SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The nonreceptor tyrosine kinase c-Src is frequently over-expressed or hyperactivated in various human cancers and contributes to cancer progression in cooperation with up-regulated growth factor receptors. However, Src-selective anticancer drugs are still in clinical trials. To identify more effective inhibitors of c-Src-mediated cancer progression, we developed a new screening platform using Csk-deficient cells that can be transformed by c-Src. We found that purvalanol A, developed as a CDK inhibitor, potently suppressed the anchorage-independent growth of c-Src-transformed cells, indicating that the activation of CDKs contributes to the c-Src transformation. We also found that purvalanol A suppressed the c-Src activity as effectively as the Src-selective inhibitor PP2, and that it reverted the transformed morphology to a nearly normal shape with less cytotoxicity than PP2. Purvalanol A induced a strong G2-M arrest, whereas PP2 weakly acted on the G1-S transition. Furthermore, when compared with PP2, purvalanol A more effectively suppressed the growth of human colon cancer HT29 and SW480 cells, in which Src family kinases and CDKs are activated. These findings demonstrate that the coordinated inhibition of cell cycle progression and tyrosine kinase signaling by the multi-selective purvalanol A is effective in suppressing cancer progression associated with c-Src up-regulation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The deregulation of tyrosine kinases has been implicated in the pathogenesis of various malignant human cancers (Hunter 1998). c-Src is a nonreceptor tyrosine kinase that is under the strict control of its negative regulator kinase, C-terminal Src kinase (Csk), and its adopter, Csk-binding protein (Cbp) (Nada et al. 1991, 1993; Kawabuchi et al. 2000). c-Src normally plays pivotal roles in the control of cell growth, differentiation, adhesion and migration (Jove & Hanafusa 1987; Brown & Cooper 1996). Although the C-SRC gene is rarely mutated, the protein levels and/or kinase activity of c-Src are frequently elevated in various malignant human cancers (Cartwright et al. 1990; Talamonti et al. 1993). Up-regulated c-Src activity contributes to tumor growth, invasion and angiogenesis (Yeatman 2004). Because c-Src serves as a critical downstream effector of growth factor receptors (e.g., EGFR), many of which are frequently amplified or mutated in human cancers, it can contribute to cancer progression when there is up-regulated receptor signaling (Ishizawar & Parsons 2004). Because of the crucial role of c-Src in human cancer, chemical inhibitors of c-Src such as dasatinib (BMS-354825), bosutinib (SKI-606) and saracatinib (AZD0530) (Kim et al. 2009) have been developed, but most of them are still in clinical trials.

However, it is known that the over-expression of c-Src in normal cells does not induce cell transformation because of the presence of the strict regulatory system enforced by Csk and Cbp (Masaki et al. 1999; Kawabuchi et al. 2000; Oneyama et al. 2008a). Therefore, to analyze the mechanism of Src-mediated transformation, transformed cells expressing v-Src or a constitutively active form of c-Src with a mutation at the regulatory site (c-Src527F) were used (Yeatman 2004). However, to analyze the precise functions of up-regulated c-Src in cancer, a more appropriate system that can reconstitute the up-regulation of endogenous c-Src is needed. For this purpose, we previously developed a model system using Csk-deficient cells that could be transformed by the expression of a limited amount of c-Src (Oneyama et al. 2008b). Using this system, we unraveled the functional differences between the active form of Src (v-Src or c-Src527F) and wild-type c-Src (Oneyama et al. 2008b). Thus, we think that this model system offers significant benefits as a platform for screening potential inhibitors of up-regulated c-Src in cancer.

Most tumors consist of heterogeneous cell populations at distinct developmental stages, making it difficult to treat them effectively with a single agent. Thus, combination therapies that target multiple signaling pathways might prove more effective. Anecdotal evidence has also suggested that agent targets multiple cancer-related pathways can be more effective (Drevs et al. 2003). Indeed, multi-kinase inhibitors such as sorafenib (Nexavar), which acts on VEGFR, cRAF and PDGFR, and sunitinib (Sutent), which targets PDGFR, VEGFR and c-KIT, have already been approved. In the case of Src-mediated transformation, multiple signaling pathways are activated because of the multiple functions of Src (Yeatman 2004). Furthermore, it has been recently reported that transient activation of Src is sufficient to activate a positive feedback loop that can maintain the transformed phenotype (Iliopoulos et al. 2009). This intriguing finding suggests that inhibition of the pathways downstream of c-Src is necessary to achieve more effective suppression of Src-mediated tumor growth.

