 272, 32436–32442). Furthermore, maintaining a normal intracellular potassium concentration represses the cell death process by inhibiting the activity of apoptotic nucleases and suppressing the activation of effector caspases (Hughes, F. M., Jr., Bortner, C. D. Purdy, G. D., and Cidlowski, J. A. (1997) J. Biol. Chem. 272, 30567–30576). We have now investigated the relationship between cell shrinkage, ion efflux, and changes in the mitochondrial membrane potential, in addition to the role of caspases in these apoptotic events. Treatment of Jurkat cells with a series of inducers which act via distinct signal transduction pathways, resulted in all of the cell death characteristics including loss of cell viability, cell shrinkage, K+ efflux, altered mitochondrial membrane potential, and DNA fragmentation. Interestingly, only cells which shrunk had a loss of mitochondrial membrane potential and the other apoptotic characteristics. Treatment of Jurkat cells with an anti-Fas antibody in the presence of the general caspase inhibitor z-VAD, abrogated these features. In contrast, when Jurkat cells were treated with either the calcium ionophore A23187 or thapsigargin, z-VAD failed to prevent cell shrinkage, K+ efflux, or changes in the mitochondrial membrane potential, while effectively inhibiting DNA degradation. Treatment of Jurkat cells with various apoptotic agents in the presence of either the caspase-3 inhibitor DEVD, or the caspase-8 inhibitor IETD also blocked DNA degradation, but failed to prevent other characteristics of apoptosis. Together these data suggest that the cell shrinkage, K+ efflux, and changes in the mitochondrial membrane potential are tightly coupled, but occur independent of DNA degradation, and can be largely caspase independent depending on the particular signal transduction pathway.

Apoptosis is characterized by a distinct set of morphological and biochemical features which differ substantially from those observed during necrosis (1)(2)(3). Characteristics of apoptosis include cell shrinkage, protein degradation, nuclear condensation and fragmentation, and eventual budding of the cells into what have been termed apoptotic bodies (2,3). In apoptosis, diverse external or internal signals trigger the cell death process, which is then executed by the activation of a central or common death pathway. Upon receipt of an apoptotic signal, cells activate numerous signaling pathways whose function is to modulate the cell death process. Modulators of cell death include proteins of the Bcl-2 family, p53, and various kinases and phosphatases, which condemn the cell to death through their action on intracellular effector molecules (4). Notable effector molecules include cytochrome c, proteases, and nucleases which ultimately cleave protein and DNA substrates, respectively. The pathway that leads to cell shrinkage, alterations in mitochondrial membrane potential, and DNA degradation are not yet completely understood.
Recently, the study of mitochondria and changes in the mitochondrial membrane potential have become a focus of apoptosis regulation. A loss of mitochondrial membrane potential has been shown to occur in a variety of apoptotic model systems (5)(6)(7). The opening of mitochondrial permeability transition (MPT) 1 pores, located on the inner mitochondrial membrane, is thought to underlie the loss of mitochondrial membrane potential (8,9). Activation of these MPT pores permits the redistribution of molecules across the inner mitochondrial membrane, thus disrupting the membrane potential of this organelle (10 -12). Induction of the MPT has been suggested to result in the release of several apoptotic factors, including apoptotic proteases (apoptotic-inducing factor; Ref. 13) and cytochrome c (14). Other experimental observations suggest that disruption of electron transport, oxidative phosphorylation, and adenosine triphosphate (ATP), as well as the generation of reactive oxygen species are thought to play an important roles in apoptosis (15)(16)(17). Currently, however, the exact role that the MPT plays during apoptosis is not apparent and its relationship to other apoptotic events has not been clearly defined.
