Transactivation of ErbB-2 induced by tumor necrosis factor α promotes NF-κB activation and breast cancer cell proliferation
- First Online:
- Cite this article as:
- Rivas, M.A., Tkach, M., Beguelin, W. et al. Breast Cancer Res Treat (2010) 122: 111. doi:10.1007/s10549-009-0546-3
- 174 Views
Tumor necrosis factor alpha (TNFα) is a pleiotropic cytokine which, acting locally, induces tumor growth. Accumulating evidence, including our findings, showed that TNFα is mitogenic in breast cancer cells in vitro and in vivo. In the present study, we explored TNFα involvement on highly aggressive ErbB-2-overexpressing breast cancer cells. We found that TNFα induces ErbB-2 phosphorylation in mouse breast cancer C4HD cells and in the human breast cancer cell lines SK-BR-3 and BT-474. ErbB-2 phosphorylation at Tyr877 residue was mediated by TNFα-induced c-Src activation. Moreover, TNFα promoted ErbB-2/ErbB-3 heterocomplex formation, Akt activation and NF-κB transcriptional activation. Inhibition of ErbB-2 by addition of AG825, an epidermal growth factor receptor/ErbB-2-tyrosine kinase inhibitor, or knockdown of ErbB-2 by RNA interference strategy, blocked TNFα-induced NF-κB activation and proliferation. However, the humanized monoclonal antibody anti-ErbB-2 Herceptin could not inhibit TNFα ability to promote breast cancer growth. Interestingly, our work disclosed that TNFα is able to transactivate ErbB-2 and use it as an obligatory downstream signaling molecule in the generation of mitogenic signals. As TNFα has been shown to be present in the tumor microenvironment of a significant proportion of human infiltrating breast cancers, our findings would have clinical implication in ErbB-2-positive breast cancer treatment.
Tumor necrosis factor alpha (TNFα) is a pleiotropic cytokine originally characterized to cause hemorrhagic necrosis of tumors at high doses . However, it is now widely accepted that TNFα, acting locally, induces the growth of certain tumor types such as ovary and breast [2–5]. In particular, TNFα has been shown to be produced by malignant or host cells in the tumor microenvironment of human infiltrating breast cancer and to be associated with increasing malignancy . Treatment of cancer cells with TNFα enhances cell proliferation in vitro [7, 8] and, as we have recently demonstrated, also supports murine breast cancer growth in vivo . In addition, we have shown that TNFα induces in vitro proliferation of breast cancer cells through a mechanism which requires activation of p42/p44 mitogen-activated protein kinase (MAPK), c-jun NH2-terminal kinase (JNK), Akt and NF-κB transcriptional activation . However, the exact mechanism by which TNFα enhances tumor growth and its involvement in different breast cancer subtypes remains elusive.
ErbB-2, a transmembrane tyrosine kinase receptor, is overexpressed in nearly 30% of human breast cancer and has been associated with enhanced tumor aggressiveness and poor clinical outcome . These tumors also display NF-κB activation . ErbB-2 is an orphan receptor belonging to the family of type I tyrosine kinase receptors, which signals by forming heterodimers with epidermal growth factor receptor (EGFR), ErbB-3 and ErbB-4  in response to ligands including heregulins . After ligand binding, all ErbB receptors are phosphorylated, serving as docking sites for the recruitment of cytoplasmic adaptor proteins, initiating signaling cascades that control multiple cellular processes. Herceptin™ (Trastuzumab) is a monoclonal antibody which binds to the extracellular domain of the receptor  and is administrated to breast cancer patients whose tumors overexpress ErbB-2. However, the clinical efficacy of Herceptin is limited to 30% of these patients. An important reason for this is that other tumor-cell alterations may influence the response to ErbB-2-targeted inhibitors . Thus, understanding the mechanisms by which ErbB-2 can be activated through non-classical receptors and ligands is relevant in order to design a new therapeutic approach and to predict patients’ response. Ligands of G protein-coupled receptors which act through c-Src kinase activation , as well as hormones such as prolactin, acting through Janus kinase 2 activation , and cytokines such as interleukin-6  have all been demonstrated to transactivate ErbB-2. Several groups have so far shown transactivation of EGFR by TNFα through the activation of matrix metalloproteinases (MMPs) which are able to release EGFR ligands from the cell membrane [18–20]. In contrast, there are no reports demonstrating TNFα ability to transactivate ErbB-2 in breast cancer.