In this study, to identify effective inhibitors of c-Src-mediated tumor growth, we developed a new screening system using Csk-deficient cells transformed by c-Src. Interestingly, we found that purvalanol A, a conventional CDK inhibitor, could effectively prevent c-Src-mediated transformation by inhibiting both cell cycle progression and c-Src signaling, and we also found that it effectively suppressed the anchorage-independent growth of some human cancer cells in which c-Src is up-regulated. Therefore, we propose that the multi-selective drug purvalanol A could serve as a potent seed compound for the treatment of various human cancers associated with c-Src up-regulation.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Identification of purvalanol A as an effective inhibitor of the growth of c-Src-transformed cells

To search for effective inhibitors of c-Src-mediated tumor progression, we examined the effects of 288 compounds supplied from the Screening Committee of Anticancer Drugs (SCAD) on the anchorage-independent growth of Csk-deficient cells transformed by c-Src (Csk−/−/c-Src cells) (Oneyama et al. 2008b). The cells were cultured in 96-well plates precoated with poly-HEMA (Fukazawa et al. 1995, 1996), and cell viability was determined by the MTT assay (Fig. 1a). The effects of the compounds on the anchorage-dependent growth and cell morphology of Csk−/−/c-Src cells was monitored to verify the reversion of transformed phenotype, and the compounds’ effects on normal mouse embryonic fibroblasts (MEFs) were also observed to assess their cytotoxicity to normal cells (Fig. 1a). Before screening the compounds, we examined the dose dependency of the effects of PP2, a widely used inhibitor of Src family kinases, in our screening system (data not shown). The results showed that the optimal concentration of PP2 is 3 μm for selectively inhibiting the anchorage-independent growth of Csk−/−/c-Src cells without influencing normal growth of MEFs. Therefore, in the first screening, we used each compound at a concentration of 3 μm to compare its efficacy with that of PP2. Of these compounds, we found that 22 selectively inhibited the anchorage-independent growth of Csk−/−/c-Src cells with apparent morphological reversion and less cytotoxicity to normal MEFs (Fig. 1b). As expected, the inhibitors of Src family kinases PP1 and PP2 were included in these candidates, and the others included inhibitors of protein kinases implicated in cancer progression and cell cycle control. Interestingly, we found that purvalanol A, originally developed as an inhibitor with high selectivity for CDK1 and CDK2 (Gray et al. 1998, 1999), effectively inhibited anchorage-independent growth and induced the most dramatic reversion of cell morphology (Fig. 1c). The suppressive effects of cell cycle inhibitors on c-Src-mediated transformation suggest that activation of the cell cycle machinery contributes to c-Src transformation and that CDK inhibitors such as purvalanol A would be effective for the treatment of c-Src-mediated tumor progression.

image

Figure 1.  Identification of purvalanol A as a compound that inhibits c-Src-induced cell growth. (a) A flow chart of the screening strategy. Cells were seeded on both poly-HEMA-coated and noncoated 96-well plates, and the inhibitors were added to each well at a concentration of 3 μm. After incubation for 2 days, cell viability was quantified by MTT assay, and cell morphology was observed under a light microscope. (b) The relative viabilities of cells treated with the effective inhibitors compared to those of mock-treated cells are shown as percentages. (c) Soft-agar colony formation assays of transformed Csk−/−/c-Src cells treated with varying concentrations of purvalanol A. Colonies were stained with MTT 9 days after plating (upper panels), and were observed under a light microscope (middle panels). Representative dishes from three independent experiments are shown. The morphology of cells on tissue culture plates was also observed (lower panels).