Along with changes in the mitochondrial membrane potential, expression of a number of genes has also been reported to be pivotal in apoptosis. Studies in Caenorhabditis elegans have shown that the cysteine protease Ced-3 is essential for programmed cell death (18). Sequence similarity between Ced-3 and the human interleukin-1␤-converting enzyme (19) led to the identification of a family of interleukin-1␤-converting enzyme-related cysteine proteases that become activated during apoptosis. These proteases, collectively known as caspases, have been shown to play a critical role in apoptosis (20). Caspases are synthesized as inactive proenzymes, which are activated following cleavage at specific aspartate residues, to become mature active enzymes. Interestingly, many of the pro-caspase cleavage sites are also caspase recognition sites (21), raising the possibility that some caspases sequentially activate others, thus establishing a hierarchy of caspase activation (22)(23)(24). This cascade-like activation of activator and effector caspases permits the amplification of the proteolytic cleavage process during apoptosis only if the normal intracellular K ϩ levels are reduced (25,26). Currently the precise mechanism by which these proteases contribute to the classical apoptotic characteristics are still being elucidated.
A fundamental characteristic of apoptosis that occurs in all model systems of programmed cell death is the loss of cell volume. We have previously shown that the loss of cell volume correlates with the population of cells which have degraded DNA (25,27). Additionally, we have determined that cell shrinkage is associated with a dramatic decrease in intracellular ion concentration, particularly potassium, and that the ionic environment within cells can affect both the activation and activity of apoptotic enzymes (25,26). We have now evaluated the relationship between cell shrinkage and loss of the mitochondrial membrane potential during apoptosis and examined the role of caspases in the regulation of cell volume loss, the ion efflux, and alterations in the mitochondrial membrane potential. We show that when the death response is activated through different apoptotic signaling pathways, cell shrinkage, K ϩ efflux, and the loss of mitochondrial membrane potential are tightly linked, but all three can occur independently of both caspase activation and DNA degradation.

MATERIALS AND METHODS
Cell Culture and Reagents-Jurkat cells (human lymphoma) were cultured in RPMI 1640 media containing 10% heat-inactivated fetal calf serum, 4 mM glutamine, 31 mg/liter penicillin, and 50 mg/liter streptomycin at 37°C, 7% CO 2 atmosphere. Jurkat cells (4 ϫ 10 5 cells/ml) were initially treated with 10 ng/ml anti-human Fas IgM (Kamiya Biomedical) for 4, 8, and 12 h. In latter studies, Jurkat cells were treated with either 10 ng/ml anti-human Fas IgM, 1 M A23187 (Calbiochem), or 7.5 M thapsigargin (Sigma) for 24 h to achieve a similar amount of loss of membrane integrity, in the presence or absence of 25 M of the various caspase inhibitors (z-VAD-fmk, DEVD-fmk, and IETD-fmk; Kamiya Biomedical) unless indicated otherwise. Cell viability was determined by trypan blue exclusion.
Determination of Cell Size by Flow Cytometry-Cell size and changes in the light scattering properties of the cells were determined by flow cytometry as described previously using a Becton Dickinson FACSort equipped with CellQuest software (27). For each sample, 10,000 cells were examined by exciting the cells with a 488 nm argon laser and determining their position on a forward scatter versus side scatter dot plot. Light scattered in the forward direction is roughly proportional to cell size, while light scattered at a 90 o angle (side scatter) is proportional to cell density (28). Therefore, as a cell shrinks or loses cell volume, a decrease in the amount of forward scattered light is observed, along with a slight change in side scattered light. A gate based on the properties of the control cells was set on each forward scatter versus side scatter dot plot to separate the normal and apoptotic populations of cells, and remained constant throughout the analysis. The percent of apoptotic cells was determined by statistical analysis of the dot plots using CellQuest software. The data was converted to three-dimensional plots for presentation.
Determination of Intracellular Potassium by Flow Cytometry-Intracellular potassium concentrations were determined as described previously using a Becton Dickinson FACSVantage (25). Briefly, Jurkat cells treated in the presence or absence of various apoptotic agents and caspase inhibitors were loaded with the potassium-sensitive fluorescent dye potassium-binding benzofuran isophthalate (PBFI-AM; Molecular Probes) to a final dye concentration of 5 M for 1 h at 37°C, 7% CO 2 atmosphere prior to examination. Immediately prior to flow cytometric analysis, propidium iodide (PI, Sigma) was added to each sample to a final concentration of 10 g/ml. Ten thousand cells were analyzed by sequential excitation of the cells containing PBFI-AM and PI at 340 -350 and 488 nm, respectively. Gates were set on a PBFI (K ϩ ) versus PI dot plot to individually examine cells which had a reduced potassium concentration. Cells which were PI positive, indicating a loss of mem-brane integrity, were excluded from ion analysis. The data were converted to three-dimensional plots using CellQuest software for presentation.
