Rac1 protects epithelial cells against anoikis.

Rho family members play a critical role in malignant transformation. Anchorage-independent growth and the ability to avoid apoptosis caused by loss of anchorage (anoikis) are important features of transformed cells. Here we show that constitutive activation of Rac1 inhibits anoikis in Madin-Darby canine kidney (MDCK) epithelial cells. Constitutively active Rac1-V12 decreases DNA fragmentation and caspase activity by 50% in MDCK cells kept in suspension. In addition, expression of Rac1-V12 in MDCK cells in suspension conditions causes an increase in the number of surviving cells. We also investigated the signaling pathways that are activated by Rac1 to stimulate cell survival. We show that expression of Rac1-V12 in MDCK cells in suspension stimulates a number of signaling cascades that have been implicated in the control of cell survival, including the p42/44 ERK, p38, protein kinase B, and nuclear factor kappaB pathways. Using specific chemical or protein inhibitors of these respective pathways, we show that Rac1-mediated cell survival strongly depends on phosphatidylinositol 3-kinase activity and that activation of ERK, p38, and NF-kappaB are largely dispensable for Rac1 survival signaling. In conclusion, these studies demonstrate that Rac1 can suppress apoptosis in epithelial cells in anchorage-independent conditions and suggest a potential role for Rac1-mediated survival signaling in cell transformation.


Rho family members play a critical role in malignant transformation. Anchorage-independent growth and the ability to avoid apoptosis caused by loss of anchorage (anoikis) are important features of transformed cells. Here we show that constitutive activation of Rac1 inhibits anoikis in Madin-Darby canine kidney (MDCK) epithelial cells. Constitutively active Rac1-V12 decreases DNA fragmentation and caspase activity by 50% in MDCK cells kept in suspension. In addition, expression of Rac1-V12 in MDCK cells in suspension conditions causes an increase in the number of surviving cells.
We also investigated the signaling pathways that are activated by Rac1 to stimulate cell survival. We show that expression of Rac1-V12 in MDCK cells in suspension stimulates a number of signaling cascades that have been implicated in the control of cell survival, including the p42/44 ERK, p38, protein kinase B, and nuclear factor B pathways. Using specific chemical or protein inhibitors of these respective pathways, we show that Rac1-mediated cell survival strongly depends on phosphatidylinositol 3-kinase activity and that activation of ERK, p38, and NF-B are largely dispensable for Rac1 survival signaling. In conclusion, these studies demonstrate that Rac1 can suppress apoptosis in epithelial cells in anchorage-independent conditions and suggest a potential role for Rac1-mediated survival signaling in cell transformation.
Members of the Rho family of small GTP-binding proteins mediate a large number of biological processes that are stimulated by growth factors, cytokines, and adhesion molecules. The functions that are regulated by these GTPases include the organization of the actin cytoskeleton and the regulation of lipid metabolism, gene transcription, and vesicle trafficking (1).
Rho family GTPases also play an important role in cell transformation (2,3). For example, we showed that expression of a constitutively active (hydrolysis-defective) Rac1 mutant in Rat1 fibroblasts induces serum-and anchorage-independent growth and is tumorigenic in nude mice (4). Conversely, expression of a dominant negative Rac1 mutant that is impaired in GTP binding inhibited the serum-and anchorage-independent growth of Ras-transformed Rat1 cells (5). Similar findings have been made by a number of other laboratories (6,7).
Ras-induced activation of Rac1 is thought to occur via phos-phatidylinositol 3-kinase (PI3K) 1 (8). The PI3K-Rac1 branch of Ras signaling appears to function largely independently from the Raf/MEK/ERK and the RalGDS/Ral pathways (8,9). In addition, constitutively active versions of Rac1 and Raf strongly synergize with each other in cell transformation (4,6). Interestingly, Rac1 can also cooperate with Raf to activate ERK (10,11), suggesting one possible mechanism that could contribute to Rac1-induced cell transformation. Expression of constitutively active Rac1 strongly activates the transcription factor NF-B (12,13), suggesting that NF-B may play a role in Rac1-stimulated transformation. In line with this idea is the observation that NF-B is required for transformation by oncogenic Ras (14) and that Rac is essential for Ras-induced NF-B activation (12). Rac1 also activates and/or modulates a number of other signaling events that have been implicated in growth control. These include the c-Jun N-terminal kinase (JNK) and p38 pathways and the transcription factor serum response factor (SRF) (1). The precise role of any of these signaling elements in Rac1-stimulated cell proliferation remains to be elucidated (1,3).
