Src inhibition enhances paclitaxel cytotoxicity in ovarian cancer cells by caspase-9-independent activation of caspase-3 Free

In the United States, ovarian cancer accounts for nearly 15,000 deaths per year (1). The prognosis for ovarian cancer is generally very poor with high mortality rates. This is largely the result of late presentation of patients due to clinically silent symptoms until disseminated and metastatic disease has been well established. Following tumor debulking, patients generally receive chemotherapy treatment with paclitaxel, platinum-based agents, or a combination of both (2, 3). These agents act via different mechanisms. Paclitaxel binds tubulin, enhancing tubulin polymerization and inhibiting microtubule disassembly, thus preventing completion of mitosis and producing a G₂-M block (4, 5). Cisplatinum, in contrast, binds DNA causing DNA adducts, preventing DNA replication and elongation, inducing cell cycle arrest, and apoptosis (6, 7). Unfortunately, even in patients in whom there has been an initial positive clinical response to chemotherapy, development of recurrent chemoresistant disease is a common outcome (8). The development of multidrug resistance is multifactorial, but probably includes the activation of growth factor signal transduction cascades (9).

Src tyrosine kinase is the prototypical member of a family of cytoplasmic, membrane-associated, nonreceptor tyrosine kinases. Cellular Src is expressed in a wide variety of cell types, including several ovarian cancer cell lines (10), and has recently been identified to be overexpressed and activated in a high proportion of ovarian cancers examined (11). Cellular Src is involved in a variety of cell signaling events, regulating both cell proliferation (12, 13) and differentiation (14). We have recently described the expression and signaling of cellular Src in mouse and human ovarian cancer cell lines (15). Inhibition of Src was associated with decreased activation of cell growth (Ras-Raf-MEK-ERK) and survival (PI3-kinase-Akt-FOXO1) pathways. Furthermore, cellular Src inhibition in ID8 mouse ovarian cancer cells was found to enhance the cytotoxicity of paclitaxel and restore sensitivity in a paclitaxel-resistant mouse ovarian cancer cell line (15). The cell death pathways activated by paclitaxel in at least one ovarian cancer cell line has been reported to be caspase-9- and caspase-3-independent (16). In the current study, we investigated the cell death pathways activated by paclitaxel alone or in combination with Src inhibition. We find that enhancement of paclitaxel cytotoxicity by Src inhibition is mediated by markedly increased caspase-3 activity. Caspase-3 activation seems to be caspase-9-independent.

All liquid media and LipofectAMINE 2000 transfection reagent were purchased from Invitrogen Life Technologies (Carlsbad, CA). MitoTracker red and 4′,6-diamidino-2-phenylindole dihydrochloride were from Molecular Probes (Eugene, OR). Caspase antibodies were purchased from Cell Signaling (Beverly, MA). An antibody specific for cytochrome c was purchased from PharMingen (San Diego, CA). Enhanced green fluorescent protein (EGFP) expression plasmid (EGFP-N1) was from Clontech (Palo Alto, CA). A Src dominant-negative (Src^dn) fusion construct with K296R and Y528F mutations was purchased from Upstate Biotechnologies (Lake Placid, NY). RNA duplexes for RNA interference were purchased from Dharmacon Research, Inc. (Lafayette, CO). The Src inhibitors 4-amino-5(4-chlorophenyl)-7-(t-butyl)pyrazolo(3,4-d)pyrimidine (PP2), SU6656, the caspase activity assay kits and paclitaxel were obtained from Calbiochem (La Jolla, CA). Enhanced chemiluminescence Western blotting detection reagents were purchased from Amersham Life Science, Incorporated (Arlington Heights, IL). Cisplatin and all other reagents were purchased from Sigma (St. Louis, MO).