In the present work, we explore the effect of TNFα on breast cancer cells that overexpress ErbB-2 and found that TNFα induces not only ErbB-2 autophosphorylation, but also phosphorylation of its Tyr877 residue through the activation of the tyrosine kinase c-Src. We further found that it promotes association between ErbB-2 and ErbB-3. ErbB-2 transactivation by TNFα is a rapid event that does not involve either ligand release or MMPs activation, and leads to cell proliferation even in the presence of Herceptin. Interestingly, we showed that TNFα regulates ErbB-2 phosphorylation as a requisite to activate NF-κB and cell proliferation in human and murine breast cancer cells. This is the first demonstration of ErbB-2 transactivation by TNFα in breast cancer cells, which may be one of the mechanisms by which ErbB-2-overexpressing tumors show resistance to anti-ErbB-2 monoclonal antibodies therapy.
Materials and methods
Animals and tumors
Experiments were carried out in virgin female Balb/c mice, raised at the Instituto de Biología y Medicina Experimental of Buenos Aires. All animal studies were conducted as described . C4HD mouse mammary tumor expresses progesterone and estrogen receptors, lacks EGFR expression, overexpresses ErbB-2 and exhibits high levels of ErbB-3 and low expression of ErbB-4 [21–23].
Antibodies to the following proteins were used: Neu/ErbB-2 (C-18), ErbB-3 (C-17), phosphotyrosine (PY99), p85 phosphatidyl inositol 3-kinase (PI3-K), p42/p44 MAPK (C-14), phospho-p42/p44 MAPK (E-4), JNK (N-18) and phospho JNK (G-7) all from Santa Cruz Biotechnology (Santa Cruz, CA, USA), Akt, phospho Akt (Ser 473), phospho IκBα (Ser32/36), IκBα, phosphotyrosine c-Src (Tyr416), c-Src (36D10), phospho-ErbB-2 (Tyr1221/1222), phospho-ErbB-2 (Tyr877) and EGFR from Cell Signaling (Beverly, MA, USA), v-Src (ab-1) from Calbiochem (La Jolla, CA, USA) and actin (Clone ACTN05) and cyclin D1 from Neomarkers (Fremont, CA, USA).
Cell culture and treatments
Primary cultures of epithelial cells from the mouse mammary tumor C4HD, growing in medroxyprogesterone acetate (MPA)-treated mice, were performed as previously described [5, 24]. As C4HD cells are sensitive to progestin, all the experiments were performed in Dulbecco’s Modified Eagle’s Medium/F12 without phenol red (Sigma, St. Louis, MO, USA) (DMEM) + 0.1% charcoal-stripped fetal calf serum (ChFCS). Human breast cancer cell lines SK-BR-3 and BT-474 were obtained from the American Type Culture Collection and maintained in Mc Coy’s 5A + 10% FCS (Gen SA, Buenos Aires, Argentina) and RPMI 1640 + 10% FCS, respectively. Experiments were performed in their respective medium + 1% FCS. Cells were treated for the indicated times with 20 ng/ml of murine or human TNFα (mTNFα, hTNFα, respectively) (Cell Sciences, Canton, MA, USA) or with 20 ng/ml of recombinant human β1 heregulin (HRG, Upstate, Millipore, Bedford, MA, USA). The following inhibitors were added to cells 60 min before incubation with TNFα: AG825, EGFR/ErbB-2 tyrosine kinase inhibitor from the benzylidene malononitrile family; PP2, 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d]pyrimidine, c-Src inhibitor; GF109203X, PKC inhibitor; GM6001, a broad MMPs inhibitor, and its corresponding negative control (all from Calbiochem); Dasatinib, c-Src inhibitor (LC Laboratories, Woburn, MA, USA); GW2974, EGFR/ErbB-2 tyrosine kinase inhibitor from the indazolyamino quinazoline family and Bay 11-7082, inhibitor of IκB phosphorylation (both from Sigma). To perform certain experiments, cells were pre-incubated for 16 h with 10 μg/ml of the anti-ErbB-2 antibody Herceptin™ (Hoffmann-La Roche Ltd, Basel, Switzerland) for Western blot analysis, and 1 h for proliferation and reporter assays. Cell proliferation was evaluated by [3H]-thymidine incorporation assay. Cells were incubated with TNFα for 48 h. 1 μCi [3H]-thymidine (NEN, Dupont, Boston, MA, USA; specific activity 20 Ci/mmol) was added at hour 24 of said culture. Cells were then trypsinized and harvested and radioactivity was counted using standard scintillation procedures . Assays were performed in octuplicate. In earlier experiments we demonstrated that [3H]-thymidine uptake correlated with the number of cells/well . Cell viability was performed by triplicate by Trypan blue exclusion at 48 h of treatment with the corresponding inhibitors. For C4HD cells, cultured with 100 μM AG825, it was of 67.6 ± 4.7%, for BT-474 cells cultured with 100 μM AG825, 10 μg/ml Herceptin, 1 μM GW2974, 0.5 μM Dasatinib and 1 μM Bay 11-7082, it was of 62.3 ± 7.2, 70.9 ± 7.6, 73.7 ± 10.3, 77.0 ± 1.3 and 73 ± 8.3%, respectively, and for SK-BR-3 cells cultured with 100 μM AG825 and 10 μg/ml Herceptin it was of 55.8 ± 9.3 and 65.6 ± 1.3%, respectively.