Many mammalian cell types, including epithelial and endothelial cells, are dependent on integrin-mediated attachment to extracellular matrix for cell survival and proliferation. The acquisition of anchorage-independent growth is considered to be a hallmark of cell transformation (15). In normal epithelial and endothelial cells, prevention of cell attachment to extracellular matrix results in the rapid activation of programmed cell death (apoptosis) (16,17). This form of apoptosis is also termed anoikis (16). Thus, anoikis may represent an important defense mechanism of the organism against aberrant cell proliferation and migration (18,19).
Rac1 has been implicated in the regulation of cell survival, but its precise role in survival signaling remains highly controversial. For instance, studies in hematopoietic cells have shown that expression of constitutively active Rac in some conditions can induce apoptosis (20 -22) but in other conditions enhances cell survival (23). A similar situation holds for fibroblasts (24 -28).
In this paper, we investigate the role of Rac1 in epithelial cell survival in suspension. For this study we use the Madin-Darby canine kidney (MDCK) epithelial cell line, a well established system for the study of anoikis (16, 29 -31). Our results show that expression of a constitutively active mutant of Rac1 protects MDCK cells against anoikis and that PI3K plays an important role in this Rac1-mediated survival function.

EXPERIMENTAL PROCEDURES
Cell Culture-The cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal calf serum and penicillin/streptomycin, supplemented with 20 ng/ml doxycycline (DC) (Sigma) at 37°C in a humidified atmosphere containing 5% CO 2 . To achieve transgene expression, MDCK cells were cultured in media containing 0 -200 pg/ml DC 24 h prior to the experiments.
Cell Line Establishment-The establishment and characterization of MDCK cell lines expressing constitutively active Rac1-V12 and dominant negative Rac1-N17 in a tetracycline-repressible fashion have been described previously (32). To establish MDCK cell lines that inducibly express an IB␣ super-repressor mutant (IB␣M), we generated a tetracycline-repressible expression construct containing the IB␣-S32A,S36A double mutant cDNA (33)(34)(35) by cloning it in the pTRE2 vector (CLONTECH). The pTRE2-IB␣M construct was co-transfected with the zeocin resistance marker, pCMV-ZEO (Invitrogen), into MDCK-Rac1-V12 and the T23 tet-off parental cells. MDCK-Rac1-V12 cells transfected which failed to integrate the IB␣M construct were used as a selection marker control. Clonal populations were selected in 500 g/ml Zeocin (Invitrogen) and analyzed by Western blotting with a rabbit polyclonal anti-IB␣ antibody (UBI).
DNA Fragmentation Assay-Cells were induced for 24 h as a monolayer and subsequently trypsinized and cultured in suspension on ultra low cluster plates (Costar) at a density of 1.5 ϫ 10 5 cells/ml for 0 -18 h. Cells were then pelleted and lysed for 30 min. The level of DNA fragmentation was quantified using the Cell Death ELISA kit (quantifying histone-associated DNA fragments) using the protocol suggested by the manufacturer (Roche). Lysates assayed were equivalent to 7.5 ϫ 10 4 cells.
Caspase Assays-Cells were induced and cultured as for the DNA fragmentation assay, except that extracts were collected from 6 ϫ 10 5 cells and harvested at 0, 4 and 6 h in suspension. Extracts were assayed for caspase activity using the ApoAlert Caspase-8 Fluorescent Assay Kit using the protocol supplied by the manufacturer (CLONTECH). This kit measures the fluorescence of the IETD-7-amino-4-trifluoromethylcoumarin (AFC) cleavage product and in addition to caspase 8 also reflects the activities of caspases 9 and 10 (36).
