The E3-ligase E6AP Represses Breast Cancer Metastasis via Regulation of ECT2-Rho Signaling Free

Deregulated tumor cell invasion contributes to the metastatic phenotype of cancer (1, 2). To metastasize, tumor cells detach from the primary tumor, invade surrounding tissues (3), intravasate into lymphatic and blood vessels, and extravasate into distant organs to initiate the formation of secondary tumors (4–6). Invasion is initiated and maintained by signaling pathways that control cytoskeletal dynamics and the regulation of cell–cell junctions (7). The Rho GTPase family of proteins, including Rac1, Cdc42, and RhoA, have a fundamental role in cell invasion and migration by regulating cytoskeletal remodeling and cell adhesion (8–10). Increased activity and aberrant protein levels of the Rho GTPases are observed in human cancers and are involved in all stages of cancer progression (11–13). The Rho GTPases cycle between an active GTP-loaded state and inactive GDP-loaded state (14) controlled by guanosine exchange factors (GEF) and GTPase-activating proteins (GAP; ref. 15). Epithelial cell transforming-2 (ECT2) belongs to the DBL family of GEFs for the Rho GTPases (16). The best understood biologic function of ECT2 is regulation of cytokinesis (17, 18). ECT2 can catalyze in vitro guanine nucleotide exchange on RhoA, Rac1, and Cdc42 (19). Both ECT2 and pbl, its Drosophila ortholog, are regulated by E6AP (Ube3a; ref. 20).

E6AP is an E3-ubiquitin ligase best known for its role in promoting proteasomal degradation of p53 in human papillomavirus (HPV)-infected cells (21, 22). We demonstrated that E6AP knock-out (KO) mouse embryo fibroblasts (MEF) exhibit enhanced proliferation and evasion of the cellular senescence (23). Importantly, Ras-transformed E6AP-KO MEFs form tumors in mice that are more invasive than their WT counterparts, indicating a role for E6AP in controlling tumor dissemination (23). In addition, recent studies indicate a role for E6AP in human breast cancer, where invasive ductal carcinoma has lower E6AP protein levels than normal breast in situ (24). Here, we report that loss of E6AP enhances invasiveness of the cells and their ability to seed distant metastases. Concomitantly, enhanced E6AP levels inhibited breast cancer metastasis via the regulation of ECT2/Rho GTPase pathway.

HA-WT-E6AP(pCDNA3-E6AP) and C820-E6AP(pCDNA3-C820E6AP) were gifts from Z. Nawaz (Baylor College of Medicine, Houston, TX). pEGFP-ECT2, pLL5.0-cherry, and pLL5.0-cherry-shECT2 were provided from Ratheesh and colleagues (25); pGIPs-shECT2 (Clone V3LHS-391452/391450/391448) was from Dharmacon, FH1t-shE6AP, FH1t-E6AP were cloned into FH1t-UTG using PacI site (26, 27).

mAbs against E6AP (clone-330; Sigma-Aldrich), Rac1, RhoA, and Cdc42 (BD Biosciences), and EpCAM (Abcam) were used. Polyclonal antibodies against E6AP (Clone-H-182), ECT2 (Millipore), Phospho-Myosin-Light Chain(Ser19; Thermo Fisher Scientific), and HSP60 (Santa Cruz Biotechnology) were used. HRP-labeled secondary antibodies, Alexa-488, Alexa-647, and Rhodamine-phalloidin were from Thermo Fisher Scientific. MG132 was from Calbiochem, EGF from Invitrogen, and doxycycline from Sigma-Aldrich.

Breast cancer lines [MCF10, SKBR3, T47D, HBL100, SUM159PT and human kidney epithelial 293T (HEK293T)] were purchased from ATCC. WT/E6AP-KO MEFs and mammary epithelial cells (MEC) were isolated and cultured as described previously (26, 28). MDA-MB-231-HM cells were obtained from X.Z. Chang (Fudan University, Shanghai, P.R. China; ref. 29). All human cell lines were authenticated by Cell Bank Australia in 2012 using STR profiling and tested for Mycoplasma by a reference laboratory (VIDRL).

MECs (EpCAM⁺) were resuspended in Matrigel and cultured in medium supplemented with 1% FCS as described previously (28). Acini formation and differentiation was induced using DME-HAM media containing 1 mmol/L glutamine, 5 mg/mL insulin, 500 ng/mL hydrocortisone, and 1% FCS. Acini were counted under DIC microscope as described in Supplementary Material.

