Dissociation of Staurosporine-induced Apoptosis from G₂-M Arrest in SW620 Human Colonic Carcinoma Cells: Initiation of the Apoptotic Cascade Is Associated with Elevation of the Mitochondrial Membrane Potential (Δψ_m)¹

Although maturation of colonic epithelial cells in vivois governed by the spatial and temporal integration of complex genetic programs that regulate growth arrest, differentiation, and apoptosis of cells as they migrate up the crypt-villus axis (reviewed in Ref.1), components of maturation can be recapitulated in undifferentiated human colonic carcinoma cells in culture by exposure to various agents. For example, the short-chain fatty acid butyrate is a normal constituent of the colonic contents generated as a consequence of fiber fermentation by endogenous bacteria (2, 3). Butyrate promotes differentiation and apoptosis of colonic epithelial cells in vivo (4), is chemopreventive against colon cancer (5), and is an inhibitor of histone deacetylase (6). We have demonstrated that treatment of SW620 human colonic carcinoma cells with butyrate initiates a transient G₂-M arrest followed by sustained arrest in G₀-G₁,Δψ _m (mitochondrial membrane potential)dissipation, and apoptosis, and have shown that the cells that survive or escape apoptosis exhibit differentiation along the absorptive lineage (7, 8). Thus, butyrate induces all of the components of colonic epithelial cell maturation.

In contrast,TSA,³which is also an inhibitor of histone deacetylase (9, 10),and sulindac, which is a nonsteroidal anti-inflammatory and chemopreventive agent (11), induce G₀-G₁ arrest, dissipation of the Δψ_m, and apoptosis of SW620 cells(12). However, neither agent induces markers of differentiation.⁴

Cells exposed to the cAMP-dependent protein kinase A activator,forskolin, undergo G₀-G₁arrest and differentiate along the secretory lineage but do not exhibit alterations in the Δψ_m or apoptosis⁵ (13, 14).

Finally, the dietary pigment and spice curcumin, another agent with chemopreventive properties (15, 16), induces arrest of SW620 cells in G₂-M but does not induce differentiation, Δψ_m disruption, or apoptosis (12).⁴

Therefore, in addition to demonstrating that components of maturation can be induced in colonic epithelial cells in vitro, these studies have established that growth arrest can be induced without activating either differentiation programs or apoptotic cascades, and that apoptosis can be triggered in growth-arrested cells in the absence of differentiation. In addition, these studies suggest that the events involved in the process of Δψ_mdisruption play a role in integrating apoptosis with cell cycle arrest programs.

Here, we investigated the effect of the potent cAMP-dependent protein kinase A (PKA) and phospholipid/Ca²⁺-dependent kinase (PKC) inhibitor stsp on growth arrest and apoptosis of SW620 human colonic carcinoma cells. Consistent with its effects in other colonic carcinoma cells (17), stsp induced G₂-M arrest and apoptosis. However, in contrast to the paradigm that cells undergo apoptosis in response to growth arrest, cell cycle arrest was not detected until after the induction of significant apoptosis. Moreover, inhibition of the process ofΔψ _m dissipation, through its pharmacologically mediated collapse, was ineffective in altering stsp-induced apoptosis or cell cycle arrest. In contrast, elevation of the Δψ_m enhanced the initiation of apoptosis but blocked G₂-M arrest.

The generation of ROS has been reported to play a role in colonic epithelial (18, 19) and stsp(20, 21, 22)-mediated apoptosis. However, stsp-induced apoptosis of SW620 cells was unaffected by the inhibition of de novo RNA, cytosolic, or mitochondrial protein synthesis, or by three structurally and functionally distinct antioxidants.

Finally, although the liberation of cytochrome c from the mitochondria to the cytosol has been linked to dissipation of theΔψ _m and caspase-3 activation(23, 24, 25), stsp induced activation of caspase-3 prior to an increase in cytosolic cytochrome c. Moreover, activation was accompanied by transient elevations in both theΔψ _m and mitochondria-associated cytochrome c.

