Introduction

Breast carcinoma affects 12% of women, 40% of who eventually die from metastatic disease [1]. Breast cancer metastasis results from the accumulation of multiple genetic alterations in mammary epithelial cells, and many of the complex molecular pathways underlying this pathological process remain to be characterized [2].

In our previous studies, we started by screening a human mammary cDNA library in order to investigate new mechanisms involved in breast carcinoma growth and progression. From these studies, the receptor for activated C-kinase 1 (RACK1) emerged as a target for future experiments, and its potential as a diagnostic and prognostic factor for breast carcinoma was initially observed [3, 4]. RACK1 is a scaffold protein with a propeller-like structure of seven WD-40 repeats. It was originally identified on the basis of its ability to bind the activated form of protein kinase C, stabilize this protein and facilitate its trafficking within the cell [5]. Previous studies have shown that the WD-40 repeats play a role in complex protein–protein interactions between signaling molecules such as integrins, phosphodiesterase 4D5, and Src tyrosine kinase, as well as protein kinase C (PKC) [6]. As a result, it may be involved in diverse cellular processes, including signal transduction, immune response, as well as cell growth, migration and differentiation [7, 8].

Accumulating evidence suggests that activation of small GTPase proteins such as RhoA and Rho kinase, a downstream effector, are critical for tumor invasion and/or metastasis [9–11]. RhoA stimulates the assembly of contractile actomyosin filaments and associated focal adhesion complexes [12]. However, the role of RhoA/Rho kinase activation in breast carcinoma is still unclear.

In this article, we demonstrate that RACK1 promotes breast carcinoma migration/metastasis by interacting with RhoA and activating the RhoA/Rho kinase pathway in both breast carcinoma cell lines and 160 breast cancer patient samples. RACK1 is a prognostically important factor that promotes breast carcinoma migration/metastasis in vitro and in vivo.

Materials and methods

Cell culture and transfection

Four human mammary carcinoma cell lines with different invasion abilities were used in our experiment: MCF7 and T-47D with low invasion abilities as well as MDA-MB-231 and the MDR (multidrug resistant) counterpart of MCF7, MCF7/ADR [13], with relatively higher invasion abilities [3]. All cells were grown in the recommended medium supplemented with 10% fetal bovine serum (Gibco, Grand Island, USA), 2 mmol/l glutamine, 50 units/ml penicillin, and 50 units/ml streptomycin; cells were maintained in a 37°C atmosphere.

Transient transfection of cells was achieved with Lipofectamine 2000 reagent (Invitrogen, Carlsbad, USA). Cell lines stably expressing RACK1 were generated by transfection and selection with G418 antibiotic (Gibco) as described previously [3]. We used siRNAs targeting the human RACK1 (GNB2L1) gene sequence, 5′ CAGATTGTCTCTGGATCTCGA 3′, which were purchased from QIAGEN (Duesseldorf, Germany).

Immunoprecipitation and immunoblotting

Forty-eight hours after transfection, cells were collected, washed with phosphate-buffered saline (PBS) (pH 7.4), and lysed in modified RIPA buffer (50 mM Tris (pH 7.8), 150 mM NaCl, 5 mM EDTA, 15 mM MgCl2, 1% NP-40, 0.5% sodium deoxycholate, 1 mM DTT, and 20 mM N-ethylmaleimide) supplemented with 1 tablet/50 ml of Complete Protease Inhibitor Cocktail (Roche Molecular Biochemical, Indianapolis, USA). Lysates were cleared by centrifugation (10,000×g for 15 min at 4°C) and incubated on ice for 2 h with 1–2 μg of the relevant antibody (Ab) (anti-GFP, Roche; anti-HA, anti-RhoA, Sigma–Aldrich, St. Louis, USA; anti-RACK1, BD Biosciences, San Jose, USA). Next, 20 μl of slurry protein A/G PLUS-Agarose immunoprecipitation reagent (Santa Cruz Biotechnology, Santa Cruz, USA) was added to each lysate, and the mixtures were incubated with rotation for at least 2 h at 4°C. The beads were retrieved by centrifugation and washed (by vortex and short spin) three times with RIPA buffer and once with PBS. Proteins bound to the beads were eluted by boiling in 2× electrophoresis sample buffer and separated by SDS-PAGE as described previously [3].