Download figure to PowerPoint

Inhibition of CDK2 contributes to the suppression of c-Src-induced transformation

Purvalanol A has greater specificity for CDK1 and CDK2 than other CDKs. Therefore, to determine the target of purvalanol A in these cells, we investigated the contribution of CDK1 and CDK2 to c-Src-induced transformation. First, we compared the expression and activity of cell cycle regulatory proteins in Csk−/−/c-Src cells with those in normal MEFs and Csk−/− cells by Western blot analysis (Fig. 2a). Although the phosphorylation of the CDK1 substrate PP1α was not prominently altered, the levels of the CDK1/2 inhibitor p27kip1 (Belletti et al. 2005) and cyclin E were both substantially down-regulated in Csk−/−/c-Src cells. Because it is known that cyclin E serves as a substrate of activated CDK2 (Welcker et al. 2003) and is rapidly degraded because of the phosphorylation by CDK2, the down-regulation of cyclin E protein indirectly indicates that the activity of CDK2 is up-regulated by c-Src transformation. It should be noted here that cyclin E, but not p27kip1, was also down-regulated in pretransformed Csk−/− cells, in which endogenous c-Src is activated to a degree below the threshold for full transformation (Nada et al. 1993; Oneyama et al. 2008b). Csk−/− cells have higher ability of cell growth compared with normal MEFs, and they can be easily transformed by the expression of small amounts of exogenous c-Src (Oneyama et al. 2008b). Therefore, it is likely that the down-regulation of cyclin E is more sensitive to c-Src activity than that of p27kip1 and may be associated with cell growth promoted by the activation of c-Src signaling.

image

Figure 2.  Inhibition of CDK2 contributes to the suppression of c-Src-induced transformation. (a) The expression and activity of cell cycle regulatory molecules in the indicated cell types were analyzed by Western blotting. Representative blots from three independent experiments are shown. (b) The phosphorylation of PP1α as a CDK1 substrate and the expression of cyclin E as a CDK2 substrate in Csk−/−/c-Src cells expressing a dominant negative CDK2 (CDK2T160A), a dominant negative CDK1 (CDK1T161A) or p27kip1 were analyzed by Western blotting. Representative blots from three independent experiments are shown. (c) The effect of the expression of CDK2T160A, CDK1T161A or p27kip1 on anchorage-independent growth was analyzed by soft-agar colony formation assay (upper panels). Colonies were scored 9 days after plating, and the mean number of colonies per cm2 ± SD obtained from three independent experiments is shown (lower panel). (d) The effect of the expression of CDK2T160A, CDK1T161A or p27kip1 on anchorage-dependent growth was analyzed by MTT assay (left) and growth curve analysis (right). Means ± SD were obtained from three independent assays. (e) The morphology of Csk−/−/c-Src cells expressing the indicated inhibitory proteins was visualized by F-actin staining with Alexa Fluor 594 phalloidin.

Download figure to PowerPoint

We next examined whether the activation of CDK1 or CDK2 contributes to the anchorage-independent growth of c-Src-transformed cells. The expression of a dominant negative CDK2 (CDK2T160A) or p27kip1 in c-Src-transformed cells restored cyclin E expression more effectively than a dominant negative CDK1 (CDK1T161A). This result further supports the notion that CDK2 is dominantly activated in these cells and its activity is suppressed by the expression of CDK2T160A or p27kip1 (Fig. 2b). Consistent with these results, the expression of CDK2T160A or p27kip1 suppressed the anchorage-independent growth of c-Src-transformed cells more effectively than CDK1T161A (Fig. 2c). Notably, the suppression of CDK2 by these inhibitor proteins did not affect the anchorage-dependent growth of these cells, suggesting that CDK2 activation is required for cells to acquire the capacity for anchorage-independent growth (Fig. 2d). However, we observed that the suppression of CDK2 by CDK2T160A or p27kip1 did not induce as remarkable a reversion of cell morphology as purvalanol A (Fig. 2e). These findings suggest that CDK2 becomes activated in c-Src-transformed cells, and that inhibition of CDK2 can suppress the anchorage-independent growth of c-Src-transformed cells; however, additional pathway(s) are involved in the morphological reversion induced by purvalanol A.