Immunofluorescence staining and confocal microscopy
SK-BR-3, BT-474 and C4HD cells grown on glass coverslips were treated with TNFα 20 ng/ml for 30 min. Cells were fixed and permeabilized in ice-cold methanol and ErbB-2 was localized using ErbB-2 9G-6 (Santa Cruz Biotechnology) followed by incubation with a rhodamine conjugated anti-mouse secondary antibody (The Jackson Laboratory, Bar Harbor, ME, USA). Stained cells were analyzed using a Nikon C1 confocal laser scanning microscope.
Flow cytometry analysis
SK-BR-3 and BT-474 cells treated with hTNFα for the indicated times were harvested with PBS + EDTA 1% and incubated with anti-ErbB-2 9G-6 antibody followed by incubation with anti-mouse phycoerythrin (PE)-conjugated antibody (Santa Cruz Biotechnology). A total of 104 cells/sample was analyzed using FACSAria cytometer (Becton–Dickinson, La Jolla, CA, USA). Background staining was evaluated in cells incubated with an isotype control IgG followed by anti-mouse PE-conjugated antibody. Data analysis was performed using WinMDI software (J. Trotter, Scripps Research Institute, San Diego, CA, USA). Delta mean fluorescence intensity (MFI) values were obtained by subtracting the MFI of the cells incubated with control isotype antibody from the MFI of cells incubated with ErbB-2 antibody.
For cell cycle analysis, BT-474 cells were subjected to the different treatments, harvested at 36 and 48 h, and fixed in 70% ethanol for 24 h at 4°C, as we previously described . They were washed twice with PBS, followed by RNA digestion (RNAse A 50 U/ml) and propidium iodide (20 μg/ml) staining for 30 min at room temperature in the dark. Cell cycle analysis was performed using a FACScalibur flow cytometer (Becton–Dickinson) and Modfit LT software.
Western blot and immunoprecipitation
Lysates were prepared from cells subjected to the different treatments described in each experiment, as previously detailed, and proteins were subjected to SDS-PAGE . Association among ErbB-2, ErbB-3 and p85 PI3-K was studied by performing co-immunoprecipitation experiments as previously described . Briefly, 500 μg of protein lysates was incubated with 2 μg of rabbit anti-EGFR, ErbB-2 or ErbB-3 antibody and the immunocomplexes were captured by adding protein A-agarose (Santa Cruz Biotechnology). Beads were washed, boiled in sample buffer, and proteins were electroblotted as described above. As negative control, normal rabbit serum was used.
In vitro cold phosphorylation assay
C4HD cells were treated with TNFα for 5 min or preincubated before TNFα stimulation for 60 min with PP2 (10 μM). Cells were lysed in kinase lysis buffer (20 mM HEPES pH 7.5, 10 mM EGTA, 1% NP-40, 2.5 mM MgCl2) and Src was immunoprecipitated from 500 μg protein extracts using an anti-v-Src antibody. ErbB-2 was immunoprecipitated from 500 μg protein extract from unstimulated C4HD cells. The immunoprecipitated ErbB-2 was then subjected to an in vitro phosphorylation assay with Src immunoprecipitated from each treatment as previously described . Proteins were separated by electrophoresis; gels were transferred onto nitrocellulose and immunoblotted with anti-Tyr877/927 ErbB-2 and anti-total c-Src antibodies and ErbB-2.
