Rapamycin and UCN-01 synergistically induce apoptosis in human leukemia cells through a process that is regulated by the Raf-1/MEK/ERK, Akt, and JNK signal transduction pathways

The bacterial macrolide rapamycin was isolated >20 years ago as an antifungal agent (1). More recently, it has been identified as a potent immunosuppressant and has been used in this role to inhibit graft rejection following renal transplantation (2, 3) as well as to prevent restenosis (4). Rapamycin and two related compounds, CCI-779 and everolimus, are also currently being developed as anticancer agents in view of their antiproliferative activity against a variety of tumor cell types (5, 6) as well as xenografts (7, 8). The mechanism by which rapamycin and related compounds inhibit tumor cell proliferation involves inhibitory effects on mammalian target of rapamycin (mTOR). mTOR is a 290-kDa member of the phosphatidylinositol 3′-kinase–like family that possesses serine/threonine kinase activities (9). The best described function of mTOR is its participation in the activation of p70S6K and inhibition of the 4E-binding protein-1, which result in transcriptional activation of a subset of proteins involved in cell proliferation and survival. Rapamycin and its derivatives bind to a ubiquitous intracellular protein of 12-kDa FKBP12 (10, 11). The rapamycin/FKBP12 complex potently inhibits mTOR and subsequent downstream signaling to p70S6K and 4E-binding protein-1 (12). The activity of mTOR in mammalian cells is regulated by nutrients and growth factors (13). Although the precise mechanism is not entirely clear, recent evidence suggests that Akt, acting via negative regulation of the tuberous sclerosis complex, a repressor of mTOR, represents an important determinant of mTOR activity (14).

UCN-01 (7-hydroxystaurosporine) is a staurosporine derivative that was originally developed as a specific inhibitor of protein kinase C (15) but was subsequently found to inhibit multiple other kinases, including cyclin-dependent kinases (16), Chk1 (17), and, most recently, Akt (18). By inhibiting Chk1, UCN-01 blocks the phosphorylation and proteosomal degradation of the Cdc25C phosphatase (17). Through interference with Cdc25C degradation, UCN-01 opposes inhibitory phosphorylations of p34^cdc2 on Thr¹⁴ and Tyr¹⁵ sites and thus abrogates G₂-M checkpoint control (19). UCN-01-mediated checkpoint abrogation is felt to be responsible for synergistic interactions in cells exposed to various DNA-damaging agents, including cisplatin, mitomycin C, 1-β-d-arabinofuranosylcytosine, etc. (20–22). In preclinical studies, UCN-01 is a potent inducer of apoptosis in human leukemia cells, a phenomenon associated with dephosphorylation of cyclin-dependent kinases 1 and 2 (23). Recently, our group has reported that exposure of human leukemia and myeloma cells to UCN-01 triggers activation/phosphorylation of the MAPK (ERK) and that interference with this process [i.e., by pharmacologic inhibitors (e.g., U0126 and PD184352) of MEK1/2] results in a dramatic increase in mitochondrial damage and apoptosis (24–26). Several phase I and II trials of UCN-01, either alone or in combination with established cytotoxic agents, are under way, and preliminary evidence of activity against certain hematopoietic malignancies has been reported (27).

The role of the phosphatidylinositol 3′-kinase/Akt pathway in promoting tumor cell survival is well documented (28). It has also been shown that agents that interfere with the phosphatidylinositol 3′-kinase/Akt signal transduction pathway (e.g., LY294002) potentiate the lethal effects of cytotoxic drugs (29, 30). In light of recent evidence suggesting that UCN-01 may disrupt this signaling cascade (18) and the fact that in some settings mTOR is a downstream effector of this pathway, the possibility arose that rapamycin and UCN-01 might cooperate to antagonize tumor cell survival. To test this notion, we have examined interactions between rapamycin and UCN-01 in human leukemia cells, emphasizing effects on signaling pathways and induction of apoptosis. Here, we report that these agents interact in a highly synergistic manner to induce mitochondrial damage, caspase activation, and apoptosis in several human leukemia cell types. Evidence is also presented, suggesting that disruption of the Raf-1/MEK/MAPK and Akt signaling pathways, as well as activation of the stress-related c-Jun NH₂-terminal kinase (JNK) cascade, plays key functional roles in synergistic interactions between UCN-01 and rapamycin in these cells.

U937 human leukemia cells were purchased from American Type Culture Collection (Rockville, MD). The Jurkat and Raji cell lines were purchased from the German Collection of Microorganisms and Cell Cultures (Braunschweig, Germany). All cells were cultured in RPMI 1640 supplemented with sodium pyruvate, MEM non-essential, amino acids, penicillin, streptomycin, and 10% heat-inactivated FCS (Hyclone, Logan, UT). They were maintained in a 37°C, 5% CO₂, fully humidified incubator, passed thrice weekly, and prepared for experimental procedures when in log-phase growth (cell density <6 × 10⁵ cells/mL).

UCN-01 was kindly provided by the Developmental Therapeutics Program, National Cancer Institute (Bethesda, MD), formulated in DMSO (Sigma Chemical Co., St. Louis, MO) as 10⁻³ mol/L stock solutions, and stored at −20°C. Rapamycin and DiOC₆ were purchased from Sigma Chemical, dissolved in DMSO, and stored at −20°C. The pan-caspase inhibitor BOC-D-fmk was purchased from Enzyme System Products (Livermore, CA) and dissolved in DMSO. SP600125 was purchased from Biomol (Plymouth Meeting, PA), dissolved in DMSO, and stored at −20°C.

