Expression of Kinase-defective Mutants of c-Src in Human Metastatic Colon Cancer Cells Decreases Bcl-x L and Increases Oxaliplatin- and Fas-induced Apoptosis*

Tumor resistance to current drugs prevents curative treatment of human colon cancer. A pressing need for effective, tumor-specific chemotherapies exists. The non-receptor-tyrosine kinase c-Src is overexpressed in > 70% of human colon cancers and represents a tractable drug target. KM12L4A human metastatic colon cancer cells were stably transfected with two distinct kinase-defective mutants of c-src. Their response to oxaliplatin, to SN38, the active metabolite of irinotecan (drugs active in colon cancer), and to activation of the death receptor Fas was compared with vector control cells in terms of cell cycle arrest and apoptosis. Both kinase-defective forms of c-Src co-sensitized cells to apoptosis induced by oxaliplatin and Fas activation but not by SN38. Cells harboring kinase-defective forms of function point mutations in in function altered Oxaliplatin-induced potentiated by kinase-defective c-Src on activation of caspase with of the kinase-defective mutants of c-Src were had reduced levels of Bcl-x inhibition of c-Src activity by in did the oxaliplatin over did it Bcl-x L that longer term suppres- sion of kinase activity may be required to lower Bcl-x L and sensitize cells to oxalipla-

Tumor resistance to current drugs prevents curative treatment of human colon cancer. A pressing need for effective, tumor-specific chemotherapies exists. The non-receptor-tyrosine kinase c-Src is overexpressed in >70% of human colon cancers and represents a tractable drug target. KM12L4A human metastatic colon cancer cells were stably transfected with two distinct kinasedefective mutants of c-src. Their response to oxaliplatin, to SN38, the active metabolite of irinotecan (drugs active in colon cancer), and to activation of the death receptor Fas was compared with vector control cells in terms of cell cycle arrest and apoptosis. Both kinasedefective forms of c-Src co-sensitized cells to apoptosis induced by oxaliplatin and Fas activation but not by SN38. Cells harboring kinase-defective forms of c-Src carrying function blocking point mutations in SH3 or SH2 domains were similarly sensitive to oxaliplatin, suggesting that reduction in kinase activity and not a Src SH2-SH3 scaffold function was responsible for the observed altered sensitivity. Oxaliplatin-induced apoptosis, potentiated by kinase-defective c-Src mutants, was dependent on activation of caspase 8 and associated with Bid cleavage. Each of the stable cell lines in which kinase-defective mutants of c-Src were expressed had reduced levels of Bcl-x L. However, inhibition of c-Src kinase activity by PP2 in vector control cells did not alter the oxaliplatin response over 72 h nor did it reduce Bcl-x L levels. The data suggest that longer term suppression of Src kinase activity may be required to lower Bcl-x L levels and sensitize colon cancer cells to oxaliplatin-induced apoptosis.
Colon cancer, often diagnosed at an advanced stage, is generally treated by surgical resection followed by adjuvant treatment with 5-fluorouracil plus leucovorin (1). However, because only 26% of patients respond, the need for improved therapies is pressing. Oxaliplatin, a platinum agent that directly damages DNA, and irinotecan, an inhibitor of topoisomerase I, are currently in clinical trial for colon cancer and are showing some significant activity and promise (2). The cellular response to anticancer drugs is dependent on molecules that couple druginduced damage to those that regulate the engagement of apoptosis (3,4). Indeed, the suppression of apoptosis is one of the six hallmarks of cancer (5). Several oncogenes and mutated tumor suppressor genes relevant to colon cancer, such as Kirsten (Ki) Ras and p53, modulate the threshold for druginduced apoptosis, prompting drug discovery programs directed at these molecules and at the pathways that they regulate (6). The non-receptor-tyrosine kinase c-Src is overexpressed in Ͼ70% of human colon cancers (7,8), and recent studies suggest that its activity may also modulate apoptosis (9,10).
Cellular Src (c-Src) is ubiquitously expressed, and its activity is tightly controlled in healthy cells. It is normally inactivated via phosphorylation of tyrosine 527 by C-terminal Src kinase (11). Phosphorylated Tyr-527 binds its own SH2 domain, which is tightly coupled to its SH3 domain, and this inhibits kinase activity (the closed conformation). Activating c-Src involves both dephosphorylation at Tyr-527 (e.g. by protein-tyrosine phosphatase PTP␣) or the binding of tyrosine-phosphorylated cellular proteins to the SH2 domain to destabilize the phosphorylated Tyr-527 intermolecular interaction with the SH2 domain (12). These events allow c-Src to adopt an open conformation that results in autophosphorylation at Tyr-416 and c-Src kinase activity. C-terminal Src kinase is down-regulated in a subset of human colon cancers, providing one mechanism explaining the high activity of c-Src in this disease (11), and c-Src is mutated and activated in a subset of late-stage colon cancer (13). The roles c-Src plays in signaling changes in cellular adhesion and motility together with its ability to promote proliferation are well established (14 -16), but surprisingly little has been reported regarding its role in the regulation of apoptosis in colon cancer cells or on the effect of c-Src on sensitivity to drugs used to treat this disease. Studies carried out in transformed fibroblasts expressing the viral homologue v-Src suggest that activated Src can signal cell survival via phosphatidylinositol 3-kinase and Ras-dependent pathways but can also prime for apoptosis via an unknown mechanism (9, 10). Shp-2 has been shown not only to be necessary for morphological transformation by v-Src but also for activation of AKT via phosphatidylinositol 3-kinase (17). Transformation of mammalian fibroblasts by v-Src specifically induces constitutive activation of STAT3 (18), and this is thought to require activation of p38 MAPK (mitogen-activated protein kinase) and c-Jun N-terminal kinase, (19). v-Src and c-Src have also been shown to up-regulate the anti-apoptotic protein Bcl-x L in fibroblasts and in intestinal epithelial cells (20 -22). These studies suggest a connection between Src-mediated STAT3 phosphorylation and bcl-x L transcription, although there is no direct evidence to link the two events. Here we assess the impact of expressing kinase-defective mutants of c-Src on drug-induced apoptosis in human metastatic colon cancer cells in vitro.