Our previous studies have established the dissociation of growth arrest pathways from those inducing differentiation of human colonic epithelial cells⁴ (12). Here, we further establish the dissociation of apoptosis from growth arrest. Moreover, we have identified an apoptotic cascade initiated in SW620 cells by stsp that is distinctly different from the cascades induced in colonic epithelial cells by other agents (7, 8, 12, 14, 18, 19) and by stsp in other cells (20, 21, 22). Therefore, these data emphasize not only the intricate integration of programs regulating colonic epithelial cell maturation but also the complexity of death by apoptosis.

stsp (Calbiochem) was used at 1 μm; DEVD.CHO (Calbiochem)was used at 5 μm; JC-1(5,5′6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolcarbocyanineiodide;Molecular Probes, Eugene, OR) was used at 1 μm;actinomycin D, emetine, chloramphenicol, N-acetyl-l-cysteine (NAC), l-ascorbic acid (l-AA) and DMSO (Sigma,St. Louis, MO) were used at 5 μg/ml, 2 μm, 500μg/ml, 10 mm, 1 mm, and 2%, respectively;valinomycin and nigericin (Sigma) were each used at 5 μm.

SW620 human colonic carcinoma cells (26) were obtained from the American Type Culture Collection and maintained as we have described previously (8).

In experiments using valinomycin or nigericin, fresh media supplemented with pyruvate (Life Technologies, Inc.) and uridine(Sigma) at 1 mm and 5 μg/ml, respectively, were added to confluent cultures 24 h prior to and during exposure of cells to the agents, as we have described previously (7, 27).

Treatment of cells with DEVD.CHO, valinomycin, or nigericin was for 24 h. Therefore, when cells were simultaneously exposed to stsp and treated with these agents, or initially exposed to stsp for 4, 8,16, or 24 h and then treated with the agents (in the presence of stsp), the experimental end points were 24, 28, 32, 40, or 48 h,respectively. Because stsp-treated and untreated controls were included in each experiment at each end point, the effect of stsp alone could be compared with untreated cells as well as with cells treated with the agents coincident with, or subsequent to, exposure to stsp.

Similar to untreated cells, >90% of cells were viable after treatment with each of these agents as determined by trypan blue exclusion.

Cells were stained with PI and analyzed as we have described previously(7, 27, 28) The percentage of cells in the S phase,G₀-G₁, and G₂-M regions was determined using ModFIT software(Verity Software house, Topsham, ME).

Terminal apoptosis was determined by quantitation of percentage of DNA fragmentation (8, 27) or percentage of PI-stained cells containing subdiploid amounts of DNA as determined by flow cytometry(7, 27, 28). Apoptosis was confirmed by visualization of nonrandom DNA fragmentation using HaeIII-restricted φX174 RF DNA as a size reference, as described previously(8), and by nuclear morphology after DAPI staining.

For DAPI staining, cells were grown and treated on glass coverslips,fixed in cold methanol, acetone-permeabilized, stained with 300 nm DAPI (Sigma) for 5 min, followed by rinsing and mounting. Cells were visualized using a BX60 fluorescence microscope(Olympus America, Melville, NY) equipped with a standard DAPI filter set (Chroma Technology, Brattleboro, VT) and a ×40 Plan Fluor 0.75 numerical aperture objective. Images were acquired with a SPOT real time cooled charged coupled device camera(Diagnostic Instruments, Sterling Heights, MI) and SPOT RT software.

Cells were stained with JC-1and analyzed by flow cytometry as we have described previously (7, 28). Briefly, cells were harvested, washed in PBS, resuspended in phenol-free MEM containing JC-1 at 1 μm and incubated at 37° for 10 min. Stained cells were then washed once with PBS prior to analysis.

Specimens were visualized using a BX60 fluorescence microscope (Olympus America) equipped with a ×60 PlanApo, 1.4 numerical aperture objective and High Q fluorescence band-pass filter sets (Chroma Technology). Images were captured with a SPOT RT cooled CCD camera (Diagnostic Instruments) and SPOT RT software. Fields were imaged using a 10-ms exposure time in both the “green” and “red” emission channels(filter sets 41001 and 41007, respectively). The 12-bit acquired images were rescaled identically and superimposed, using PhotoShop (Adobe, San Jose, CA), to create color images.

