Rac1 contributes to trastuzumab resistance of breast cancer cells: Rac1 as a potential therapeutic target for the treatment of trastuzumab-resistant breast cancer

Overexpression of ErbB2, a member of the ErbB family of receptor tyrosine kinases, is associated with breast cancer progression, metastasis, poor prognosis, and poor response to therapy (1, 2). Trastuzumab is a humanized monoclonal antibody targeting the ErbB2 receptor extracellular domain. Clinical studies with trastuzumab, either alone or in combination with adjuvant chemotherapy, have shown favorable responses in ErbB2-positive breast cancer patients (3–5). Proposed mechanisms of action for trastuzumab include (a) down-regulation of ErbB2 through endocytic uptake of trastuzumab/ErbB2 complex, (b) disruption of downstream signaling pathways, (c) induction of G1 cell cycle arrest and apoptosis, (d) suppression of angiogenesis, (e) induction of antibody-dependent cellular cytoxicity, and (f) inhibition of ErbB2 extracellular domain proteolysis (1, 2).

Although trastuzumab treatment significantly improves outcomes for women with ErbB2-positive breast cancers, many patients who achieve an initial response to trastuzumab also acquire resistance to trastuzumab (2, 6, 7). Understanding the mechanisms contributing to trastuzumab resistance and identifying novel therapeutic targets is critical for improving the survival of this population of breast cancer patients. Several studies have advanced our understanding of molecular mechanisms of trastuzumab resistance. First, trastuzumab resistance is associated with reduced p27^kip1, elevated cyclin-dependent kinase 2 activity, and increased heterodimerization of ErbB2 receptor with IGF-IR receptor in trastuzumab-resistant SKBR3 cells (8, 9). Second, in a BT-474 cell line model of trastuzumab resistance, significant increases in the levels of epidermal growth factor (EGF) receptor (EGFR), transforming growth factor α, heparin-binding EGF, and Heregulin occur (10). Third, in a cell line (JIMT-1) established from a patient who failed to respond to trastuzumab and adjuvant chemotherapy, resistance to trastuzumab is associated with heterodimerization of ErbB2 to glycoprotein MUC4, which prevents ErbB2 from binding to trastuzumab (11, 12). Fourth, analysis of clinical samples from trastuzumab responders and nonresponders indicates that loss of PTEN tumor suppressor correlates with a significant decrease in response to trastuzumab therapy (13).

Because multiple mechanisms likely contribute to trastuzumab resistance in breast cancer, combining trastuzumab with other therapeutic agents that target either ErbB family receptors or signaling molecules downstream of ErbB receptors may increase the magnitude and duration of trastuzumab response. Lapatinib, which inhibits the tyrosine kinase activity of both EGFR and ErbB2, is currently undergoing clinical testing for patients whose cancers have progressed while receiving trastuzumab (14). Furthermore, synergistic growth inhibition occurs in breast cancer cell lines treated with both Lapatinib and trastuzumab (15). Additionally, IGF-IR–targeting agents have been introduced into pharmaceutical testing and are being assessed in preclinical trastuzumab-resistant models (2).

Rho family GTPases belong to the Ras superfamily of small GTPases that serve as molecular switches. Activation of Rho family GTPases regulates a wide range of cellular functions, including cytoskeletal dynamics, cell polarity, adherens junctions, and member trafficking (16–20). It has been very well documented that Rac1, Cdc42, and RhoA have the ability to control normal cell growth and, when improperly regulated, contribute to tumorigenesis, tumor invasion, and metastasis (17–21). Several lines of evidence indicate that Rac1 and Cdc42 function downstream of ErbB family receptors. For instance, EGF-coupled signaling is sufficient to activate Cdc42 and Rac1, and treatment of cells with Heregulin, an ErbB3 ligand, increases Rac1 activity (16, 22–25). Increased Rac1 activity occurs in MCF10A human mammary epithelial cells overexpressing ErbB2 receptor (26). There is also evidence that Rac1 is overexpressed or hyperactivated in breast cancer tissue (27).

