Rho- and Rac-GTPases are known to be implicated in cancer cell aggressivity. We established previously that RhoA is over-expressed and spontaneously activated in MDA-MB-231, and found that blockage of RhoA expression inhibited cancer cell invasion and tumour growth by more than 80% in a mouse model . The role of Rac1 in cancer progression has also been clearly shown . In contrast, the role of Rac3 in the aggressiveness of breast cancer cells is not well established: in some studies Rac3 was found to be involved mainly in cell proliferation , whereas in others Rac3 has been implicated in cell invasion but not migration . Despite 92% amino acid homology between Rac1 and Rac3 proteins, they differ in their C-terminal membrane targeting regions, which are known effector-binding regions, suggesting that differential effector binding could occur for Rac1 versus Rac3 in different cell types . Moreover, in neuronal cells , Rac1 and Rac3 are known to have opposing functions in cell morphology and adhesion to the ECM.
Our aim in this study was to examine the effects of Rac3 in breast cancer cell aggressivity in both the invasive MDA-MB-231, which have a spontaneous activation of RhoA, and in the non-invasive MCF-7, which do not. Furthermore, we determined that even after stimulation, the activated Rho found in MCF-7 represented 48% of that seen in MDA-MB-231 (data not shown). The criteria we defined as factors of aggressiveness were proliferation, migration, invasion, adhesion on extracellular matrix in flow conditions, and vasculogenic mimicry. To examine the role of Rac3, we analyzed the consequences of Rac3 depletion by using siRNA technology.
Firstly, to validate our work, we studied the efficacy and specificity of Rac3 siRNA. Under the conditions used, the siRNA treatment is indeed efficient. This effect is specific, because anti-Rac3 siRNA did not inhibit the expression of Rac1, RhoA or Cdc42 (Figure 1C). Moreover, the possible interrelations between the different Rho and Rac proteins led us to examine the effects of Rac3 inhibition on activation of other Rho proteins that could potentially influence the aggressiveness cancer cells. In both cell lines, activated Cdc42 was not modified by Rac3 depletion, whereas RhoA activation was slightly increased. This increase did not seem to be enough to counterbalance the Rac3-depletion-induced cell inhibition effects. Therefore, the effects of Rac3 depletion on MDA-MB-231 and MCF-7 cells cannot be attributed to a modification of Rho or Cdc42 expression or activity.
Secondly, we found that Rac3 depletion in MDA-MB-231 cells inhibited cell spreading and lamellipodia extension (Figure 2A), leading to cell rounding associated with a disorganisation of the actin cytoskeleton. The rounding of MDA-MB-231 is consistent with a loss of lamellipodia, which are involved in cell adhesion to the ECM by connecting the ECM to actin filaments. Our results are in agreement with the results of Joyce and Cox , who reported that both Rac1 and Rac3 strongly stimulated lamellipodia formation in fibroblasts. However, the effect is much clearer in MDA-MB-231 than in MCF-7 cells. As the actin network is more organized in MDA-MB-231 than in MCF-7, this can explain the differences of impact of Rac3 inhibition on cell morphology because Rac3 regulates actin cytoskeleton organization (Figure 5F). At any rate, this disappearance of lamellipodia and cell rounding in MDA-MB-231 when Rac3 is depleted provides an explanation for the important decrease in their adhesion to collagen type I under flow conditions. This cell adhesion has been implicated in the development and progression of the majority of cancers . Moreover, the observed decrease in adhesion might be expected to prevent metastatic dissemination to bone, since collagen type I is the most abundant protein in the bone matrix, and cell adhesion on collagen is considered to be a marker of tumour invasiveness in bone .
In addition, the depletion of Rac3 induced an important decrease in MDA-MB-231 migration and invasion through Matrigel. The cell rounding we observed upon Rac3 siRNA treatment may be partially responsible for the decreased cell invasion (Figure 2D) and speed of wound repair in the scratch test (Figure 2B), since it has been reported that small pseudopodia are required for cell motility and to bear focal complexes [30, 31]. The decrease in MMP-9 secretion observed in Rac3-depleted cells can also contribute to the reduction of cell invasion through Matrigel.
Moreover, we tested the effect of Rac3 depletion on capillary-like structure formation (VM) by cancer cells plated on Matrigel rich in growth factors. This cell property is currently considered as being a sign of great aggressiveness, possibly by contributing to the blood supply in the tumours . This is confirmed by our observation showing that, whereas the aggressive cells (MDA-MB-231) were able to form capillary-like tubes in Matrigel, the poorly aggressive cells, MCF-7, did not. To form channels, cancer cells must migrate and modify their morphology to become elongated. The inhibition of Rac3 expression in MDA-MB-231 cells blocked the VM; this provides yet another argument for thinking that Rac3 plays a role in tumour aggressiveness, even in cells where the small GTPase RhoA is overexpressed and spontaneously activated. The decrease in MMP-9 secretion following treatment with siRNA anti-Rac3 can also be involved in its inhibitory effect on VM in MDA-MB-231 cells, as it was described that both MMP-2 and MMP-9 play an important role in VM in cancers .