Modified Boyden chamber assays were used as described previously (30) and in Supplementary Material. Briefly, cells were seeded in Matrigel, cultured in 0.1% FCS medium, and allowed to invade toward 10% FCS gradient. For transendothelial migration (TEM) assay, human umbilical vein endothelial cells (HUVEC) were grown to confluency on cell culture inserts and GFP tumor cells were added to the endothelial layer and left to transmigrate for 24 hours. Transmigrated cells were counted under the microscope (Olympus BX-51).

Transfections of cDNA were completed as described previously (31). Cells were lysed in 50 mmol/L Tris pH 7.4, 150 mmol/L NaCl, 1% Triton X100, 0.1% SDS, and protease inhibitors (Complete, Roche). For immunoprecipitation assay, 2 mg of total protein was used. Immunoprecipitates and lysates were analyzed by Western blot analysis with appropriate antibodies as described in Supplementary Material.

Rac, Cdc42, and RhoA activation assays were completed by affinity precipitation following the protocol in the activation assay kit from Cytoskeleton, Inc.

Cells were fixed in 4% paraformaldehyde, permeabilized, and stained with indicated antibodies as described in Supplementary Material. For IHC of tissues, the antigen retrieval was performed on FFPE sections of mammary fat pads in 1× citrate buffer or Tris-EDTA pH 9. Samples were blocked and stained with primary antibodies and Alexa Fluor before mounting in ProLong Gold with DAPI (Invitrogen) as described previously (28).

Stable cell lines expressing or downregulating E6AP were generated by lentiviral infection using HEK293T packaging cells and sorted on the basis of the expression of reporter EGFP or mCherry using FACS and as described in ref. 31 and Supplementary Material.

Ube3a/E6AP WT and KO mice were maintained on C57Bl/6 background as described previously (31). Female NOD-SCID-IL2Rgamma (NSG) mice were bred in-house. Doxycycline was injected intraperitoneally (i.p.) at a dose of 2 mg/kg. In addition, mice were provided water containing doxycycline (2 mg/mL). All animal experiments were completed with the approval of the Peter MacCallum Animal Ethics Experimentation Committee.

These techniques have been described in Supplementary Material. Briefly, tumor cells were injected into the tail vein for experimental metastasis assays or into the mammary gland for monitoring primary tumor growth and spontaneous metastasis after resection of the primary at 500 mm³. Metastatic burden was measured by gPCR and BLI, as described previously (32).

Data with clinical outcome (relapse) were obtained from the METABRIC (Molecular Taxonomy of Breast Cancer International Consortium) cohort (33). Data analysis is described in ref. 34 and in Supplementary Material.

Tissue microarrays were constructed from archival patient samples (Peter MacCallum Cancer Centre). The cores were immunophenotyped into luminal, basal, and HER2 subtypes. TMAs were immunostained for E6AP as described previously (35) and for ECT2 (20, 25, 36, 37) and scored as described in Supplementary Material.

All results were reported as mean ± SEM or mean ± SD as indicated in the figure legends. Metastasis data have been log₁₀ transformed for visualization. For comparisons of central tendencies, normally distributed datasets were analyzed using unpaired Student t test. All data including Kaplan–Meier survival data were analyzed using GraphPad Prism 5.

To investigate the function of E6AP in breast cancer, we first evaluated the prognostic value of E6AP in a breast cancer dataset from 1,992 patients (METABRIC; ref. 33), and found that low E6AP expression correlated significantly with poorer survival (Fig. 1A). Similar results were found by interrogating the BreastMark plotter database (Supplementary Fig. S1A). The HR from METABRIC samples confirmed the favorable prognostic value of high E6AP expression in breast cancer (Supplementary Fig. S1B) and that E6AP expression was reduced in the basal-like breast cancer subtype, but increased in luminal A and luminal B breast cancer subtypes compared with normal breast tissue (Fig. 1B).

The reduced expression of E6AP was confirmed at the protein level in a tissue microarray (TMA) of 100 human breast cancer samples. E6AP levels were categorized into low, intermediate, and high groups (Fig. 1C). Low and intermediate E6AP was found in 74% of the samples, with only 26% revealing high E6AP levels (Fig. 1D). The proportion of tumors with high E6AP protein decreased with increasing aggressiveness, as classified by tumor grade (Supplementary Fig. S1C).