It is becoming more evident that the activities of Rho family GTPases contribute to multiple aspects of oncogenic phenotypes in breast cancer cells, including resistance to chemotherapy (28). Overexpression of Cdc42 and Rac1 has been reported in human breast cancers (29, 30). Up-regulated Rac1 and Cdc42 activities are believed to be associated with breast cancer progression, invasion, and metastasis (29–32). However, it has not been reported whether Rho family GTPases are involved in trastuzumab resistance in breast cancer cells.

SKBR3 breast cancer cells express high levels of ErbB2 and are sensitive to trastuzumab treatment (33). Here, we investigate the mechanisms contributing to trastuzumab resistance using SKBR3 cells and a SKBR3 trastuzumab-resistant derivative clone, known as Clone 3 cells, which were created by chronically exposing parental SKBR3 cells to trastuzumab in culture media (9, 33). We show that Rac1 activity is up-regulated in Clone 3 cells compared with trastuzumab-sensitive parental SKBR3 cells and that, in Clone 3 cells, actin stress fibers are disrupted, resulting in altered cell morphology. Furthermore, trastuzumab-mediated endocytic down-regulation of ErbB2 is impaired in Clone 3 cells. This leads to the accumulation of ErbB2 on the cell surface and increased ErbB2 signaling in Clone 3 cells. Clone 3 cells treated with the Rac1 inhibitor NSC23766 reassemble actin-based stress fibers and display a cellular morphology similar to parental SKBR3 cells. Additionally, inhibition of Rac1 activity by NSC23766 restores endocytotic uptake of the trastuzumab/ErbB2 complex and trastuzumab-mediated ErbB2 degradation. Overexpression of Rac1-activated mutant in parental SKBR3 promotes partial resistance to trastuzumab-induced growth arrest. These findings highlight an important role for Rac1 in trastuzumab resistance of breast cancer cells and provide evidence that targeting Rac1 may be a potential therapeutic approach for the treatment of trastuzumab-resistant breast cancer.

Trastuzumab (Genentech, Inc.) was purchased from the pharmacy at NIH. NSC23766 was obtained from Tocris Cookson, Inc. Antibodies against Cdc42, Rac1, EGFR, and ErbB2 were purchased from BD Transduction Laboratories. Antibodies against p44/42 extracellular signal-regulated kinase 1/2 (ERK1/2), phosphorylated 44/42 ERK1/2, phosphorylated stress-activated protein kinase (SAPK)/c-Jun NH₂ terminal kinase (JNK; Thr¹⁸³/Tyr¹⁸⁵), phosphorylated c-Jun (Ser⁶³) II, phosphorylated ErbB2 (Tyr⁸⁷⁷), and phosphorylated ErbB2 (Tyr^1221/1222) were obtained from Cell Signaling Technology. Polyclonal anti-actin antibody was obtained from Sigma. Antibody against EEA1 were obtained from Santa Cruz Biotechnology. Anti-Rho antibody was purchased from Pierce. Monoclonal anti-HA antibody was obtained from Covance. Trypan blue was obtained from Lonza. Rhodamine-conjugated phalloidin and Alexa Fluor 594 donkey anti-mouse antibody were obtained from Invitrogen Molecular Probes. Cy2-conjugated donkey anti-human antibody and rhodamine-conjugated donkey anti-goat secondary antibody were obtained from Jackson ImmunoResearch. EZ-detect Rho activation kit and EZ-detect Rac1 activation kit were obtained from Pierce.

Parental SKBR3 cells (American Type Culture Collection) and Clone 3 cells, which were described previously (33), were obtained from Dr. Esteva of M.D. Anderson Cancer Center and cultured in the standard culture media unless otherwise indicated. The standard culture media for parental SKBR3 cells is DMEM/F12 supplemented with 10% fetal bovine serum. The standard culture media for Clone 3 cells is DMEM/F12 supplemented with 10% fetal bovine serum and trastuzumab (4 μg/mL). JIMT-1 cell line was purchased from DSMZ (German Collection of Microorganisms and Cell Cultures) and described by Tanner and colleagues (12). The JIMT-1 cells are cultured in DMEM supplemented with 10% fetal bovine serum. Transient transfections using electroporation were done according to the manufacturer's protocol (Amaxa, Inc.).