Purvalanol A inhibits the cellular functions of c-Src

To elucidate the mode of purvalanol A’s action on Src-transformed cells, we compared the effects of purvalanol A and roscovitine on the growth of c-Src-transformed cells (Fig. 3a). Roscovitine is a more selective CDK2 inhibitor that, like purvalanol A, is a 2,6,9-trisubstituted purine derivative (De Azevedo et al. 1997). Purvalanol A showed potent suppressive effects on the anchorage-independent growth of these cells in a dose-dependent manner, whereas roscovitine had a weaker suppressive effect at these concentrations (Fig. 3b). Treatment with purvalanol A induced a dramatic accumulation of cyclin E and p27kip1 even at a low dose (3 μm), and it induced an accumulation of CDK2 in a dose-dependent manner. These observations suggest that purvalanol A is able to strongly inhibit CDK2 activity as well as degradation of p27kip1, and that the subsequent accumulation of the inactivated CDK2 is well correlated with the suppressive effects of purvalanol A on the anchorage-independent growth of these cells (Fig. 3c). In contrast, roscovitine induced an accumulation of cyclin E in a dose-dependent manner by inhibiting CDK2 activity, but it failed to induce the accumulation of CDK2 probably because the expression of p27kip1 could not be restored by roscovitine. This may account for its weaker effects on the anchorage-independent cell growth (Fig. 3c). Furthermore, unlike purvalanol A, roscovitine treatment failed to induce the morphological reversion of c-Src-transformed cells even at the highest dose (Fig. 3d). As shown earlier, the suppression of CDK2 activity by the expression of its dominant negative form or p27kip1 was not sufficient to induce morphological reversion (Fig. 2f). These data suggest that purvalanol A may target additional molecules besides CDK2.

image

Figure 3.  Purvalanol A inhibits the activity of both CDKs and c-Src. (a) Structural formulas of purvalanol A, roscovitine and PP2. (b) The effect of purvalanol A and roscovitine on the anchorage-independent growth of Csk−/−/c-Src cells was analyzed by soft-agar colony formation assay (upper panels). Colonies were scored 9 days after plating, and the mean number of colonies per cm2 ± SD obtained from three independent experiments is shown (lower panel). (c) Csk−/−/c-Src cells were treated with purvalanol A (Pur A) and roscovitine (Ros) for 48 h, and the cell lysates were subjected to Western blotting for the indicated antigens. Representative blots from three independent experiments are shown. (d) The effects of purvalanol A and roscovitine on cell morphology. Csk−/−/c-Src cells were treated with the indicated agents for 48 h, and their morphology was observed by F-actin staining with Alexa Fluor 594 phalloidin. Nontreated Csk+/+ (MEF) and Csk−/− cells are also shown as normal controls. (e–g) Csk−/−/c-Src cells were treated with purvalanol A (Pur A), PP2 or roscovitine (Ros) for 48 h, and the cell lysates were subjected to Western blotting for the indicated antigens. Representative blots from three independent experiments are shown. (g) The inhibitory effect of Purvalanol A on c-Src activity was evaluated by in vitro kinase assay. c-Src was immunoprecipitated from Csk−/−/c-Src cells, followed by incubation with recombinant cortactin as a substrate. Phosphorylation was detected by Western blotting using anti-cortactin pY421. Representative blots from three independent assays are shown.

Download figure to PowerPoint

Because purvalanol A has a purine skeleton similar to PP2, the most frequently used Src-selective inhibitor (Fig. 3a), it seemed likely that it may directly target c-Src itself. Therefore, we compared the effects of purvalanol A on c-Src activity in Csk−/−/c-Src cells with those of PP2 and roscovitine. Treatment with purvalanol A or PP2 markedly reduced the tyrosine phosphorylation levels of total cellular proteins and of Src Tyr418, an autoactivation site of c-Src (Fig. 3e). Furthermore, the phosphorylation of physiological substrates of c-Src, such as annexin II, FAK and cortactin, was reduced by treatment with purvalanol A or PP2 (Fig. 3f). In addition, these two compounds suppressed the down-regulation of p27kip1, whose degradation is known to be a hallmark of Src activation (Fig. 3g) (Chu et al. 2007; Grimmler et al. 2007). The inhibitory effects of purvalanol A on c-Src function occurred at a concentration range similar to that of PP2, suggesting that purvalanol A inhibits c-Src activity in cells as effectively as PP2 (Fig. 3e,f). Moreover, in vitro kinase assay showed that purvalanol A could directly inhibit the c-Src activity toward cortactin, although its efficacy was somewhat lower than that of PP2 (Fig. 3h). By contrast, roscovitine exhibited only a moderate inhibitory effect on tyrosine phosphorylation events under these conditions (Fig. 3e,f) and did not induce the accumulation of p27kip1 (Fig. 3h). Overall, these results indicate that purvalanol A has (at least) a dual inhibitory effect on both CDK2 and c-Src even in cells, whereas PP2 appears to act more selectively on c-Src.