Logarithmically growing cells were placed in 12- or 6-well sterile plastic culture plates (Nalge Nunc International, Naperville, IL) to which the designated drugs were added. At the end of the incubation period, cells were transferred to sterile centrifuge tubes, pelleted by centrifugation at 400 × g at room temperature for 10 minutes, and prepared for analysis as described below.

After drug exposures, cytocentrifuge preparations were stained with Wright-Giemsa and viewed by light microscopy to evaluate the extent of apoptosis (i.e., cell shrinkage, nuclear condensation and fragmentation, formation of apoptotic bodies, etc.) as described previously (31). For these studies, the percentage of apoptotic cells was determined by evaluating 200 to 500 cells per condition. To confirm the results of morphologic analysis, Annexin V/propidium iodide (PI; BD PharMingen, San Diego, CA) staining was used as per the manufacturer's instructions. For these experiments, 2 × 10⁵ cells per condition were harvested and analysis was carried out using a Becton Dickinson FACScan cytofluorometer (Mansfield, MA). Measurement of 7-amino actinomycin D (7-AAD) by flow cytometry as an indicator of leukemic cell apoptosis has been described previously in detail (31).

Mitochondrial membrane potential (ΔΨ_m) was monitored using DiOC₆ (31). For each condition, 4 × 10⁵ cells were incubated in 1 mL of 40 nmol/L DiOC₆ at 37°C for 15 minutes and subsequently analyzed using a FACScan cytofluorometer. Control experiments documenting the loss of ΔΨ_m were done by exposing cells to 5 μmol/L carbamoyl cyanide m-chlorophenylhydrazone (Sigma Chemical), an uncoupling agent that abolishes the ΔΨ_m, for 15 minutes at 37°C.

U937 cells were harvested after drug treatment as described previously (31) by centrifugation at 600 × g for 10 minutes at 4°C and washed in PBS. Cells (4 × 10⁶) were lysed by incubating for 3 minutes in 100 μL lysis buffer (75 mmol/L NaCl, 8 mmol/L Na₂HPO₄, 1 mmol/L NaH₂PO₄, 1 mmol/L EDTA, 350 μg/mL digitonin). The lysates were centrifuged at 12,000 × g for 5 minutes and the supernatant was collected and added to an equal volume of 2× Laemmli buffer [1× = 30 mmol/L Tris (pH 6.8), 2% SDS, 2.88 mmol/L β-mercaptoethanol, 10% glycerol]. The protein samples were quantified and separated by 15% SDS-PAGE.

Immunoblotting was done as described previously (31). In brief, drug-treated cells were pelleted by centrifugation, lysed immediately in Laemmli buffer, and briefly sonicated. Homogenates were quantified using Coomassie protein assay reagent (Pierce, Rockford, IL). Equal amounts of protein were boiled for 10 minutes, separated by SDS-PAGE, and transferred to nitrocellulose membrane. After blocking in TBST and 5% milk at room temperature for 1 hour, the blots were incubated in fresh blocking solution with an appropriate dilution of primary antibody at 4°C overnight. The source of antibodies were as follows: Bcl-2, mouse monoclonal (DAKO, Carpinteria, CA); Bcl-x_L and Bid (Santa Cruz Biotechnology, Santa Cruz, CA); X-linked inhibitor of apoptosis, rabbit polyclonal (Cell Signaling Technology, Beverly, MA); Mcl-1, mouse monoclonal (BD PharMingen); Raf-1 and phospho-Akt (Ser⁴⁷³), mouse monoclonal (Santa Cruz Biotechnology); MEK and phospho-MEK (Cell Signaling Technology); ERK1/2 and phospho-ERK1/2 (Thr²⁰²/Tyr²⁰⁴), rabbit polyclonal (Cell Signaling Technology); phospho-JNK, mouse monoclonal (Santa Cruz Biotechnology); p34^cdc2 and phospho-p34^cdc2, rabbit polyclonal (Cell Signaling Technology); Akt, rabbit polyclonal (Cell Signaling Technology); cytochrome c, mouse monoclonal (Santa Cruz Biotechnology); Smac/DIABLO (Upstate Biotechnology, Lake Placid, NY); AIF, mouse monoclonal (Santa Cruz Biotechnology); poly(ADP-ribose) polymerase (PARP), mouse monoclonal (Calbiochem, La Jolla, CA); caspase-3, mouse monoclonal (BD Transduction Laboratories, Lexington, KY); caspase-8 and caspase-9, rabbit polyclonal (BD PharMingen); and α-tubulin, mouse monoclonal (Calbiochem). Membranes were washed 3 × 15 minutes in TBST and then incubated with a 1:2,000 dilution of horseradish peroxidase–conjugated secondary antibody (Bio-Rad Laboratories, Hercules, CA) at room temperature for 1 hour. Blots were subsequently washed 3 × 15 minutes in TBST and then developed by enhanced chemiluminescence (Pierce).