Immunofluorescence analysis

Cells seeded on 6-well chamber slides were fixed in 4% paraformaldehyde, permeabilized in 0.1% Triton X-100 and blocked in 5% BSA. Protein levels were detected using the indicated primary antibodies, followed by incubation for 45 min with Cy3-conjugated or FITC-conjugated secondary antibodies (Amersham Biosciences). The coverslips were washed, mounted in phosphate-buffered saline containing 50% glycerol, and viewed on a Leica laser scanning confocal microscope (or the automated microscope Leica DMRXA2) equipped with a Photometrics Cool SnapES N&B camera driven by MetaMorph software (Universal Imaging Corporation, Downingtown, USA).

Migration analysis

The scratch test and modified Boyden dual chamber assay were performed as previously described [3]. The Rho kinase inhibitor, Y-27632, was purchased from Calbiochem (Darmstadt, Germany).

Determination of the level of GTP-bound RhoA

GTP-Rho pull-down assays were performed according to the supplier’s protocol (Upstate Biotechnology, Lake Placid, NY). Briefly, cells treated with the indicated chemicals were washed twice with ice-cold TBS, scraped off the plates in lysis/wash buffer (50 mM Tris–HCl at pH 7.4, 2 mM MgCl2, 1% lgepa, 10% glycerol, 100 mM NaCl, supplemented with 1 mM dithiothreitol, 10 μg/ml leupeptin, and 10 μg/ml aprotinin) on ice. The lysates were incubated with the Rhotekin RBD-agarose slurry at 4°C for 45 min with gentle agitation and then washed three times with wash buffer. The bound proteins were boiled and eluted with 2× sample buffer and detected by SDS-PAGE as described above.

Patient characteristics and immunohistochemistry

Paraffin-embedded tissue samples from 160 mammary carcinomas were prepared as previously described [4]. All patients with primary mammary carcinoma underwent curative surgery at the Huashan Hospital of Fudan University between January 1999 and December 2002. None of the patients received chemotherapy or radiation therapy before surgery, and after successful radical mastectomy, all 160 patients only received four cycles of cyclophosphamide, methotrexate, and 5-fluorouracil (CMF). All patients were followed after surgery (until December 31, 2009), and detailed and complete clinicopathological data were collected on each patient. The follow-up time ranged from 1.5 months to 130 months, with a median time of 81 months. At the end of the follow-up period, 132 patients were still alive and 28 had died of the disease.

Analysis of RACK1 protein expression (including positive and negative control samples) was carried out and scored as previously described using a RACK1 mouse monoclonal antibody (Lifespan, Seattle, USA) [4]. Briefly, total RACK1 staining was scored based on the intensity and percentage of cells with RACK1 cytoplasmic staining on the following scale: score 0, negative/weak staining for all tumor cells or moderate staining less than 30%; score 1, moderate staining in more than 30% but less than 70% of tumor cells or strong staining within 30% of tumor cells; score 2, moderate staining in more than 70%, or strong staining in more than 30% of all tumor cells. RhoA staining using the rabbit polyclonal antibody (Santa Cruz Biotechnology) was scored as previously described [9].

The clinical studies included the essential elements of “Reporting recommendations for tumor marker prognostic studies (REMARK)” [14].

Statistical evaluation

All data were representative of at least three independent experiments with similar results. The results are expressed as the mean (±1 SEM) from multiple experiments. A student’s t test was used to determine significant differences (two-tailed, P < 0.05). Pearson’s correlation coefficients were used to determine whether two prognosis-related factors were correlated to each other over all cases. Kaplan–Meier survival analysis was used to estimate the prognostic relevance of RACK1, and the disease-free survival difference between groups was assessed by the log-rank test. The COX proportional hazards model was used for multivariate analysis of survival. The software SPSS 15.0 for Windows was used for statistical analyses, and P < 0.05 was considered statistically significant.

Results

Up-regulation of RACK1 promotes migration of breast carcinoma cells

Cell lines stably expressing RACK1 were generated as described previously [3]. The effects of RACK1 on breast carcinoma cell migration (MCF7, clone M-R4; T-47D, clone T-R2) were assessed by scratch tests (Fig. 1a) and transwell assays (Fig. 1b). As shown in Fig. 1a, b, upregulation of RACK1 in both MCF7 and T-47D induced a significant increase in the cells’ migration capacity.