Measurement of Mitochondrial Membrane Potential-Changes in the mitochondrial membrane potential were measured by flow cytometry using JC-1 (Molecular Probes). Thirty minutes prior to cytometric analysis, JC-1 was added to 1-ml of cells to a final concentration of 10 M and incubated at 37°C, 7% CO 2 atmosphere. At the designated time, 10,000 cells were examined for each sample on a FL-1 (530 nm) versus FL-2 (585 nm) dot plot on a Becton Dickinson FACSort. JC-1 has dual emission depending on the state of the mitochondrial membrane potential. JC-1 forms aggregates in cells with a high FL-2 fluorescence indicating a normal mitochondrial membrane potential. Loss of the mitochondrial membrane potential results in a reduction in FL-2 fluorescence with a concurrent gain in FL-1 fluorescence as the dye shifts from an aggregate to monomeric state. Therefore, retention of the dye in the cell can be monitored through the increase in FL-1 fluorescence. The data were converted to density plots using CellQuest software for presentation.
DNA Analysis by Flow Cytometry-The DNA content for each sample was determined as described previously (27). Briefly, 5 ml of cells were pelleted from the culture medium and fixed by the slow addition of cold 70% ethanol to a volume of approximately 1.5-ml. The volume of each sample was adjusted to 5-ml with cold 70% ethanol, and the cells were stored at 4°C overnight. For flow analysis, the fixed cells were pelleted, washed once in 1 ϫ phosphate-buffered saline and stained in 1 ml of 20 g/ml of PI, 1 mg/ml RNase in 1 ϫ phosphate-buffered saline for 20 min. Seven thousand five hundred cells were examined by flow cytometry by gating on an area versus width dot plot to exclude cell debris and cell aggregates. The percent of degraded DNA was determined by the number of cells with subdiploid DNA divided by the total number of cells examined under each experimental condition.

The Loss of Intracellular Potassium Is Associated with the Shrunken Population of Apoptotic Cells-Previous studies from
our laboratory have shown that DNA degradation and loss of intracellular potassium (K ϩ ) occurs only in the shrunken apoptotic cells (25,27). We have now kinetically analyzed the relationship between K ϩ efflux and the loss of cell volume during apoptosis in Jurkat cells treated with anti-Fas. As shown in Fig. 1A, forward scatter versus side scatter threedimensional plots showed an increase in the number of cells which lose cell volume, which occurred in a time-dependent manner. Cells which have a decrease in cell volume have a reduced ability to scatter light in the forward direction. Furthermore, these apoptotic cells also showed a slight increase in their ability to scatter light at a 90°angle, indicating a concurrent increase in cellular density. We next analyzed these cells for the loss of intracellular K ϩ at the single cell level by flow cytometry utilizing the fluorescent potassium indicator dye PBFI-AM and propidium iodide to eliminate cells which loss their membrane integrity. The loss of intracellular potassium, as determined by a decrease in PBFI (K ϩ ) fluorescence in the viable shrunken population of anti-Fas-treated Jurkat cells, was also observed in a time-dependent manner with similar kinetics (Fig. 1A). To explore the relationship between cell shrinkage and the decrease in K ϩ concentration, we evaluated the K ϩ concentration in the population of shrunken and nonshrunken cells obtained from the kinetic study shown in Fig.  1A. Fig. 1B shows that cells which fail to shrink show no reduction in intracellular K ϩ concentration. In contrast, only shrunken cells had a decrease in intracellular K ϩ concentration, indicating that decreased K ϩ concentration and cell shrinkage are tightly coupled events in apoptosis.