A Becton Dickinson FACScan (Becton Dickinson Immunocytometry Systems,San Jose, CA) was used to quantitate J-aggregate formation in a minimum of 10,000 cells per sample. Data were acquired in list mode and evaluated using WinList software (Verity Software House). Forward and side scatter, FSC and SSC, respectively, were used to gate the viable population of cells. JC-1 monomers emit at 527 nm (FL-1 channel) and J-aggregates at 590 nm (FL-2 channel).

Two probes were used to measure the production of ROS:(a) H₂DCFDA, specific for H₂O₂ detection; and(b) DHE, specific for O₂⁻ detection (Molecular Probes). Both probes are reduced, nonfluorescent dyes that can be oxidized to yield the fluorescent parent dyes. After treatment, cells were trypsinized, washed, and loaded with 20 μmH₂CFDA and 10 μm DHE for 30 min at 37°C. Cells were then washed in PBS, followed by analysis on a Becton Dickinson FACScan. Fluorescence was measured in a minimum of 10,000 cells on a log scale in FL-1 (H₂DCFDA)or FL-2 (DHE). Data were acquired in list mode and were analyzed using WinList software.

Cleavage of procaspase-3 was determined using an affinity-purified,PE-conjugated polyclonal rabbit anti-active caspsase-3 antibody(PharMingen, San Diego, CA) according to the manufacturer’s protocol. Briefly, cells were fixed and permeabilized using Cytofix/Cytoperm solution (PharMingen), washed in Perm/Wash buffer (PharMingen), and incubated with anti-active caspase-3 antibody for 30 min at 4°C. Samples were analyzed using a Becton Dickinson FACScan, measuring logarithmic PE fluorescence in the FL-2 channel in a minimum of 10,000 cells.

Caspase-9 activity was measured using the Caspase-9 Colorimetric Assay kit from R&D Systems (Minneapolis, MN), according to manufacturer’s protocol. Briefly, cells were lysed at 0°, and protein concentrations were determined (29). Cell lysates were incubated with caspase-9 colorimetric substrate (LEHD-pNA) for 2 h at 37°C. Absorbance at 405 nm was determined, and activity of caspase-9 was expressed relative to micrograms of protein per reaction.

Partially purified mitochondrial and cytosolic protein fractions were prepared as we have described (28). Briefly, cells were harvested in PBS at 4°, resuspended in 5 volumes of buffer A [20 mm HEPES-KOH (pH 7.5), 10 mm KCl, 1.5 mm MgCl₂, 1 mm EDTA, 1 mm EGTA, 1 mm DTT, 0.1 mmphenylmethylsulfonyl fluoride, 1% aprotinin, 10 μg/ml leupeptin, 1μg/ml pepstatin A, and 250 mm sucrose) and held at 4°for 10 min prior to their disruption with a sandpaper homogenizer. Nuclei and cellular debris were removed by centrifugation at 4°. Partially purified mitochondria were extracted from the supernatant by centrifugation at 10,000 × g for 15 min at 4°, resuspended in buffer A, and held at −80°.

Cytosolic proteins were prepared from the resulting supernatant by centrifugation at 100,000 × g for 60 min at 4°. The final supernatants were concentrated using Amicon centriplus concentrator units (Amicon Inc., Beverly, MA) and were aliquoted and stored at −80°.

Twenty μg of cytosolic proteins or 25 μg of mitochondrial proteins were size-fractionated on 15% acrylamide gels by SDS-PAGE(Bio-Rad, Hercules, CA), blotted onto Hybond ECL nitrocellulose(Amersham, Arlington Heights, IL), and blocked according to manufacturer’s protocol. Blots were then incubated for 60 min at room temperature with mouse monoclonal antibodies directed against human cytochrome c (PharMingen; clone 7H8.2C12) at 1 μg/ml;human cytochrome c oxidase, subunit II (COII; Molecular Probes; clone 12C4-F12) at 1 μg/ml; or actin (Boehringer Mannheim,Indianapolis, IN; clone C4) at 1 μg/ml. Reactions were detected by enhanced chemiluminescence (Amersham) according to the manufacturer’s instructions and were quantified by densitometry using a Personal Densitometer SI and ImageQuant software (Molecular Dynamics). Variations in protein loading were standardized using COII for mitochondrial proteins and actin for cytosolic proteins, and data were expressed relative to untreated controls.