Investigation of the correlation between E6AP levels and the tumor subtype revealed that of the cancers with low E6AP, 75% were ER⁺/PR⁺ and 20% were basal/triple-negative breast cancer (TNBC; Fig. 1E). Conversely, of the samples with high E6AP, 50% were ER⁺/PR⁺ and only 8% were basal (Fig. 1E). Importantly, reduced E6AP occurred most frequently in the basal subtype, with more than 40% of these cancers exhibiting low or intermediate E6AP protein levels (Fig. 1E). Consistent with this, low E6AP was significantly associated with reduced disease-free survival in TNBC (Fig. 1F). Collectively, our clinical data demonstrate that loss of E6AP is associated with aggressiveness and invasiveness of breast cancer, especially in TNBC.

We next investigated the contribution of E6AP to cell invasion by comparing the invasive capacity of normal MECs isolated from either E6AP-KO or WT mice. E6AP-KO MEC invaded Matrigel more efficiently than WT (Fig. 2A). WT cells formed acini with a hollow lumen by day 6–12, but developed aberrant morphogenesis after 16 days in culture (Fig. 2B). In contrast, 38% of E6AP-KO MEC formed aberrant acini after 6 days, with 58% failing to create a lumen by 12 days and forming membrane protrusions with invasive morphology (67%) after 16 days in culture (Fig. 2C).

We next determined whether E6AP is correlated with the invasive potential of human breast cancer cell lines by measuring E6AP protein levels in a series of lines for which the invasive capacity has been reported (38). We found an inverse association between the level of E6AP and the invasive capacity of these lines (Fig. 2D). Importantly, E6AP levels were reduced substantially in the highly metastatic TNBC MDA-MB-231-HM cells (29), consistent with an inhibitory role of E6AP in invasion. To test this association functionally, MDA-MB-231-HM cells with inducible expression of E6AP were generated (Fig. 2E). Induction of E6AP by doxycycline resulted in a significant decrease in cell invasion through Matrigel (Fig. 2F), but had only a small impact on proliferation (Fig. 2G). Doxycycline treatment alone did not alter proliferation of parental MDA-MB-231-HM cells (Supplementary Fig. S2A). Together, our results demonstrate that E6AP regulates breast cancer cell invasion in vitro.

We next tested whether high E6AP suppresses the metastatic potential of MDA-MB-231-HM tumors in mice. These cells express both EGFP and luciferase to assist in imaging and determination of metastatic burden. Enforced expression of E6AP (confirmed by E6AP immunostaining after resection of the primary tumors (Supplementary Fig. S2B) and Western blotting (Supplementary Fig. S2C) of the primary tumors, did not substantially alter primary tumor growth (Fig. 3A), consistent with the minimal impact on proliferation (Fig. 2G). Spontaneous metastatic dissemination after primary tumor resection at 0.5 g was monitored by bioluminescence (Fig. 3B), revealing a significant reduction in whole body luminescence at 35 days after resection in mice treated with doxycycline to induce E6AP (Fig. 3C). Fewer macrometastases were evident at autopsy in both lungs and livers of E6AP-expressing tumors (Supplementary Fig. S3A), confirmed by histology (Fig. 3D). Quantification of tumor burden by genomic PCR revealed a significant reduction in metastasis to the lung (Fig. 3E), liver (Fig. 3F), femur (Fig. 3G), and spine (Fig. 3H). Doxycycline treatment of mice bearing parental tumors showed no significant changes in metastatic burden in these organs (Supplementary Fig. S3B–S3E). To verify that the reduction in metastasis is mediated by E6AP, we stained both lung and liver for EGFP and E6AP. EGFP-positive cells were detected in lung and liver macrometastases, and E6AP expression was still evident at day 35 in the E6AP-inducible tumor group (Supplementary Fig. S4A).

To confirm that restoration of E6AP expression can prolong survival, we monitored mice for metastasis-free survival after primary tumor resection on day 10. Induction of E6AP expression in the MDA-MB-231-HM tumors led to significantly prolonged disease-free survival (Fig. 3I). Taken together, these results demonstrate that E6AP expression can suppress metastasis and prolong the survival of MDA-MB-231-HM tumor-bearing mice.