The Loss of Mitochondrial Membrane Potential Is Associated with the Shrunken Population of Apoptotic Cells-Since the loss of cell volume is a fundamental characteristic of apoptosis, we next determined if other features of apoptosis were restricted to the shrunken population of cells. Recently, several laboratories have suggested that nuclear features of apoptosis are preceded by alterations in mitochondrial structure and transmembrane potential (5, 8, 29 -34). The fluorescent probe JC-1 has been shown to be most specific for measuring changes in the mitochondrial membrane potential (MMP) (35), and was used for our analysis of the MMP. JC-1 forms either J-aggregates or monomers, depending on the state of the mitochondrial membrane potential, with the emissions of the two dye forms detectable by flow cytometry at 585 or 530 nm, respectively (35). The high mitochondrial membrane potential of normal cells loaded with JC-1 allows for the formation of J-aggregates, detected by a peak in fluorescence at 585 nm. As the mitochondrial membrane potential is loss, these aggregates dissipate into monomers, which are detected as a shift in fluorescence from 585 to 530 nm. We initially examined JC-1 along with DiOC 6 (Molecular Probes) for their ability to detect acute changes in both the MMP and plasma membrane potential (PMP). As shown in Fig. 2, JC-1 responds only to an acute uncoupling of the MMP using CCCP (Sigma). In contrast, DiOC 6 responds to both acute changes in the MMP and PMP. These data suggest that JC-1 measures only changes in the MMP whereas DiOC 6 measures both MMP and PMP. This observation cautions against the use of DiOC 6 to specifically examine mitochondrial events. Fig. 3A shows the time-dependent loss of the MMP in anti-Fas-treated Jurkat cells, as the population of cells shifts to a lower JC-1 aggregate state and a concurrent higher JC-1 monomer state. To determine the relationship between the loss of cell volume and changes in the MMP, we compared the response of the mitochondrial dye JC-1 to changes in cell size which occur during apoptosis by flow cytometry. Examination of each individual JC-1 state (J-aggregate and monomeric) versus cell size in control and anti-Fas-treated Jurkat cells showed that only the shrunken population of cells had a decrease in JC-1 aggregates and an increase in JC-1 monomers (Fig. 3A). The increase in JC-1 monomeric fluorescence with this dye suggests that the loss of JC-1 aggregate fluorescence is not due to an overall loss of this dye from the cell. Similar results were observed in A23187 and thapsigargin-treated Jurkat cells (Fig. 3B). Therefore, the loss of cell volume and MMP in apoptotic cells appears to be tightly coupled and is independ- A, flow cytometry was used to assess the light scattering properties of the cells and the loss of intracellular K ϩ . Ten thousand cells were examined on a FACSort flow cytometer for the ability of cells to scatter light in the forward direction (forward scatter), which indicates cell size, and at a 90°angle (side scatter), which indicates cell density. Cells which have a decrease in forward scatter light have a decrease in cell size. Cells which have an increase in side scatter light have an increase in cell density. For intracellular K ϩ analysis, 2 l of a 2.5 mM PBFI-AM stock (5 M final) was added to 1 ml of cells for each sample 1 h prior to the time of examination. Incubation was then continued at 37°C, 7% CO 2 . Immediately prior to cytometric examination, PI was added to a final concentration of 10 g/ml. Samples were analyzed on a FACSVantage flow cytometer examining 10,000 cells per sample on a PBFI (K ϩ ) versus PI fluorescent dot plot to eliminate the PI positive (dead) cells. Viable cells were then examined on a forward scatter versus PBFI (K ϩ ) three-dimensional plot. B, each viable population of cells (normal and shrunken) were analyzed by flow cytometry on a forward scatter versus PBFI (K ϩ ) fluorescence contour plot. Only the apoptotic or shrunken population of cells had a decrease in potassium concentration. All data shown represents one of at least three independent experiments. ent of the mechanism used to signal cell death.