Treated and untreated control cells from at least three individual experiments were compared with two-sample Student’s t tests using individual groups and pooled variances (GB-Stat Computer-Aided Statistics, Version 1.0; Dynamic Microsystems).

We began these investigations by determining the kinetics of stsp-induced cell cycle arrest and apoptosis of SW620 human colonic carcinoma cells. Cells were treated with stsp for 2–48 h, stained with PI, and evaluated for cell cycle parameters and terminal apoptosis by flow cytometry as we have described previously (7, 27, 28). As shown in Fig. 1,A, in which the percentage of stsp-treated cells in each phase of the cell cycle and in terminal apoptosis are expressed relative to untreated cells, the first effect detected was a significant increase in the number of apoptotic cells, seen ∼16 h after starting stsp treatment. The extent of apoptosis increased in cells treated for 24 h, becoming relatively stable for up to 48 h. Nuclear fragmentation (Fig. 1,C), visual detection of nonrandom fragmentation of nuclear DNA (Fig. 1,D), and an increase in the percentage of fragmented DNA (see Fig. 3, Band D) corroborated apoptosis in cells treated with stsp for 16 h.

As shown in Fig. 1,A, and illustrated by the representative flow cytometry profiles shown in Fig. 1 B, alterations in cell cycle were not detected until cells had been exposed to stsp for∼24 h, at which point a significant number were lost from G₀-G₁ and accumulated in the S and G₂-M phases of the cell cycle. With continued exposure to stsp, the arrest of cells in G₂-M was maintained for up to 48 h.

Therefore, consistent with its effect in other human colonic carcinoma cells (17), stsp induced G₂-M arrest and apoptosis of SW620 cells. However, in these cells, significant apoptosis was induced at least 8 h before detection of changes in cell cycle parameters.

The accumulation of cells at cell cycle checkpoints provides a pause in proliferation during which—depending on the cell type, environment,and other factors—cells either remain arrested or enter pathways of mitosis or apoptosis. Therefore, growth arrest typically triggers apoptosis. However, our data, which demonstrated stsp-induced apoptosis of SW620 cells prior to G₂-M arrest, suggested initiation of an apoptotic cascade independent of cell cycle arrest. Because our previous work linked dissipation of theΔψ _m to the transition from growth arrest to apoptosis (28), we asked how two agents that target theΔψ _m, valinomycin and nigericin, effected stsp-induced apoptosis and G₂-M arrest of SW620 cells.

The electrochemical gradient that is generated across the mitochondrial inner membrane, as a result of electron transport coupled with the pumping of protons from the mitochondrial matrix to the intermembranous space, is composed of two interrelated components: aΔψ _m and, because protons determine acidity, a pH gradient (30). As a K⁺ ionophore,valinomycin disrupts the Δψ_m component of the electrochemical gradient, which results in its collapse. In contrast,by inducing an electrically neutral exchange of protons for K⁺, the K⁺/H⁺ ionophore nigericin decreases the pH gradient, inducing a compensatory elevation of theΔψ _m (30, 31).

To validate their effects on the Δψ_m, SW620 cells were exposed to valinomycin or nigericin before being stained with the lipophilic fluorescent dye JC-1. In the absence of, or at low Δψ_m, JC-1 exists as a monomer with emission at 527 nm, within the green spectrum of visible light (32, 33). An example of the dye in its monomeric state is shown in JC-1-stained valinomycin-treated cells, which appear green (Fig. 2 A; Refs. 7 and 28).