Luciferase Assays-NF-B promoter activity was assayed using two types of luciferase constructs driven either by a wild-type or mutant E-selectin-promoter (37) or a promoter containing five NF-B binding elements (pNF-B luc, Stratagene). As an internal control, reporter constructs were co-transfected with a plasmid expressing Renilla luciferase downstream of the cytomegalovirus promoter (pRL-CMV, Promega). MDCK cells were plated on 35-mm dishes and were 50 -60% confluent on the day of transfection. Transfections with the E-selectinpromoter constructs were carried out using the GenePorter transfection reagent (Gene Therapy Systems). Reporter and control plasmids were mixed with GenePorter reagent at a ratio of 1:3 and added to the cells in the absence of serum for 4 h. Transfections with pNF-B luc were carried out with Effectine reagent (Qiagen) using a protocol that is optimized for MDCK cells. The cells were incubated for 4 -5 h in the presence of serum using a ratio of 1 g of total DNA (950 ng of pNF-B luc and 50 ng of pRL-CMV) to 5 l of Effectine reagent, yielding an estimated transfection efficiency of 20 -25% as determined using a green fluorescent protein control plasmid.
For both protocols, transfected cells were divided into two pools (kept in either 0 or 20 ng/ml DC) that were incubated in monolayer for 24 h and subsequently split and cultured in monolayer or suspension at a density of 1.5 ϫ 10 5 /ml for an additional 18 h. Subsequently, cells were harvested and assayed for NF-B activity by measuring the level of reporter and control luciferase activity using the dual luciferase reporter assay (Promega).
In Vitro Kinase Assays-In vitro kinase assay kits (New England Biolabs) were used to assess the levels of p38 and PKB/Akt activity in MDCK suspension cells. Extracts were made from1.5 ϫ 10 6 cells that were induced and cultured in suspension for 2 h as described above, using the lysis buffer as supplied by the manufacturer. Immunoprecipitation of either activated p38 or total PKB was carried out for at least 3 h at 4°C. Immunoprecipitates were washed twice with lysis buffer and twice with kinase buffer, and kinase reactions were carried out at 30°C for 30 min in the presence of 200 M ATP and 2 g of ATF-2 or 1 g of GSK-3 to assay for p38 and PKB activity, respectively. The reactions were then terminated by adding 2ϫ SDS sample buffer and loaded onto a SDS-polyacrylamide gel. Proteins were transferred as described above and blotted with either anti-phospho ATF-2 or antiphospho GSK-3 rabbit polyclonal antibodies (New England Biolabs).

Constitutively Active Rac1 Suppresses MDCK Cell
Anoikis-We have studied the role of Rac1 in cell survival in anchorage-independent conditions using stable cell lines that express constitutively active Rac1-V12 or dominant negative Rac1-N17 under the control of a tet-repressible promoter in MDCK cells (38,39). These cell lines have the dual advantage of displaying relatively homogenous levels of expression after induction of the transgene, in addition to allowing regulation of transgene expression levels by titration of the concentration of DC in the medium (32). Upon complete removal of DC from culture medium for 42-48 h, these cells show expression levels of Rac1-V12, which are similar to that of endogenous Rac1 (Fig.  1A). Culture of these MDCK cells in suspension does not significantly alter transgene expression levels in the absence of DC (data not shown).
MDCK cells readily undergo apoptosis when cultured in suspension conditions, as indicated by a number of criteria, including increased DNA fragmentation, annexin V staining, and caspase activation (16,30,31). In the absence of Rac1-V12 expression, MDCK cells cultured in suspension for 18 h show significant DNA fragmentation (an ϳ40-fold increase over monolayer conditions) (Fig. 1B). Cells expressing Rac1-V12 exhibit decreased levels of DNA fragmentation, and this decrease correlates with Rac1-V12 expression levels ( Fig. 1B). Complete removal of DC lowers DNA fragmentation to about 50% of that obtained in cells cultured in the presence of 20 ng/ml DC. This Rac1-V12-induced inhibition of DNA fragmentation is detectable as early as after 8 h in suspension conditions (Fig. 1C).