Diverse Caspase Inhibitors Protect against DNA Degradation in Jurkat Cells Treated with Numerous Apoptotic Agents-We next examined the role that caspase activity plays in the generation of several apoptotic features by using various caspase inhibitors which are known to block both activator and effector caspases. We examined cells after 24 h of apoptotic treatment to ensure complete activation of the cell death program. Since DNA degradation has been a persistent characteristic of apoptosis and is thought to occur downstream of caspase activation during cell death, we utilized this apoptotic end point to evaluate the effectiveness of our caspase inhibitors. Jurkat cells treated with an anti-Fas antibody, A23187, or thapsigargin showed all of the classical characteristics of apoptosis, including DNA degradation, as determined by an increase in the number of cells with a subdiploid peak of DNA by flow cytometry (Fig. 4). z-VAD, a cell-permeable, broad spectrum inhibitor of caspase activity, has previously been shown to prevent apoptotic death in a number of apoptotic model systems by inhibiting both early and late proteases (reviewed in Ref. 8). When Jurkat cells were treated with various apoptotic agents for 24 h, the presence of 25 M z-VAD completely inhibited DNA degradation (Fig. 4, top). Similar results were observed when either 25 M DEVD (a caspase-3 inhibitor) or IETD (a caspase-8 inhibitor) were used (Fig. 4, middle and bottom), indicating that the concentration of caspase inhibitors used in these experiments were effective in preventing a late apoptotic event.
Loss of Cell Viability Occurs in the Presence of Caspase Inhibition in a Signal-specific Manner-Having determined the efficacy of the caspase inhibitors in this model system, we next examined the effects of these various caspase inhibitors on their ability to prevent the loss of cell viability. In the absence of caspase inhibition, a significant increase in trypan blue positive cells was observed following a 24-h treatment of Jurkat cells with anti-Fas, A23187, or thapsigargin (Fig. 5). z-VAD was effective in preventing the loss of cell viability in the anti-Fas-treated cells (Fig. 5, top). However, z-VAD was only marginally effective in preventing the loss of cell viability in cells treated with A23187, and was decidedly ineffective in cells treated with thapsigargin (Fig. 5, top). Interestingly, the loss of cell viability occurred in the absence of DNA degradation (see Fig. 4), suggesting that membrane permeability changes, as assessed by trypan blue exclusion, can occur independent of DNA degradation and in a caspase-independent manner. Similarly, DEVD and IETD significantly blocked the loss of cell viability in the anti-Fas-treated cells. However, these inhibitors were ineffective where apoptosis was induced using a non-Fas signaling pathway (Fig. 5, middle and bottom). Thus although all caspase inhibition was effective in preventing DNA degradation under each apoptotic condition, the caspase inhibitors employed only blocked the loss of cell viability induced via Fas-receptor signaling. Higher concentration of the caspase inhibitors (100 M) were similarly ineffective in preventing the loss of cell viability except when anti-Fas antibodies were the death signal (data not shown). These data show that the characteristics of apoptosis can be separated from caspases activity and cell death appears to be linked to activation of the death signaling process.

Cell Shrinkage Occurs in a Signal-specific Manner in the Presence of Various Caspase Inhibitors-We have previously
shown that only the shrunken population of apoptotic cells contain degraded DNA (25,27). Since caspase inhibition prevented DNA degradation in response to numerous apoptotic stimuli, we wished to determine if cell shrinkage was caspase dependent. We thus examined the loss of cell volume in Jurkat cells treated with various apoptotic agents in the presence or absence of caspase inhibitors. Jurkat cells activated to undergo apoptosis with anti-Fas, A23187, or thapsigargin in the absence of caspase inhibition showed a significant reduction in cell size (Fig. 6, top row). In the presence of 25 M z-VAD, anti-Fas-treated Jurkat cells did not show a change in cell size, suggesting the dependence of caspase activity for cell shrinkage under this apoptotic condition, consistent with the data on cell viability. In contrast, z-VAD only marginally prevented the alterations in cell size in Jurkat cells treated with either A23187 or thapsigargin (Fig. 6, second row). Jurkat cells treated with anti-Fas in the presence of either 25 M DEVD or

FIG. 3. Changes in the mitochondrial membrane potential occur only in the shrunken population of cells.