κB-Luc vector (κB sites from HIV promoter) and cyclin D1 promoter-Luc vector were kindly provided by Dr M. Bell (Mayo Clinic, Rochester, MN, USA) and Dr R. Pestell (Northwestern University Medical School, Chicago, IL, USA), respectively. Human ErbB-2 wild type, kinase-negative (KN) ErbB-2 and ErbB-2-Y877F were kindly provided by Dr T. Akiyama (Gumma University, Gumma, Japan), Dr A. Gertler (Protein Laboratories Rehovot, Israel) and Dr O. Segatto (Centro Recerca Sperimentale, Rome, Italy). C4HD cells were transfected for 24 h in DMEM supplemented with 10 nM MPA and 2.5% ChFCS, and cell lines in the corresponding growth medium without antibiotics. FuGENE 6 transfection reagent technique (Roche Biochemicals, Indianapolis, IN, USA) was used in accordance with the manufacturer’s instructions. Cells were transiently co-transfected with 1 μg of κB-Luc or 1 μg cyclin D1-luc construct plus 10 ng Renilla luciferase expression vector CMV-pRL (Promega, Madison, WI, USA) used to correct variations in transfection efficiency. As control, cells were transfected with a pGL3-basic reporter lacking κB. Transfected cells were lysed and luciferase assays carried out using the Dual-Luciferase Reporter Assay System (Promega).
siRNAs targeting mouse ErbB-2 mRNAs were synthesized by Dharmacon Inc. (Lafayette, CO, USA) (ErbB-2 siRNA#02: antisense, 5′-GAUGUCCUCCGUAAGAAUA-3′, ErbB-2 siRNA#03 antisense 5′-GAUGGUGCUUACUCAUUGA-3′, ErbB-2 siRNA#04 antisense 5′-GGAAUCCUAAUCAAACGAA-3′). The non-silencing siRNA oligonucleotide from Dharmacon, which does not target any known mammalian gene, was used as negative control. Transfection of siRNA plus 1 μg of empty vector (pcDNA3.1) or 1 μg of human ErbB-2 expression vector and 1 μg of κB-Luc vector plus 10 ng Renilla luciferase expression vector was performed using the DharmaFECT Duo transfection reagent following the manufacturer’s directions, using 25 nM of siRNA for 3 days. Cell treatment and reporter activity were measured as described above. In some experiments, siRNA at a final concentration of 25 nM was transfected with Dharmafect 1 reagent following the manufacturer’s directions.
The differences between control and experimental groups were analyzed by ANOVA followed by Tukey t test among groups. Kolmorgorov–Smirnov test was used for flow cytometric studies.
TNFα-induced breast cancer cell proliferation requires ErbB-2 phosphorylation
We examined ErbB-2 phosphorylation levels after TNFα treatment. Figure 1b shows that TNFα augmented total levels of ErbB-2 tyrosine phosphorylation from 2 to 15 min of treatment and decline thereafter. Then, we addressed the specific mouse ErbB-2 tyrosine phosphorylation at Tyr1272, analogous to human Tyr1222 ErbB-2. There was an increase in Tyr1272 phosphorylation 2 min after TNFα treatment in C4HD cells, with maximum phosphorylation at 5 min (Fig. 1c). Addition of AG825 blocked TNFα-induced Tyr1272 phosphorylation in a concentration-dependent manner (Fig. 1d). Since our goal was to determine whether TNFα-induced ErbB-2 transactivation was taking place also in human breast cancer, we then used SK-BR-3  and BT-474 cell lines  that are widely used models of ErbB-2-overexpression. While TNFα induced SK-BR-3 and BT-474 cell proliferation, addition of AG825 blocked TNFα mitogenic effect (Fig. 1e). Similar effects were obtained with HRG-treated cells. In addition, TNFα induced phosphorylation of the Tyr1222 residue after 2 min stimulation, with a peak at 5 min (Fig. 1f) which was inhibited by addition of AG825. Comprehensively, our present findings indicate that TNFα is able to transactivate ErbB-2 in breast cancer cells.