We also used caspase activation as a read-out for apoptosis. Prevention of attachment of MDCK cells rapidly stimulates caspase activity, measured by the fluorescence of caspase 8 substrate (see "Experimental Procedures") ( Fig. 1D). Expression of Rac1-V12 inhibits suspension-induced caspase activation by ϳ50% (Fig. 1D). This level of inhibition is comparable with the reduction in the extent of DNA fragmentation caused by Rac1-V12 (Fig. 1B). The inhibitory effect of Rac1-V12 on anoikis was confirmed by TdT-mediated dUTP nick end labeling stain (data not shown). We also verified that DC itself had no effect on apoptosis. Removal of DC from parental MDCK cells has no effect on either the extent of suspension cultureinduced DNA fragmentation or caspase activation (Fig. 1, C  and D).
To confirm that Rac1 enhances the survival potential of MDCK cells in suspension, we performed a viability assay. Cells expressing Rac1-V12 and controls were kept in suspension for 18 h and then replated onto normal tissue culture dishes. Cell viability was quantified by the sulforhodamine B assay (40). Whereas only 40% of the control cells are viable, up to 80% of the RacV12 expressing cells survived after culture in suspension (data not shown). In contrast, continuous expres-sion of the Rac1-V12 transgene had no significant effect on the growth properties of adherent MDCK cells in monolayer conditions over a 72-h period (data not shown).
Dominant Negative Rac1 Enhances Anoikis in MDCK Cells-We also examined the effect of inhibiting the activation of endogenous Rac1 on anoikis. MDCK cells induced to express Rac1-N17 transgene display a 2-fold higher DNA fragmentation in suspension relative to control cells ( Fig. 2A). Rac1-N17 expression also led to a marked increase in apoptosis in adherent cells. These data, together with our results that Rac1-V12 increases cell survival in suspension conditions, strongly indicate a role for Rac1 in the regulation of anoikis.
Activation of NF-B Is Dispensable for Rac1-mediated Inhibition of Anoikis-We next investigated the signaling mechanisms that are utilized by Rac1 to stimulate cell survival in suspension. Rac1 has been shown to stimulate the activity of NF-B (12,13), and this transcription factor plays an important role in the control of cell survival (41). To test whether Rac1-V12 can activate NF-B in MDCK cells in suspension conditions, we used a NF-B-dependent luciferase reporter, E-selectin-luc (37). Whereas the NF-B activity in control cells was dramatically attenuated in suspension conditions, cells that express Rac1-V12 in suspension maintain a level of NF-B activity that is similar to that observed in adherent conditions (Fig. 3A). An E-selectin promoter construct in which two of the three NF-B binding sites are mutated and that is no longer responsive to NF-B (37) was largely unaffected by Rac1-V12 expression (Fig. 3A). Similar results were obtained using a different NF-B-responsive luciferase construct, containing a minimal promoter with five upstream canonical NF-B elements (Fig. 3B).
In resting cells, NF-B is kept in the cytosol as a complex with IB inhibitory proteins. Activation of NF-B involves phosphorylation of IB on two serine residues by the IB kinase complex, leading to proteolytic degradation of IB. Subsequently, free NF-B can enter the nucleus and activate gene transcription. To test whether NF-B contributes to Rac1-mediated inhibition of anoikis, we established a set of MDCK cell lines that inducibly express a super-repressor mutant of IB␣ (IB␣-S32A,S36A or IB␣M) (33-35) from the same tetracycline-repressible promoter used for the Rac1-V12 transgene (see "Experimental Procedures"). IB␣M cannot be phosphorylated and remains constitutively bound to NF-B in the cytoplasm, thereby preventing nuclear entry of NF-B and subsequent transcription of NF-B target genes.