Jurkat cells treated with: A, an anti-Fas antibody as described in the legend to Fig.  1, or B, 2 M A23187 or 10 M thapsigargin (Thaps) for 10 h, were examined for changes in their mitochondrial membrane potential using JC-1. Ten thousand cells were examined per sample on a JC-1 aggregate versus JC-1 monomer contour plot using a FACSort flow cytometer. JC-1 forms aggregates in cells with a high mitochondrial membrane potential, which is detected at 585 nm. The formation of monomers, indicative of decreased mitochondrial membrane potential, is detected as a loss of aggregate fluorescence, and an increase in JC-1 monomers detected at 530 nm. Each individual mitochondrial potential state was also examined versus forward scattered light (cell size). All data shown represents one of at least three independent experiments. IETD also had a significant effect on preventing the changes in cell size, such that cell shrinkage was reduced by 50 and 70%, respectively (Fig. 6, second column). Both DEVD and IETD at concentrations which effectively inhibited DNA degradation were completely ineffective in preventing cell shrinkage when either A23187 or thapsigargin was used to induce apoptosis (Fig. 6, third and fourth columns). These results suggest cell shrinkage can occur independently of caspase activation and activity except when anti-Fas is used to activate apoptosis.
K ϩ Efflux Can Occur Independently of Caspase Activity-The loss of intracellular ions, particularly potassium, has been proposed to be responsible for cell shrinkage during apoptosis (26, 27, 36 -39). Additionally, the inhibition of intracellular K ϩ loss was shown to prevent the loss of cell volume in several model systems, and in response to numerous apoptotic stimuli (26, 27, 36 -39). Therefore, we wanted to determine if the viable, shrunken population of cells observed in the presence of caspase inhibition also had a decrease in intracellular potassium. As shown in Fig. 7 (top row), when Jurkat cells were treated with anti-Fas, A23187, or thapsigargin, only the viable, shrunken population of cells had a decrease in intracellular potassium. Similar to the results observed for cell size, the presence of 25 M z-VAD completely prevented the loss of intracellular potassium in anti-Fas-treated Jurkat cells, along with the loss of cell size, but was only marginally effective when either A23187 or thapsigargin was used as the apoptotic stimulus (Fig. 7, second row). Both DEVD and IETD provided significant protection from potassium loss upon anti-Fas treatment, but failed to inhibit the loss of potassium in A23187-or thapsigargin-treated cells (Fig. 7, third and fourth rows). Thus caspase inhibition prevents the loss of ions and cell shrinkage when the Fas receptor pathway is used to stimulate apoptosis, but not when apoptosis is activated by alternative signaling pathways.
Changes in the Mitochondrial Membrane Potential Accompany the Loss of Cell Viability and Cell Size-To determine the effect of caspase inhibition on other characteristics of apoptosis in response to various apoptotic stimuli, changes in the MMP were examined using the mitochondrial specific dye JC-1. Treatment of Jurkat cells with anti-Fas, A23187, or thapsigargin in the absence of caspase inhibitors resulted in an increase in the percent of cells which showed a loss of MMP (Table I). The presence of z-VAD prevented only the change in MMP in anti-Fas-treated Jurkat cells. Both DEVD and IETD showed some protection from a loss of MMP in the anti-Fas-treated cells, however, these caspase inhibitors were ineffective when either A23187 or thapsigargin was used to stimulate apoptosis. The effect of caspase inhibition on MMP is consistent with other data in this study, being protective mainly in anti-Fastreated cells, but ineffective in preventing the change in membrane potential in cells treated with either A23187 or thapsigargin.

