The goal of this study was to elucidate whether four different anti-cancer drugs (vincristine, paclitaxel, cisplatin and etoposide) could prompt invasive ability of tumor cells. We studied cellular invasive ability and intracellular signaling using these anti-cancer drugs in MKN45 cells, and report four main findings here. First, only vincristine, but not the other anti-cancer drugs, enhanced cellular invasive ability in MKN45 cells (Figure 1). Second, it induced the formation of membrane blebs and amoeboid-like motility (Figure 2). Third, it induced GEF-H1/RhoA/ROCK/MLC signaling (Figures 3-6). Fourth, GEF-H1 and ROCK mediated vincristine-induced cellular invasive ability (Figures 4 and 5). These results indicated that vincristine enhanced amoeboid-like motility via GEF-H1/RhoA/ROCK/MLC signaling, thereby promoting invasive ability in MKN45 cells.
Vincristine is widely used in the treatment of leukemia, lymphoma, myeloma, glioma, solid tumors of childhood often combined with other drugs. Lung, breast, and cervical cancer are also treated with vincristine. Vincristine is used at up to 2 mg/body in one administration. When 2 mg vincristine is injected i.v. to a patient, its blood concentration is estimated to be within 10–400 nM for a few hours . Furthermore, it was reported that vincristine accumulated to some tissues such as spleen, thyroid, large and small intestine, and the local concentrations in these organs could reach to 6–70 times higher than that in blood . These data suggest that the maximum concentration of vincristine is in the range of 0.06-28 μM in some organs of a patient treated with 2 mg vincristine. Therefore, it is conceivable that 15 μM vincristine, the maximum dose we used in this study, is physiologically-achievable.
We found that vincristine enhanced cellular invasive ability of MKN45 cells in a concentration-dependent manner. Previously, Zhao et al. have shown that the IC50 of vincristine (72 h treatment) in MKN45 cells is about 7 μM . Because we observed the invasion-stimulating effect by vincristine at as low as 1 μM after 24 h treatment (Figure 1A), which is lower than IC50, we assume that this is not a non-specific effect of vincristine. However, this finding contradicts the data that microtubule depolymerizers inhibit cellular invasion observed in other studies [7–9]. The concentrations of vincristine used in our study were higher than the concentrations in the other studies [7–9]. Therefore, we consider that one potential reason for this discrepancy in the effect of vincristine on cellular invasiveness could be due to the concentration of vincristine used, and only high concentration vincristine would be able to induce GEF-H1/RhoA/ROCK/MLC signaling, leading to high cellular invasiveness. In addition, since the cell types and the assay systems to measure cellular invasion used in these studies are different from those used in this study, we cannot exclude the possibility that these factors also contributed to the outcome. To test the differences in the effect of vincristine on cellular invasion among cell types, we examined it using human lung adenocarcinoma A549 cells and human cervical adenocarcinoma HeLa cells. Whereas vincristine stimulated cellular invasive ability in A549 cells similar to MKN45 cells, it was not increased in HeLa cells (data not shown). These results suggest that the enhancement of cellular invasive ability by vincristine is at least in part cell type-specific.
In the present study, we observed no significant difference in the cell viability in 0.1 and 15 μM vincristine-treated cells (Figure 1B). This result is supported by the data reported by Warlters et al. showing that 0.1 and 11 μM vincristine exhibited the same level of cell toxicity in MKN45 cells . On the other hand, we observed that the effects on invasive ability were significantly different between 0.1 and 15 μM vincristine. These results suggested that vincristine enhanced cellular invasive ability in a concentration-dependent manner without affecting the viability in MKN45 cells.
In contrast to vincristine, paclitaxel had a strong inhibitory effect on cellular invasive ability (Figure 1A). Paclitaxel has been shown to inhibit RhoA activity . Because RhoA activity is required not only for amoeboid-like motility but also for general cellular motility [22, 30, 45, 46], it is possible that paclitaxel attenuated cellular invasion by inhibiting RhoA activity. Although both vincristine and paclitaxel act on microtubules as anti-cancer drugs, our results indicate that they influence cellular motility differently depending on the effect on RhoA activity. In addition, microtubule depolymerization is shown to activate GEF-H1 [37–39]. Therefore, paclitaxel may inhibit GEF-H1 activity through the inhibition of microtubule depolymerization, thereby inhibiting the signaling pathway leading to cellular motility.
MLC phosphorylation induces actomyosin contraction, which is required for the formation of membrane blebs [17, 25, 27, 35]. As shown in Figure 2A, 15 μM vincristine induced the formation of membrane blebs, which were not observed in control cells or in the cells treated with 0.1 μM vincristine. Consistent with this result, 15 μM vincristine induced MLC phosphorylation whereas 0.1 μM vincristine did not (Figure 3B). Therefore, we assume that the difference in the effects of vincristine on the formation of membrane blebs is attributable to MLC phosphorylation induced by GEF-H1/RhoA/ROCK signaling. As mentioned above, microtubule depolymerization activates GEF-H1 [37–39], promoting RhoA/ROCK/MLC signaling. It is thus possible that severe depolymerization of microtubules by 15 μM vincristine, but not by 0.1 μM vincristine, stimulates GEF-H1/RhoA/ROCK/MLC signaling, resulting in the formation of membrane blebs (Figure 2A).
The functions of microtubules in amoeboid-like motility are not well understood . In this study, we showed that vincristine enhanced amoeboid-like motility. Because vincristine is a microtubule depolymerizer, our results may provide evidence that amoeboid-like motility does not require structural functions of microtubules. This concept will be clarified by performing further studies such as the live-cell fluorescence imaging of microtubules in vincristine-induced amoeboid-like moving cells.
To the best of our knowledge, this is the first report to provide evidence that GEF-H1 can regulate amoeboid-like motility. Previous studies have reported that GEF-H1 regulates the interaction of actin and microtubule at the leading edge and focal adhesion turnover [47, 48] that are involved in mesenchymal motility [49, 50], suggesting the involvement of GEF-H1 in this mode of cellular motility. Considering these findings together with the role of GEF-H1 in amoeboid-like motility that we presented in this study, it seems likely that GEF-H1 regulates not only mesenchymal motility but also amoeboid-like motility depending on the situation. In recent studies, tumor necrosis factor-β and TGF-β have been reported to promote cellular invasion and metastasis [51–54]. These cytokines have been reported to activate or up-regulate GEF-H1 [55–57]. Additionally, radiation and doxorubicin have been shown to induce metastasis and invasion of tumor cells via TGF-β, . Therefore, pathophysiological conditions that increase these cytokines such as inflammation might stimulate cellular invasion via the activation and/or up-regulation of GEF-H1.
As described above, vincristine has been reported to accumulate in some organs at higher concentration than in blood after administration . Given the fact that vincristine is widely used in cancer treatment, we surmise that vincristine treatment to cancer patients could adversely induce the invasion of tumor cells in some organs when its local concentration increases in the clinical setting. If this is the case, it would be beneficial to inhibit GEF-H1/RhoA/ROCK/MLC signaling pathway when treated with vincristine to prevent tumor metastasis.