We also analyzed the effect of Rac3 inhibition on two other functions important in cell aggressivity: cell proliferation and resistance to apoptosis. siRNA anti-Rac3 induced a slight decrease in cell proliferation in MDA-MB-231 cells between 72 and 96 hours after treatment. To understand this decrease of cell number we first analyzed the effects of Rac3 inhibition on cell apoptosis. We did not observe any modification of the apoptotic index under basal conditions for either MDA-MB-231 or MCF-7 cell lines. However, in MDA-MB-231 the Rac3 inhibition increased sensitivity to TNFα-induced apoptosis. This effect on apoptosis does not occur in MCF-7. Moreover, when we monitored the progression of treated cells through the cell cycle, we found that MDA-MB-231cells were blocked in S phase but no modification was observed for MCF-7. This can explain the more profound effect of Rac3 inhibition on proliferation in MDA-MB-231compared to MCF-7.
All these results intrigued us because, although MCF-7 cells express Rac3, they are poorly aggressive and non-invasive compared to MDA-MB-231. Consequently, to understand why Rac3 plays a positive role in MDA-MB-231 aggressiveness but does not do so in MCF-7 cells, we analyzed the effects of Rac3 depletion on cell signaling. We found that, in both cell types, Rac3 inhibition led to a decrease of ERK and phospho-ERK protein levels. The decrease of the level of ERK protein in the cell could be due to an inhibition of ERK expression or ERK stabilisation. We therefore analyzed the consequences of ERK inactivation for MDA-MB-231 and MCF-7 cells.
Knowing that ERK is critical for the activation of NF-κB, we postulated the intervention of a cell signal cascade, Rac3/ERK/NF-κB, analogous to that previously proposed for Rac1 . Indeed, activated ERK-1 and ERK-2 are known to act upstream of IκB kinase activation, initiating the degradation of IκB through the ubiquitin/proteasome pathway, leading to NF-κB activation .
In MDA-MB-231 cells, we found that Rac3 does indeed induce cell signaling leading to NF-κB activation by an increase in IκBα phosphorylation and degradation. This NF-κB activation is important both for protection against apoptosis and for MMP-9 secretion , which is implicated in cancer cell invasion  and vasculogenic mimicry, two major criteria of cancer cell aggressiveness; and activated NF- κB represses apoptosis by inducing the expression of anti-apoptotic genes including cIAPs, FLIP, TRAF-1, TRAF-2, Bcl-2, and Bcl-xL . In contrast, in MCF-7, Rac3 does not increase cancer aggressivity despite the fact that Rac3 activates ERK in these cells. This could be explained by the fact that NF-κB p50 and p65 proteins are both inactive in MCF-7 (Figure 5B). Thus although Rac3 is expressed and does activate ERK in MCF-7, this does not result in increased NF-κB activation, because of the very small amounts of the two active subunits of NF-κB expressed in these cells. This also helps explain why the MCF-7 expressing Rac3 are not very aggressive, and show relatively weak invasion, migration and VM.
Finally, we analyzed the importance of ERK/NF-κB signaling in the Rac3-depletion-induced modification of cytokine secretion in MDA-MB-231 cells. We found that IL-6, IL-8, GRO, GRO-α and IL-10 were all down-regulated in Rac3-depleted cells. These cytokines are known to be secreted by immune cells; however, it has been reported that they can also be secreted by cancer cells, and their secretion is usually associated with a poor prognosis [39–41].
Interestingly, IL-6, IL-8 and GRO, are involved in tumour aggressivity by favoring a higher invasiveness potential of cancer cells and a proangiogenic activity [42, 43]. The decrease of the secretion of these three cytokines by Rac3-depleted MDA-MB-231 could be explained by the role of Rac3 in ERK-2/NF-κB signaling. In fact, the levels of IL-6, IL-8 and GROα are under the control of NF-κB [37, 44, 45]. Moreover, a recent work has underlined the essential role of IL-8 signaling for the acquisition and/or maintenance of the mesenchymal and invasive features of overexpressing tumour cells and shows that IL-8 secreted by tumour cells undergoing epithelial-mesenchymal transition can potentiate tumour progression . The decrease in IL-10 secretion induced by anti-Rac3 siRNA can also be explained by the action of Rac3 on ERK-2/NF-κB signaling, but its role in cancer aggressivity has been reported to be due mainly to an immunosuppression and also to its angiogenic activity .
We also studied the effect of Rac3 knockdown in breast epithelial cells and showed in these cells that Rac3 inhibition had not effect on the cell proliferation, morphology, actin organization and cell cycle evolution. These results were probably due to the low level of Rac3 protein expressed in these cells. Mira et al. has previously demonstrated the crucial role of Rac3 in breast cell proliferation by introduction of hyperactive Rac3 into normal breast epithelial cell line and breast cancer cell lines. They showed that hyperactive Rac3 contributes to the aggressive growth of epithelial cancer cells . Therefore, it appears that the non-aggressive behaviour of normal epithelial cells can be related to absence of hyperactivity or low expression of Rac3. Inversely, aberrant Rac3 expression and activation, along with the effectiveness of its downstream pathways in cancer cells, contribute to cancer aggressiveness.