The development of ID8 cells from the ovarian surface epithelium of C57BL6 mice has previously been described (17). ID8 cells are cultured in DMEM supplemented with 4% fetal bovine serum, insulin (10 μg/mL), transferrin (5 μg/mL), sodium selenite (7 ng/mL), and Hepes (15 mmol). Human SkOV3, CaOV3, and OvCar3 ovarian cancer cell lines were obtained from the Lombardi Cancer Center Tissue Culture Shared Resource and maintained in DMEM supplemented with fetal bovine serum (10%) and Hepes (15 mmol).

Relative viable cell number was determined as previously described (18) by incubation of cells with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide and measuring the production of formazan at an absorbance at 570 nm, which occurs only in the mitochondria of viable cells. Briefly, cells were seeded at 5,000 cells per well in 96-well plates and allowed to attach overnight. After attachment, medium was removed and replaced with fresh full growth medium and treatments (paclitaxel ± inhibitors) were initiated. Following the treatment period, relative cell number was assessed by adding 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (1 mg/mL) and measuring formazan production following a 3-hour incubation. In other experiments, direct cell counts were utilized. For statistical analyses, control values were set to 100%. Treatment effects were analyzed by an ANOVA with differences between individual means compared by Fisher's protected least significant differences test.

For immunofluorescence, ID8 cells were seeded on glass coverslips in six-well culture plates. After the treatment period (described in Results), cells were fixed with 4% paraformaldehyde for 10 minutes at room temperature. Cells were permeabilized with 0.1% NP40 in PBS for 20 minutes at room temperature. Coverslips were blocked with 10% preimmune serum (in PBS) from the species in which the secondary antibody was raised. Following blocking, coverslips were incubated with primary antibody (1:200) in PBS-1% normal serum at room temperature for 2 hours. Negative controls consisted of coverslips incubated with preimmune immunoglobulin G. Coverslips were then washed extensively with PBS, incubated with a Cy2 conjugated secondary (1:200) antibody for 2 hours at room temperature, washed in PBS, counterstained with 4′,6-diamidino-2-phenylindole (0.1 μg/mL PBS) for 15 minutes and then washed with PBS. Coverslips were mounted with 90% glycerol and sealed. Cells were visualized by laser scanning confocal microscopy.

At the end of the treatment period, cells were collected and washed in PBS then resuspended in CHAPS lysis buffer [50 mmol PIPES/NaOH (pH 8.5), 2 mmol EDTA, 0.1% CHAPS, 5 mmol DTT, 10 μg/mL of leupeptin, 10 μg/mL pepstatin, 10 μg/mL aprotinin, and 1 mmol phenylmethylsulfonyl fluoride]. Resuspended cells were frozen and thawed thrice to disrupt cell membranes. Insoluble material was cleared by centrifugation (14,000 × g for 20 minutes at 4°C). Protein concentration was determined by a modified Bradford protein assay and equalized across all samples. Soluble protein was mixed with an equal volume of 2× Laemmli sample buffer. Samples were heated to 95°C for 5 minutes and then subjected to SDS-PAGE. Protein was electrotransferred to polyvinylidene difluoride membranes. Membranes were blocked with TBST-5% milk [10 mmol Tris-HCl (pH 8.0), 150 mmol NaCl, 0.05% Tween 20, and 5% nonfat dry milk] for 1 hour at room temperature prior to overnight incubation with the appropriate specific antibody in TBST-5% milk (1:1,000 dilution) at 4°C. Membranes were washed extensively with TBST, incubated with the appropriate peroxidase-conjugated secondary antibody in TBST-5% milk for 1 hour and then washed with TBST. Proteins were visualized by enhanced chemiluminescence. For some experiments, bands were quantified by densitometric analysis using the NIH Image J software.

Generation of a Src dominant-negative-EGFP fusion construct has been previously described (15). ID8 cells were transfected with expression vectors (Src dominant-negative-EGFP fusion construct or EGFP) with LipofectAMINE 2000 transfection reagent (Invitrogen Life Technologies) using the manufacturer's protocol. For caspase-9 RNA interference, a pool of four different duplexes (SMARTpool, Dharmacon) targeted against caspase-9 was used. Twenty-four hours after plating, cells were transfected with RNA duplexes using LipofectAMINE 2000 transfection reagent. Twenty-four hours after transfection, treatments were initiated and 24 hours later, cells were collected in lysis buffer for subsequent analysis of protein expression. ID8 cells expressing EGFP were transfected with RNA duplexes against EGFP for determination of transfection efficiency and to serve as RNA interference controls.