Cell Culture
Cells were cultured in Eagle's minimal essential medium with Earle's salts supplemented with minimal essential medium vitamins (ϫ2), non-essential amino acids, L-glutamine 2 mM, and sodium pyruvate (1 mM) (all from Invitrogen) in the presence of 10% fetal bovine serum. All cells were routinely maintained in log phase in a humidified incubator at 37°C with 5% CO 2 .

Drug Preparation
Oxaliplatin was purchased from Alexis Biochemicals (Nottingham, UK) and was dissolved at 12.5 mM in sterile phosphate-buffered saline. SN38 was from Aventis Pharma (West Malling, UK) and was dissolved in Me 2 SO (2 mM). Caspase inhibitors IETD CHO and DEVD CHO, the c-Src destabilizing agent herbimycin A, and the Src kinase inhibitor PP2 were from Calbiochem and were dissolved in Me 2 SO as stock solutions. In selected experiments cell monolayers were washed, and medium containing freshly prepared PP2 at 100 nM-10 M was added daily for up to 72 h. Low concentrations of PP2 were used to minimize potential inhibition of other kinases (23).

Cell Viability
Cells were seeded at 5 ϫ 10 5 cells/well in 6-well plates before drug treatment. Detached and adherent cells were collected and pelleted, cell pellets were resuspended in Vectorshield® containing 4,6-diamidino-2phenylindole (Vector Laboratories, Burlington, CA) before examination of nuclear morphology by fluorescent microscopy. Apoptotic cells were identified with nuclear condensation and fragmentation. The {2,3bis(2-methoxy-4-nitro-5-sulfophenyl)-5[(phenylamino)carbonyl]-2Htetrazolum hydroxide} assay was performed as previously described (24). For clonogenic assays, cells were seeded at 5 ϫ 10 3 cells/well in a 6-well plate and allowed to adhere overnight. Medium was removed and replaced with warm, fresh medium or medium containing 50 or 100 M oxaliplatin, and cells were left at 37°C for 1 h. After incubation, medium was removed, cells were washed once in medium, and fresh medium was replenished. Cells were left to grow for 10 -14 days after which colonies were fixed in 70% methanol, stained with methylene blue, photographed, and scored using Total Lab colony counter v1.11 from Phoretix (Newcastle, UK).