To further understand how E6AP impacts on metastasis, we used a model of experimental metastasis whereby MDA-MB-231-HM cells were injected directly into the tail vein to measure lung colonization and growth. Similar to our results from the spontaneous metastasis experiments, doxycycline-induced E6AP expression resulted in a significant reduction of metastatic burden as measured by luminescence (Fig. 4A and B). To confirm the luminescence measurements, lungs and livers were weighed at autopsy as a measure of metastatic burden. Consistent with our previous findings, tumors in which E6AP was induced had significantly lower lung (Supplementary Fig. S4B) and liver (Supplementary Fig. S4C) weights compared with doxycycline-free animals. Doxycycline treatment of mice bearing control tumors showed no alteration in organ weight (Supplementary Fig. S5A and S5B). Taken together, these data indicate that restoring E6AP expression reduces the ability of tumor cells to colonize secondary organs.

As E6AP-mediated metastatic suppression is apparently not due to slower tumor growth, we next investigated the ability of tumor cells to extravasate into secondary tissues. Control and E6AP-expressing cells with or without prior in vitro doxycycline treatment were injected intravenously into NSG mice and their ability to lodge in the lungs was measured by bioluminescence at early time points (Fig. 4C). Dox-induced E6AP-expressing cells exhibited a 6-fold decrease in their capacity to lodge in the lung at 30 minutes after inoculation, indicating that E6AP is greatly reducing subsequent tumor cell colonization (Fig. 4D). To further analyze the impact on extravasation, we investigated the effect of E6AP on lung colonization of SKBR3 cells that express high levels of E6AP. Inducible downregulation of E6AP expression in SKBR3 cells (Fig. 4E) increased their ability to lodge in the lung (Fig. 4F and G). Thus, we have shown using two different breast cancer lines that E6AP regulates the early events of lung colonization.

Migration of cancer cells through the endothelial lining of the blood vessels is a critical step of extravasation (39, 40). To obtain more direct evidence for the contribution of E6AP to tumor cell extravasation, we measured their ability to cross an endothelial barrier using the transendothelial migration assay. As shown in Fig. 4H, doxycycline-induced E6AP expression in MDA-MB-231-HM resulted in a significant decrease in the number of cells traversing a confluent monolayer of HUVECs. In contrast, downregulation of E6AP expression in SKBR3 augmented their capacity to transmigrate (Fig. 4I). Doxycycline treatment did not alter the transmigratory capacity of either MDA-MB-23-1HM or SKBR3 cells bearing the control vector (Supplementary Fig. S5C and S5D).

To explore the potential mechanism by which E6AP regulates metastasis, we examined the effect of E6AP expression on cell morphology and on organization of the actin cytoskeleton. Gain of E6AP altered the morphology of MDA-MB-231-HM cells, resulting in a significant increase in an epithelial-like appearance with increased numbers of cuboidal cells (Fig. 5A and B). In contrast, downregulation of E6AP expression in SKBR3 cells increased the number of cells with an elongated morphology (Fig. 5A and B). Doxycycline treatment alone did not affect the morphology of either cell line (Supplementary Fig. S6A).

Loss of E6AP also altered the morphology of MECs derived from E6AP-KO or WT mice. E6AP-KO MEC displayed a spindle-shaped appearance and formed loose clusters (Supplementary Fig. S6B), whereas WT MEC retained the typical epithelial features of cuboidal morphology and formed tight clusters (Supplementary Fig. S6B). A similar phenotype was observed in fibroblasts. E6AP-KO MEFs are more elongated, with a 50% increase in the number of elongated cells (Supplementary Fig. S6C and S6D), and with a 35% increase in the number of long membrane protrusions (filopodia-/lamellipodia-like structures) compared with WT MEFs (Supplementary Fig. S6E and S6F).

Consistent with these findings, loss of E6AP altered the ductal structure of mammary gland, resulting in disorganized luminal epithelium, with an increased number of fibroblasts and increased fibrillar collagen deposition (Fig. 5C, left). In E6AP-KO mammary epithelium, the expression of the mesenchymal marker vimentin was unchanged and E-cadherin remained enriched at the cell–cell adhesion boundaries when compared with WT epithelium (Fig. 5C, right). Immunostaining of cytokeratin-8 and alpha-SMA of the E6AP-KO mammary epithelium revealed abnormal multilayers of both luminal and myoepithelial cells as well as a disrupted apical staining of pERM compared with WT (Fig. 5D, left). Thus, in the absence of E6AP, the normal tissue architecture of the mammary gland is disrupted.