At higher Δψ_m, JC-1 forms physiological pH range-independent complexes, or “J-aggregates,” within mitochondria. Such aggregates exhibit fluorescence emission at 590 nm, within the orange range of visible light(32, 33, 34). As shown in Fig. 2 A,nigericin-treated cells that are stained with JC-1 appear as a more intense orange than untreated cells, which suggests an increase in the extent of J-aggregate formation and, consequently, an elevation of theΔψ _m (31).

The extent of J-aggregate formation in valinomycin- or nigericin-treated cells was quantitated using the mean fluorescence channel-2(FL-2) of JC-1 emission measured by flow cytometry. As shown in Fig. 2 B, cells exposed to valinomycin for 4 or 24 h exhibit significant decreases in mean FL-2 emission or collapse of theΔψ _m. In contrast, FL-2 emission of cells exposed to nigericin for 4 and 24 h is significantly higher than that of untreated cells, confirming an elevation in theΔψ _m.

Having validated the contrasting effects of valinomycin and nigericin on the Δψ_m of SW620 cells, we asked how collapse or elevation of the Δψ_m altered stsp-induced G₂-M arrest and/or apoptosis. Cells were either simultaneously treated with stsp and exposed to valinomycin or nigericin for 24 h, or were first treated with stsp for 8, 16,or 24 h followed by exposure to each agent for 24 h. Cells were then stained with PI and analyzed by flow cytometry for cell cycle parameters and terminal apoptosis, or cells were lysed for quantitation of apoptosis using the percentage of DNA fragmentation as an index.

Neither valinomycin nor nigericin alone altered cell cycle parameters or induced apoptosis in SW620 cells (not shown; Ref. 27). Moreover, as shown in Fig. 3,A, regardless of whether valinomycin was added coincident with, or after, stsp treatment, the percentage of cells in G₂-M was comparable with that in cultures treated with stsp alone. Levels of apoptosis were also comparable between cultures treated with stsp alone and those treated with stsp in conjunction with valinomycin (Fig. 3 B). Similarly,2,4-dinitrophenol, a proton ionophore that also collapses theΔψ _m in SW620 cells,⁵ was not effective in altering stsp-induced apoptosis (not shown). Thus,interfering with the process of Δψ dissipation, through its collapse, was ineffective in altering either stsp-induced G₂-M arrest or stsp-induced apoptosis.

In contrast, as shown in Fig. 3,C, exposure of cells to nigericin simultaneous with, or 8, 16, or 24 h after, treatment with stsp effectively blocked the induction of G₂-M arrest. However, elevation of theΔψ _m coincident with, or 8 h after,treatment of cells with stsp resulted in levels of apoptosis that were significantly higher than those induced by stsp alone (Fig. 3 D). When cells were treated with stsp for 16 or 24 h and then exposed to nigericin, the apoptosis-enhancing effect of nigericin was lost.

The effects of valinomycin and nigericin on stsp-induced apoptosis were confirmed by quantifying the percentage of apoptotic cells using PI-stained cells and flow cytometry, which exhibited excellent correlation with quantitation of the percentage of DNA fragmentation (Fig. 3 E). Therefore, as opposed to the other agents that we have investigated (7, 8, 12, 14), stsp induces apoptosis of SW620 cells independent of prior cell cycle arrest or the process of dissipation of the Δψ_m. Furthermore, in contrast to reports linking elevations of theΔψ _m with escape from, or delayed, apoptosis(35, 36, 37), stsp-induced apoptosis of SW620 cells was enhanced by an increase in the Δψ_m during the initial 8 h of stsp treatment, a period preceding the detection of terminal apoptosis.

Although all subcellular organelles, structural components, and cytoplasmic contents contribute to the generation of ROS, mitochondrial oxidative phosphorylation is the major endogenous source(38). Apoptosis of colonic epithelial cells has been linked to the activation of genes involved in the production of ROS(18), and stsp-induced apoptosis in some cell types is dependent on their generation (20, 21, 22). Therefore, we investigated the effects of the inhibition of de novo RNA,cytosolic, or mitochondrial protein synthesis and ROS scavengers on stsp-initiated apoptosis of SW620 cells.