Cell Cycle Analysis
Detached and adherent cells were washed in phosphate-buffered saline, fixed in 70% ethanol, incubated with RNase, and stained with propidium iodide, and DNA content was analyzed by flow cytometry as previously described (25).

Drug Uptake and DNA Damage
Total Intracellular Platinum Levels-Control and oxaliplatin-treated cells were washed twice with phosphate-buffered saline to remove unreacted drug and then harvested by centrifugation. Cell pellets (1 ϫ 10 6 cells) were resuspended in ice-cold deionized distilled H 2 0. 90 l of sonicated sample was solubilized with 10 l of 10 M NaOH (Sigma) and incubated at 60°C for 1 h, whereas 10 l was retained for protein quantification. 900 l of 1 M HNO 3 was added, and samples were incubated at 70°C overnight and then stored at Ϫ20°C before inductively coupled plasma mass spectrometry analysis. HNO 3 was Spec-trosoL grade (69%, BDH) and diluted 1:20 in deionized distilled H 2 O for a final concentration of 1 M.
Oxaliplatin DNA Adducts-DNA extraction was performed using the Qiagen Genomic-tip system according to the manufacturer's protocol. Lysed pellets were sonicated after the addition of buffer G2 and subjected to RNase digestion for 45 min at 37°C before digestion with proteinase K for 1 h at 50°C. Quantification of DNA was performed by UV-visible spectrophotometry with ratios of A 260 nm /A 280 nm of 1.8.

Inductively Coupled Plasma Mass Spectrometry
To aliquots of DNA solution, equal volumes of nitric acid (7% w/v) were added, and the solutions then incubated at 70°C for 18 h before analysis. Whole cell lysates were diluted with nitric acid (3.5% w/v) to give platinum concentrations appropriate for analysis. The platinum contents of these preparations were analyzed on a PerkinElmer Life Sciences Sciex Elan 6000 inductively coupled plasma mass spectrometer using a standard cross-flow nebulizer and Scott-type double-pass spray chamber. The method involved monitoring four isotopes of platinum ( 194 Pt,195 Pt,196 Pt,198 Pt) and correction for isobaric interferences from mercury as described previously (25). Platinum concentrations were derived with reference to a standard solution calibration curve and are the mean of the concentrations derived from monitoring of the 194 Pt, 195 Pt, and 196 Pt isotopes.
Activation of c-Src was assessed in three ways. First, a phosphospecific antibody reactive with the 416 phosphorylated form of c-Src (catalog #44-660, BIOSOURCE International, Inc., Camarillo, CA) was used to assess c-Src activation status. Second, the effect of c-Src phosphorylation of its substrate paxillin was assessed. Cell extracts were prepared in radioimmune precipitation assay buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS), and paxillin immunoprecipitates were prepared from 1 mg of protein with 4 g of anti-paxillin antibody (Transduction Laboratories, Lexington, KY). Immunocomplexes were collected using protein A-Sepharose and washed extensively with radioimmune precipitation assay buffer. Proteins were resolved by 7% SDS-PAGE gels before transfer to nitrocellulose membranes and probed with anti-phosphotyrosine antibody (pY20, Transduction Laboratories). Blots were subsequently stripped and reprobed with anti-paxillin antibody (27). Last, c-Src kinase activity was assessed using a standard in vitro kinase assay using a kit from Upstate Biotechnology according to the manufacturer's instructions.  The increased levels of c-Src protein are shown in the cells overexpressing the kinase-defective forms (251-6, MF10, MFW, and MFR) (Fig. 1B). The activity of endogenous c-Src in these cell lines was assessed using two different methods. First, the phosphorylation status of the Src substrate paxillin was examined using a phospho-specific antibody. This showed a decrease in Paxillin phosphorylation in cells expressing MF-10 and 251-6 compared with the empty vector control (Fig. 1C, i and  ii). Second and confirming these results, c-Src was immunoprecipitated, and its activity was assessed using an in vitro kinase assay (Fig. 1Ciii). There was decreased Src kinase activity in cells expressing either kinase-defective mutant compared with the vector controls. No compensatory rise in the protein expression levels of other Src family members Fyn and Yes was observed after expression of the kinase-defective c-Src mutants (Fig. 1D).