Co-expression of IB␣M with Rac1-V12, for a representative set of cell lines tested (C3, C4, and A6) strongly inhibits Rac1-V12-induced activation of NF-B when compared with a Rac1-V12-expresssing control cell line (C4) (Fig. 3B). The level of inhibition of NF-B activation roughly correlates with IB␣M expression levels (Fig. 3B, inset). The basal level of NF-B activity, measured in the presence of doxycycline, is similar in the lines that co-express IB␣M with Rac1-V12 and the one that only expresses Rac1-V12. In two of the cell lines (MDCK-Rac1-V12/IB␣M-C3 and -C4), co-expression of IB␣M inhibits NF-B transcriptional activity to below the level that is obtained in the absence of Rac1-V12, indicating that the expression level of IB␣M in these lines is sufficient to inhibit basal, nonstimulated transcriptional activity.
Interestingly, however, cell lines that co-express Rac1-V12 and IB␣M show a similar degree of resistance to anoikis to that of cell lines that express Rac1-V12 only (Fig. 3C). In particular, a cell line that displays total inhibition of Rac-V12induced NF-B activation (Rac1-V12/IB␣M-C4) still shows strong protection from anoikis. These experiments indicate that activation of NF-B is dispensable for Rac1-mediated inhibition of anoikis.
Rac1 Survival Signaling Is Independent of p42/44 ERK Activation-The Raf-MEK-p42/44 ERK cascade is an extensively characterized effector pathway that is activated by the Ras proto-oncogene. Activation of this pathway contributes to cell cycle progression as well as cell survival (42,43). Although there is no evidence that Rac1 can directly activate the ERK pathway, Rac1 can synergize with Raf to activate ERK (10,11). Rac1 also potentiates epidermal growth factor-induced ERK activation (11). In light of these studies, we examined the role of ERK in Rac1-mediated inhibition of anoikis.
In adherent MDCK cells, ERK appears to be activated in a constitutive fashion, as suggested by high levels of phosphorylated ERK (Fig. 4A). This high level of ERK activation persists in cells that are serum-starved for 48 h (data not shown). MDCK cells expressing Rac1-V12 do not significantly increase the level of phospho-ERK. Interestingly, cell detachment of control MDCK cells from extracellular matrix in the presence of 10% serum leads to rapid down-regulation of ERK activation, followed by a slow recovery (Fig. 4A). Rac1-V12-expressing cells, however, maintain a much higher level of phospho-ERK in comparison with cells that are kept in doxycycline, indicating that expression of Rac1-V12 strongly stimulates ERK activity in suspension conditions.
To examine the role of ERK signaling in Rac1-mediated cell survival, we used the MEK1/2-specific inhibitor U0126 (44). In the presence of 10 M U0126, ERK activation in suspension cultures was almost completely inhibited in both MDCK-Rac1-V12 and control cells (Fig. 4B). We verified that at this concentration, Rac1-V12 transgene expression was not significantly affected (Fig. 4C). At higher concentrations (30 -100 M), however, Rac1-V12 transgene expression was strongly inhibited (data not shown). Even in the absence of ERK activation, Rac1 still fully protects suspension cells from anoikis (Fig. 4D), in-dicating that ERK activation is dispensable for Rac1-induced cell survival in suspension.
Rac1 Survival Signaling Is Partially Dependent on p38 -Activation of Rac1 has been shown to stimulate both the c-Jun N-terminal kinase and p38 stress kinase cascades (45,46). Because Rac1-induced stimulation of p38 pathway has been implicated in cell survival (23,27), we examined whether Rac1 can stimulate p38 in nonadherent conditions. p38 was immunoprecipitated from MDCK cells in suspension, followed by an in vitro kinase assay using the p38 substrate ATF-2. As shown in Fig. 5A, expression of Rac1-V12 causes a significant increase in p38 activity.