Given the central role of the Rho GTPases in actin cytoskeletal remodeling (14), we measured the levels of the activated Rho GTPase family members using RBD and PAK-pull down assays (13, 41). E6AP-KO MEFs displayed a higher GTP loading of RhoA at a steady state compared with WT cells (Fig. 5E and Supplementary Fig. S7). After EGF stimulation, both Rac1 and Cdc42 were highly activated in the E6AP-KO cells, with no activation of RhoA (Fig. 5E and Supplementary Fig. S7). We next measured the levels of activated Rac1 and Cdc42 in the tumor cells with modulated E6AP levels. E6AP expression in MDA-MB231-HM resulted in reduced steady state and EGF-induced GTP-bound Rac1 and Cdc42 levels (Fig. 5E, left). Conversely, downregulation of E6AP in SKBR3 cells increased the GTP loading of both Rac1 and Cdc42 at steady state and after EGF stimulation (Fig. 5E, right). Thus, E6AP regulates Rho GTPase activity, specifically that of Rac1 and Cdc42.

Previously, it was reported that the Drosophila ortholog of E6AP (UBE3A) regulates the expression of the Rho-GEF Pbl, the Drosophila ortholog of human ECT2 (20). Therefore, we asked whether E6AP regulates ECT2 in mammalian cells. We found that ECT2 levels were elevated in the E6AP-KO MEFs compared with WT MEFs (Fig. 6A). Increased accumulation of ECT2 in the E6AP-KO MEFs was also evident by immunofluorescence (Fig. 6B). Consistent with our previous results showing that E6AP regulates the Rho GTPases, the increase of ECT2 was associated with increased phosphorylation of myosin light chains (MLC; Fig. 6A) and cytoskeletal reorganization in MEFs, as revealed by filamentous actin (F-actin) staining using rhodamine-phalloidin. E6AP KO MEFs displayed an extensive network of stress fibers (Fig. 6B) compared with WT MEFs.

To assess the effect of E6AP on steady-state levels of ECT2, HEK293 cells were transfected with expression plasmids for GFP-ECT2 alone or together with increasing amounts of either WT-E6AP or the catalytically inactive E6AP mutant (C820A). ECT2 protein levels were reduced in cells cotransfected with WT-E6AP, indicating that E6AP regulates the expression of ECT2 (Fig. 6C). In contrast, the E6AP C820A mutant was less able to downregulate ECT2, indicating the requirement for catalytically active E6AP. As E6AP is an E3-ligase, we proposed that E6AP promotes the degradation of ECT2 through the ubiquitin-proteasome system. HEK293 cells transfected with ECT2 together with increasing amounts of WT-E6AP were treated with the proteasome inhibitor, MG132. As shown in Fig. 6C, blocking proteasome activity prevented E6AP-induced degradation of ECT2, indicating that E6AP controls the proteasome-dependent degradation of ECT2.

To demonstrate directly the ubiquitination of ECT2 by E6AP, we utilized an in vivo ubiquitination assay in SKBR3 cells that express a doxycycline-inducible knockdown of E6AP. Ubiquitinated species of ECT2 were readily detected when endogenous ECT2 was immunoprecipitated under denaturing conditions from MG132-treated SKBR3 cells. However, the level of ubiquitinated ECT2 was greatly reduced in E6AP-depleted cells (Fig. 6D). Consistent with previously published data describing the critical role of E3-ligase UBE3A/E6AP in the regulation of Pbl in Drosophila (20), we demonstrate here that E6AP also promotes the ubiquitination of ECT2 in mammalian cells.

To establish whether ECT2 is required for E6AP-mediated regulation of Rho GTPases activity, we measured Rho GTPase activation in SKBR3 cells depleted of ECT2 (Fig. 6E). Consistent with previous results in Fig. 5E, knockdown of E6AP increased levels of GTP-bound Rac1 and Cdc42 (Fig. 6F). However, downregulation of ECT2 resulted in reduction of both basal and EGF-induced RhoA and Cdc42 activity, regardless of E6AP expression (Fig. 6F). Surprisingly, knockdown of ECT2 did not alter either the basal or the EGF-induced GTP-bound Rac1 (Fig. 6F), indicating that E6AP regulates Rac1 activity through GEFs other than ECT2.