Expression of c-Src Kinase-defective Mutants Increased Sensitivity to Oxaliplatin Downstream of Drug-induced Damage-
The response of MF-10 and 251-6 (and 251-13, not shown) to oxaliplatin in vitro was assessed by 3 day {2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5[(phenylamino)carbonyl]-2H-tetra-zolumhydroxide} viability assay ( Fig. 2A)  c-Src Impact on Drug Sensitivity in Human Colon Cancer Cells controls. These data were mirrored by a pronounced increase in the percentage of apoptotic cells in KM12L4A clones expressing the kinase-defective forms of c-Src 24 -72h after treatment with oxaliplatin (150 M). Expression of both types of kinasedefective c-Src also reduced the IC 50 concentrations for cisplatin that were 70 M for MF-10, 251-6 and 251-13 and 90 M for empty vector controls. The increased sensitivity of MF-10 and 251-6 cells to oxaliplatin in comparison with vector controls was also observed in a clonogenic assay; oxaliplatin-treated MF-10 and 251-6 cells displayed 33 and 21% clonogenicity relative to untreated cells (taken as 100%). In comparison, oxaliplatin treatment resulted in 76.9% clonogenicity relative to untreated vector control cells (Fig. 2C).
To determine whether the enhanced sensitivity to oxaliplatin mediated by kinase-defective c-Src was upstream or downstream of oxaliplatin-induced cellular damage, we examined changes in the cell cycle (Table I), the cellular uptake of oxaliplatin (Fig. 3A), and the degree of DNA damage inflicted (