Cells were simultaneously treated with stsp and the agents listed in Table 1, were lysed, and the percentage of DNA fragmentation was determined. As shown in Fig. 4, none of these agents were effective in altering stsp-induced terminal apoptosis, which was demonstrated by levels of DNA fragmentation that were comparable with those induced by stsp alone. In addition,quantitation of relative levels of H₂O₂ and O₂⁻ in SW620 cells that were treated with stsp for 24 h were comparable with the levels in untreated cells (0.86 ± 0.08 and 0.95 ± 0.02, respectively). Consequently, despite their role in some colonic epithelial cell and stsp-induced cascades, the production of ROS is not essential for stsp-induced apoptosis of SW620 cells.

Because these data suggested induction of apoptosis through a pathway that was distinctly different from the cascades initiated in colonic epithelial cells by other agents (7, 8, 12, 14, 18) and from the cascade induced by stsp in other cell types(20, 21, 22), we dissected initiation of stsp-induced apoptosis of SW620 cells.

stsp-induced apoptosis of other cell types involves the activation of caspase-3 (39, 40). Therefore, we investigated the effects of DEVD.CHO, a cell-permeable competitive inhibitor of caspase-3 activity (41), on stsp-initiated apoptosis of SW620 cells. Cells were either simultaneously treated with stsp and exposed to DEVD.CHO for 24 h or were first treated with stsp for 4, 8, or 16 h and then exposed to the inhibitor for 24 h. After PI staining, cells were analyzed by flow cytometry for terminal apoptosis.

As shown by Fig. 5 A (○), the addition of DEVD.CHO coincident with,or up to 8 h after, treating cells with stsp significantly blocked the induction of apoptosis. However, cells became refractory to the inhibitory effects of DEVD.CHO when it was added 16 h after stsp.

Activation of caspase-3 within the initial 8 h of stsp treatment was confirmed by quantifying the binding of a fluorescent labeled antibody directed against cleaved procaspase-3 using flow cytometry(Fig. 5 B). Thus, stsp-induced activation of caspase-3 coincides with the apoptosis-enhancing effect of nigericin.

Through its cytosolic interaction with Apaf-1 and the subsequent activation of caspase-9, the liberation of cytochrome c from the mitochondrial intermembranous space has been implicated in a predominant pathway leading to caspase-3 activation (23, 24). Therefore, we investigated the effect of stsp on levels of cytosolic and mitochondria-associated cytochrome c, and on caspase-9 activation.

Cells were treated with stsp for 2–16 h, and immunoblots were generated from isolated cytosolic or mitochondrial proteins. Blots of mitochondrial proteins were probed with anti-cytochrome cand an antibody directed against the mitochondrial-encoded and-synthesized polypeptide COII to standardize protein loading. Blots generated from cytosolic proteins were also probed with anti-cytochrome c and anti-COII, using anti-actin to standardize loading. Because COII was not detected on these blots (not shown), it is unlikely that the cytosolic proteins were contaminated with mitochondrial proteins. Reactions were quantitated by scanning densitometry, standardized to reference proteins, and data are expressed as the mean relative to untreated cells.

As shown in Fig. 6,A, the level of cytosolic cytochrome c distinctly decreased 4 h after exposure of cells to stsp before significantly increasing after 16 h. In contrast, as shown in Fig. 6 B, the mean relative level of mitochondria-associated cytochrome c significantly increased 4 h after exposure of cells to stsp, followed by its return to levels comparable with those in untreated cells by 8 and 16 h.

To determine its effect on caspase-9 activation, cells were treated with stsp for 16 h, incubated with substrate and analyzed as described in “Materials and Methods.” Despite the detection of significant terminal apoptosis and significant accumulation of cytochrome c in the cytosol at 16 h, the level of caspase-9 activation in stsp-treated cells was comparable with that in untreated cells (1.09 ± 0.10, relative to untreated cells).