Fig. 1
figure 1

Up-regulation of RACK1 promotes migration in breast carcinoma cells. Scratch tests and migration assay (using a modified Boyden chamber) showing increased migration in RACK1 upregulated breast cancer cell lines (a and b) and verification using siRNA specifically targeted to RACK1 (b and c). The increase in migration caused by RACK1 in breast carcinoma cells (MCF7, clone M-R4; T-47D, clone T-R2) was significantly inhibited after administering the Rho kinase inhibitor, Y-27632 (10 μM) (d and e)

Full size image

In order to verify the effects of RACK1, we introduced siRNAs specifically targeting RACK1 into breast carcinoma cells (MCF7/ADR and MDA-MB-231) and repeated the migration assay (Fig. 1b, c). These data confirmed that RACK1 promotes migration in breast carcinoma cells.

RACK1 promotes migration of breast carcinoma cells through interaction with RhoA

In order to explore possible mechanism(s) underlying RACK1 regulation, we pretreated breast carcinoma cells (MCF7, clone M-R4; T-47D, clone T-R2) for 1 h with the Rho kinase inhibitor Y-27632 (10 μM) and plated them. Both the scratch test and migration test (using the transwell technique) were performed. Compared to the solvent control, inhibition of Rho kinase dramatically reduced the increase in cell migration induced by RACK1 (Fig. 1d, e).

Accordingly, we investigated interactions between RACK1 and RhoA using MCF7 and MCF7/ADR via immunoprecipitation (Fig. 2a) and immunofluorescence analyses (Fig. 2b). Co-immunoprecipitation was performed in both MCF7 cells stably expressing RACK1 and MCF7/ADR cells transiently transfected with pEGFP-N3-RACK1. As shown in Fig. 2a, RACK1 was co-immunoprecipitated with RhoA. This interaction was further confirmed using MCF7 and MCF7/ADR cells transiently transfected with pcDNA3.0-HA-RhoA (Fig. 2b).

Fig. 2
figure 2

RACK1 interacts with RhoA. a MCF7 and MCF7/ADR cells were transfected with pEGFP-N3-RACK1 or pcDNA3.0-HA-RhoA. Cell lysates were subjected to immunoprecipitation (IP) with anti-GFP followed by immunoblotting (IB) with anti-RhoA or IP with anti-HA followed by IB with anti-RACK1. b Co-localization of RACK1 and RhoA in MCF7-RACK1 (clone M-R4) and MCF7/ADR cells. The panels show representative fields. Co-localization is shown by yellow cellular staining and occurred mainly in the cytoplasm

Full size image

Furthermore, immunofluorescence analysis revealed that RhoA co-localized with RACK1 mainly in the cytoplasm of MCF7 and MCF7/ADR cells, especially in the cytoplasm of MCF7 cells stably transfected with pEGFP-RACK1 (Fig. 2c).

Up-regulation of RACK1 promotes the RhoA activation

Rearrangement of the actin cytoskeleton, which precedes the formation of pseudopods, is regulated by Rho family small G proteins including RhoA [15]. We tested the possibility that RACK1 activates this small G protein in breast carcinoma cells. First, we examined the involvement of RhoA in the cell line stably expressing RACK1. The level of active, GTP-bound RhoA significantly increased in MCF7-RACK1 and T-47D, whereas total RhoA remained unchanged (Fig. 3a). Next, we found that the level of GTP-bound RhoA was significantly decreased when siRNA targeting RACK1 was introduced in MCF7/ADR and MDA-MB-231 cell lines as well as when Y-27632 was introduced in MCF7 and T-47D cells (Fig. 3b). These results show that RACK1 enhanced the activation of RhoA, which is a mechanism underlying the promotion of cell migration by RACK1.

