Tel.: 302-831-2669, Fax: 302-831-2279
Cancer Cell Biology
Farnesyl transferase inhibitor treatment of breast cancer cells leads to altered RhoA and RhoC GTPase activity and induces a dormant phenotype
Article first published online: 26 OCT 2010
Copyright © 2010 UICC
International Journal of Cancer
Volume 129, Issue 1, pages 61–69, 1 July 2011
How to Cite
Chatterjee, M. and van Golen, K. L. (2011), Farnesyl transferase inhibitor treatment of breast cancer cells leads to altered RhoA and RhoC GTPase activity and induces a dormant phenotype. Int. J. Cancer, 129: 61–69. doi: 10.1002/ijc.25655
- Issue published online: 26 APR 2011
- Article first published online: 26 OCT 2010
- Accepted manuscript online: 7 SEP 2010 10:21AM EST
- Manuscript Accepted: 19 AUG 2010
- Manuscript Received: 1 JUL 2010
- Department of Defense Breast Cancer Research Program. Grant Numbers: W81XWH-06-1-0495, W81XWH-08-1-0356
- farnesyl transferase inhibitor;
- breast cancer dormancy;
- RhoA GTPase;
- RhoC GTPase
Farnesyl transferase inhibitors (FTIs) were shown to be effective in modulating tumor growth in Ras-transformed tumor cells. Recent studies have focused on Rho GTPases as putative targets of FTI action. Previously, we demonstrated that FTIs were effective in inhibiting the growth and invasiveness of RhoC GTPase-overexpressing inflammatory breast cancer (IBC) cells however, RhoC activity was increased. In this study, we examine the mechanisms of FTI action on breast cancer cells in culture through modulation of RhoC and RhoA GTPases. We found that FTI inhibition of breast cancer cell growth was reversible and resembled what has been described for an in vitro model of tumor cell dormancy. On FTI treatment, levels of active RhoA decreased significantly, whereas levels of active RhoC increased 3.8-fold. We studied the role of these two GTPases in a fibronectin and basic FGF-induced model of breast cancer cell dormancy. Hypoactivation of RhoA and hyperactivation of RhoC were seen to induce morphology and growth changes consistent with tumor cell dormancy in culture. In addition, the JNK/SAPK pathway was induced on FTI treatment. A pharmacologic inhibitor of the JNK/SAPK pathway significantly reduced the number of dormant cells. This study has implications for the use of FTIs as therapeutic agents as well as potential mechanisms for breast cancer cell dormancy.
Breast cancer is the leading cause of cancer deaths among women in the United States.1 Recent progress in understanding the molecular mechanisms of breast tumor development, coupled with improvements in multi-modality treatments have led to significant increases in the survival of breast cancer patients.1 Inflammatory breast cancer (IBC) is a particularly aggressive form of locally advanced breast cancer that carries 5- and 10-year disease-free survival rates of 38% and 20%, respectively.2 The dismal prognosis of IBC is due to its propensity to invade the dermal lymphatics,3 which is due to expression and activation of RhoC GTPase.3, 4
RhoC is a member of the Rho subfamily of monomeric GTP-binding proteins and is a metastatic switch for a number of cancers.3, 5, 6 In an attempt to develop new therapies to target IBC metastasis, we demonstrated that treatment of the SUM149 IBC cell line with the farnesyl transferase inhibitor (FTI) L-744,832, led to a reversion of the cells growth and invasive capabilities.7 However, we found that inhibition of the SUM149 cells was reversible on FTI withdrawal.
We observed that FTI-treated IBC cells had a distinct flattened morphology, which was accompanied by a dramatic increase in RhoC GTPase activity levels.7 The FTI-induced morphologic changes are reminiscent of what is described for a fibronectin and basic fibroblast growth factor (FGF)-induced in vitro model of breast tumor cell dormancy where a decrease in RhoA GTPase activity is observed.8–10 Although they are highly homologous to each other RhoA and RhoC have distinct functions and can antagonize one another.11, 12 Thus, the possibility of a hypoactive RhoA and a hyperactive RhoC leading to a dormant phenotype in breast cancer cells is viable.
In this study, we set out to determine if altered RhoA and RhoC activity drives dormancy in breast cancer cells, particularly with FTI treatment. We compared the MCF-7 and MDA-MB-231 cells and found that the low metastatic MCF-7 cell line readily entered into dormancy due to hypoactivated RhoA and hyperactivated RhoC. Manipulating Rho levels changed the cells ability to enter or exit dormancy.