To establish whether ECT2 is required for E6AP-mediated suppression of metastasis, we evaluated the effect of downregulating ECT2 expression on the extravasation capacity of SKBR3 cells using a transendothelial migration assay. Loss of E6AP resulted in a significant increase in the number of cells traversing the monolayer of endothelial cells. Importantly, downregulation of ECT2 expression in SKBR3 cells abolished their capacity to transmigrate, regardless of level of their E6AP expression (Fig. 6G and H). Our results support the proposal that ECT2 is involved in tumor cell extravasation and that downregulation of E6AP is required for the increased ability of tumor cells to colonize secondary organs, mediated, at least in part, by ECT2.

To gain insight into the link between E6AP and ECT2 in breast cancer, we evaluated the prognostic value of ECT2 in clinical samples. By Kaplan–Meier analysis, high ECT2 expression was correlated significantly with poorer survival in the METABRIC dataset of breast tumors from 1,992 patients (Fig. 7A). The prognostic power of ECT2 was also evident in patients with basal-like tumors (Fig. 7B). ECT2 expression was increased in all breast cancer subtypes, with the basal subtype exhibiting the highest levels compared with normal breast tissue (Fig. 7C). To determine whether E6AP regulates ECT2 transcription, we examined the relationship between E6AP and ECT2 transcript levels in the METABRIC dataset. However, no correlation between E6AP and ECT2 was found in the whole cohort (Fig. 7D) nor in patients with basal-like tumors (Fig. 7E), further indicating that E6AP regulates ECT2 at the protein level. Importantly, immunostaining of ECT2 and E6AP in a TMA of 100 human breast cancer samples revealed an inverse correlation between E6AP and ECT2 proteins in all breast cancer subtypes (Fig. 7F), confirming this mode of regulation. Collectively, our analysis of the clinical data clearly supports the preclinical results revealing a link between E6AP and ECT2 in breast cancer metastasis. A model for the proposed function of E6AP and ECT2 in regulating breast cancer metastasis is shown in Fig. 7G.

We have identified E6AP as a novel metastasis suppressor in human breast cancer and revealed a dependence on ECT2 for this activity of E6AP. E6AP is an E3-ligase that regulates the transcript and protein levels of the steroid hormone receptors (42–45). Overexpression of ligase-defective E6AP in the mammary gland initiates mammary tumor development. Reduced E6AP expression correlates with more invasive breast carcinomas (24). Taken together, these previous findings support a tumor-suppressive function for E6AP in breast cancer. Here we demonstrate that E6AP-mediated metastasis suppression is achieved by inhibiting cell invasion and colonization, which are critical stages of metastatic dissemination.

Our data show that loss of E6AP is essential for the acquisition of metastatic properties and that E6AP loss in both normal and tumor cells promotes the acquisition of elongated morphology and membrane protrusions. Changes in cell morphology driven by actin cytoskeleton remodeling is a hallmark of cell invasion (46, 47). Accordingly, E6AP deficiency promotes cell invasion in vitro, and leads to aberrant acinar morphogenesis. E6AP does not affect primary tumor growth, but is essential for blocking tumor cell dissemination and colonization of secondary tissues. Notably, downregulation of E6AP expression enhances the ability of tumor cells to colonize the lung and to transmigrate through a layer of endothelial cells in vitro. Conversely, restoration of E6AP expression constrains the ability of tumor cells to lodge in the lung and impairs their transendothelial migratory behavior.

Increased Rho GTPase activity stimulates the formation of actin cytoskeletal structures that promote migration and invasion (48). Loss or knockdown of E6AP stimulates both Rho GTPase steady-state activity and in response to EGF stimulation, in normal and tumor cells. This high Rho GTPase activity is consistent with the pronounced invasive and migratory capacity of the cells. Conversely, expression of E6AP abolished Rac1 and Cdc42 activity, rendering the cells less invasive and less metastatic. In cancer cells, Rho GTPases are often activated by indirect mechanisms such as increased RhoGEF and/or decreased RhoGAP activity (8). The GEF ECT2 has multiple substrates including Rac1, RhoA, and Cdc42 (19). Here we have demonstrated that E6AP is a negative regulator of ECT2 in breast cancer, with increased expression in the absence of E6AP and enhanced ubiquitination and proteasomal degradation of ECT2 when E6AP is restored. This is in agreement with previous findings that expression of both ECT2 and its Drosophila ortholog Pbl are regulated by Ube3a (E6AP homolog) in Drosophila neurons (20). Together, our data demonstrate that E6AP regulates ECT2 and its downstream Rho GTPase signaling pathway. Surprisingly, we found that ECT2 is required for E6AP regulation of EGF-induced RhoA and Cdc42 activation but not Rac1 activation. As E6AP downregulation still increases Rac1 activity, other GEFs are likely to mediate Rac1 activation and indeed, others GEFs are highly expressed in human breast cancers (49). For example, the GEF Tiam1 is highly expressed in high-grade breast tumors, where it mediates migration and invasion (50, 51). Vav1 and Vav3 are overexpressed in breast tumors and regulate lung metastasis (52). Likewise, P-Rex is highly expressed in metastatic breast tumors and mediates motility and tumorigenesis driven by ErbB receptors (53). It is possible that the effect of E6AP on Rac1 activation may be mediated by these other GEFs. Collectively, our findings indicate that ECT2 mediates some but not all of the effects of E6AP on Rho GTPase activation.