After electroporation of SKBR3 cells with plasmid encoding HA-tagged Rac1G12V and pcDNA3.1 vector carrying neomycin resistance gene, the stable cell line selection was achieved using G418 (21).

Cells were cultured in 60-mm plates for 12 h in the standard culture media. Thereafter, cells were treated with either trastuzumab or trastuzumab plus NSC23766, or left untreated. At the indicated times, cells were harvested and 20 μg whole-cell lysate (WCL) was subjected to SDS-PAGE. ErbB2 protein levels were assessed by Western blot analysis with an anti-ErbB2 antibody.

Cells were plated in either 12-well or 24-well plates and cultured in the standard culture media for 12 h. Cells were either not treated, treated with trastuzumab, or treated with trastuzumab plus NSC23766. At the indicated times, cells were trypsinized and counted. Cell viability was determined using trypan blue staining in a standard hemocytometer setup.

Cells were seeded overnight in chamber slides at 1 × 10⁵ per well and cultured in standard culture media. Cells were then rinsed with PBS, fixed, permeabilized, and stained, as previously described (20).

Both parental SKBR3 and Clone 3 cells were cultured in the standard culture media. Parental SKBR3 cells were either treated with trastuzumab or not treated for the indicated times, whereas Clone 3 cells remained in the standard culture media for the indicated times. Internalized or surface-bound trastuzumab was detected by Cy2-conjugated donkey anti-human antibody. EEA1 was detected using a mouse EEA1 monoclonal antibody. For the effects of NSC23766 on trastuzumab internalization, the experimental procedures were essentially the same as that described above, except that NSC23766 was used to treat the cells for the indicated times. For EGFR down-regulation assay, the parental and Clone 3 cells were plated and treated with 100 ng/mL EGF. At the indicated times, cells were harvested and WCL were subjected to SDS-PAGE. The levels of EGFR were assessed by Western blot analysis with anti-EGFR antibody.

GTPase activity pull-down assays were done using either glutathione S-transferase (GST)–p21-binding domain (PBD) of p21-activated kinase or GST-Rho-binding domain (RBD) of Rhotekin, as described previously (20). Briefly, GST-PBD or GST-RBD was immobilized by binding to glutathione-Sepharose beads. The immobilized GST-PBD was used to precipitate activated Cdc42 or Rac1 from WCL. The immobilized GST-RBD was used to precipitate activated Rho from WCL.

We observed profound differences in cell morphology between parental SKBR3 and Clone 3 cells. Whereas parental SKBR3 cells had a flattened morphology and were translucent by phase-contrast microscopy, Clone 3 cells exhibited a round cell shape with long protrusions extending from the cell body and were refractive on phase-contrast microscopy (Fig. 1A). Clone 3 cells also differed from parental cells, in that Clone 3 cells lacked stress fibers and had strong circular actin staining at the cell edges (Fig. 1A). We then questioned whether the significant cytoskeleton rearrangement displayed by Clone 3 cells was correlated with the changes in the activities of Rho family GTPases. As shown in Fig. 1B, the levels of active Rac1 were increased in Clone 3 cells compared with that in parental SKBR3 cells. Active Cdc42 and Rho proteins were comparable in these two cell lines, and there was no detectable background binding in either of the GST control assays (Fig. 1B). In addition, there were no detectable differences in the protein levels of Rho, Cdc42, and Rac1 between the parental SKBR3 cells and Clone 3 cells in the WCL, which was used for both PBD and RBD pull-down assays (Fig. 1C). Thus, the activity of Rac1 is increased in Clone 3 cells compared with parental SKBR3 cells, which might contribute to the morphologic changes displayed by Clone 3 cells.