Following drug exposure, 6 × 10⁶ cells were lysed in lysis buffer [20 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 1 mmol/L Na₂-EDTA, 1 mmol/L EGTA, 1% Triton X-100, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L β-glycerophosphate, 1 mmol/L sodium vanadate). Protein samples were centrifuged at 12,800 × g for 5 minutes, the supernatant was collected, and protein concentration was determined. Protein (200 μg per condition) was used to immunoprecipitate p34^cdc2 using a mouse monoclonal p34^cdc2 antibody conjugated to agarose beads (Santa Cruz Biotechnology). The cdc2 kinase activity in the immunoprecipitates was determined using a Cdc2 Kinase Assay kit (Upstate Biotechnology) according to the manufacturer's instructions. Briefly, 10 μL of immunoprecipitate were incubated with 400 μg/mL histone H1, 2 μCi [γ-³²P]ATP, and 10 μL inhibitor cocktail in assay dilution buffer (total volume, 50 μL) at 30°C for 10 minutes. A 25 μL aliquot of reaction mixture was transferred onto P81 paper. After washing thrice with 0.75% phosphoric acid and once with acetone, counts per minute of [γ-³²P] incorporated into histone H1 were monitored using TRI-CARB 2100TR liquid scintillation analyzer (Packard Instrument Co., Downers Grove, IL).

Stable Jurkat clones inducibly expressing myristoylated Akt (myr-Akt) and MEK (Upstate Biotechnology), which are constitutively active, were generated as follows. Myc-tagged myr-Akt and hemagglutinin-tagged MEK (HA-MEK) were separately subcloned into the pTRE2-hyg expression vectors (Clontech Laboratories, Inc., Palo Alto, CA) by standard techniques. Jurkat “Tet-On” cells that stably express reverse tetracycline transactivator regulator protein (Clontech Laboratories) were transfected with myr-Akt-pTRE2-hyg and HA-MEK-pTRE2-hyg by electroporation (600 V, 60 ms) using 0.4 μm cuvettes. Stable clones derived from a single cell were selected in RPMI 1640 supplemented with 10% of tetracycline system–approved fetal bovine serum (Clontech Laboratories) in the presence of 400 μg/mL hygromycin. To test for the induced expression of the myr-Akt and HA-MEK, stable clones were left untreated or treated for 24 hours with 2 μg/mL doxycycline (Sigma Chemical), harvested, and analyzed for expression of the appropriate protein by Western blot as described above.

The constitutively active constructs for Akt (Akt-CA) and MEK1 (MEK-CA) as described above were cloned into a pUSEamp vector containing G418 selection marker and transfected into U937 cells using an Amaxa Nucleofector as described above in the Tet-On session. The stable clones (A3 for Akt-CA and M22 for MEK-CA) were selected by culturing cells in medium containing G418 and isolated by limiting dilution. Clones were monitored for expression of phospho-ERK or phospho-Akt by Western analysis, and those displaying maximal increases in ERK or Akt activation compared with empty vector controls were used in the indicated studies (32).

The significance of differences between experimental conditions was determined using the two-tailed Student's t test. To characterize synergistic or antagonistic interactions between agents, median dose effect analysis (33) was employed using a commercially available software program (Calcusyn, Biosoft, Ferguson, MO).

To assess interactions between UCN-01 and rapamycin, U937 cells were exposed to rapamycin (10 nmol/L) and UCN-01 (100 nmol/L) alone or in combination. For combination drug studies, rapamycin was added 1 hour before the addition of UCN-01 or simultaneously with UCN-01, after which the percentage of cells exhibiting morphologic features of apoptosis was determined at 24-, 30-, 36-, and 48-hour intervals. The two schedules yielded equivalent results. Compared with control cells, rapamycin alone (10 nmol/L) was essentially nontoxic, whereas UCN-01 (100 nmol/L) only minimally increased cell death over the 48-hour time course (Fig. 1A). However, in cells exposed to both agents, a dramatic increase in apoptosis was observed, which approached 80% of cells after 48 hours. Flow cytometric analysis of Annexin V/PI–stained cells revealed similar levels of cell death in rapamycin/UCN-01–treated cells (i.e., ∼60% after 36 hours; Fig. 1B). Loss of ΔΨ_m is often associated with and used as a measure of apoptosis. Examination of the effects of combined drug treatment on the loss of ΔΨ_m was therefore assessed by monitoring uptake of DiOC₆ by flow cytometry (Fig. 1B). Consistent with morphologic and Annexin V analysis of apoptosis after 36 hours, 10 nmol/L rapamycin had essentially no effect and 100 nmol/L UCN-01 only modestly increased the loss of ΔΨ_m when given individually. In marked contrast, combined treatment resulted in a substantial increase in the number of cells exhibiting loss of ΔΨ_m, suggesting that apoptosis induced by rapamycin and UCN-01 is accompanied by pronounced mitochondrial dysfunction.

Further characterization of the lethality of this drug combination revealed that the effects of both UCN-01 (Fig. 2A) and rapamycin (Fig. 2B) were concentration dependent. Near-maximal potentiation of cell death was observed for a UCN-01 concentration of 100 nmol/L. Interestingly, rapamycin concentrations as low as 0.5 nmol/L significantly increased the lethality of 100 nmol/L UCN-01, whereas near-plateau levels were observed at a rapamycin concentration of 2 nmol/L. To document the extent of synergistic interactions between UCN-01 and rapamycin over a range of drug concentrations, median dose effect analysis (33) was employed. Using apoptosis induction as an end point, combination index values <1.0 were obtained, corresponding to a highly synergistic interactions (Fig. 2C).