Purvalanol A has a potent suppressive effect on the growth of c-Src-transformed cells

We next examined the dose dependence of the effects of purvalanol A, roscovitine and PP2 on the growth of Csk−/−/c-Src cells (Fig. 4a). Purvalanol A and PP2 showed potent suppressive effects on the anchorage-independent growth of these cells at concentrations ranging from 3 to 10 μm, although higher concentrations of purvalanol A were required for full suppression. These results suggested that PP2 was more effective at the same concentration range. However, we observed that PP2 induced an apoptosis-like morphological change in normal MEFs, whereas neither purvalanol A nor roscovitine did (Fig. 4b). Indeed, PP2 more strongly induced caspase-3 cleavage than purvalanol A or roscovitine (Fig. 4c). As mentioned earlier, studies on cell morphology showed that c-Src-transformed cells treated with purvalanol A reverted to nearly normal cell morphology, whereas PP2-treated cells reverted only partially (Fig. 4b). These results suggest that purvalanol A is less toxic to normal cells than PP2. We then analyzed the effects of purvalanol A treatment on cell cycle progression by flow cytometry. PP2 and roscovitine induced a weak G1 arrest, and PP2 also induced an accumulation of the sub-G1 fraction even in MEFs, which is consistent with the induction of apoptosis by PP2. By contrast, purvalanol A specifically induced a strong G2-M arrest (Fig. 4d), raising the possibility that purvalanol A may also inhibit the CDK1-dependent G2 check-point machinery and/or other components that promote G2-M transition. These findings suggest that purvalanol A exerts its potent inhibitory effects on c-Src-induced transformation not only by inhibiting CDK2 but also by blocking cell cycle progression at the G2-M transition.

image

Figure 4.  Purvalanol A is less toxic than PP2. (a) The effect of purvalanol A, PP2 and roscovitine on the anchorage-independent growth of Csk−/−/c-Src cells was analyzed by soft-agar colony formation assay (upper panels). Colonies were scored 9 days after plating, and the mean number of colonies per cm2 ± SD obtained from three independent experiments is shown (lower panel). (b) The effect of purvalanol A, PP2 and roscovitine on cell morphology. Csk+/+ MEFs and Csk−/−/c-Src cells were treated with the indicated agents for 48 h, and their morphology was observed under a light microscope (upper gray pictures) and by F-actin staining with Alexa Fluor 594 phalloidin (lower color pictures). F-actin stainings of nontreated Csk+/+ MEFs and Csk−/− cells are also shown as normal controls. (c) Csk+/+ MEFs and Csk−/−/c-Src cells were treated with the indicated agents for 48 h, and total cell lysates were subjected to Western blotting with anti-cleaved caspase-3 antibody. Representative blots from three independent experiments are shown. (d) The effects of purvalanol A, PP2 and roscovitine on cell cycle progression were analyzed by flow cytometry. Csk+/+ (MEFs) and Csk−/−/c-Src cells were treated with the indicated agents at a final concentration of 10 μm. The cells were harvested 48 h after treatment, labeled with propidium iodide and analyzed for their DNA content by flow cytometry. The quantitative data are shown in the lower table. Representative data from three independent experiments are shown.

Download figure to PowerPoint

Purvalanol A has a stronger inhibitory effect than PP2 on the growth of human cancer cells

We next examined the effects of purvalanol A on human cancer cells. Because the causes of human cancers vary substantially depending on cell type, we investigated the levels of expression and activity of c-Src and CDK2 in various cancer cell lines (Fig. 5a). Most cancer cells had lower levels of p27kip1 expression than normal human embryonic fibroblasts (HEFs), suggesting that CDK2 is widely activated in these cancer cells. Of these cells, we selected the colon cancer cell lines HT29, HCT116 and SW480 to analyze the effects of purvalanol A because they have different c-Src activities and they show a clear inverse correlation between p27kip1 expression and cellular CDK2 activity as indicated by CDK2-dependent phosphorylation of RB (pT821). HT29 cells displayed the highest c-Src levels and activity resulting in high levels of tyrosine phosphorylation of cellular proteins (Fig. 5a). HCT116 cells displayed higher levels of CDK2 activity but lower levels of c-Src. SW480 cells displayed the lowest levels of c-Src and the highest levels of CDK2 activity. However, SW480 cells contained relatively high levels of tyrosine phosphorylation of cellular proteins, probably because of the activation of other members of the Src family kinases that were detected by the anti-pY418 antibody (Fig. 5a).