Rac1 has been characterized as a downstream effector of ErbB receptors and has recently been shown to associate with ErbB2 in MCF10A cells (16, 22, 26, 36). In light of these data, we investigated whether Rac1 interacted with ErbB2 in Clone 3 cells. Specific binding of ErbB2 to the PBD-Rac1-GTP protein complex was observed in Clone 3 cells, but not in parental SKBR3 cells (Fig. 1D). Binding of ErbB2 to activated Rac1 was further confirmed by transient overexpression of constitutively activated mutants of Rac1 in parental SKBR3 cells. As shown in Fig. 1E, both HA-tagged Rac1F28L, a fast cycling mutant (37), and HA-tagged Rac1Q61L associated with endogenous ErbB2 in parental SKBR3 cells. We next examined whether Rac1 was colocalized with ErbB2 in cells. As shown in Fig. 1F, ErbB2 proteins were mainly localized (closed arrows) at the cell edges in parental SKBR3 cells, whereas Rac1 exhibited a cytosolic staining. No significant colocalization between ErbB2 and Rac1 was detected at the cell edges in parental SKBR3 cells (Fig. 1F). In contrast, both ErbB2 and Rac1 colocalized at the cell edge membrane ring region in Clone 3 cells (Fig. 1F, open arrows). These data support idea that the activated Rac1 is associated with ErbB2 in Clone 3 cells.

Trastuzumab, a humanized monoclonal antibody, directly binds to the extracellular domain of ErbB2 receptor to mediate internalization and degradation of ErbB2 receptor (34). As shown in Fig. 2A, after treatment of parental SKBR3 cells with 10 μg/mL trastuzumab for 4 days, trastuzumab was endocytosed as visualized by donkey anti-human Cy2-conjugated antibody (green punctate structures) and partially colocalized with EEA1, an early endosomal marker (top), consistent with other reports (34, 35). In contrast, trastuzumab-untreated control SKBR3 cells, for which trastuzumab was only used to recognize ErbB2, did not show any significant colocalization between trastuzumab and EEA1. The trastuzumab was mainly located at the cell surface, and no significant green punctate structures were observed in control SKBR3 cells (Fig. 2A, bottom row). The enlarged boxed region in Fig. 2B depicts the partial colocalization between the trastuzumab and EEA1. These data indicate that trastuzumab is internalized in parental SKBR3 cells. Under identical trastuzumab treatment conditions, Clone 3 cells differed from parental SKBR3 cells, in that the trastuzumab accumulated at the cell surface, was not detected in the intracellular compartments, and had no colocalization with EEA1 (Fig. 2A, second row from top). We next asked whether Clone 3 cells were capable of down-regulating ErbB2. After being incubated with trastuzumab (10 μg/mL) for 5 days, little or no ErbB2 was detected in parental SKBR3 cells (Fig. 2C, left). In contrast, ErbB2 levels were not reduced in Clone 3 cells (Fig. 2C, right). Taken together, these data indicate that trastuzumab is able to mediate endocytic down-regulation of ErbB2 in parental SKBR3 cells, and trastuzumab-mediated endocytic down-regulation of ErbB2 is impaired in Clone 3 cells.

The results described in Fig. 2C lead to the question of whether EGF-induced down-regulation of EGFR is also impaired in Clone 3 cells. After the addition of EGF to parental SKBR3 cells, the total receptor levels as detected by Western blotting with an anti-EGFR antibody were significantly reduced over a time period of 0 to 3 hours (Fig. 2D). However, after a 3-hour incubation of EGF with Clone 3 cells, there was no significant reduction in the total levels of EGFR (compare lane 5 with lane 8). These data, when taken together with Fig. 2A and C, suggest that the significant cytoskeleton disorganization described in Fig. 1 may disrupt clathrin-dependent endocytosis and subsequently impair lysosomal degradation of ErbB receptors in Clone 3 cells.