Attempts were then made to extend these findings to other human leukemia cell types. As shown in Fig. 3A, treatment of Jurkat T-lymphoblastic cells with rapamycin (10 nmol/L) alone was nontoxic, whereas 150 nmol/L UCN-01 alone induced 20% apoptosis. As observed in U937 cells, the two drugs in combination induced a marked increase in apoptosis after 48 hours. Similar results were obtained in Raji B-lymphoblastic leukemia cells exposed to a somewhat higher concentration of UCN-01 (300 nmol/L). Thus, rapamycin and UCN-01 interacted synergistically in several human leukemia cell types, including those of both myeloid and lymphoid lineage.

The effects of combined exposure of U937 cells to rapamycin and UCN-01 were then examined in relation to expression and the phosphorylation status of various apoptotic regulatory proteins. Treatment of U937 cells with rapamycin alone for 36 hours, an interval at which drug synergism was readily apparent (Fig. 1A), did not affect the levels of pro-caspase-9 or pro-caspase-3 and did not result in the cleavage of PARP, caspase-8, or Bid. UCN-01 (100 nmol/L) alone resulted in a slight increase in PARP and caspase-8 cleavage, consistent with the modest induction of apoptosis observed previously (Fig. 1A). In contrast, combined exposure to both agents resulted in marked cleavage of pro-caspase-9, substantial cleavage of pro-caspase-3, virtually complete degradation of full-length (115-kDa) PARP, and cleavage of caspase-8 and Bid (Fig. 4A). Rapamycin and UCN-01 given alone had little or only modest effects on release of the proapoptotic mitochondrial proteins cytochrome c, Smac/DIABLO, or AIF into the cytosolic S-100 cell fraction (Fig. 4B). Consistent with effects on apoptosis, combined treatment with rapamycin and UCN-01 resulted in a substantial increase in cytosolic accumulation of these proteins. Thus, combined treatment with rapamycin and UCN-01 was associated with a striking increase in the release of mitochondrial proapoptotic proteins and activation of both intrinsic and extrinsic caspase cascades.

The phosphorylation status of p70S6K is commonly employed to assess inhibition of mTOR by rapamycin; consequently, this substrate was examined using several phosphospecific antibodies (Fig. 5A). p70S6K was down-regulated in cells treated with the rapamycin and UCN-01 combination. The presence of a low molecular weight band not present in control or single drug–treated cells suggested that p70S6K was cleaved during apoptosis. Western blot analysis using an antibody specific for the Thr⁴²¹/Ser⁴²⁴ phosphorylated form of p70S6K revealed that rapamycin alone almost completely inhibited phosphorylation of p70S6K at these sites and that effects were similar in cells exposed to rapamycin and UCN-01. These results indicate that consistent with previous findings rapamycin inhibits the phosphorylation/activation of p70S6K at very low concentrations and argue against the possibility that antileukemic synergism between rapamycin and UCN-01 stems from enhanced inhibition of the pathway.

Effects of these agents alone and in combination on the Raf-1/MEK/ERK, JNK, and Akt signaling pathways, Bcl-2 family members, and proteins involved in cell cycle regulation were then examined in U937 cells (Fig. 5). A 36-hour exposure to rapamycin or UCN-01 alone did not affect the expression of MEK1/2 or ERK1/2. Rapamycin alone induced a modest reduction in the levels of Raf-1, whereas UCN-01 alone increased phosphorylation of MEK and ERK1/2 as we have reported previously (24). Strikingly, cells treated with the combination of rapamycin and UCN-01 exhibited a marked reduction in the expression of Raf-1 and MEK1, whereas the levels of phospho-MEK and phospho-ERK1/2 were also clearly reduced (Fig. 5B). Examination of JNK revealed that rapamycin alone slightly reduced expression of JNK1, whereas UCN-01 alone induced a small increase in JNK1 phosphorylation. In marked contrast, combined drug treatment resulted in a dramatic increase in JNK1 phosphorylation (Fig. 5C). Individual drug treatments had no effect on total Akt expression or phosphorylation, whereas both were diminished in cells treated with the combination of rapamycin and UCN-01 (Fig. 5D). In separate studies, exposure to rapamycin with or without UCN-01 did not result in significant changes in levels of p38 MAPK (data not shown). Thus, coexposure of human leukemia cells to rapamycin plus UCN-01 resulted in a marked reduction in activation of the cytoprotective Raf-1/MEK/ERK and Akt pathways and a reciprocal increase in activation of the stress-related, proapoptotic JNK pathway.

Exposure to rapamycin or UCN-01 individually did not modify expression of Bcl-x_L, Mcl-1, or Bcl-2 (Fig. 5E). However, combined treatment resulted in clear down-regulation of Bcl-x_L and Mcl-1 but not Bcl-2. Individual drug treatment also resulted in modest reductions in expression of cyclin D1 as well as levels of both total and phospho-p34^cdc2, whereas effects were more pronounced with combined exposure (Fig. 5F). In separate studies, it was determined that cotreatment with rapamycin did not significantly modify p34^cdc2 kinase activity in cells exposed to UCN-01 (e.g., 1,193 ± 295 counts per minute of γ-³²P incorporated into histone H1 versus 1,314 ± 114 for UCN-01 alone; P > 0.05). Thus, combined exposure to rapamycin and UCN-01 induced down-regulation of several antiapoptotic Bcl-2 family members (e.g., Bcl-x_L and Mcl-1) as well as cyclin D₁. It also resulted in diminished expression of total p34^cdc2 and phospho-p34^cdc2, although significant changes in p34^cdc2 activity (relative to values for UCN-01-treated cells) were not detected.