To investigate the role of Rac1-induced p38 activation in the inhibition of anoikis that is caused by Rac1, we used the pyridinyl imidazole compound SB203580, a specific inhibitor of p38 (47). SB203580 at a concentration of 2 M, a concentration that strongly inhibits p38 in other systems (23,48), caused a modest but significant (p Ͻ 0.01) reduction in cell survival in the presence of Rac1-V12 (Fig. 5B) without affecting Rac1-V12 transgene expression levels (data not shown). Control cells did not appear to be significantly affected by 2 M SB203580 (Fig.  5B). These data suggest that Rac1 stimulation of p38 may play a minor role in cell survival in suspension conditions. Activity of PI3K Is Required for Rac1 Protection from Anoikis-Although it is generally accepted that PI3K works upstream of Rac in a wide range of biological systems, Rac1 has also been shown to stimulate some cellular functions in a PI3K-dependent fashion. These include invasion of T47D cells in three-dimensional collagen gels (49), adhesion of T lymphocytes to fibronectin (50), and protection of murine hematopoietic cells from interleukin-3 withdrawal-induced apoptosis (23). We therefore investigated the role of PI3K in Rac1-mediated survival in suspension conditions, using the specific inhibitor LY294002. Low concentrations of LY294002 (2 M) completely
These results suggested the possibility that Rac1 may stimulate the activity of PI3K in suspension conditions. To examine this, we tested whether Rac1-V12 can stimulate PKB activity in a PI3K-dependent fashion. In vitro kinase assays performed on total PKB immunoprecipitated from MDCK cells show that Rac1-V12 significantly stimulates the kinase activity of PKB in suspension conditions (Fig. 6C). In addition, Rac1-induced stimulation of PKB is inhibited by 2 M of LY294002. These results suggest that Rac1 may protect epithelial cells from anoikis by activating the PI3K-PKB pathway. DISCUSSION In this paper, we have examined the role of Rac1 in the survival of epithelial cells in suspension conditions. Our study shows that inducible expression of Rac1-V12 significantly inhibits anoikis in MDCK cells. Our results also suggest that Rac1-induced activation of PI3K plays an important role in protection against anoikis.
We have studied the role of Rac1 in anoikis using a number of independent methods: quantitation of histone-associated DNA fragments, caspase activation, TdT-mediated dUTP nick end labeling staining, and the measurement of cell viability using sulforhodamine B staining. The results of these assays are in strong agreement with each other; expression of Rac1-V12 potently enhances the survival of MDCK cells in the absence of attachment, whereas Rac1-N17 inhibits cell survival in these conditions.
Expression of Rac1-V12 in MDCK cells in suspension conditions causes the activation of a number of signaling pathways that have been implicated in the control of cell survival in other systems. These include the ERK and p38 mitogen-activated protein kinase cascades, PKB, and the transcription factor NF-B. Our results using specific inhibitors of these pathways indicated that the ERK or NF-B pathway do not significantly contribute to Rac1-mediated protection against anoikis, whereas the p38 pathway may play a minor role in Rac1 survival signaling in these conditions.
The activity of PI3K, however, appears to be critical for Rac1-mediated protection against anoikis. This is somewhat surprising because several earlier reports strongly indicate that Rac1 functions downstream of PI3K (8,(51)(52)(53). Interestingly, more recently a number of biological functions that are regulated by Rac1, including cell survival, have been shown to be inhibited by pharmacological inhibitors of PI3K (23,28,49,50). In addition, similar to the results that we obtained in MDCK cells, expression of constitutively active forms of Rac1 in hematopoietic cells can stimulate PKB in a PI3K-dependent fashion (23,54).
It appears therefore that PI3K may function both upstream and downstream of Rac, suggesting the existence of a positive feedback loop. Such a feedback loop has been postulated to function in the establishment of front-to-back polarity during chemotaxis (55), but our results suggest that it could also play a role in integrin-mediated cell survival signaling.
The precise function of PI3K in cell survival signaling downstream of Rac remains to be clarified. One possible mechanism is that Rac recruits PI3K to the plasma membrane and/or stimulates the enzymatic activity of PI3K. In support of this, it has been shown that Rac1 can bind to the p85 subunit of PI3K in a GTP-dependent manner, although whether or not this leads to enhanced phosphatidylinositol 3,4,5-trisphosphate generation in vivo remains to be established. Alternatively, Rac could activate PI3K in an indirect manner, by stimulating the production of autocrine factors, which in turn bind to receptors that cause the activation of PI3K. The signaling elements that act downstream of PI3K in the Rac-stimulated survival pathway remain to be identified. The prominent role played by PKB in PI3K-mediated cell survival, however, strongly suggests that PKB may also be important for the protection against anoikis that is stimulated by Rac.