Fig. 3
figure 3

Elevated RACK1 promotes the activation of RhoA. GTP-Rho pull-down assays showing the activation of RhoA in RACK1-upregulated breast cancer cells and verification using RACK1 siRNA. The activation of RhoA caused by RACK1 in stably transfected breast carcinoma cells (clone M-R4 and T-R2) could be significantly inhibited after the administration of the Rho kinase inhibitor, Y-27632

Full size image

RACK1 significantly correlates with indicators of clinical spread and RhoA in 160 breast carcinoma patients

In order to verify the in vitro results mentioned above, we performed immunohistochemical staining for RACK1 and RhoA in 160 breast carcinoma samples. Table 1 shows the correlation of RACK1 and patients’ major clinicopathological characteristics. Both the RACK1 and RhoA were readily detected in the cytoplasm (Supplementary Fig. S1). Moreover, we observed a significant correlation between RACK1 expression and tumor size (P = 0.046), lymph node metastasis (P = 0.027) (Fig. 4a, Table 1), and percentage of carcinoma in situ (the percentage of in situ components in all carcinoma tissues per patient; r = –0.691, P < 0.001; Fig. 4a, b), which are commonly used indicators closely related to metastasis in breast carcinoma. More intriguingly, RACK1 significantly correlated with RhoA (r = 0.814, P < 0.001; Table 2). Further, a Kaplan–Meier survival analysis of 160 carcinoma cases revealed a correlation between higher RACK1 expression levels and shorter survival times (P < 0.001, Fig. 4c).

Table 1 Correlations of RACK1 and major clinical pathologic factors (n = 160)
Full size table
Fig. 4
figure 4

The correlations of RACK1 and clinicopathological factors. RACK1 correlates with tumor size, lymph node metastasis (a), percentage of carcinoma in situ (a and b) (all P < 0.05). c Kaplan–Meier analysis of disease-specific survival in 160 breast carcinoma patients stratified by RACK1 expression level (P < 0.001)

Full size image
Table 2 The significant correlation of RACK1 expression with RhoA
Full size table

Discussion

In this study, a series of experiments (in vitro and in vivo) were designed to evaluate the role of RACK1 in breast carcinoma migration/metastasis. First, we found that upregulation of RACK1 significantly promoted breast carcinoma migration/metastasis by interacting with RhoA and activating the RhoA/Rho kinase pathway.

In our previous studies, we screened a human mammary cDNA library (using the PI3K p110α kinase domain as bait) to identify novel PI3K p110α-interacting proteins. From the candidate genes, we identified RACK1 as a putative target in human breast carcinoma cells. Further experimentation indicated that RACK1 promotes breast carcinoma proliferation in vitro and in vivo. In addition, these studies suggested the potential involvement of RACK1 in breast carcinoma invasion/metastasis [3, 4].

In this study, we performed migration assays including the scratch test and modified Boyden dual chamber assay in four breast carcinoma cell lines with different migratory properties accompanied with different expression levels of RACK1 [3]. In order to test the results obtained from RACK1 upregulation, we also applied siRNA targeting RACK1 in MCF7/ADR and MDA-MB-231 and the results were the same as we have observed in MCF7 [3]. Immunohistochemistry and statistical analyses of 160 breast carcinoma samples showed that increased RACK1 expression correlated with tumor size, lymph node metastasis, and percentage of carcinoma in situ, all of which are clinical indicators for tumor spread. More intriguingly, RACK1 significantly correlated with RhoA. Furthermore, Kaplan–Meier survival analysis indicated a significant correlation between higher RACK1 expression levels and shorter overall survival times. Our in vitro and in vivo experiments demonstrated that RACK1 promotes breast carcinoma migration/metastasis. In recent years, a number of studies have focused on the role of RACK1 in cell migration control mechanisms; however, these studies have been limited to different mammalian cell lines [6]. Cells that over-express RACK1 demonstrate enhanced spreading, an increased number of actin stress fibers, focal contacts, and enhanced tyrosine phosphorylation of both focal adhesion kinase and paxillin [8, 16]. Conversely, reduction of RACK1 expression in NIH3T3 cells by antisense depletion blocked cell spreading [8].

Another break-through of our present research is that genuine RACK1/RhoA interaction was demonstrated for the first time using immunoprecipitation and immunofluorescence in MCF7 and MCF7/ADR cells. The Rho kinase inhibitor further demonstrated the involvement of the RhoA/Rho kinase pathway. Multiple reports have reported that versatile interactions exist between RACK1 and central signaling molecules including integrins, phosphodiesterase 4D5, Src tyrosine kinase, and protein kinase C (PKC) [5, 6, 17]. The basis of these complex protein–protein interactions lie in the WD-40 repeats. RACK1 is a 36 kDa cytosolic protein with a propeller-like structure consisting of seven WD-40 repeats that are compatible with the consensus WD motif used to identify members of the WD repeat family. Its amino acid sequence is 100% identical in humans, rats, chickens [18], mice [19], and cows [20], and RACK1 is ubiquitously expressed in a wide range of tissues, including the brain, liver, and spleen [21]. This spatial–temporal conservation has prompted a suggestion that RACK1 is important in a variety of biological functions including signal transduction, immune response, as well as cell growth, migration, and differentiation [6].