c-Src Impact on Drug Sensitivity in Human Colon Cancer Cells
KM12L4A cells to Fas-induced apoptosis. Cells were exposed to the CH-11 Fas-activating antibody (25 ng/ml), and apoptosis was assessed up to 72 h later. Cells expressing either form of kinase-defective c-Src were significantly more sensitive to Fasinduced apoptosis with 90% cell death at 72 h compared with 20% cell death in the empty vector control cells (Fig. 4A). However, no differences were observed between the KML124A clones in the protein levels of Fas, Fas ligand, the adaptor FADD, or the negative regulator of the death-inducing signaling complex FLICE-inhibitory protein (data not shown).
The role of the initiator caspase, caspase 8, in the co-sensitization to oxaliplatin and Fas by dominant negative c-Src was then examined. Oxaliplatin-induced apoptosis, the cleavage of caspase 8, and of the downstream effector caspase, caspase 3, was examined in control and MF-10 cells in the presence and absence of the caspase 8-specific inhibitor IETD CHO or the caspase 3 (and 7) inhibitor DEVD CHO (Fig. 4, A and B). Oxaliplatin-induced cleavage of both caspase 8 and caspase 3 was observed in MF-10 but not vector control cells. That IETD CHO suppressed both the oxaliplatin-induced cleavage of caspase 8 and of caspase 3 suggested that it acted upstream of caspase 3 in the oxaliplatin-induced apoptotic pathway (Fig.  4B). Consistent with these data, the dominant negative c-Srcmediated sensitization to oxaliplatin-induced apoptosis was completely abrogated by caspase 8 inhibition (Fig. 4C). Caspase 3 inhibition also prevented the detection of cleaved caspase 8 (Fig. 4B), implying that caspase 3 may be acting as an amplifier of caspase 8 cleavage. The pro-apoptotic BH-3-only Bcl-2 family member Bid is cleaved by caspase 8 to generate truncated Bid that directly or indirectly activates Bax and/or Bak at the mitochondrial surface to facilitate release of apoptogenic factors from mitochondria in a pathway that ultimately activates effector caspases (32). Oxaliplatin treatment resulted in increased full-length Bid in both vector control and MF-10 cells (Fig. 4D). However, Bid was cleaved only in oxaliplatin-treated MF-10 cells (Fig. 4D). In MF-10 cells the appearance of trun- c-Src Impact on Drug Sensitivity in Human Colon Cancer Cells cated Bid was suppressed by inhibition of caspase 8 or caspase 3, again suggesting that caspase 3 activity served to amplify a caspase 8-initiated apoptotic pathway.
The Sensitivity to Oxaliplatin in c-Src-defective Mutants Was Not Attributable to a Kinase-independent Scaffold Function of Src-The data above show that increased oxaliplatin sensitivity was observed in KM12L4A cells that stably overexpress full-length kinase-defective Src in an open conformation (i.e. in MF-10) and in those which overexpress truncated Src mutants without the kinase domain but with open available SH3 and SH2 domains (i.e. 251-6 and 251-13). To determine whether the enhanced oxaliplatin sensitivity was due to the SH3-SH2 scaffold function of c-Src, the response to oxaliplatin was tested in KM12L4A cells overexpressing the same truncated c-Src mutants in which the additional function blocking point mutations were present in either the SH3 or SH2 domain (i.e. MFW and MFR). Fig. 5 shows that, like MF10 and 251-6 cells, MFW and MFR cells exhibited enhanced sensitivity to oxaliplatin compared with vector controls. The point mutations in the SH3 and SH2 domains have been previously shown to abrogate proteinprotein interactions with c-Src, suggesting that these interactions are not required for the observed oxaliplatin sensitivity (33).
Inhibition of c-Src Kinase by PP2 Did Not Sensitize KML4A Control Cells to Oxaliplatin-induced Death-To confirm that c-Src kinase activity modulates oxaliplatin-induced apoptosis in KM12L4A cells, we examined the impact of the small molecular weight Src TK inhibitor PP2 on oxaliplatin response in KM12L4A vector control cells. In contrast to the stable expression of kinase-defective c-Src mutants (e.g. MF-10) PP2 (1-10 M) when added every 24 h for up to 72 h had no effect on oxaliplatin response (Fig. 6A), although it clearly reduced c-Src kinase activity as determined by reduced reactivity with the phospho-specific antibody reactive with the tyrosine 416-phosphorylated form of c-Src (Fig. 6B). The lack of effect of PP2 on oxaliplatin-induced apoptosis was not a reflection of altered cell growth kinetics over the 72-h period (data not shown). Similar studies using herbimycin A at the minimal concentration that inhibited Src kinase activity but at a time point at which there was no decrease in c-Src expression also failed to increase the levels of oxaliplatin-induced apoptosis (data not shown).

The Role of Bcl-x L in Cell Lines Expressing Kinase-defective Mutants of c-Src or Those Treated with a Small Molecule Src
Kinase Inhibitor-Src is implicated in signaling for the upregulation of Bcl-x L in a variety of cell types including intestinal epithelial cells (18, 20 -22), and it was possible that the sensitivity to oxaliplatin might reflect Bcl-x L down-regulation in KM12L4A cells that stably overexpressed the kinase-defective c-Src mutants that act in a dominant negative fashion. Indeed this was observed in all KM12L4A cell lines expressing c-Src kinase-defective mutants compared with vector controls (Fig. 6C). However, daily exposure to 1-10 M PP2, although reducing c-Src activity, failed to reduce Bcl-x L levels in KM12L4A vector control cells, consistent with its inability to sensitize these cells to oxaliplatin (Fig. 6B).

c-Src Kinase-defective Mutants Have No Effect on the Response to SN-38, the Active Metabolite of Irinotecan-Irinote-
can is a topoisomerase I inhibitor that shows promise in the treatment colon cancer (2). We examined the sensitivity of the KM12L4A cell lines to the active metabolite of irinotecan SN-38 over a range of concentrations but found no difference between the response of vector control cells and those expressing the kinase-defective forms of c-Src (data not shown). There was also no effect of kinase-defective c-Src mutants on SN-38 mediated G 2 /M arrest (data not shown). We next determined whether or not caspase 8 inhibition would alter the apoptotic response to SN-38 in vector control and MF-10 cells and found that it was without effect in either cell line (data not shown). In contrast, inhibition of the effector caspase 3 by DEVD CHO reduced SN-38-induced apoptosis in both vector and MF-10 cells (not shown).