NSC23766 inhibits Rac1 function by preventing Rac1 activation by Tiam 1 and TrioN (38). We used NSC23766 to test whether the changes in cell morphology were associated with the up-regulated Rac1 activity in Clone 3 cells. Cells were cultured in the standard culture media and treated with or without NSC23766 for the indicated times. Cell morphology was assessed using phase-contrast microscopy. As shown in Fig. 3A, NSC23766 treatment of Clone 3 cells for either 2 or 4 days led to significant changes in cell shape, such that cells became flattened and spread out and tended to resemble parental SKBR3 cell morphology. This suggested that the rounded morphology in Clone 3 cells was due to increased Rac1 activity. Moreover, stress fibers were restored in Clone 3 cells treated with NSC23766 (Fig. 3B). Quantification of stress fiber formation indicated a dose-dependent increase in stress fiber restoration (4.6–67.6%) in Clone 3 cells after cells were treated with different concentrations of NSC23766 (Fig. 3B, graph). Taken together, these data suggest that the up-regulated Rac1 activity likely contributed to the cytoskeletal and morphologic changes in Clone 3 cells.

We next asked whether NCS23766 had the ability to restore trastuzumab-mediated ErbB2 internalization and degradation in Clone 3 cells. Clone 3 cells were cultured in the standard culture media, which contains trastuzumab, and were then treated with the indicated concentrations of NSC23766 or cultured in the absence of NSC23766 for 2 days. As shown in Fig. 3C, trastuzumab, as detected by Cy2-conjugated donkey anti-human antibody, was internalized and localized to intracellular compartments in NSC23766-treated Clone 3 cells, whereas trastuzumab exhibited cell surface staining in NSC23766-untreated Clone 3 cells. Furthermore, internalization of trastuzumab in Clone 3 cells treated with NSC23766 was increased in a dose-dependent manner (Fig. 3C, graph).

We next addressed the question of whether NSC23766 treatment restored trastuzumab-mediated ErbB2 degradation in Clone 3 cells. Clone 3 cells were incubated in the standard culture media and treated with or without NSC23766. ErbB2 levels in WCL were detected by Western blot analysis using an antibody against ErbB2. As shown in Fig. 3D, ErbB2 levels were not changed at day 2 and day 4 when Clone 3 cells were not treated with NSC23766. In contrast, addition of 75 or 100 μmol/L NSC23766 for either 2 or 4 days led to a significant decrease in ErbB2 protein levels. Taken together, these data further support the idea that the increased Rac1 activity in Clone 3 cells leads to significant changes in the actin cytoskeleton, which may, in turn, impair trastuzumab-mediated ErbB2 internalization and degradation in Clone 3 cells. Reduction of Rac1 activity in Clone 3 cells by NSC23766 restores cell morphology and actin cytoskeleton to that similar to parental SKBR3 cells and, therefore, restores ErbB2 internalization and degradation.

Rac1 has been reported to play an essential role in the regulation of cell growth and cancer progression (39, 40). We tested whether NSC23766 was able to affect the growth of parental SKBR3 and Clone 3 cells. Cells were incubated in the standard culture media and treated or not treated with NSC23766 at the indicated concentrations. Although NSC23766 treatment inhibited growth of both parental SKBR3 and Clone 3 cells, our data revealed differential inhibition of cell growth and loss of cell viability between parental SKBR3 cells and Clone 3 cells, such that parental SKBR3 cells were relatively more sensitive to the NSC23766 treatment than Clone 3 cells. As shown in Fig. 4A, when treated with 50 μmol/L NSC23766, Clone 3 cell numbers increased ∼4-fold at day 4 (bottom graph), whereas the cell number for parental SKBR3 cells was only slightly increased under the same condition (top graph). Moreover, 75 μmol/L NSC23766 eliminated the cell growth of parental SKBR3 cells (Fig. 4A, top graph). In contrast, numbers for Clone 3 cell doubled by day 4 when grown in media containing 75 μmol/L NSC23766 (Fig. 4A, bottom graph). Cell growth for both parental SKBR3 and Clone 3 cells was inhibited when treated with 100 μmol/L NSC23766.