ID8 and ID8^TaxR mouse ovarian cancer cells were treated as indicated in Results and collected by mild trypsin digest. Cell number was determined and then adjusted to 1 to 2 × 10⁶ cells per treatment group, washed in PBS and then resuspended in 0.1 mL of citrate/DMSO buffer. DNA cell cycle phase analysis was done by the Lombardi Comprehensive Cancer Center Flow Cytometry and Cell Sorting Shared Resource using a Becton Dickinson FACStar Plus dual laser system.

Because Src is involved in the activation of survival pathways, the effect of Src inhibition in combination with paclitaxel was investigated. Src inhibition with the Src-selective inhibitor PP2 enhanced the cytotoxic effects of paclitaxel in both mouse (ID8) and human (SkOV3 and CaOV3) ovarian cancer cell lines across a range of concentrations (Fig. 1A). Two other Src-selective inhibitors (SU6656 and herbimycin A) had similar effects (data not shown). In order to determine whether the effects of these pharmacologic inhibitors were a direct result of Src inhibition, ID8 mouse ovarian cancer cells were transfected with a Src dominant-negative EGFP fusion construct. These cells were used due to their high transfection efficiency. Cytotoxicity resulting from paclitaxel treatment was significantly greater in ID8 cells expressing the Src dominant-negative fusion construct compared with mock-transfected cells or cells transfected with EGFP (Fig. 1B), indicating that the increased cytotoxicity can be attributed to Src inhibition.

The human SKOV3 ovarian cancer cell line was less sensitive to paclitaxel than the CaOV3 human and ID8 mouse ovarian cancer cell lines. Interestingly, pharmacologic inhibition of Src with PP2 sensitized SKOV3 cells to paclitaxel (>10-fold increase; Fig. 1A). SKOV3 cells have been reported to be paclitaxel-resistant (19, 20). As previously reported (15), an ID8 paclitaxel-resistant cell line has been developed (ID8^TaxR). These cells are completely resistant to up to 2 μmol/L paclitaxel. In addition, the ID8^TaxR cells have cross-resistance to cisplatin. Inhibition of Src with either PP2 or SU6656 resensitizes ID8^TaxR cells to both paclitaxel and cisplatin (Fig. 2A). A recently derived paclitaxel-resistant human CaOV3 cell line showed similar results (data not shown). Expression of the Src dominant-negative fusion construct also resensitizes ID8^TaxR cells to paclitaxel, confirming the specificity of Src inhibition in the resensitization (Fig. 2B). Thus, Src inhibition seems to enhance the cytotoxicity of paclitaxel in both drug-sensitive and drug-resistant ovarian cancer cells.

As a first step in determining the mechanism by which Src inhibition enhances paclitaxel cytotoxicity, ID8 mouse ovarian cancer cells were exposed to paclitaxel (0.2 μmol/L), PP2 (10 μmol/L), or both overnight, after which caspase-3 processing was determined by immunoblot analysis. Treatment with either paclitaxel or PP2 alone resulted in little or no detectable cleaved caspase-3. In contrast, combination treatment resulted in accumulation of cleaved/processed caspase-3. Interestingly, similar results were obtained with colchicine (Fig. 3A), a microtubule-disrupting agent, indicating that Src inhibition sensitizes cells to the G₂-M arrest rather than microtubule stabilization per se, or some other paclitaxel-specific effect. A time course experiment showed that cleaved caspase-3 began to accumulate by 12 hours of paclitaxel plus PP2 combination treatment, with even greater accumulation at 18 hours (Fig. 3B). Interestingly, paclitaxel alone induced caspase-3 processing and accumulation of cleaved caspase-3 at 18 hours of treatment, although accumulation was not as great as the combination treatment. The dose-response showed that paclitaxel plus PP2 combination treatment induced ∼3- to 4-fold greater accumulation of cleaved caspase-3 compared with paclitaxel alone at all concentrations tested (Fig. 3C and D). Finally, combination treatment also resulted in increased cleaved poly-ADP ribose polymerase, a substrate of activated caspase-3 and marker of apoptosis, compared with paclitaxel alone (Fig. 3E).