DISCUSSION
Elevated c-Src signaling has been implicated in development of carcinomas of the colon, breast, lung, esophagus, skin, parotid, cervix, ovary, and stomach (8). The participation of c-Src in a variety of signal transduction pathways results in deregulated cellular behavior that may be important for the development of malignancy, including cell proliferation, cell survival, invasion, and metastasis. However, the impact of c-Src signaling on the response to clinically relevant chemotherapeutic drugs has not been characterized in appropriate cellular contexts that reflect the human tumors in which c-Src functions are important.
We have examined drug responses in the human metastatic colon cancer cell line KM12L4A. This cell line was derived after intrasplenic injection of KM12C cells (derived from a Dukes' B human colon cancer patient) into the nude mouse and represents the fourth passage of the KM12C cell through the spleen and liver (34). These cells are mismatch repair-defective and harbor mutant Ki-Ras and mutant p53. They also display higher levels of total and activated c-Src than their non-metastatic predecessor KM12C (27,35), consistent with previous reports of this cell line and in human colon cancer in vivo (36).
Recent studies by ourselves and others of v-Src signaling in fibroblasts suggested to us that activated c-Src might modulate drug-induced apoptosis. In LA29 Rat-1 cells, v-Src had a proapoptotic effect under low serum conditions, possibly via upregulation of c-Myc (10), an oncoprotein that promotes both cell proliferation and apoptosis (37). In a separate study, v-Src suppressed apoptosis in transformed Rat-2 fibroblasts via the synergistic activation of phosphatidylinositol 3-kinase and of