image

Figure 5.  Effects of purvalanol A, PP2 and roscovitine on human colon cancer cells. (a) Total cell lysates (TCL) from the indicated cell lines were subjected to Western blotting with the indicated antibodies (left panels). The lysates were immunoprecipitated with anti-Src, and the immune complexes were probed with the indicated antibodies (upper right panels). TCL from the indicated cell lines were subjected to Western blotting with anti-CDK2, anti-RB pT821 and anti-p27kip1 antibodies (lower right panels). Representative blots from three independent experiments are shown. (b) The effects of purvalanol A, PP2 and roscovitine on the morphology of cancer cells. Cancer cells were treated with the indicated agents for 48 h, and their morphology was observed under a light microscope. (c) Western blot analysis of c-Src and CDK2 activity in cancer cells treated with the indicated agents. The cell lysates prepared from the cultures shown in (b) were blotted with the indicated antibodies. Representative blots from four independent experiments are shown.

Download figure to PowerPoint

Suppression of c-Src activity by PP2 induced morphological changes from well-spread shapes to tightly compacted cell aggregates, which is largely attributed to the suppression of c-Src-mediated cell adhesion and cytoskeletal organization (Rengifo-Cam et al. 2004). Purvalanol A induced similar morphological changes in HT29 and HCT116 cells, but it induced different changes in SW480 cells (Fig. 5b). Roscovitine also induced morphological changes in SW480 cells to cell shapes similar to those treated with purvalanol A. As SW480 cells have much higher activity of CDK2 compared with other cells (Fig. 5c), these observations raise the possibility that the unique morphological changes in SW480 cells may be attributed to the inhibition of CDK2. However, additional studies are necessary to verify this possibility, because the functional link between CDK2 and cell morphology is thoroughly unknown. Even in these cancer cells, purvalanol A was able to inhibit both c-Src and CDK2 activities, whereas PP2 and roscovitine selectively inhibited c-Src (pY418 phosphorylation) and CDK2 (RB phosphorylation), respectively (Fig. 5c). A colony formation assay for these cancer cells showed that purvalanol A suppressed the anchorage-independent growth of HT29 and SW480 cells, both of which contain relatively high levels of Src family activity, more effectively than PP2 or roscovitine (Fig. 6). These results suggest that the multi-selective purvalanol A may be more effective for human cancers in which Src family kinases are up-regulated than Src-selective inhibitors such as PP2.

image

Figure 6.  Purvalanol A has a stronger inhibitory effect than PP2 on the anchorage-independent growth of HT29 and SW480 human colon cancer cells. Colonies (left panels) were scored 9 days after plating, and the mean number of colonies cm2 ± SD obtained from three independent experiments is shown (right panels).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Using Csk-deficient cells transformed by wild-type c-Src, we identified purvalanol A as an anticancer agent that effectively suppresses c-Src-promoted cell growth. Although purvalanol A was originally developed as a CDK inhibitor, we found that it could suppress c-Src function as effectively as the widely used SFK-selective inhibitor PP2. It was previously reported that purvalanol A was co-crystallized with the Src kinase domain and it had the potential to inhibit the Src activity in vitro (Breitenlechner et al. 2005). However, there were no reports on the effects of purvalanol A at a cellular level. In this study, we demonstrated for the first time that purvalanol A is applicable for inhibiting c-Src even in living cells. We also found that purvalanol A attenuated cell cycle progression by inhibiting CDK2, the activation of which is associated with c-Src-induced transformation. As a potential consequence of these multiple functions, purvalanol A reverted the transformed cell morphology to nearly normal with lower side effects than PP2, which had significant toxic effects as indicated by the enhanced apoptosis of normal cells. The results obtained in this model system suggest that purvalanol A is a suitable seed compound for the development of a more effective and less toxic anticancer drug for the treatment of human cancers involving c-Src up-regulation.