Material and Methods
All breast cancer cell lines were obtained from American Type Culture Collection (Manassas, VA) and maintained as previously described.13 Cell lines were validated for authenticity by the Johns Hopkins Genetics Resource Core Facility. FTI L-744,832 was kindly provided by Dr. George Prendergast (Lankenau Institute for Medical Research) and treatment was performed as previously described.7 For cells treated with the JNK inhibitor SP600125 (EMD Biosciences, Gibbstown, NJ): 1000 MCF-7 and MDA-MB-231 cells were plated at a clonogenic density of 1,000 cells per well in six well plates and treated with 25 μM of FTI for 2 days, 25 μM SP600125, pharmacological inhibitor of JNK, was added in each well in 2 ml of medium along with 0.25% sterile dimethyl sulphoxide (DMSO). As control, 0.25% DMSO in 2 ml of medium was added to the FTI-alone treated cells.
Induction of in vitro dormancy and counting of dormant clones
MCF-7 and MDA-MB-231 breast cancer cells were cultured in DMEM/10% FBS ±10 ng/ml recombinant human (rh)FGF-2 (R&D Systems, Minneapolis, MN) on plastic or extracellular matrix (ECM)-coated plates (BioCoat, Becton Dickinson, Lincoln Park, NJ). Cells were seeded at a clonogenic density of 1,000 cells/well in 6-well plates, visualized at 5, 10 and 15 days by staining with 0.1% crystal violet in 2% ethanol/10 mM sodium borate (pH 9.0) and growing (≥8, 24 or 100 cells on days 5, 10 or 15, respectively) or dormant (≤10 cells) clones were counted. Data are represented as percent growing versus percent dormant colonies with 500 colonies being counted.
Rho activation assays
Active levels of RhoA were measured using the G-LISA™ RhoA Activation Assay Kit, which is based on luminescence (Cytoskeleton, Denver, CO) per the manufacturer's specifications. A RhoC activation assay was performed as previously described using a GST-C214, 14 and detecting with a chicken anti-RhoC developed by our laboratory.
Specific shRNAs for human RhoA and RhoC were designed using Ambion's shRNA target finder. Hairpins targeting specific sequences of each GTPase were synthesized and cloned into pSilencer2.1-U6 neo (Ambion, Austin, TX). Details of these hairpin sequences for scrambled control (shScr), RhoA and RhoC were previously described and validated.11 Specific and significant downregulation of RhoA and RhoC was demonstrated using qPCR and western blot analysis. For the shRhoC experiments, cells were plated onto human Fn-coated 22 mm round coverslips (BD Biocoat) and treated with ±10 ng/ml rhFGF-2 (R&D Systems, Minneapolis, MN). Numbers of growing and dormant clones were counted after 5 days in culture. After 5 days, shRhoC was introduced into these cells using FuGENE HD transfection reagent (Roche, Indianapolis, IN), the numbers of dormant and growing clones were counted at 7 days (5 days in culture + 2 days post-transfection) and 10 days (5 days in culture + 5 days post-transfection). For shRhoA experiments, the constructs were transfected into cells by using FuGENE HD transfection reagent (Roche, Indianapolis, IN).
Western blot analysis
Western blotting was performed as previously described.15 Briefly, protein was harvested from cells using RIPA buffer. Lysates were separated by SDS-PAGE on a 4-20% (w/v) gel (Biorad, Hercules, CA) transferred to nitrocellulose, blocked and probed with a monoclonal antibodies for phosphorylated and total JNK, p38, ERK 1/2 (Cell Signaling, Beverly, MA) and β-actin (Sigma, St. Louis, MO). After incubation with a goat anti-rabbit-horseradish peroxidase (HRP) or goat anti-mouse-HRP antibody (Santa Cruz), immunoblots were developed with ECL, exposed to Hyperfilm (Amersham, Piscataway, NJ) and images recorded on an Alpha Image 90 documentation system (Alpha Innotech, San Leandro, CA).
Dormancy was induced for 5 days, the cells fixed with 4% paraformaldehyde and incubated with either Alexa-Fluor488 phalloidin (Cytoskeleton) or mouse anti-Ki67 (Lifespan Biosciences, Seattle, WA). For Ki-67 staining an AlexaFluor 555 goat anti-mouse IgG (Invitrogen Molecular Probes, Carlsbad, CA) secondary antibody was used. Immunoflourescence was performed on a Zeiss LSM5 High-speed Live confocal microscope housed in Delaware Biotechnology Institute. X–Z scans were performed on the same microscope using the same conditions as immunofluorescence.