Because proteins involved in signal transduction can themselves be substrates for apoptotic caspases (34), an attempt was made to characterize the hierarchy of events following treatment of cells with rapamycin and UCN-01. To this end, U937 cells were exposed for 36 hours to 10 nmol/L rapamycin plus 100 nmol/L UCN-01 in the presence or absence of the pan-caspase inhibitor BOC-D-fmk (20 μmol/L), after which the extent of apoptosis was examined. As shown in Fig. 6A, administration of BOC-D-fmk significantly reduced the level of apoptosis induced by rapamycin and UCN-01 as measured by Annexin V/PI staining. Similar results were obtained when apoptosis was measured by monitoring DiOC₆ uptake or cellular morphology (data not shown).

Parallel studies were done to assess the role of caspases in UCN-01/rapamycin–mediated mitochondrial injury. As shown in Fig. 6B, coadministration of BOC-D-fmk substantially inhibited the cytosolic release of cytochrome c, AIF, and Smac/DIABLO. These findings indicate that caspase activation occurs upstream of the mitochondrial release of cytochrome c, AIF, and Smac/DIABLO in UCN-01/rapamycin–treated U937 cells.

Western blot analysis of total cell extracts prepared from cells treated as described above revealed that BOC-D-fmk significantly reduced the cleavage of PARP and caspase-8 and the generation of cleaved caspase-3, consistent with the reduction in apoptosis (Fig. 7A). In light of the preceding results demonstrating that caspase activation occurred upstream of cytochrome c as well as a recent report indicating that caspase-2 may act upstream of mitochondrial injury (35, 36), cleavage of caspase-2 in drug-treated cells was monitored. As shown in Fig. 7A, BOC-D-fmk inhibited the generation of cleaved caspase-2, raising the possibility that caspase-2 activation may contribute to mitochondrial release of cytochrome c, AIF, and Smac/DIABLO in rapamycin/UCN-01–treated cells. Incubation of cells in the presence of BOC-D-fmk inhibited the down-regulation of MEK, Akt, and p70S6K (Fig. 7B and C), suggesting that these proteins undergo caspase-dependent cleavage in rapamycin/UCN-01–treated cells. In marked contrast, BOC-D-fmk failed to block the down-regulation of phospho-MEK, phospho-Akt, or phospho-p70S6K (Fig. 7B and C), indicating that diminished expression of the phosphorylated forms of these proteins in rapamycin/UCN-01–treated cells cannot be solely attributable to caspase-mediated degradation. Similarly, caspase inhibition did not attenuate the reduction in expression of Raf-1, phospho-ERK1/2, and Mcl-1 or the increase in phospho-JNK (Fig. 7B and C). Lastly, down-regulation of, cyclin D₁ and p34^cdc2 was also unperturbed by BOC-D-fmk (data not shown).

These findings are consistent with those of separate studies done at early intervals (24 hours) when the extent of apoptosis mediated by the combined treatment was only modest (∼25%). As shown in Fig. 7D, coadministration of rapamycin (10 nmol/L) and UCN-01 (100 nmol/L) in U937 cells resulted in a marked decrease in the phosphorylation level of Akt and p70S6K, an increase in JNK phosphorylation, and a decrease in protein levels of Raf-1, Mcl-1, and Bcl-x_L. Parallel results were also noted at a very early interval (e.g., 16 hours) when induction of apoptosis by the rapamycin/UCN-01 regimen was minimal (Fig. 1A). For example, marked down-regulation of Raf-1 and Mcl-1 by the drug combination was readily discernible at this early time point (Fig. 7E). As expected, rapamycin treatment alone suppressed p70S6K phosphorylation on Thr⁴²¹/Ser⁴²⁴ sites and this inhibition was not appreciably altered by addition of UCN-01, consistent with results obtained at later intervals (Fig. 5A). Taken together, these findings indicate that rapamycin/UCN-01–mediated inhibition of the Raf-1/MEK/ERK and Akt pathways, activation of JNK, and down-regulation of Mcl-1 and Raf-1 do not solely represent a consequence of engagement of the caspase cascade.

To gain insights into the functional role of down-regulation of the MEK/ERK and Akt pathways in rapamycin/UCN-01–mediated lethality, Jurkat cells inducibly expressing either MEK-CA (Jurkat MT6) or Akt-CA (Jurkat Akt/29) were employed. As shown in Fig. 8A, a 24-hour incubation with doxycycline dramatically induced expression of HA-MEK in MT6 cells. Cells not incubated with doxycycline and treated with single drugs displayed low levels of phosphorylated (activated) ERK1/2, which was essentially abrogated following combination treatment, as observed in U937 cells. When cells were cultured in the presence of doxycycline, a pronounced increase in the level of ERK1/2 activation was observed, whereas levels of total ERK1/2 remained unaffected by doxycycline treatment. Significantly, enforced expression of MEK-CA circumvented the blockade of ERK activation by rapamycin/UCN-01 exposure. To determine if prevention of ERK inhibition affected apoptosis, cells were treated with rapamycin (10 nmol/L) or UCN-01 (150 nmol/L) alone or in combination in the presence or absence of doxycycline, and after 48 hours, apoptosis was determined using Annexin V/PI staining. Activation of ERK partially but significantly protected cells from rapamycin/UCN-01–induced apoptosis (P < 0.05; Fig. 8B). Similar results were obtained with two other MEK-inducible clones (Jurkat MEK/17 and MEK/7; data not shown).