We found that purvalanol A successfully suppressed the growth of cells in which the Src family is up-regulated (HT29 and SW480 cells), whereas it was ineffective on cells with no significant activation of the Src family kinases (HCT116 cells). By contrast, PP2 was ineffective for all of these human cancer cells, although it was effective for c-Src-transformed MEFs. These results suggest that the multiple actions of purvalanol A, particularly on Src and CDKs, are beneficial for suppressing the growth of these human cancers in which multiple signaling pathways are deregulated. It is well established that many types of malignant cancer cells contain activated Src family kinases (Cartwright et al. 1990; Talamonti et al. 1993) and that other oncogenic molecules, such as EGFR, cooperate with c-Src to exert their oncogenic functions (Ishizawar & Parsons 2004). These results suggest that multi-selective purvalanol A, as a functional Src signaling inhibitor, has the potential to serve as a more effective anticancer agent than Src-selective inhibitors for a wide range of human cancers.

Purvalanol A is a 2,6,9-trisubstituted purine derivative, a number of which have been identified as CDK family kinase inhibitors. However, AP23464 and AP23451, which are both 2,6,9-trisubstituted purine derivatives as well, have been developed as Bcr-Abl and Src family kinase inhibitors (Dalgarno et al. 2006). Furthermore, purvalanol A could be co-crystallized with the Src kinase domain, providing direct evidence for the interaction between purvalanol A and Src (Breitenlechner et al. 2005). These results indicate that 2,6,9-trisubstituted purine derivatives can serve as inhibitors of Src-related tyrosine kinases. However, the inhibitory effects on Src and CDK2 substantially differ among these compounds: AP23464 and AP23451 have a higher affinity for Src than CDK2 (AP23464; Src IC50 = 0.00045 μm, CDK2 IC50 = 20.9 μm, AP23451; Src IC50 = 0.067 μm, CDK2 IC50 = 2.01 μm), whereas purvalanol A has a lower affinity for Src (Src IC50 = 0.24 μm, CDK2 IC50 = 0.070 μm). Nevertheless, we found that purvalanol A can effectively inhibit c-Src function in cells with only moderate cytotoxicity, suggesting that the lower affinity for Src might be rather favorable for reducing toxic effects. The development of purvalanol A derivatives with a different spectrum of affinity for Src and CDK2 might further improve the anticancer effects of this agent.

Multi-drug resistance (MDR) is a major problem in the successful treatment of cancer. Human ABCG2, a member of the ATP-binding cassette transporter superfamily, plays a key role in MDR by protecting cancer stem cells from anticancer drugs (Peng et al. 2009). A recent report showed that purvalanol A interacts with the active site of ABCG2 and inhibits ABCG2 function (An et al. 2009), suggesting the potential use of purvalanol A for the treatments of drug resistance. Thus, it would be interesting to examine the effects of purvalanol A on MDR of cancer cells associated with c-Src up-regulation. In this study, we found that purvalanol A can induce a strong G2-M arrest, but it remains unclear whether it is a consequence of inhibition of the CDK1-dependent G2 check-point machinery or an indirect result of perturbations of other pathways. It is known that G2-M arrest allows DNA replication, which can be targeted by DNA-damaging agents. Although the underlying mechanisms for the G2-M arrest induced by purvalanol A need to be clarified, purvalanol A may also serve as a potential seed compound for combination therapy with DNA-damaging agents. Further analysis of in vivo effects of purvalanol A and/or its derivatives may open a new avenue for the treatment of a subset of human cancers.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Chemicals

SCADS inhibitor kits were kindly supplied by the Screening Committee of Anticancer Drugs (SCADS) supported by a Grant-in-Aid for Scientific Research on the Priority Area ‘Cancer’ from the Ministry of Education, Culture, Sports, Science and Technology in Japan. Poly-HEMA was obtained from Sigma. Purvalanol A, PP2 and roscovitine were purchased from Calbiochem.