Caspase-3 can be activated by either intrinsic or extrinsic apoptotic stimuli. It seemed logical that combination treatment of Src inhibition and paclitaxel would activate caspase-3 by an intrinsic mechanism, via activation of caspase-9. Interestingly, treatment of ID8 mouse ovarian cancer cells with paclitaxel, PP2, colchicine, combinations of paclitaxel plus PP2, or colchicine plus PP2 all failed to consistently result in processing of caspase-9 as determined by immunoblot analysis (Fig. 4A). Even high concentrations of paclitaxel plus PP2, which induced large accumulations of cleaved caspase-3, did not result in processing of caspase-9 (Fig. 4B). Furthermore, ID8 cell lysates from combinatorial treated cells showed considerable caspase-3 activity compared with controls as determined by in vitro caspase-3 activity assays. In contrast, caspase-9 activity showed little change in response to treatments (Fig. 4C).

Procaspase-9 has a relatively high basal activity. In addition, in some experiments, some minor processing of casapase-9 was observed (see faint band in colchicine + PP2 treated cells in Fig. 4A), although this observation was not consistent. In order to rule out a role for caspase-9, several different approaches were undertaken. Caspase-9 is activated as a result of mitochondrial cytochrome c release. Immunoblot analysis for cytochrome c failed to show any increase in cytosolic cytochrome c resulting from any of the treatments (Fig. 5A). Furthermore, laser scanning confocal microscopy was utilized in order to colocalize cytochrome c and mitochondria. Functional mitochondria were labeled with a MitoTracker red dye. Cytochrome c localization was determined by immunofluorescence with a cytochrome c–specific antibody. Cytochrome c colocalized with mitochondria in both control and paclitaxel plus PP2 treated cells (Fig. 5B). In contrast, H₂O₂ treatment was associated with a dramatic decrease in functional mitochondria and release of cytochrome c to the cytosol.

Finally, small interfering RNA directed against caspase-9 effectively knocked-down caspase-9 protein expression with no concomitant decrease in caspase-3 processing in paclitaxel plus PP2 treated ID8 mouse ovarian cancer cells (Fig. 6). These results, taken in toto, suggest that mitochondrial release of cytochrome c and caspase-9 activation is not involved in caspase-3 activation in response to paclitaxel plus PP2.

Microtubule disrupting agents such as paclitaxel and colchicine induce cell cycle arrest, which ultimately results in apoptosis. Paclitaxel-resistant cells presumably fail to respond to paclitaxel with cell cycle arrest. In order to determine whether the resensitization of ID8^TaxR cells was the result of a restoration of cell cycle arrest in response to paclitaxel, cell cycle analysis was done under various treatment conditions. ID8 paclitaxel-sensitive ovarian cancer cells treated with paclitaxel (1 μmol/L) showed a dramatic decrease in G₁ and massive apoptosis (Fig. 7A). In contrast, ID8^TaxR cells treated with paclitaxel were completely unaffected by paclitaxel in comparison to control ID8^TaxR cells (Fig. 7B). Treatment of ID8^TaxR cells with the Src inhibitor PP2 (10 μmol/L) plus paclitaxel showed a decrease in the proportion of cells in G₁ and massive apoptosis (Fig. 7B). Even more dramatically, treatment with colchicine (1 μmol/L) induced a large accumulation of ID8 cells in G₂ and extensive apoptosis (Fig. 7A); however, a 5-fold increase in colchicine (5 μmol/L) had little effect on ID8^TaxR cells (Fig. 7B). Src inhibition restored the accumulation of ID8^TaxR cells in G₂ and apoptosis in response to colchicine (Fig. 7B).