Parallel results were also obtained in U937 cells stably transfected with a MEK-CA. As shown in Fig. 8C (inset, arrow), MEK-CA expression was readily detected by anti-MEK1 Western blot analysis. Similar to results obtained in the inducible Jurkat system, expression of MEK-CA modestly but significantly attenuated rapamycin/UCN-01 lethality as determined by both DiOC₆ and 7-AAD assays (P < 0.05 in each case; Fig. 8C).

Parallel studies were done using Jurkat cells stably transfected with a tetracycline-inducible Akt-CA (Jurkat Akt/29). Cells cultured in the presence of doxycycline for 24 hours displayed a substantial increase in the expression of c-Myc-tagged Akt accompanied by a pronounced increase in the levels of phospho-Akt (Fig. 9A). Interestingly, exposure of these cells to rapamycin alone modestly diminished the phosphorylation of Akt, an effect that was slightly more pronounced in cells coexposed to UCN-01. Nevertheless, in the presence of doxycycline, phospho-Akt was clearly increased in cells exposed to the rapamycin/UCN-01 regimen compared with cells treated in the absence of doxycycline. To define the functional role of enforced activation of Akt in rapamycin/UCN-01–mediated cell death, the extent of apoptosis was monitored by Annexin V/PI staining. Although enforced activation of Akt was clearly less effective than enforced activation of MEK in blocking apoptosis induced by rapamycin/UCN-01 after 48 hours, the extent of protection was nevertheless statistically significant (P < 0.05; Fig. 9B). Comparable results were obtained in U937 cells stably transfected with a Akt-CA (Fig. 9C, inset). Specifically, U937 cells transfected with the Akt-CA construct displayed a modest but statistically significant reduction in lethality determined by loss of ΔΨ_m or 7-AAD uptake (P < 0.05 in each case) following combined exposure to UCN-01 and rapamycin. Taken in conjunction with the preceding findings, these observations suggest that interruption of both Raf-1/MEK/ERK and, to a lesser extent, Akt pathways by the rapamycin/UCN-01 regimen may each contribute to the synergistic induction of apoptosis by this drug combination.

Earlier findings suggested that rapamycin/UCN-01–induced lethality was associated not only with inhibition of ERK and Akt but also with an increase in JNK activation (Fig. 5). To investigate the functional significance of this event, the pharmacologic JNK inhibitor SP600125 (37) was employed. For these studies, U937 cells were exposed to rapamycin and UCN-01 for 36 hours in the absence or presence of SP600125 (5 μmol/L), after which apoptosis was monitored by Annexin V/PI staining. SP600125 (5 μmol/L) significantly (P < 0.05) diminished apoptosis induced by the rapamycin/UCN-01 combination (Fig. 10A). Similar results were obtained when loss of ΔΨ_m was monitored by DiOC₆ staining or apoptosis determined by analysis of cell morphology (data not shown).

To determine what position JNK activation occupied in the apoptotic hierarchy, Western blot analysis was done using extracts prepared from cells incubated for 36 hours with rapamycin/UCN-01 in the absence or presence of SP600125 (5 μmol/L; Fig. 10B). SP600125 clearly inhibited PARP degradation, caspase-8 activation, and generation of cleaved caspase-3 in cells exposed to rapamycin/UCN-01, consistent with the previously observed reduction in apoptosis (Fig. 10A). SP600125 also essentially abrogated JNK phosphorylation induced by rapamycin/UCN-01 but had little or no effect on rapamycin/UCN-01–induced down-regulation of Raf-1 or phosphorylation of MEK and ERK1/2. Inhibition of JNK also did not modify diminished p70S6K or Akt phosphorylation observed in rapamycin/UCN-01–treated cells (Fig. 10B). Finally, reductions in the expression of cyclin D₁, total p34^cdc2, and phospho-p34^cdc2 induced by rapamycin/UCN-01 treatment were also unaffected by JNK inhibition (data not shown). We did not detect any changes in protein levels of MEK1, ERK, JNK, and Akt after treating cells with SP600125 (data not shown). Because SP600125 is not a specific inhibitor of JNK, we employed a genetic approach to confirm the role of JNK activation in rapamycin/UCN-01 lethality. To this end, we employed U937 cells stably expressing a dominant-negative, transactivation domain–deficient c-Jun mutant, TAM67, which had been described previously in detail (38). As shown in Fig. 10 (inset), the truncated mutant c-Jun protein migrates faster than the full-length endogenous protein, consistent with earlier reports (38). Coadministration of rapamycin and UCN-01 was significantly less lethal, reflected by loss of ΔΨ_m or 7-AAD uptake, in TAM67 cells compared with empty vector controls (P < 0.02 in each case). Collectively, these findings suggest that JNK activation plays a significant functional role in rapamycin/UCN-01 lethality but that this phenomenon occurs downstream or independent of perturbations in the Raf-1/MEK/ERK and Akt signaling pathways.