Antibodies

Anti-Csk (C-20), anti-focal adhesion kinase (FAK) (C-20), anti-cdc2 (CDK1) (17), anti-cyclin E (M-20) and anti-actin (C-11) were purchased from Santa Cruz Biotechnology. Anti-Src pY418, anti-Src pY529, anti-FAK pY576, anti-cortactin pY421 and anti-RB pT821 were purchased from BioSource. Anti-cleaved caspase-3 and anti-PP1α pT320 were purchased from Cell Signaling. Anti-Src (Ab-1) was purchased from Oncogene. Anti-phosphotyrosine (4G10) and anti-cortactin (4F11) were purchased from Upstate Biotechnology. Anti-CDK2 and anti-p27kip1 were purchased from BD Transduction Laboratories.

Cells and cell culture

Csk−/− and Csk+/+ MEFs were a kind gift from Dr Akira Imamoto (Thomas et al. 1995). HT29, HCT116, SW480, Caco-2, MCF7, T47D, MDA-MB-231, HeLa, U2OS, HT1080 and HEK293 cells were obtained from the American Type Culture Collection (ATCC). HEF cells were kindly provided by Dr Masuo Yutsudou (Ishiwatari et al. 1994). HaCaT cells were kindly provided by Dr N. E. Fusenig (German Cancer Research Center, Heidelberg, Germany). All cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS).

Retroviral-mediated gene transfer

All gene transfer experiments were carried out with the pCX4 series of retroviral vectors (Akagi et al. 2003). Retroviral vectors carrying wild-type chicken c-Src, its active form (c-SrcYF) and v-Src were kindly provided by Dr Tsuyoshi Akagi (Akagi et al. 2003). Wild-type mouse p27kip1, a CDK1 kinase-deficient mutant (T161A) and wild-type rat Csk were subcloned into CX4bleo. The CDK2 kinase-deficient mutant (T160A) was subcloned into CX4bsr. The production of and infection with retroviral vectors were carried out as described previously (Akagi et al. 2003).

Measurement of anchorage-independent growth

Poly-HEMA-coated 96-well plates were prepared as described (Fukazawa et al. 1995). Briefly, 50 μL of poly-HEMA solution (5 mg/mL in 95% ethanol) was pipetted into the wells of 96-well plates and dried for 2 days with lids in place. Cells were inoculated in a volume of 100 μL at a density of 2 × 104 cells per well. Each inhibitor dissolved in 50% MeOH was added at a final concentration of 3 μm, and cells were cultured for 2 days. For the quantitation of cell growth, 10 μL of WST-1 reagent (Roche) was added and the absorbance at 450 nm was measured using a microplate reader.

Soft-agar colony formation assay

Single-cell suspensions of 1 × 104 cells were plated onto six-well culture dishes in 1.5 mL of DMEM containing 10% FBS and 0.36% agar on a layer of 2 mL of the medium containing 0.7% agar. Nine days after plating, colonies were stained with 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT), and photographs of the stained colonies were taken.

Cell cycle analysis and apoptosis assays

Cell lines were treated with purvalanol A, PP2, roscovitine or DMSO as a control. After 24 h, cells were collected, fixed in 70% ethanol and stained with propidium iodide. Acquisition was carried out on a FACScan flow cytometer (Becton Dickinson) to identify apoptotic cells with <2N DNA content. In addition, apoptosis was detected by Western blotting with anti-cleaved-caspase 3.

Immunoprecipitation and Western blotting

Cells were lysed in RIPA buffer (50 mm Tris/HCl, pH 8.0; 2 mm EDTA; 150 mm NaCl; 0.1% SDS; 1% sodium deoxycholate; 1% Triton X-100; 10 μg/mL aprotinin; 10 μg/mL leupeptin; 1 mm sodium orthovanadate; 1 mm PMSF). Immunoprecipitation and Western blotting were carried out as described previously (Oneyama et al. 2008a). For in vitro kinase assay, immunoprecipitated c-Src was incubated in a reaction mixture consisting of 50 mm Tris/HCl, pH7.4, 10 mm MgCl2, 50 μm ATP, 0.5 mm Na3VO4 and 2 μg GST-cortactin. After incubation for 30 min at 30 °C, cortactin phosphorylation was analyzed by Western blotting with anti-pY421 antibody.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Drs A. Imamoto, T. Akagi and H. Fukazawa for their generous gifts of reagents and technical advices. This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan and the Uehara Memorial Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References