Breast Cancer Research and Treatment

, Volume 122, Issue 3, pp 685–697

EGFR over-expression and activation in high HER2, ER negative breast cancer cell line induces trastuzumab resistance

Authors

    • Oncology, Monogram Biosciences
  • Jianhuan Zhang
    • Oncology, Monogram Biosciences
  • Phets Nhonthachit
    • Oncology, Monogram Biosciences
    • Ross University School of Medicine
  • Elicia Penuel
    • Oncology, Monogram Biosciences
  • Chris Petropoulos
    • Oncology, Monogram Biosciences
  • Gordon Parry
    • Oncology, Monogram Biosciences
Preclinical study

DOI: 10.1007/s10549-009-0592-x

Cite this article as:
Dua, R., Zhang, J., Nhonthachit, P. et al. Breast Cancer Res Treat (2010) 122: 685. doi:10.1007/s10549-009-0592-x

Abstract

HER2 is gene amplified or over-expressed in 20–25% of breast cancers resulting in elevated HER2 activation. Trastuzumab (Herceptin), a humanized monoclonal antibody, targets activated HER2 and is clinically effective in HER2-over-expressing breast cancers. However, despite prolonged survival, treated breast cancer patients develop resistance. Resistance to trastuzumab occurs upon inactivation of HER2 regulatory proteins or upon up-regulation of alternative receptors. In particular, elevated levels of EGFR, present in estrogen receptor (ER) positive, trastuzumab-resistant BT-474 xenografts caused, a trastuzumab-resistant phenotype (Ritter et al. Clin Cancer Res 13:4909–4919, 2007). However, the role of EGFR in acquired trastuzumab resistance in ER negative cell models is not well defined. In this study, SKBR3 cell line clones expressing EGFR were generated to examine the role of EGFR over-expression on trastuzumab sensitivity in an, ER-negative breast carcinoma cell line. A stable clone, SKBR3/EGFR (clone 4) expressing moderate levels of EGFR remained sensitive to trastuzumab, whereas a stable clone, SKBR3/EGFR (clone 5) expressing high levels of EGFR, became resistant to trastuzumab. Depletion of EGFR by EGFR small-interfering RNAs in the SKBR3/EGFR (clone 5) reversed trastuzumab resistance. However, the SKBR3/EGFR (clone 5) cell line remained sensitive to lapatinib, an EGFR/HER2 inhibitor. Biochemical analysis using co-immunoprecipitation and proximity-based quantitative VeraTag assays demonstrated that high levels of EGFR phosphorylation, EGFR/EGFR homo-dimerization, and EGFR/HER2 hetero-dimerization were present in the trastuzumab-resistant cells. We conclude that EGFR over-expression can mediate trastuzumab resistance in both ER positive and ER negative cells and hypothesize that a threshold level of EGFR, in the absence of autocrine ligand production, is required to induce the resistant phenotype.

Keywords

Breast cancerEGFRVeratagHER2Trastuzumab resistance

Introduction

The ErbB receptor tyrosine kinase family consists of four members: HER1/ErbB1/EGFR, HER2/ErbB2, HER3/ErbB3, and HER4/ErbB4. EGFR, HER3, and HER4 are activated by ligand binding unlike HER2, for which there is no known ligand. As a consequence, HER2 exists in a constitutively activated configuration [1]. ErbB signaling plays an essential role in cell differentiation, proliferation, and survival; however aberrant signaling through the ErbB pathway has also been implicated in a number of human cancers [2]. Similarly, the aberrant signaling by HER2 that is due to HER2 gene amplification has been observed in ~20–25% of invasive breast carcinomas [3]. HER2 targeted therapies, including trastuzumab (Herceptin), a humanized monoclonal antibody against HER2, have been used successfully in treatment of both metastatic and adjuvant breast cancer for HER2 over-expressing tumors [4, 5]. Despite the clear clinical benefit of trastuzumab in treatment of breast cancer, tumors eventually become resistant, resulting in relapse. The molecular basis of resistance to trastuzumab has been extensively studied and a number of potential resistance mechanisms proposed, including loss of PTEN [6], increased HER2 phosphorylation [7], as well as, up-regulation of c-Met [8], insulin-like growth factor-I (IGFR-I) [9, 10] or EGFR, and its ligands [11]. In most of these studies, trastuzumab-resistant cell lines were selected by exposure to trastuzumab in cell culture or in tumor xenografts. Although selecting resistant cells in this manner may mirror how patients might develop resistance, it is difficult to rule out the possibility of multiple mechanisms of drug resistance even in clonally derived resistant cell lines.