Analysis of cell death by trypan blue exclusion in cells treated with NSC23766 revealed similar trends as those of the growth inhibition shown in Fig. 4A. There was a slight increase in the cell death observed in parental SKBR3 cells compared with Clone 3 cells at days 2 and 4, when cells were treated with 50 μmol/L NSC23766 (Fig. 4B). Whereas the percentage loss in viability for parental SKBR3 cells treated with 75 μmol/L NSC23766 was significantly increased at days 2 and 4, Clone 3 cells exhibited relative resistance to NSC23766 treatment (Fig. 4B). Treatment with 100 μmol/L NSC23766 caused a significant loss in viability in both cell lines at day 4 (Fig. 4B, bottom graph). Our data indicate that NSC23766-induced growth inhibition is correlated with the activity of Rac1, such that the higher activity of Rac1 in Clone 3 cells may render the cells more resistant to NSC23766 treatment compared with parental SKBR3 cells.

JIMT-1 breast cancer cells are an ErbB2-positive cell line derived from a patient primarily insensitive to trastuzumab therapy (12). We next tested if NSC23766 was able to induce growth inhibition in JIMT-1 cells. Figure 4C shows that NSC23766 treatment caused a marked reduction in cell number in JIMT-1 cells. Taken together, data obtained from JIMT-1 cells may generalize the NSC23766-induced growth inhibition to broader trastuzumab-insensitive cell lines.

As shown in Fig. 5A, after treatment of parental SKBR3 cells with trastuzumab for 48 hours, the cell number was slightly decreased compared with untreated cells, consistent with the previous report (41). Trastuzumab exhibited significant inhibition of cell growth after incubation of cells with trastuzumab for 5 days. If the increased Rac1 activity contributes to trastuzumab resistance of Clone 3 cells, we would expect that stable expression of constitutively active Rac1G12V mutant in parental SKBR3 cells would decrease their sensitivity to the trastuzumab treatment. As shown in Fig. 5B, after the treatment of vector control cells with trastuzumab for 4 days, a 40% reduction in cell number was observed compared with untreated vector control cells. In contrast, after treatment of SKBR3 cells stably expressing constitutively active Rac1G12V (Clone 16-SKBR3 and Clone 18-SKBR3) for 4 days, no significant reduction in cell numbers was observed compared with untreated Rac1G12V expressing cells. Moreover, after trastuzumab treatment for 7 days, there was ∼60% decrease in cell number in vector control cells compared with untreated vector control cells, whereas there were 36% and 40% decreases in cell numbers after trastuzumab treatment in Clone 18-SKBR3 cells and Clone 16-SKBR3 cells, respectively, compared with untreated Rac1G12V-expressing cells (Fig. 5B). Figure 5C showed the overexpression levels of Rac1G12V proteins in different clones of SKBR3 cells. Taken together, these data suggest that overexpression of Rac1G12V decreases the sensitivity of parental SKBR3 cells to trastuzumab and provide additional evidence that increased Rac1 activity contributes to trastuzumab resistance in Clone 3 cells.

Based on the results shown in Fig. 5A, 2-day treatment of cells with trastuzumab is an appropriate time point for studying whether trastuzumab-mediated growth inhibition can be enhanced by NSC23766. Parental SKBR3 cells were seeded and cultured in media containing either no trastuzumab or 4 μg/mL trastuzumab and were treated with or without NSC23766 at the indicated concentrations for 2 days. Cells were then examined for growth, viability, and levels of ErbB2 in WCL. A 25% reduction in cell number was observed 48-hour after trastuzumab treatment (Fig. 5D, compare column 2 with column 1). Cells treated with 70 μmol/L NSC23766 for 48 hours in the absence of trastuzumab had ∼15% to 20% growth inhibition (compare column 3 with column 1). However, an additive effect on growth inhibition (∼50% decrease in cell number) was found when cells were treated with both trastuzumab (4 μg/mL) and 70 μmol/L NSC23766 (compare column 4 with column 1). The same was true when cell viability was quantified in the same experiment (Fig. 5E). We next determined the effect of NSC23766 on trastuzumab-mediated ErbB2 degradation. NSC23766 treatment significantly increased trastuzumab-mediated ErbB2 degradation in the parental SKBR3 cells (Fig. 5F). Whereas the protein levels of ErbB2 were not reduced 48 hours posttreatment with either trastuzumab or NSC23766 (Fig. 5F, lanes 2, 3, and 4), the total ErbB2 level in WCL was significantly reduced when cells were incubated in the culture media containing 70 μmol/L NSC23766 and trastuzumab for 48 hours (lane 6), indicating that inhibition of Rac1 activity by NSC23766 enhances trastuzumab-induced down-regulation of ErbB2 in parental SKBR3 cells. Taken together, these data show that trastuzumab-mediated growth inhibition became more effective when the activity of Rac1 was down-regulated by NSC23766.