The present results indicate that the protein kinase C and Chk1 inhibitor UCN-01 interacts in a highly synergistic manner with the mTOR inhibitor rapamycin to trigger mitochondrial injury and apoptosis in a variety of human leukemia cell types. Recently, considerable attention has focused on the use of signal transduction modulators to enhance the lethal effects of conventional cytotoxic agents. For example, UCN-01 has been shown to potentiate the lethal actions of camptothecin and 1-β-d-arabinofuranosylcytosine (22) and synergistic interactions between geldanamycin and paclitaxel have also been reported (39). However, there is accumulating evidence that simultaneous interruption of two cytoprotective signaling pathways represents a particularly potent apoptotic stimulus in neoplastic cells. For UCN-01, coadministration of pharmacologic MEK inhibitors (e.g., PD184352 or U0126) has been shown to induce a dramatic increase in cell death in human leukemia and myeloma cells (24, 26) as well as in malignant cells of epithelial origin (25). Analogously, the Bcr/Abl kinase inhibitor STI571 interacts synergistically with pharmacologic MEK1/2 inhibitors (40), the phosphatidylinositol 3′-kinase inhibitor LY294002 (41), and farnesyltransferase inhibitors (42) in Bcr/Abl⁺ leukemia cells. Taken together, these findings suggest that neoplastic cells seem to be ill equipped to escape the lethal consequences of interruption of more than one survival-related signaling pathway. The observation that in human leukemia cells UCN-01, an inhibitor of protein kinase C, Chk1, and cyclin-dependent kinases (15–17), interacts synergistically with rapamycin, an agent known to disrupt signaling downstream of the Akt signal transduction pathway (14, 18), provides additional support for this notion.

The mitochondria play a critical role in stress-related apoptosis in which cytochrome c released into the cytoplasm forms, in conjunction with apoptosis-inducing factor-1 and dATP, the apoptosome, which activates caspase-9 (43). Caspase-9 then activates effector caspases (e.g., caspase-3). However, it has been shown recently that under some circumstances caspase activation occurs upstream of the mitochondria, arguing against a role for caspase-9 as the initiator caspase in this pathway (35, 36, 44, 45). In the present studies, translocation of cytochrome c, AIF, and Smac/DIABLO into the cytosol following exposure of cells to UCN-01 and rapamycin was largely inhibited by the pan-caspase inhibitor BOC-D-fmk, suggesting that under these conditions release of apoptogenic factors from the mitochondria represented a caspase-dependent event. It has been shown that one of the initiator caspases that can be activated before mitochondrial cytochrome c release, including that triggered by cytotoxic drugs, is caspase-2 (35, 36). Significantly, caspase-2 was also cleaved in cells undergoing UCN-01/rapamycin–induced apoptosis, an event that, along with mitochondrial injury and apoptosis, was also inhibited by BOC-D-fmk. It is therefore tempting to speculate that caspase-2 activation may mediate mitochondrial permeabilization and subsequent release of cytochrome c (as well as other proapoptotic mitochondrial proteins) in U937 cells treated with rapamycin/UCN-01. Finally, it is noteworthy that BOC-D-fmk largely failed to prevent the down-regulation of Raf-1, ERK, and Akt observed in rapamycin/UCN-01–treated cells, which occurred at relatively early intervals (i.e., before extensive apoptosis had occurred). Such findings argue against the possibility that caspase activation is solely or primarily responsible for interruption of these signaling cascades. However, the possibility that early induction of apoptosis may lead to further down-regulation of these pathways and in doing so amplify the apoptotic response cannot be completely excluded.

Based on the present findings, it seems likely that down-regulation of the Raf-1/MEK/ERK and Akt survival signaling pathways contributes at least in part to the marked potentiation of apoptosis observed in leukemia cells exposed to rapamycin and UCN-01. Interestingly, rapamycin alone modestly reduced the levels of Raf-1, although coadministration of UCN-01, which has been shown recently to stimulate ERK activation in leukemia and myeloma cells (24, 26) markedly potentiated rapamycin-mediated Raf-1 down-regulation in a caspase-independent manner. The mechanism by which the rapamycin/UCN-01 regimen induces Raf-1 down-regulation is unclear. However, it may be relevant that Raf-1 down-regulation induced by the combination UCN-01 and 17-AAG in human leukemia cells could be inhibited by the proteasome inhibitor MG-132 (46), raising the possibility that proteosomal degradation of Raf-1 may also occur in rapamycin/UCN-01–treated cells. On the other hand, given the observation that rapamycin alone induced a small decrease in Raf-1 expression, the possibility that inhibition of translation may also play a role cannot be excluded. The pronounced down-regulation of Raf-1 was also associated with a marked reduction in phosphorylation of MEK1, a major Raf-1 substrate (47), as well as that of ERK, the primary MEK target (48). Previously, we showed that exposure of multiple malignant hematopoietic and nonhematopoietic cells to UCN-01 induces, through a yet to be defined mechanism, activation of the MEK/ERK cascade (24–26). Furthermore, interruption of these events (e.g., by pharmacologic MEK inhibitors) resulted in a marked increase in mitochondrial injury and apoptosis (24–26). Other agents, such as the Hsp90 antagonist 17-AAG, have also been shown to diminish UCN-01-induced ERK activation and potentiate apoptosis human leukemia cells, thereby mimicking the actions of pharmacologic MEK inhibitors, such as PD98059 and U0126 (46). It is therefore conceivable that rapamycin may exert a similar effect, although the mechanism by which this event occurs is unclear given evidence that mTOR and its effectors are thought to operate downstream of the ERK and Akt pathways (49, 50). One speculative possibility is that rapamycin may trigger a yet to be defined feedback mechanism, possibly involving activation of phosphatases, such as MKP1, MKP3, or PP2A (51, 52), leading to down-regulation of the Raf-1/MEK/ERK signal transduction pathway. Studies designed to address this possibility are currently in progress.