Our results show that expression of Rac1-V12 also activates NF-B in nonadherent MDCK cells. This is consistent with previous observations that Rac can stimulate NF-B in other cell types and conditions (12,13). NF-B is thought to play a key role in survival signaling (41,56). However, inhibition of NF-B activation by co-expression of the IB super-repressor did not have any effect on cell survival in suspension conditions, indicating that NF-B does not mediate for Rac1-induced protection against anoikis.
We have also investigated whether the ERK and p38 pathways mediate Rac1-stimulated cell survival in suspension conditions. These mitogen-activated protein kinase cascades have been shown to play complex roles in cell survival, in that both have been shown to have either pro-apoptotic (26,27,57,58) or anti-apoptotic (23,27,59) effects, depending on the experimental conditions. Cell survival that is stimulated by Rac1 both in hematopoietic BaF3 cells deprived from interleukin-3 and in fibroblasts prevented from attachment to extracellular matrix has been shown to be dependent on Rac1-stimulated activation of p38 (23,27). Interestingly, whereas we showed that Rac1-V12 can activate p38 in MDCK cells in suspension culture, p38 only appears to play a minor role at best in Rac1-stimulated cell survival in these conditions. We also showed that Rac1-V12 can activate ERK in a MEK-dependent fashion in MDCK suspension cells, but inhibition of Rac1-induced ERK activation does not significantly affect cell survival that is stimulated by Rac1.
An additional potential mechanism for Rac1-induced protection against anoikis that we considered is that Rac1 could enhance cell survival because it may increase cadherin-based cell-cell interactions (60). Indeed, both Rac1-V12-expressing and control MDCK cells strongly cluster when kept in suspension for prolonged periods of incubation. However, at the early time point (4 h) at which we observed significant Rac1-induced inhibition of caspase activity, cell clustering is minimal (data not shown). Furthermore, we observed that anoikis is significantly attenuated when the cells are incubated at lower densities, and cell clustering in these conditions is greatly reduced (data not shown). Together, these data suggest that Rac1induced protection against anoikis is not mediated via an effect on cell-cell adhesion.
Our results that Rac1 can protect epithelial cells against anoikis are in line with recent observations that expression of constitutively active Rac1-V12 inhibits anoikis in fibroblasts (27). In fibroblasts, however, Rac-stimulated cell survival in suspension appears to be strongly dependent on p38 activation, whereas in MDCK cells, Rac1-induced activation of p38 only plays a minor role at best in the protection against anoikis. Activated Rac1 also enhances the survival of hematopoietic BaF3 cells deprived from interleukin-3 in a p38-dependent fashion, and in this system PI3K also plays a critical role (23).
Expression of activated Rac1 also stimulates cell survival in other conditions and cell lines, including Myc-expressing Rat1 fibroblasts that are challenged by serum starvation (25) and  fibroblasts that express high levels of oncogenic Ha-Ras (26). In the latter, Rac-induced activation of NF-B was proposed to mediate the survival function of Rac. Thus, Rac stimulates a number of distinct survival pathways, and the precise contribution of these various pathways may vary depending on the nature of the challenge and cell type.
Interestingly, in other experimental systems, Rac1 has been shown to be actively involved in relaying pro-apoptotic signals. These include ceramide-, Fas-, and cytotoxic T lymphocyteinduced cell killing (22,57) and nerve growth factor withdrawal-induced cell death in rat sympathetic neurons (61). Thus, Rac1 controls both pro-and anti-apoptotic signaling pathways. This characteristic is shared with several other signal transduction elements, including ERK, p38, and NF-B (43,56), and whether these elements promote or inhibit apoptosis appears to depend on the cell type and the nature of the inducer. Our observations that activated Rac1 can protect epithelial cells from cell death that is induced by suspension conditions suggest that Rac1-stimulated survival signals contribute to cell transformation and tumorigenesis.