We also performed GTP-Rho pull-down assays to assess the role of RhoA activation in the promotion of migration induced by RACK1. We found the level of active form of RhoA significantly increased in MCF7-RACK1 and T-47D-RACK1 compared with parental cells, while the total amount of RhoA remained unchanged. Further, the RhoA activation could be significantly decreased by the introduction of siRNA targeting RACK1 into MCF7/ADR and MDA-MB-231 cells. In line with our findings, a causal effect of Rho GTPases in breast cancer has been shown using a model of breast carcinoma in rodents [22]. Other studies have indirectly shown an important role of RhoA in breast carcinogenesis [23, 24]. It is known that uncontrolled cell migration is necessary for tumor invasion and metastasis, which requires actin polymerization, depolymerization, changes in the integrin-mediated membrane attachment to the substrate, and modifications in cell shape. These changes are produced, at least in part, by rearrangements of the actin cytoskeleton, which is also a crucial step in tumor progression [25]. Several members of the Ras homologous (Rho) family of small GTPases (namely Rho, Rac, and Cdc42) act as key regulators of the actin cytoskeleton [26, 27]. RhoA, a regulator of cell migration [28], is one of the most important members of Rho GTPases. Like all members of the small GTPases superfamily, this protein cycles between an inactive GDP-bound form, and an active GTP-bound state. In their GTP-bound (active) form, RhoA GTPases are localized to membranes and are able to interact with effector molecules, thereby initiating downstream responses. RhoA regulates the organization of the actin cytoskeleton of cells and is essential for the regulation of cell shape, polarity, motility, and adhesion [29]. The cytoskeletal changes elicited by RhoA, including neurite retraction, stress fiber formation, and focal adhesion assembly are mediated mostly by Rho kinases [30]. Studying the role of Rho kinases in tumorigenesis has been greatly facilitated by several compounds that specifically target their activity [31], the prototype of which is Y-27632. This compound inhibits focus formation and growth in soft agar of Rho- and Ras-transformants [32]. There is an extensive body of evidence indicating that Rho kinase activity is directly implicated in the invasive behavior of tumor cells derived from different human tumors. For instance, inhibition of Rho kinase via Y-27632 or expression of a dominant negative mutant abrogates invasion of rat MM1 hepatoma-induced tumors in vivo [33]. Similarly, in our experiment, human breast carcinoma cells stably transfected with RACK1 treated with Y-27632 show significant inhibition of in vitro tumor migration, suggesting that the RhoA/Rho kinase pathway might be one of the mechanisms underlying breast carcinoma migration/metastasis induced by RACK1.

Moreover, it has been reported that RhoA-mediated focal adhesion kinase (FAK) phosphorylation can activate PI3K and AKT [34]. In addition, RhoA/Rho kinase pathway may mediate epithelial–mesenchymal transition (EMT) induced by TGF-β1 in rat peritoneal mesothelial cells [35]. In our previous findings, we identified RACK1 as a binding partner of both PI3K p110α [3] and p-AKT (unpublished results). RACK1 up-regulation led to PI3K/AKT pathway activation [3] and EMT in breast cancer cells, accompanied by enhanced invasive phenotype (unpublished results). These results supported a hypothesis that RACK1 exposes the propeller architecture with seven blades as a platform for interactions with versatile kinases and crosstalk between pathways such as the PI3K/AKT and RhoA/Rho kinase pathways. Applying this platform, RACK1 works as a cog in cell which can recruit specific proteins to the appropriate location for a particular function.

In summary, this study is the first to present evidence that RACK1 promotes breast carcinoma migration/metastasis through interaction with RhoA and activation of the RhoA/Rho kinase pathway. In the near future, RACK1 could be included in a broadened panel of biomarkers used for more accurate diagnosis and prognosis in breast cancer patients.