FTI treatment induces a dormant phenotype in breast cancer cells
MCF-7 and MDA-MB-231 cells were treated with 25 μM FTI L-744,832. Only the MCF-7 cells had a marked reduction in growth and exhibited a spread-out morphology as shown in Figure 1a. A small number of MDA-MB-231 cells displayed a similar flattened morphology but growth was not significantly altered. Lack of response of the MDA-MB-231 to the FTI R115,777 has been reported.16 On withdrawal of FTI, the MCF-7 cells regained a normal morphology and proliferation pattern. RhoC activation (Figure 1b) was increased on average 3.8- and 1.7-fold in the MCF-7 and MDA-MB-231 cells after FTI treatment, respectively. The increase in MCF-7 active RhoC is similar to what was observed in the IBC cells.7 Figure 1c demonstrates that there is a 4.2- and 2.1-fold reduction in active RhoA in the MCF-7 and MDA-MB-231 cells after FTI treatment, respectively. FTI withdrawal resulted in resumption of MCF-7 cell growth within 48 hr, concurrently levels of active RhoA and RhoC returned to pre-treatment levels (Fig. 1d). These data suggest a reciprocal balance of RhoA and RhoC activation after FTI treatment. Both the altered morphology and drastic changes in RhoA activation are reminiscent of what is described for an in vitro model of breast cancer cell dormancy.8, 9, 17
Establishing an in vitro model of dormancy
To determine whether altered Rho GTPase activity drives the dormant phenotype, we re-established an in vitro model of dormancy comparing the MCF-7 and MDA-MB-231 cell lines. Growth of MCF-7 cells on fibronectin (Fn) with recombinant human basic fibroblast growth factor (rhFGF-2) has been shown to induce dormancy.8, 9 We began by comparing the number of growing colonies on various substrates in the absence of rhFGF-2: plastic, type I collagen (col) and Fn. There was no significant difference in the number of growing MCF-7 or MDA-MB-231 colonies, defined as having >29 cells, on any of the substrates (data not shown). Figure 2a demonstrates that breast cancer cells grown on plastic in the presence of rhFGF-2 or on Fn alone, form few dormant colonies. However, when grown on Fn in the presence of rhFGF-2 (Fig. 2b) a significant number of MCF-7 cells become dormant. A significant number of MCF-7 cells also become dormant when grown on col in the presence of rhFGF-2; however, the level of significance is less than Fn. The dormant state of the MCF-7 cells was confirmed by the absence of Ki-67 expression, suggesting that the cells are in G0 (Supporting Information Figure S1). In contrast, fewer MDA-MB-231 cells entered dormancy on either Fn or col in the presence of rhFGF-2 and the majority of cells expressed Ki-67. The number of growing MDA-MB-231 cells on Fn was significantly higher than the number of dormant cells. Figure 2c are representative micrographs showing the morphology of MCF-7 cells grown on Fn in the presence and absence of rhFGF-2 over a 15-day period. In the presence of rhFGF-2, the cells have morphology similar to FTI treatment, suggesting involvement of the Rho proteins.7
Altered Rho GTPase activity leads to a dormant phenotype
The Rho GTPases are intimately associated with the cytoskeleton; therefore, we visualized actin organization of the MCF-7 cells (Fig. 2d). Compared with the actively growing cells, the actin filaments in the dormant cells were short and disrupted. X–Z scans show that dormant MCF-7 cells have a diameter 10-fold greater than the actively growing cells. In addition, dormant cells are an average of 5.3-fold thinner than growing cells. Reflection interference contrast (RIC) imaging shows that the dormant cells have an increased number of contact points along the cell periphery suggesting that the cells are tightly bound to their substrate.