As outlined above, the human SKOV3 ovarian cancer cell line is relatively resistant to paclitaxel. It has also recently been reported that paclitaxel-induced cytotoxicity in SKOV3 cells is both caspase-9- and 3-independent (16). We have also produced CaOV3 human ovarian cancer cells that are completely resistant to paclitaxel (CaOV3^TaxR cells) in a manner similar to ID8^TaxR cells. Treatment of ID8^TaxR and CaOV3^TaxR (Fig. 7C) cells with paclitaxel alone was associated with little or no accumulation of cleaved caspase-3. However, treatment with paclitaxel plus pharmacologic inhibition of Src tyrosine kinase resulted in significant accumulation of cleaved caspase-3 in both resistant cell types. Similar results were observed with several different Src-specific inhibitors (PP2, SU6656; data not shown). Again, this was independent of caspase-9 processing (data not shown). Interestingly, treatment of SkOV3 cells with paclitaxel alone resulted in significant cleavage of caspase-3 (Fig. 7C), even at relatively lower concentrations (0.2 μmol/L), contrary to previous reports. Paclitaxel plus pharmacologic Src inhibition did increase the accumulation of cleaved caspase-3.

Ovarian cancer is the leading gynecologic cause of death and the fifth leading cause of cancer deaths in women (1). Standard therapy includes tumor debulking followed by chemotherapy treatment with paclitaxel, platinum-based agents, or combinations of both (21). Unfortunately, the success of these treatment regimens in advanced disseminated disease is relatively poor, with 5-year survival rates <30% (1). Furthermore, development of recurrent drug-resistant disease is a common outcome in ovarian cancer. The data presented here shows that inhibition of Src tyrosine kinase, a proto-oncoprotein commonly overexpressed and constitutively activated in ovarian cancer and ovarian cancer cell lines (10, 11, 15), enhances the cytotoxicity of paclitaxel. Perhaps even more importantly, inhibition of Src tyrosine kinase resensitizes drug-resistant ovarian cancer cells to both paclitaxel and cisplatin. The question remains somewhat open concerning the Src family kinases involved in the resensitization. However, most evidence suggests that it is Src tyrosine kinase itself; expression of a Src dominant-negative or Src knockdown by RNA interference recapitulate the effects seen with pharmacologic Src kinase inhibitors (15).

The enhancement of cytotoxicity in paclitaxel-sensitive (ID8 and CaOV3), partially resistant (SkOV3), and resistant (ID8^TaxR and CaOV3^TaxR) cells seems to involve activation of caspase-3. Combination treatment of Src inhibition and paclitaxel resulted in a large accumulation of processed caspase-3, cleaved poly-ADP ribose polymerase, and increased caspase-3 activity. Interestingly, Src inhibition alone seems to induce some processing of caspase-3 in ID8, ID8^TaxR, and Skov3 cells (see Figs. 6 and 7), suggesting that Src inhibition alone promotes some caspase-3 activation. Src inhibition has been linked to enhanced apoptosis under various different apoptotic stimuli (22–25).

Caspase-3 is a distal executioner caspase that can be activated by either an intrinsic mitochondrial-cytochrome c-caspase-9 activated pathway or via extrinsic ligand (e.g., Fas ligand) activated signals. At this point, the mechanism by which caspase-3 is activated in response to Src inhibition-paclitaxel treatment is not clear. It does not seem to include a mitochondrial-caspase-9–activated pathway, as there was no mitochondrial release of cytochrome c and caspase-9 processing and activity was minimal. Most convincingly, caspase-3 was activated in response to combination treatment even in cells in which caspase-9 expression had been knocked down by RNA interference. Some caspase-9 processing was observed in some experiments, however, this was an inconsistent observation and the cleaved product largely seemed to be the 39 kDa species, possibly the result of caspase-3-mediated cleavage of caspase-9 (26).