DISCUSSION
Apoptosis is defined by a distinct set of morphological and biochemical characteristics, but it is not clear how these various apoptotic characteristics relate to each other during cell death. Additionally, it is unclear what role caspases play in the generation of each apoptotic feature. Previously, we have shown that cells which lost intracellular potassium and degraded their DNA were largely coincident with cell shrinkage (25)(26)(27). We now show that the loss of MMP is also associated with the shrunken apoptotic cells. Using JC-1, a dye reliable for detecting changes in the mitochondrial membrane potential (35), we showed that only cells which have a decrease in cell size had a loss of MMP. Additionally, we showed that several characteristics of apoptosis, including the loss of cell volume, K ϩ efflux, and the loss of MMP can be largely caspase independent depending on the agent used to initiate the cell death response.
Recent studies have suggested that early alterations of mitochondrial function may be important for apoptosis (40). In addition to alterations in mitochondrial function, the release from mitochondria of various proapoptotic factors such as cytochrome c (14) and apoptosis-inducing factor (13) have been reported to be critical for apoptosis. The loss of the MMP is thought to release cytochrome c from the outer mitochondrial compartment, thus potentiating cell death (41). However, Kluck et al. (42) suggest that release of cytochrome c is not accompanied by mitochondrial depolarization, an event which follows the onset of the mitochondrial permeability transition. Additionally, it has been suggested that release of cytochrome c, and not the loss of MMP is the required step for initiation of the cell death program (43), implying that these events may be cell type or inducer specific features of apoptosis.
We have previously shown that a loss of intracellular potassium occurs in the shrunken population of apoptotic cells (25,27). We now show that the loss of MMP is also restricted to the shrunken population of cells, suggesting that loss of cell volume, K ϩ efflux, and loss of the MMP are tightly coupled. A recent study by Dallaporta et al. (39) supports our original publication that the loss of intracellular potassium occurs in cells which have lost their MMP; however, they suggest that these events precede the cell shrinkage. However, the relationship between intracellular potassium and MMP as determine by Dallaporta et al. (39) was based on measurements using DiOC 6 as a fluorescent indicator. We show (Fig. 2), along with studies from other laboratories (35,44), that DiOC 6 measures both mitochondrial membrane potential and plasma membrane potential. Using the specific MMP dye JC-1, which did not respond to the PMP, we have shown that the loss of MMP occurs only in apoptotic cells which shrink (see Fig. 3).
Caspases play a central role during apoptosis by cleaving and thereby altering critical cellular substrates which are thought to mediate the dramatic morphological and biochem-ical changes of apoptosis (21). Shrunken apoptotic cells have increased caspase-3-like activity and degraded DNA (25)(26)(27). These studies also showed that high extracellular potassium inhibits caspase-3 activation and DNA degradation, and prevents the loss of cell viability and cell shrinkage (25,27). Caspase-3 is an effector caspase, involved in the later stages of cell death in many model systems. Thus, caspase-3 would be activated during the common or execution phase of the process which results in the classical characteristics apoptotic, such as loss of cell volume, DNA degradation, and apoptotic body formation. Interestingly, little is known about the contribution of upstream, or activator caspases to these apoptotic events. We have used several caspase inhibitors for activator and/or effector proteases to determine the relationship between caspase activity and apoptotic events. Our results show that depending on the particular signal transduction pathway employed to induce cell death, caspase activation and activity is not an essential requirement for several features of apoptosis.