c-Src Impact on Drug Sensitivity in Human Colon Cancer Cells
Ras (9). The conclusions drawn from both studies were that v-Src could both prime for and suppress apoptosis in fibroblast models but that the priming for apoptosis was only apparent when serum survival factors were limiting and/or v-Src mediated survival signaling pathways were blocked. Where drug sensitivity and resistance in v-Src-transfected cells have been examined together with the roles of Ras and phosphatidylinositol 3-kinase in drug response, a complex picture begins to emerge. v-Src expression in HAG-1 gallbladder cells resulted in resistance to cisplatin (38) that was not mediated by the Ras or phosphatidylinositol 3-kinase pathways and sensitivity to taxotere that was associated with Bcl-2 phosphorylation, phosphatidylinositol 3-kinase-independent and reversed by Ras signaling (39). Expression of v-Src in NIH3T3 fibroblasts resulted in increased etoposide-induced apoptosis that was associated with prolonged stress-activated protein kinase/c-Jun N-terminal kinase activation (40). The impact of c-Src on gemcitabineinduced apoptosis has already been demonstrated in pancreatic cancer cells, where expression of an activated c-Src mutant decreased drug response, and conversely, a dominant negative c-Src mutant (K295R) increased sensitivity (41). In these studies the small molecule Src TK inhibitor PP2, used at 10 M, enhanced gemcitabine sensitivity in contrast to our results with colon cancer cells.
Taken together these pharmacological studies illustrate the complexity of signaling pathways that may occur both downstream of and/or in parallel to v-Src signaling, where the therapeutic outcome appears both insult-and cell type-specific. Therefore, it was not readily possible to predict the role that elevated c-Src activity might have in colon cancer cell survival after treatment with anti-cancer drugs.
Here we have demonstrated that overexpression of two distinct kinase-defective forms of c-Src resulted in a caspase 8-dependent co-sensitization to oxaliplatin and to Fas in human colon cancer cells, suggestive of a common pathway. Colon carcinoma cells that metastasize to the liver induce hepatocyte cell death, facilitating penetration of the metastatic colon cancer cells into the tissue (42). Colon cells express Fas and during carcinogenesis, their expression of FasL becomes elevated (42) yet they remain resistant to Fas-induced death via an unknown mechanism. In our studies we were unable to detect changes between vector control and dominant negative c-Src-expressing cells in the levels of FasL or Fas death-inducing signaling complex-associated proteins (FLICE-inhibitory protein, FADD, or caspase 8), suggesting that c-Src may not regulate Fas resistance/sensitivity at the protein expression level. c-Src may instead alter the formation and activity of the death-inducing signaling complex, and previous studies implicate this possibil- c-Src Impact on Drug Sensitivity in Human Colon Cancer Cells ity (43,44). In this regard, the Src-like tyrosine kinase Btk (Bruton's tyrosine kinase) suppresses Fas-induced apoptosis in B-lineage lymphoid cells (43). It does so by associating with Fas and preventing the FAS-FADD interaction necessary for formation of the death-inducing signaling complex. Bruton's kinase (Btk) is activated by Src, and down-regulation of Src by C-terminal Src kinase abrogates Btk mediated suppression of Fas-induced death (43)(44)(45). Also, several recent studies show the Src-like kinase Lck to be involved in the caspase 8 pathway (46,47). In cells deficient for Lck, radiation-induced activation of caspase 8 was inhibited, although in this case Fas-induced death was only slightly slowed. A Btk-like kinase activated downstream of c-Src may, therefore, be responsible for providing resistance to Fas in vector control KM12L4A cells, and inhibition of c-Src in the MF-10 and 251-6 cells would prevent its activation. Whatever the precise mechanism linking c-Src signaling to the suppression of Fas-induced apoptosis, our data have potentially important ramifications in vivo. The increase in c-Src signaling during colon cancer progression may counteract the parallel increase in FasL expression to prevent Fasinduced apoptosis of the metastatic colon cancer cells, whereas in normal hepatocytes the relatively low expression and tightly controlled activities of c-Src may permit FasL-induced activation of the Fas pathway and apoptosis.
It was possible that overexpression of the structural domains (SH3-SH2) could result in recruitment of proteins that are involved in the regulation of apoptosis. However, function blocking point mutations in either domain within the open kinase-defective form c-Src failed to prevent the sensitization of oxaliplatin and suggested that it was a reduction in the cellular c-Src kinase activity that was responsible for the alteration in drug response. The lack of effect of PP2 on oxaliplatin response in KM12L4A cells appears at odds with this interpretation of our data. Differences in the levels of Bcl-x L in the cells constitutively expressing c-Src kinase-defective mutants compared with those seen in PP2-treated vector control cells may offer the explanation for this. Previous studies (e.g. Ref. 48) suggest that activated Src provides an anti-apoptotic stimulus through the positive regulation of Bcl-x L expression, thus elevating the "oncogenic robustness" of the transformed cell. Consistent with the activation of this pathway, a decrease of Bcl-x L level was seen in all of the cell lines constitutively overexpressing c-Src kinase-defective mutants that were oxaliplatin-sensitive. PP2 on the other hand failed to reduce Bcl-x L levels when added every 24 h for 3 days and failed to sensitize to oxaliplatin, although it clearly reduced c-Src kinase activity. Bcl-x L has a long half-life (Ͼ24 h), and we speculate that c-Src kinase activity may need to be ablated more permanently to achieve a functional down-regulation of Bcl-x L . Others have also shown that small molecule inhibitors of c-Src do not in the short term modulate the levels of anti-apoptotic proteins of the Bcl-2 family including Bcl-x L , albeit in osteoclasts (49).
The data presented here show for the first time in human metastatic colon cancer cells that attenuation of c-Src signaling sensitizes to apoptosis induced by the clinically relevant anticancer drug oxaliplatin and also to the activation of the Fas death receptor by a caspase 8-dependent mechanism. The impact of c-Src signaling on drug response is not a general phenomenon as kinase-defective mutants of c-Src had no effect on apoptosis induced by the topoisomerase I inhibitor SN-38 that exerts its cytotoxicity independently of caspase 8. Our data may help guide the choice of drugs to be used in combination with small molecule inhibitors of c-Src and also suggest that complete and long term reduction of Bcl-x L might be required for enhanced sensitivity to drug-induced apoptosis.