We have previously reported that Src inhibition is associated with decreased phosphorylation of Akt and the forkhead transcription factor, FOXO1 (15). Dephosphorylation of forkhead transcription factors allows for their translocation to the nucleus where they activate proapoptotic genes such as Fas ligand (27), resulting in activation of an autocrine extrinsic apoptotic pathway. The lack of activation of a mitochondrial-caspase-9 intrinsic apoptotic pathway in response to Src inhibition and paclitaxel combinatorial treatment might suggest that caspase-3 activation is the result of an autocrine extrinsic death signal, resulting in activation of an initiating caspase, such as caspase-8 or caspase-10. We have tried to identify the possible upstream caspase-3-activating caspases with specific caspase inhibitors; unfortunately, low concentrations were ineffective and higher concentrations have proven to be highly toxic in all of the cell lines we tested. However, a neutralizing antibody against Fas ligand failed to block the increased cytotoxicity and caspase-3 activation. Furthermore, no caspase-8 processing was detected (data not shown). Thus we have no compelling evidence for an autocrine loop activating an extrinsic apoptotic pathway. Caspase-12, a caspase activated as a result of Ca²⁺ release from the endoplasmic reticulum (28, 29) can likewise be ruled out. Combinatorial treatment in fact decreased processed caspase-12 in comparison to control cells (data not shown).

It is possible that caspase-3 is being activated by a more direct route, perhaps by blocking the inhibition of an inhibitor of apoptosis. Inhibitors of apoptosis such as XIAP (30, 31) and survivin (32) have been implicated in protection from apoptotic stimuli in ovarian cancers cells. ID8 mouse ovarian cancer cells constitutively express high levels of survivin. Src or Akt inhibition is associated with decreased survivin expression (our observations). A recent report showed that SkOV3 cells have dysfunctional apoptosome activity which could not be explained by caspase-9 and APAF-1 or by expression of inhibitors of apoptosis, such as XIAP (33). However, XIAP was only examined in the context of apoptosome-caspase-9 activity. In the current study, combination treatment seems to bypass apoptosome formation and caspase-9 activation. It is possible that an inhibitor of apoptosis such as XIAP or survivin may act directly on caspase-3 (34), and that combination treatment interferes with this interaction or inhibition, promoting autoactivation of caspase-3. Further studies are required to test this possibility.

An important aspect of the present study is the demonstration that Src inhibition resensitizes drug-resistant CaOV3^TaxR and ID8^TaxR ovarian cancer cells to two different classes of chemotherapeutics, i.e. paclitaxel and cisplatinum. The mechanism of this resensitization is as yet unknown, however, it seems to be independent of multidrug resistance protein Mdr1 expression or function (our observations). The finding that Src inhibition and paclitaxel treatment results in caspase-3 processing in CaOV3^TaxR and ID8^TaxR cells, as well as partially resistant SkOV3 cells in a manner similar to that of sensitive cells suggests the resistance is proximal to caspase-3. Our results with SkOV3 cells are somewhat in conflict with a previous report demonstrating SkOV3 cell death to be independent of caspase-3 activity. However, other studies have shown a requirement for caspase-3 in adenoviral type 5 EIA-enhanced paclitaxel cytotoxicity in SkOV3 cells (19). The reasons for this discrepancy may involve several factors; different lots or passage numbers of the original SkOV3 cell line may contribute to the difference. The present study also used a higher concentration of paclitaxel (200 nmol/L), thus we believe caspase-3 to be activated in SkOV3 cells treated with paclitaxel.

In summary, we show that inhibition of Src tyrosine kinase enhances the cytotoxicity of chemotherapeutic agents such as paclitaxel and cisplatinum in drug-sensitive ovarian cancer cells and restores sensitivity in drug-resistant cells. The increased cytotoxicity seems to be the result of a caspase-9-independent increase in caspase-3 processing and activity. These results suggest that Src family kinase inhibitors may be useful agents in the treatment of drug-resistant ovarian cancers.