Diverse stimuli induce programmed cell death through the activation of different signaling mechanisms which ultimately lead to the induction of a common apoptotic pathway. The Jurkat cells were treated with anti-Fas, A23187, or thapsigargin, in the presence or absence of z-VAD-fmk, DEVD-fmk, or IETD-fmk as described in the legend to Fig. 4 and examined for intracellular K ϩ as described in the legend to Fig. 1. A gate was set up based on the control sample on a PBFI (K ϩ ) versus a PI fluorescent dot plot to eliminate cells which have loss their membrane integrity, thus only viable cells were used for further ion analysis. Gates were then set on a PBFI (K ϩ ) fluorescence versus forward scatter (cell size) dot plot of viable cells to determine the percent of cells which had a change in PBFI (K ϩ ) fluorescence. A representative three-dimensional plot is shown for each experimental condition. Percentages shown are the average of at least three independent experiments Ϯ S.E. induction of cell death via Fas cross-linking initially signals and activates caspase-8 at the cytosolic face of the transmembrane receptor. In contrast, other apoptotic stimuli, such as the calcium ionophore A23187 or the Ca 2ϩ -ATPase inhibitor thapsigargin, directly activate intracellular signals to initiate the cell death process. In this study, all caspase inhibitors examined (z-VAD, DEVD, and IETD) completely prevented DNA degradation after 24 h of apoptotic treatment, regardless of the initial cell death signal, affirming their effectiveness and suggesting that DNA degradation is largely caspase dependent. However, our observation that caspase inhibition permitted programmed cell death as judged by membrane permeability, cell shrinkage, K ϩ efflux, and changes in the MPP suggest that caspase activity may also not be essential for apoptosis. Activation of Fas receptor transduces an extremely rapid apoptotic signal through the cytoplasmic domain of the receptor which permits the binding of several cytoplasmic proteins, including FADD and caspase-8 (FLICE), forming what has been termed the death-inducing signaling complex (45)(46)(47). z-VAD has been shown to inhibit the activation of FLICE at the level of the death-inducing signaling complex (45), suggesting that Fas-induced cell death is triggered from the death-inducing signaling complex and that the activation of caspase-8 is required for all the downstream events. Our results support these observations, as z-VAD completely prevented all anti-Fas induced characteristics of apoptosis, including loss of cell volume and changes in the MMP. In contrast, when Jurkat cells were treated with either A23187 or thapsigargin in the presence of z-VAD, a loss of cell viability, cell volume, and intracellular potassium, concurrent with changes in the MMP were manifest. Although the severity of these apoptotic features upon A23187 or thapsigargin treatment of Jurkat cells in the presence of z-VAD was not as great as observed in the absence of caspase inhibition, an alternative or secondary caspase-independent pathways must exist.
Specific protease inhibitors for caspase-3-like (DEVD) and caspase-8-like (IETD) enzymes were only marginally effective in preventing the loss of cell viability, cell volume, and potassium, along with the change in MMP that occurs in response to anti-Fas. Furthermore, these specific caspase inhibitors were ineffective in preventing several of the aforementioned characteristics of apoptosis in A23187-or thapsigargin-treated Jurkat cells, while remaining potent inhibitors of DNA degradation. A 4-fold increase in DEVD (or IETD) in combination with various apoptotic stimuli also did not prevent the occurrence of these apoptotic features (data not shown). Thus, depending on the particular signal transduction pathway activated, alternative apoptotic pathways appear to exist in the cell which can compensate for or by-pass caspase-mediated apoptosis. Our results suggest that cells exhibit an inherent commitment to cell death in response to an apoptotic signal, regardless of caspase inhibition.
The idea of caspase independent pathways for cell death have been proposed in other apoptotic model systems (48,49). The expression of the pro-apoptotic protein Bax in Jurkat cells was shown to induce cell death even in the presence of the general caspase inhibitor z-VAD (48). Bax localization to the mitochondrial membranes, along with the changes in the mitochondrial membrane potential observed during apoptosis, implies that dysfunction of this organelle may promote an alternative cell death process. Acid-induced apoptosis in tumor cells was shown to depend on the SAP kinase pathway, to be markedly enhanced by Bax, and to be caspase independent (49). Caspase-1-or caspase-3-deficient hepatocytes were also partially sensitive to Fas (50). Alternatively, these characteristics may depend on an unidentified protease, not inhibited by current caspase substrates, as has been proposed for Baxinduced cell death (48).
The present study has shown that the loss of mitochondrial membrane potential and the loss of cell volume are tightly coupled features of apoptosis. Additionally, we show that cell viability, cell volume, and efflux of intracellular potassium, along with the concomitant change in mitochondrial membrane potential, can occur either independently or dependently of caspase activity, depending on the specific signal transduction pathway activated by a particular apoptotic stimulus. While caspase inhibition effectively prevented DNA degradation under all apoptotic conditions, the presence of other classical cell death characteristics suggest that other features of apoptosis can be caspase independent.