We next investigated the activity of ErbB2 in both parental SKBR3 cells and Clone 3 cells. At steady-state, the extent of phosphorylation at Tyr⁸⁷⁷ of ErbB2 was elevated in Clone 3 cells compared with parental SKBR3 cells (Fig. 6A). Importantly, steady-state phosphorylation at Tyr^1221/2 was found in Clone 3 cells, whereas it was almost undetectable in the parental SKBR3 cells (Fig. 6A). These data indicate that Clone 3 cells contain a higher activity of ErbB2 compared with parental SKBR3 cells. Tyr^1221/2 is one of the major autophosphorylation sites in the ErbB2 receptor and is associated with the activation of Ras-Raf-MAP kinase signal pathway to mediate cell growth (42). We then asked if signaling downstream of ErbB2 was up-regulated in Clone 3 cells. Activated phosphorylated ERK1/2 in WCL was significantly enhanced in Clone 3 cells compared with both untreated and trastuzumab-treated parental SKBR3 cells (Fig. 6B). Consistent with another report (38), NSC23766 treatment reduced Rac1 activity in Clone 3 cells, as assessed by PBD pull-down assay (Fig. 6C). We next questioned whether the up-regulated ERK activity in Clone 3 cells can be reduced by NSC23766. ERK phosphorylation levels in Clone 3 cells were reduced after the addition of 75 μmol/L NSC23766 to the culture media for 2 and 4 days compared with control untreated cells (Fig. 6D, top). These data suggest that the ability of NSC23766 to decrease ERK activity may be the consequence of enhanced ErbB2 degradation induced by NSC23766 in Clone 3 cells (Fig. 3D).

It has been reported that SAPK/JNK, when activated, regulates cell growth and survival through its effects on c-Jun (28). Given that NSC23766 is selectively targeting Rac1 activity, we tested whether NSC23766 had the ability to modulate signaling downstream of Rac1 activity. Clone 3 cells were either untreated or treated with NSC23766 for the indicated times. As shown in Fig. 6D, NSC23766 treatment decreased the levels of phosphorylated SAPK/JNK and c-Jun.

These results, when taken together with results shown in Fig. 3, suggest that two mechanisms may be used by NSC23766 to mediate its growth inhibitory effects on Clone 3 cells. First, NSC23766 down-regulation of Rac1 activity results in actin cytoskeleton reorganization that subsequently restores trastuzumab-mediated ErbB2 degradation in Clone 3 cells, leading to the reduction of ERK signaling. Second, by directly targeting Rac1 activity, NSC23766 is capable of reducing JNK signaling in Clone 3 cells.

The development of therapeutic resistance to trastuzumab poses a big challenge to the treatment of human ErbB2-positive breast cancer (2, 6). Rac1 has been implicated in the downstream signaling of ErbB receptors and has been reported to be involved in the regulation of breast cancer cell metastasis and invasion (22, 43). Here, we describe a previously unappreciated relationship between increased Rac1 activity and trastuzumab resistance. We show that in Clone 3 cells, Rac1 activity is up-regulated compared with trastuzumab-sensitive parental SKBR3 cells. We also show that Clone 3 cells display a distinctive cell morphology that is accompanied by significant cytoskeleton disorganization and compromised endocytic down-regulation of ErbB receptors. This leads to the aberrant accumulation of ErbB2 on the cell surface, the up-regulated ErbB2 activity, and the enhanced ERK activity in Clone 3 cells, resulting in trastuzumab resistance. When Rac1 activity is inhibited by NSC23766 treatment, however, Clone 3 cells reassemble actin-based stress fibers and have cell morphology similar to parental SKBR3 cells. Moreover, treatment of Clone 3 cells with NSC23766 restores stress fiber formation and trastuzumab-mediated endocytic down-regulation of ErbB2. Because trastuzumab-mediated internalization and degradation of ErbB2 is one of the mechanisms of action for trastuzumab-mediated growth inhibition (3, 34), our findings now highlight an important connection between the impaired endocytic down-regulation of ErbB2 and trastuzumab resistance.