To determine whether the observed morphology and actin organization was due to altered Rho activity, we measured levels of active RhoA in the MCF-7 and MDA-MB-231 cells growing on Fn ± rhFGF-2. MCF-7 cells growing on Fn +rhFGF-2 had significantly lower levels of active RhoA GTPase as compared with those without rhFGF-2 (Fig. 3a). Under same conditions, the MDA-MB-231 cells did not display a reduction in RhoA activity. Next we measured RhoC activity. Because of technical limitations we were not able to directly perform a RhoC activation assay on the dormant tumor cells. Alternatively, we introduced a RhoC-specific shRNA into the cells and determined the number of dormant and growing colonies over a 5-day period post-transfection. Figure 3b demonstrates that the number of growing colonies significantly increased in shRhoC-transfected MCF-7 cells. Although there was a slight increase in the number of growing shRhoC-treated MDA-MB-231 colonies, it was not significant. These data suggest that active RhoC contributes to dormancy. The MDA-MB-231 cells are one of the few cell lines in which RhoA and RhoC function appears redundant.13, 18, 19 This may explain why these cells are resistant to FTI treatment and do not readily exhibit a dormant phenotype.
To demonstrate the role of the Rho GTPases in dormancy, we transfected the MCF-7 cells with a constitutively active RhoC (RhoCG14V) and/or shRhoA and plated them on either glass or Fn (Fig. 3c). Because there are concerns that dnRhoA is not absolutely specific for RhoA and may have effects on RhoC, we decided to use shRhoA (unpublished observations). Similar problems in specificity were observed with dnRac1, which affected Rac3 activity.20 There was no significant difference between cells that were untreated and those that were treated with the scrambled control (shScr). A significant increase in dormant colonies was observed when cells were grown on Fn and had either forced increased RhoC activity or decreased RhoA expression. When RhoCG14V and shRhoA were introduced together, a significant increase in dormant colonies was observed regardless of substrate. This suggests that both increased RhoC and decreased RhoA activity are required for dormancy and that attachment to Fn impinges on both the RhoA and RhoC pathway. Figure 3d demonstrates the efficacy of the short hairpin RNAs used to deplete RhoC and RhoA GTPases. There was an 83% and 65% decrease in RhoC and RhoA protein levels, respectively, with the introduction of shRNAs.
Next, we transfected the MCF-7 cells with shRhoC and treated with FTI. Downregulation of RhoC expression led to a decreased number of cells responding to FTI treatment, suggesting a role for RhoC in FTI action (Fig. 4a). We then treated the MCF-7 and MDA-MB-231 cells with FTI to determine which signaling pathways were being activated. Since the mitogen activated protein kinases are common downstream effectors of Rho GTPases and have been implicated in tumor cell dormancy, we investigated the activation levels of p38MAPK, ERK1/2. No changes were observed in activation of the p38 or ERK 1/2 pathways (Supporting Information Figure S2) therefore we next examined the JNK stress-activated protein kinase pathway. As shown in Figure 4b, activation of the JNK/SAPK pathway in the MCF-7 was increased, whereas remaining relatively unchanged in the MDA-MB-231 cells.
Finally, to determine whether JNK/SAPK signaling is involved in FTI-induced dormancy, we treated cells with SP600125, a pharmacologic inhibitor of JNK and FTI. Figure 4c shows that pharmacologic inhibition of JNK/SAPK leads to a significant 50% decrease in nongrowing MCF-7 cells with a significant reciprocal increase in growing cells as compared with cells treated with FTI alone. In contrast, JNK inhibition did not change the ratio of nongrowing to grown MDA-MB-231 cells.
FTIs were designed to treat Ras-based tumors; however, their clinical application proved to be disappointing.21 The effect of FTI on Ras-induced transformation was later shown to be because of interference of Rho activity, specifically RhoB GTPase.22, 23 Our group demonstrated that IBC cells were susceptible to FTI treatment, with a reversion of the RhoC GTPase phenotype.7 FTI reversion of the RhoC-induced IBC phenotype could be phenocopied with the introduction of a RhoB GTPase that could only be geranylgeranylated.7 The morphology of the FTI-treated cells resemble that of cells induced in vitro to undergo dormancy.