Analysis of clinical data revealed that reduced E6AP expression is associated with TNBC. Likewise, at the protein level, we observed an inverse correlation between E6AP and ECT2 and a favorable prognosis for patients with high E6AP–expressing tumors and a poor prognosis for those with high ECT2–expressing tumors. Thus, our clinical data support our experimental data showing that E6AP is an ECT2/Rho GTPase signaling repressor and provides further evidence that E6AP acts to suppress breast cancer progression and metastasis.

We have now demonstrated that ECT2 is a key factor in the progression of breast cancer. These findings have important implications for prognosis and therapy of this disease. The level of expression of ECT2 may represent a biomarker for TNBC, predicting aggressiveness and metastasis. ECT2 is a common GEF for Rho GTPases and therefore drugs that inhibit ECT2 activity may lead to the blockade of Rac1/RhoA/Cdc42 downstream pathways. This targeting approach may be a very effective and potent option for the inhibition of the Rho GTPase signaling.

In summary, we report here for the first time, to the best of our knowledge, the involvement of E6AP and ECT2 in breast cancer metastasis. We demonstrate that reduced E6AP expression in breast cancer increases ECT2 and Rho GTPase activity to promote metastasis, whereas expression of E6AP inhibits cell invasion and colonization. We conclude that E6AP, through its negative regulation of ECT2, serves as a suppressor of metastasis. Notably, E6AP loss is common in TNBC. This event may therefore represent a major driving force for primary tumors to initiate local invasion and subsequent metastatic progression. Our findings highlight the E6AP–ECT2 axis as a potential target for therapeutic intervention in breast cancer metastasis.

No potential conflicts of interest were disclosed.

Conception and design: M. Mansour, S. Haupt, S. Loi, O. Bernard, R.L. Anderson, Y. Haupt

Development of methodology: M. Mansour, A.-L. Chan, N. Godde, C.N. Johnstone, Y. Levav-Cohen, S.B. Fox, O. Bernard, R.L. Anderson, Y. Haupt

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Mansour, S. Haupt, N. Godde, A. Rizzitelli, M. Bishton, C.N. Johnstone, B. Monahan, Y.-H. Jiang, S.B. Fox, Y. Haupt

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Mansour, A.-L. Chan, S. Loi, F. Caramia, S. Deb, S.B. Fox, O. Bernard, Y. Haupt

Writing, review, and/or revision of the manuscript: M. Mansour, S. Haupt, A.-L. Chan, N. Godde, A. Rizzitelli, S. Loi, S. Deb, C.N. Johnstone, A.S. Yap, O. Bernard, R.L. Anderson, Y. Haupt

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Mansour, A.-L. Chan, A. Rizzitelli, E.A. Takano, S.B. Fox, Y. Haupt

Study supervision: Y. Haupt

The authors thank their laboratory colleagues for their support and advice and all colleagues who provide gifts, reagents, and support, notably Cristina Gamell, Mark Devlin, Christophe Mintoff, Zafar Nawaz, Aparna Ratheesh, and Marco Herold. The authors are grateful for support from the Peter MacCallum Cancer Centre Microscopy, Histology and Animal Core Facilities. R.L. Anderson and Y. Haupt acknowledge support from the National Breast Cancer Foundation (NBCF) of Australia.

This work was supported by grant 1049179 from the National Health and Medical Research Council (NHMRC) of Australia (M. Mansour and Y. Haupt) and by NHMRC fellowship 628426 and grants, 1049179, 1026988, 1026990, 1037320 (Y. Haupt). M. Mansour is also supported by the CASS Foundation. R.L. Anderson is supported by a fellowship from the NBCF of Australia. A.S. Yap is supported by NHMRC fellowship1044041.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.