SAPK/JNK activation, which occurs downstream of Rac1, is believed to regulate gene transcription to mediate effects on cell growth and survival (44, 45). Whereas we did not find a difference in JNK activity between parental SKBR3 and Clone 3 cells (data not shown), the activity of JNK is reduced by NSC23766 in trastuzumab-resistant SKBR3 cells, suggesting that Rac1-mediated JNK signaling may be required for Rac1 to contribute to trastuzumab resistance of SKBR3 cells. A model shown in Fig. 6E depicts the role of Rac1 in contributing to trastuzumab resistance in SKBR3 cells.

The activity of Rac1 is correlated with ErbB signaling in many human breast cancers (25, 46). Furthermore, in MCF10A human mammary epithelial cells that overexpress ErbB2, ErbB2-mediated up-regulation of Rac1 activity is blocked by treatment with trastuzumab (26). This suggests that the initial escape from trastuzumab-mediated inhibition of Rac1 activity may be an important mechanism leading to the high levels of Rac1 activity in Clone 3 cells. In support of this model, here, we show that overexpression of Rac1G12V mutant in parental cells reduced the sensitivity to trastuzumab. Whereas we did obtain data indicating that knocking down Rac1 reduced cell growth of parental SKBR3 cells (data not shown), mimicking the growth inhibitory effect of NSC23766 on parental SKBR3 cell, unfortunately, we experienced technical difficulties in silencing the expression of Rac1 in Clone 3 cells, although we attempted several different techniques.

It has been reported that activated Rac1 associates with ErbB2 leading to prolonged activation of Rac1 (36). We also find that activated Rac1 is able to form a stable complex with ErbB2 in Clone 3 cells, but not in parental SKBR3 cells. This may be due to higher levels of activated Rac1 in Clone 3 cells. Binding of Rac1 to ErbB2 may stabilize Rac1 activity, which in turn provides a positive feedback signal to ErbB2-ERK signaling in Clone 3 cells.

Trastuzumab treatment significantly improves outcomes for women with ErbB2 positive breast cancer (2, 7, 47, 48). However, many patients with metastatic breast cancer who initially respond to trastuzumab develop resistance (2). The clinical problem of trastuzumab resistance is becoming increasingly important, as recent studies strongly support a role for trastuzumab in the adjuvant setting for ErbB2 overexpressing early-stage breast cancer (6, 48, 49). Our studies reveal a novel molecular mechanism contributing to trastuzumab resistance and suggest that Rac1 and ERK may be novel therapeutic targets for the treatment of trastuzumab-resistant human breast cancer. Furthermore, whereas primary resistance to cetuximab, a therapeutic monoclonal antibody to EGFR, is associated with KRAS mutations of colorectal cancers, increased Rac1 activity may provide a predictive marker for acquired trastuzumab resistance of ErbB2-overexpressing breast cancers (50). Our data show that NSC23766 inhibits Rac1 activity and has the ability to reverse cellular phenotypes of trastuzumab resistance. Further investigations, including preclinical animal models and clinical trials to study the efficacy and toxicity of NSC23766 in trastuzumab-resistant breast cancer patients, are warranted.

No potential conflicts of interest were disclosed. The information presented in this article reflects the views of the authors and does not necessarily represent the policy of the U.S. Food and Drug Administration.

We thank Dr. Francisco J. Esteva for providing parental SKBR3 cells and trastuzumab-resistant SKBR3 cells (Clone 3 cells) and Drs. David Frucht, Sarah Kennett, Wendy Weinberg, and Kathleen Clouse for the critical review of this manuscript.