There is abundant evidence that activation of the Akt pathway exerts antiapoptotic effects, although the mechanism(s) by which this phenomenon occur is not entirely understood and may vary with cell type. These include modulation of the activity of other signaling pathways [e.g., those related to nuclear factor-κB, GSK3, and p70S6K (53)] or phosphorylation of apoptotic regulatory proteins, such as Bad (54) or pro-caspase-9 (55). In this context, treatment of U937 cells with rapamycin/UCN-01 induced a marked caspase-independent decrease in Akt activation (phosphorylation). Recently, Sato et al. reported that UCN-01 blocked Akt phosphorylation through a PDK1-dependent mechanism (18), raising the possibility that rapamycin may enhance this effect through a yet to be determined mechanism. However, it should be noted that whereas enforced activation of Akt significantly protected U937 and Jurkat cells from UCN-01/rapamycin–mediated lethality this effect was quite modest, arguing against a major role for interruption of the Akt pathway in synergistic interactions between these agents. On the other hand, it is possible that disruption of Akt signaling may have a greater impact on cell death under conditions in which MEK/ERK inactivation has occurred.

In addition to down-regulation of cytoprotective signaling pathways, the present data also suggest that activation of the stress-activated JNK cascade plays a key role in UCN-01/rapamycin antileukemic synergism. Specifically, pharmacologic inhibition of JNK by SP600125 significantly attenuated caspase activation and apoptosis by this drug combination. Furthermore, expression of the dominant-negative c-Jun blocked UCN-01/rapamycin synergism. The mechanism by which JNK activation promotes apoptosis may vary with the cell type and inciting stimulus but may involve potentiation of mitochondrial cytochrome c release (56) or phosphorylation and inactivation of antiapoptotic proteins, such as Bcl-2 (57) and Mcl-1 (58). It is noteworthy that inhibition of JNK activation had no effect on down-regulation of Raf-1 or rapamycin/UCN-01–induced inhibition of MEK, ERK, or Akt phosphorylation, suggesting that JNK activation occurs downstream or independent of these events. Recent evidence indicates that rapamycin induces apoptosis and JNK activation through a process mediated by 4E-binding protein-1, the downstream effector of mTOR (59). Although significant differences exist between the present findings and the latter studies, which involved rhabdomyosarcoma cells cultured under serum-free conditions, the possibility that coadministration of UCN-01 promotes rapamycin-induced JNK activation through 4E-binding protein-1 or other pathways seems plausible.

The present findings also indicate that rapamycin/UCN-01 treatment induces the caspase-independent down-regulation of Mcl-1, Bcl-x_L, cyclin D₁, and p34^cdc2. Numerous studies have shown that Mcl-1 and Bcl-x_L oppose apoptosis in diverse settings, particularly in malignant hematopoietic cells (60, 61). Cyclin D₁ has also been identified recently as a mediator of growth and proliferation downstream of mTOR (62) as well as apoptosis regulation (63, 64). In this context, down-regulation of cyclin D₁ has been linked to apoptosis induced by the cyclin-dependent kinase inhibitor flavopiridol (65). Finally, inappropriate activation of p34^cdc2 is a potent apoptotic stimulus (66) and has been implicated in the dramatic induction of cell death exhibited by leukemia and myeloma cells exposed to UCN-01 in conjunction with MEK inhibitors (24, 26). In this context, rapamycin is known to induce G₁ arrest through a cyclin D₁–dependent process (67). Thus, it is tempting to speculate that such an event, coupled with UCN-01-mediated checkpoint abrogation, provides the cell with conflicting signals that result in apoptosis. However, in the present studies, diminished phosphorylation of p34^cdc2 in rapamycin/UCN-01–treated cells was accompanied by a parallel reduction in total p34^cdc2 expression; moreover, differences in p34^cdc2 kinase activity in rapamycin/UCN-01–treated versus UCN-01-treated cells could not be detected. Collectively, such findings argue against a significant role for inappropriate p34^cdc2 activation in rapamycin/UCN-01–induced apoptosis.

In summary, the present findings indicate that coadministration of the mTOR inhibitor at low nanomolar concentrations with UCN-01 leads to the highly synergistic induction of mitochondrial damage and apoptosis in human leukemia cells. These events are associated with multiple perturbations in signaling, cell cycle, and apoptotic regulatory proteins as well as the caspase-dependent induction of mitochondrial injury. Given the recent interest in rapamycin and its analogues (e.g., CCI-779) as anticancer agents (5–8) and demonstration of the feasibility of achieving plasma UCN-01 concentrations in excess of those employed in the present study (27), these findings could have translational implications for the development of novel antileukemic strategies. In a broader sense, such findings provide further support for the evolving concept that simultaneous interruption of multiple signal transduction/cell cycle regulatory pathways represents a highly potent apoptotic stimulus in neoplastic cells. In accord with this notion, synergistic antileukemic interactions between rapamycin and the Bcr/Abl kinase inhibitor imatinib mesylate or the FLT3 inhibitor PKC412 have been described very recently (68). It will be of considerable interest to determine whether and to what extent this strategy can be extended to other neoplastic cell types, particularly those of epithelial origin, as well as to primary tumor cells. Accordingly, such studies are currently under way.