To examine the effect of TNFα stimulation on ErbB-2 expression on the surface of SK-BR-3 and BT-474 cells, we performed immunofluorescence and flow cytometry analysis. ErbB-2 plasma membrane expression was observed to decrease, reaching its minimum at 30 min (Fig. 1g) and staying low for at least 1 h of TNFα treatment in both cell lines. To confirm these data, we performed confocal microscopy studies. ErbB-2 was localized primarily to the plasma membrane in unstimulated cells. TNFα treatment for 30 min led to a significant increase in ErbB-2 localization in the cytoplasm of C4HD, SK-BR-3 and BT-474 cells (Fig. 1h) without affecting ErbB-2 content in whole cell extracts. Comprehensively, these results show for the first time that TNFα is able to activate ErbB-2 inducing its phosphorylation and internalization into the cytoplasm of breast cancer cells.
The finding that TNFα is able to induce ErbB-2 transactivation, prompted us to explore the mechanism underlying this phosphorylation. It is known that TNFα can induce EGFR transactivation through MMP-dependent EGFR ligand release. Treatment of C4HD and SK-BR-3 cells with 10 μM GM6001, a broad MMPs inhibitor, did not modify TNFα-induced ErbB-2 phosphorylation (Supplemental Fig. 3a). In addition, the use of blocking antibodies to ErbB-3 and ErbB-4 to impede ligand binding did not modify TNFα-induced cell growth (Supplemental Fig. 3b) although both antibodies were able to inhibit HRG proliferative effect in C4HD cells . Thus, these data suggest that TNFα does not activate ErbB-2 through ligand release in these breast cancer cells.
TNFα-induced c-Src kinase activation mediates ErbB-2 phosphorylation
To further explore whether activated c-Src can phosphorylate Tyr877 ErbB-2 in vitro, we performed a cold phosphorylation assay. For this purpose, we immunoprecipitated c-Src from C4HD cells treated or not with TNFα for 5 min, and from C4HD cells treated with TNFα and PP2. We also immunoprecipitated ErbB-2 from unstimulated cells and used it as a source of unphosphorylated ErbB-2 in the assay. As shown in Fig. 2c, c-Src activated by TNFα was able to phosphorylate ErbB-2 at Tyr927 residue. Neither c-Src obtained from control cells nor c-Src inactivated by PP2 increased ErbB-2 phosphorylation (Fig. 2c). Moreover, it is known that after TNFα binding to its receptors, protein kinase C (PKC) is involved in c-Src phosphorylation . Addition of 10 μM GF109203X, a PKC inhibitor, effectively abolished TNFα ability to induce c-Src phosphorylation and subsequently ErbB-2 phosphorylation at Tyr927 in C4HD cells (Fig. 2d). Taken together, these results strongly suggest that TNFα induces c-Src activation which phosphorylates ErbB-2 at Tyr877/927 residue.
TNFα induces ErbB-2/ErbB-3 heterodimerization and PI3-K/Akt pathway activation
ErbB-2 tyrosine kinase inhibitor but not Herceptin blocks TNFα-induced NF-κB transcriptional activation, cyclin D1 expression and cell proliferation
To further explore the role of ErbB-2 on NF-κB transcriptional activation, we used different human ErbB-2 mutants. SK-BR-3 cells were co-transfected with a kinase-negative (KN)-ErbB-2 expression vector or an ErbB-2 expression plasmid in which tyrosine 877 was point mutated to phenylalanine (ErbB-2-Y877F), together with a κB-Luc reporter vector. The presence of the mutated ErbB-2 constructs inhibited endogenous ErbB-2 phosphorylation, revealing a dominant negative activity exerted on the wild-type ErbB-2 molecules (Fig. 4b). Our results showed that KN-ErbB-2 or ErbB-2-Y877F blocked TNFα ability to induce NF-κB transcriptional activation (Fig. 4b), suggesting that phosphorylation at Tyr877 residue and the ErbB-2 autophosphorylation are indispensable events for TNFα-induced NF-κB transcriptional activation. Silencing ErbB-2 expression with ErbB-2 siRNAs completely inhibited TNFα-stimulated NF-κB transcriptional activation and IκBα phosphorylation in C4HD cells (Fig. 4c). Co-transfection with a human ErbB-2 expression vector completely recovered TNFα-induced NF-κB transcriptional activation that had been lost by siRNA to mouse ErbB-2 (Fig. 4c; Supplemental Fig. 4).