Breast tumor cell dormancy is a significant clinical problem that faces every woman with breast cancer.24, 25 Many have hypothesized that the microenvironment influences whether a tumor cell will grow or become senescent over an indefinite period of time and that the Rho GTPases and cytoskeleton are involved.8, 17, 26 Here, we demonstrate that FTI treatment or induction of dormancy with Fn +rhFGF-2, reversibly perturbs RhoA and RhoC GTPase activation; specifically hypoactivating RhoA and hyperactivating RhoC. Recent reports from Wieder's group suggest that decreased RhoA GTPase activity is necessary but not sufficient to induce dormancy.9, 10 In these studies, general Rho inhibitors such as C3 exoenzyme and a dnRhoA were used, which could also drive down the activation of RhoC GTPase thereby supporting our observations. Interestingly, MDA-MB-231 cells are reported to be refractory to FTI R115,777 treatment.16 In this study, we saw that they have a limited response to FTI L-744,832. This may be due to an apparent redundancy in RhoA and RhoC function in these cells.13, 18, 19
RhoC GTPase is involved in conferring a metastatic phenotype to a number of cancers.27 Our laboratory has also shown that RhoC expression is a prognostic marker for metastatic spread of small breast cancers that are <1 cm in size.28, 29 Thus, a small breast cancer that may appear to have a good prognosis, may have already spread. RhoA GTPase is also suggested to play a role in progression to metastasis in a number of cancers.30, 31 However, overwhelming evidence suggests that increased activation of RhoA leads to stable stress-fiber formation and reduced motility.32–34 Therefore, to be effective in producing a motile cell the Rho GTPases must dynamically and transiently cycle from a “GDP-off” to a “GTP-on” back to a “GDP-off” state.34–36 This transient cycling can be influenced by the microenvironment of the cancer cell.36 Thus, the concept that Rho GTPases can control the senescent state of a cell is viable. Rho-induced cytoskeletal shape change could alter gene transcription, potentially activating cellular programs such as those for autophagy.37, 38 Indeed, it is shown that signaling via ECM or changes in tissue architecture alters gene transcription by rearranging the nuclear lamina (reviewed in Ref.39). This in turn leads to expression of autophagy genes.38 We are in the process of exploring this hypothesis in FTI treated and Fn + rhFGF-2 treated breast cancer cells.
We found that the JNK/SAPK pathway was activated in the FTI treated cells and that inhibition of JNK decreased the number of dormant cells in vitro. It has been suggested that in the case of angiogenic dormancy, increased p38 relative to ERK 1/2 signaling is indicative of dormancy.40 Previous work from our laboratory suggests that RhoC GTPase signals for motility through p38 signaling in pancreatic and IBC cells.15, 41 Therefore, it is interesting that increased RhoC activity did not affect phospho-p38 levels in either the MCF-7 or MDA-MB-231 cells. Previous reports suggest that treatment of MCF-7 cells with rhFGF-2 induces apoptosis.42 Therefore, Fn may be acting as the main external signal to promote survival of the MCF-7 cells.43 Induction of the JNK/SAPK pathway is also shown to be required for the induction of autophagy through different pathways.44, 45 Much of the literature points to Rac1 GTPase in the activation of the JNK/SAPK pathway, particularly in ECM signaling in MCF-7 cells.46 However, there are reports that RhoC GTPase can induce JNK signaling in 293T cells.47 We have previously shown in PC-3 prostate cancer cells that RhoC activity negatively modulates Rac1 expression and activation.11, 48 A reciprocal relationship between RhoA and Rac1 was also previously reported by John Collards group.49 We therefore are beginning to determine whether there is a similar interplay between RhoC, RhoA and Rac1 in FTI treated or Fn + rhFGF-2 treated breast cancer cells.
Taken together our data suggests that FTI treatment of some breast cancer cells may work by placing the cells in a state of dormancy. In vitro, dormancy can be driven by perturbing RhoA and RhoC GTPase activation. These data may suggest new applications for FTI treatment of breast cancer, possibly identifying patients that would benefit from adjuvant FTI treatment.50 Potentially, FTI-induced dormancy could synchronize tumor cells; FTI withdrawal would allow growth making the cells more susceptible to chemotherapeutics. In addition, this study may shed light on the mechanisms of breast cancer dormancy.
- 1American Cancer Society. Cancer Facts & Figures 2010 ed., 2010. pamphlet.
- 3The current understanding of the molecular determinants of inflammatory breast cancer metastasis. Clin Exp Metastasis 2006; 66: 615–20., .
- 10FGF-2-induced breast cancer dormancy in an in vitro model is maintained through integrin alpha5beta1 signaling. Am Assoc Cancer Res 2007; 100., .
- 19Intravenous delivery of anti-RhoA small interferring RNA loaded in nonoparticles of chitosan in mice: safety and efficacy in xenografted aggressive breast cancer. Hum Mol Genet 2008; 17: 1019–26., , , , , , , , , , .
Additional supporting information may be found in the online version of this article.
|IJC_25655_sm_suppfig-1.tif||2704K||Supporting Figure 1|
|IJC_25655_sm_suppfig-2.tif||16906K||Supporting Figure 2|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.