Therefore, stsp induces an early, transient increase in mitochondria-associated cytochrome c, which coincides with caspase-3 activation. The subsequent decline in mitochondrial cytochrome c is accompanied by its accumulation in the cytosol, detected coincident with significant terminal apoptosis of SW620 cells. Despite the increase in cytosolic cytochrome cand terminal apoptosis, activation of caspase-9 was not detected.

Because the liberation of cytochrome c has been linked to dissipation of the Δψ_m, (23, 24, 25), we asked how stsp affected the Δψ_m of SW620 cells. Cells were treated with stsp for 2–16 h, stained with JC-1, imaged, and analyzed by flow cytometry. As shown in representative photomicrographs(Fig. 7,A), similar to nigericin-treated cells, cells exposed to stsp for 4 h appear as a brighter orange than do untreated cells. Analysis by flow cytometry demonstrates that, compared with untreated cells, cells exposed to stsp for up to 8 h exhibit significant increases in mean FL-2 emission. After 16 h of treatment, the extent of J-aggregate formation decreased, returning to levels comparable with those of untreated cells (Fig. 7 B).

Therefore, stsp-mediated activation of caspase-3 is accompanied by early transient elevations in both mitochondria-associated cytochrome c and the Δψ_m. The return of theΔψ _m to untreated levels coincides with the accumulation of cytochrome c in the cytosol and the detection of significant terminal apoptosis.

Finally, because the stsp-induced increase inΔψ _m was accompanied by an increase in mitochondria-associated cytochrome c, we asked whether the nigericin-induced elevation in the Δψ_m was similarly associated with an accumulation of cytochrome c in the mitochondria. Cells were exposed to valinomycin or nigericin for 4 h, mitochondrial proteins were extracted, and the levels of cytochrome c were quantitated using immunoblots as described above. As shown in Fig. 8, the mean relative levels of mitochondria-associated cytochrome c in both valinomycin- and nigericin-treated cells were similar to those in untreated cells (P = 0.79 and 0.88, respectively). Therefore, despite enhancing initiation of stsp-induced apoptosis and increasing theΔψ _m, nigericin alone does not induce apoptosis or increase the levels of mitochondria-associated cytochrome c.

The normal colonic mucosa is a dynamic tissue that is dependent on the spatial and temporal integration of pathways governing growth arrest, lineage-specific differentiation, and apoptosis—the principle end points of colonic epithelial cell maturation. Previous work has demonstrated that maturation can be induced in colonic carcinoma cell lines by diverse agents including butyrate (7, 8), TSA,sulindac, curcumin (12), and forskolin (4, 13). In addition, these studies have established that cell cycle-arrest pathways may be dissociated from those that induce differentiation. Moreover, the studies demonstrated that dissipation of the Δψ_m marks the initiation of, and a cellular commitment to, an apoptotic cascade (7, 28, 42, 43, 44, 45, 46), which makes it likely that the process ofΔψ _m dissipation typically plays a role in linking growth arrest and apoptosis programs (7). In support of the relationship between Δψ_mdisruption and the initiation of apoptosis, elevations in theΔψ _m have been reported to result in escape from, or delayed, apoptosis (35, 36, 37).

Here we investigated the effects of the protein kinase inhibitor stsp on SW620 human colonic carcinoma cells. In marked contrast to the pathways induced by butyrate, TSA, sulindac, curcumin, or forskolin,stsp induced apoptosis prior to its induction of G₂-M arrest in the population that survived or escaped apoptosis. Furthermore, whereas collapse of theΔψ _m was ineffective in altering either apoptosis or cell cycle arrest, elevation of theΔψ _m enhanced the initiation, but not progression, of stsp-induced apoptosis but inhibited the induction of G₂-M arrest.

stsp-mediated apoptosis of SW620 cells is characterized by early caspase-3 activation accompanied by transient elevations in both theΔψ _m and the mitochondria-associated cytochrome c. It is unclear how, or whether, this increase in cytochrome c is linked to the elevation inΔψ _m or whether it reflects enhanced mitochondrial import or retention of the protein. However, we have shown that neither the nigericin-mediated elevation nor the valinomycin-induced collapse of Δψ_maltered the levels of mitochondria-associated cytochrome cor induced apoptosis. Therefore, similar to the initiation of apoptosis(7), it is likely that the mechanism by which theΔψ _m is altered, rather than the extent,governs the activation and/or coordination of events leading to modulations in the levels of mitochondria-associated cytochrome c.