According to our results in which c-Src acts as an upstream activator of ErbB-2, we reasoned out that Herceptin should not inhibit c-Src activation by TNFα. Indeed in the presence of Herceptin, TNFα induced c-Src activity to the same degree as in control cells (Fig. 5d). These results led us to assess whether c-Src was involved in the mitogenic effect of TNFα. We performed a proliferation assay that showed that Dasatinib completely blocked TNFα ability to induce BT-474 proliferation (Fig. 5e). These results disclose that c-Src is a key player on TNFα-induced signaling cascade leading to breast cancer cell growth.
Previous studies have shown that c-Src activation is dependent on ErbB-2, but it has recently been reported that c-Src itself plays a role upstream on ErbB-2 phosphorylation. c-Src interacts directly with the catalytic domain of ErbB-2 [33, 34] and phosphorylates ErbB-2 at Tyr877/927 residue within the activation loop of the kinase domain, stabilizing it in an open and extended conformation which increases its intrinsic kinase activity . Here, we found that TNFα induces Tyr877/927 ErbB-2 phosphorylation through c-Src activation, in a PKC-dependent manner, in murine C4HD and human SK-BR-3 and BT-474 cells. We observed that TNFα induces Tyr416 phosphorylation of c-Src, and that this activated kinase phosphorylates ErbB-2 Tyr877/927 residue in vitro. Moreover, we also demonstrated that blockage of c-Src with the clinically used inhibitor, Dasatinib, inhibits TNFα-induced BT-474 cell proliferation. This piece of information would be useful in future evaluations of breast cancer specimens with the expression of TNFα and activated c-Src, of patients undergoing anti-ErbB-2 therapy. It was recently demonstrated that TNFR1 is constitutively associated with c-Src, and that TNFα induces recruitment of additional c-Src and increases its activity in HEK293 and MCF-7 cell lines . During the preparation of this manuscript, Yamaoka et al.  demonstrated that Src kinase activity as well as transactivation of EGFR and ErbB-2 is required for TNFα-induced survival of young adult mouse colon (YAMC) epithelial cells. In the present study, we also confirmed that c-Src is the kinase responsible for ErbB-2 transactivation, and we proved that c-Src mediates the phosphorylation on ErbB-2 Tyr877 residue. Moreover, we demonstrated that ErbB-2 transactivation is responsible for TNFα-induced proliferation of breast cancer cells overexpressing ErbB-2, and that TNFα elicits heterodimerization of ErbB-2 with ErbB-3. However, we could not detect any participation of EGFR. These differences might be attributed to tissue-specific signaling. Another interesting finding is that, although both TNFα and HRG stimulate p42/p44 MAPK activation, blockage of ErbB-2 by AG825 did not inhibit the activation of p42/p44 MAPK induced by TNFα. In contrast, AG825 blocked HRG-induced p42/p44 MAPK activation. These results reveal different signaling pathways induced by ErbB-2 transactivation caused by a classic ErbB-3/ErbB-4 ligand (HRG) or a “non-classic” activator (TNFα). On the other hand, transactivation of EGFR by TNFα is a well-documented mechanism that involves MMPs stimulation; in particular, of the TNFα converting enzyme, which releases EGFR ligands [18–20]. However, in this work we could not detect any evidence supporting the fact that ErbB-2 transactivation by TNFα was induced by cleavage of membrane-tethered ErbB ligands, since pretreatment with GM6001, a broad spectrum inhibitor of MMPs, did not modify TNFα-induced ErbB-2 phosphorylation either in C4HD cells or in SK-BR-3 cells.