Despite the association of increased Δψ_m with the escape from apoptosis (35, 36, 37), elevations in theΔψ _m, accompanied by ROS production, are involved in the initiation of some apoptotic cascades (47, 48), and ROS generation is essential for stsp-induced apoptosis of some cell types (20, 21, 22). However, we were unable to detect increases in either H₂O₂ or O₂⁻ in stsp-treated SW620 cells, and apoptosis was unaltered by three structurally and functionally distinct antioxidants, N-acetyl-l-cysteine, l-ascorbic acid, or DMSO. It is important to note that, whereas these antioxidants were ineffective in altering stsp-induced apoptosis of SW620 cells, some apoptotic cascades are enhanced by the presence of reactive oxygen scavengers(19), which emphasizes a complex relationship between ROS generation and cell death.

The induction of genes associated with ROS production has been implicated in the apoptosis of colonic carcinoma cells(18). However, neither the inhibition of RNA synthesis by actinomycin D, nor the inhibition of cytosolic protein synthesis by emetine, nor the specific inhibition of mitochondrial protein synthesis by chloramphenicol was effective in altering stsp-induced apoptosis of SW620 cells.

The stsp-induced early caspase-3 activation and coincident transient elevations in the Δψm and mitochondria-associated cytochrome c were followed by the accumulation of cytochrome c in the cytosol and significant terminal apoptosis,independent of caspase-9 activation. Our inability to detect a significant increase in cytosolic cytochrome c until approximately 8 h after initial caspase-3 activation could reflect the relative insensitivity of immunoblots compared with FACS detection of a labeled antibody. Alternatively, it could indicate caspase-3 activation independent of cytosolic cytochrome c.

The liberation of cytochrome c from the mitochondria to the cytosol has been implicated in caspase-3 activation by way of its binding to Apaf-1, the subsequent activation of procaspase-9, and,finally, the caspase-9-mediated activation of procaspase-3 (23, 24). However, recent studies have detected procaspases-3 and -9,as well as the Caenorhabditis elegans homologue of Apaf-1, CED-4, in the mitochondrial fraction of untreated cells. Moreover, active caspases-3 and -9 have been identified in the cytosolic and mitochondrial fractions of apoptotic cells(49, 50, 51). Therefore, during the early stages of stsp treatment, each of the components having a role in caspase-3 activation, including an increased amount of cytochrome c,may have been associated with mitochondria. Consequently, combined with our inability to detect an increase in cytosolic cytochrome c until approximately 8 h after caspase-3 activation,it is likely that stsp-mediated activation of caspase-3 occurred at the mitochondrial, rather than the cytosolic, level.

In summary, in contrast to apoptosis induced in other cell types by stsp and other apoptotic pathways induced in colonic epithelial cells,stsp-induced apoptosis of SW620 cells independent of the generation of ROS or de novo synthesis of RNA, cytosolic, or mitochondrial proteins. Moreover, as opposed to butyrate, TSA, sulindac, or curcumin,the stsp-initiated apoptotic cascade in SW620 cells is dissociated from both cell cycle arrest and the process of Δψ_mdissipation. Finally, contrary to the association between the elevation of the Δψ_m and escape from apoptosis, the increase in Δψ_m in stsp-treated SW620 cells is linked to the efficient initiation of apoptosis, possibly potentiating the recruitment of cells to enter the cascade.

By characterizing an alternative apoptotic cascade in colonic epithelial cells, we have emphasized the complexity of cell death by apoptosis. Appreciating and exploiting the unique and shared characteristics of apoptotic pathways is critical, however,because effective chemoprevention and chemotherapy are tightly linked to preserving or inducing apoptosis.

We thank Shailesh Shenoy for superb technical advice.