An exciting novel finding of this study has been the demonstration that TNFα utilizes ErbB-2 as a downstream signaling partner in the generation of mitogenic signals. Our data supporting the fact that TNFα-induced proliferation of C4HD cells is dependent on ErbB-2 kinase activity was also confirmed in SK-BR-3 cells, where TNFα has already been reported as mitogenic , and in BT-474 cells. Our findings suggest that NF-κB activation induced by TNFα requires a functional ErbB-2 because (a) the inhibition of ErbB-2 by AG825, (b) the protein knockdown of ErbB-2 by siRNA strategy, (c) the fact that transfection with KN-ErbB-2 or ErbB-2-Y877F plasmids, impaired TNFα capacity to activate NF-κB and (d) the fact that co-transfection with siRNA to mouse ErbB-2 and reconstitution of ErbB-2 expression with a vector encoding human ErbB-2, restored TNFα-induced NF-κB activation. However, in NIH 3T3 cells, which have low levels of ErbB-2, transiently transfected with ErbB-2 plasmid, we observed no difference of TNFα effect on NF-κB activation as compared to the cells transfected with the empty vector (data not shown). These data suggest that ErbB-2-overexpressing cells have other/s signaling molecule/s that allow/s the cross talk between TNFα receptors and ErbB-2. We observed that either inhibition of ErbB-2 with AG825 or silencing ErbB-2 by siRNA treatment inhibited TNFα-induced activation of PI3-K/Akt and IκBα phosphorylation. These results are in line with our previous report showing that Akt behaves as an upstream molecule in the NF-κB activation cascade and this activation proceeds through the canonical pathway . Although it has long been known that amplified expression of ErbB-2 induces breast cancer resistance to TNFα-induced cytotoxicity , the mechanisms that promote this effect have scarcely been characterized. ErbB-2 overexpression has been found to constitutively activate Akt/NF-κB anti-apoptotic pathway, conferring resistance to TNFα in cancer cell lines . Yamaoka et al.  determined that loss of EGFR expression suppressed TNFα activation of Akt and induce apoptosis in YAMC cells, although they did not observe any modulation of IκB. On the other hand, Biswas et al.  demonstrated that NF-κB activation is predominantly found in the estrogen receptor (ER)-negative/ErbB-2 positive subclass in comparison with ER-positive human breast cancer tumors. Recently, this group demonstrated that NF-κB activation in this subclass of breast cancer is essential for tumorigenesis and tumor progression . In the present work, we illustrate that TNFα activates ErbB-2 and uses it as an intermediate to promote NF-κB activation, adding yet another layer of complexity in ErbB-2 overexpressing tumor signaling.
An extensive body of evidence has established ErbB-2 as a key mediator of tumor cell growth and survival, since ErbB-2 overexpression in nearly 30% of human breast cancers predicts an aggressive course of disease and poor prognosis . The first anti-ErbB-2 agent used in clinical practice is the humanized monoclonal antibody Herceptin . Given alone as first line treatment to metastatic breast cancer patients, Herceptin shows an overall response rate of 38% . It has been suggested that one likely mechanism of acquired resistance to Herceptin can be the amplification of ligand-induced activation of the ErbBs receptor . Consequently, small-molecule inhibitors of EGFR/ErbB-2 have progressed to clinical trials . Here, we report TNFα as a new ErbB-2 transactivating factor unexplored before, and demonstrate that TNFα-induced ErbB-2 transactivation leading to NF-κB activation and cell proliferation can be blocked at the level of ErbB-2 phosphorylation through AG825 or GW2974, specific EGFR/ErbB-2 tyrosine kinase inhibitors, but not at the level of antibody blockage of the receptor. Interestingly, Herceptin increased TNFα-induced NF-κB activity and IκBα phosphorylation in BT-474 but not in SK-BR-3 cells (Fig. 4a). Since BT-474 cell line expresses estrogen and progesterone receptors (PR) and as we have demonstrated that PR is involved in ErbB-2 phosphorylation , further studies have to be carried out to shed light on the above-mentioned finding.
In this work, we provide a suitable molecular explanation for the resistance to Herceptin as observed in clinical practice, and hypothesize that TNFα signaling may have augmented in patients unresponsive to monoclonal antibody therapy. Even when TNFα has been shown to be expressed in a significant proportion of human infiltrating breast cancers , the eventual worth of TNFα signaling as a prognostic factor in anti-ErbB-2 therapy is yet to be determined.
This work was supported by grants IDB 1728/OC-AR PICT 2006 0211 and PICT 2004 05-25301, both from the National Agency of Scientific Promotion of Argentina, PIP 5391 from the Argentine National Council of Scientific Research (CONICET) and by Oncomed-Reno CONICET 1819/03, from the Henry Moore Institute of Argentina and by grant KG090250 from the Susan G. Komen for the Cure. The authors wish to thank Dr Alfredo A. Molinolo (NIH, Bethesda, MD) for his constant help and support. We thank Dr C. Lanari for providing the MPA-induced mammary tumor model.