Understanding whether EGFR over-expression contributes to trastuzumab resistance is clinically important, as potentially, approved EGFR drugs could be used to target trastuzumab resistant, EGFR over-expressing tumors. The up-regulation of EGFR and EGFR ligands responsible for trastuzumab resistance was identified in an in vivo, estrogen receptor (ER) positive BT474 cell line model [11]. Despite a clear role of EGFR in acquired resistance in this ER-positive cell line model, a specific role of EGFR in acquired resistance in an in vitro ER-negative cell line model has not been evaluated. This is significant as ErbB family signaling pathways exhibit crosstalk with the estrogen receptor [12]. Approximately 50% of HER2 over-expressing tumors are ER negative [13] with both ER-negative and ER-positive tumors have a similar clinical benefit on trastuzumab with ER-positive patients having a longer time to progression as compared to ER-negative patients [14]. It is possible that ER and ErbB family pathway interactions may contribute to resistance and the contribution of EGFR up-regulation in ER-negative verses ER-positive cells may be quite different. Interestingly, in a recent study, trastuzumab treatment of breast cancer cells leads to ER up-regulation [15]. Consequently, we sought to examine the role of EGFR over-expression in an ER negative, trastuzumab-resistant SKBR3 cell line. To gain better understanding of the mechanistic role of ErbB family receptors, we also quantified the protein expression and activation states of ErbB family members by proximity-based VeraTag assays [16, 17]. The results are described in the present manuscript.

Materials and methods

SKBR3 breast cancer cell line was obtained from American Type culture collection and maintained in DMEM/F12 (50:50) containing 10% fetal bovine serum (Invitrogen). Geneticin (G418) was obtained from Invitrogen. Plasmid pUSEamp-EGFR was purchased from Millipore. Cell titer-Glo Luminescent cell viability kit was from Promega. The active erlotinib was extracted from tablets as described [18]. Lapatinib was obtained from American Custom Chemical Corporation. Trastuzumab was a gift from Patrick Joseph, San Ramon, California, USA. VeraTag reporter molecules were synthesized as described [16]. Antibody-fluorescent VeraTag and antibody-biotin conjugates were prepared as described before [17].

Generation of EGFR over-expressing SKBR3 cell lines

SKBR3 cells were transfected with 2 μg of pUSEamp-EGFR plasmid using 6 μl of fugene-6 reagent. After 3 days of transfection, medium was replaced with DMEM/F12 (50:50) containing 10% FBS and 400 μg/ml G418. Cells were grown for 3–4 weeks. Several drug resistant clones were selected and screened for high EGFR expression by western blotting and VeraTag quantitative assays. SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) were selected for further experiments.

Proliferation assays

SKBR3 cells were seeded in triplicate at 5000 cells per well in a 96-well plate. After overnight incubation at 37°C, cells were treated in the absence or presence of varying concentration of drug. After 5 days at 37°C, viability was measured using Cell titer-Glo cell viability kit according to manufacturer’s instructions. The results represents mean ± SE of triplicate data points expressed as a percentage of relative luminescent counts compared with the same cell line grown in the absence of the drug.

siRNA transfections

siRNA smart pools specific for EGFR and non-targeted siRNA controls were obtained from Dharmacon. Cells were seeded at 50,000–100,000 cells/well in a 6-well tissue culture plate. Cells were transfected with siRNA (4 nM final concentration) using Dharmafect-1 reagent (Dharmacon). Three days post transfection, medium was replaced with fresh DMEM/F12 (50:50) containing 10% FBS in the absence or presence of 20 μg/ml trastuzumab. Cells were grown at 37°C for 5 days, trypsinized, and counted with a Coulter counter. The results represent mean ± SE of triplicate data points expressed as a percentage of cell number compared with the same cell line grown in the absence of the trastuzumab.

Western blot analysis and co-immunoprecipitation

Cells were seeded in tissue culture plates and incubated at 37°C until 60–75% confluence. Cells were serum starved overnight and treated with drug for 4 h at 37°C. Following drug treatment, cells were stimulated with either 16 nM heregulin for 10 min or 16 nM of transforming growth factor α for 10 min. Cells were washed with cold PBS and lysis was done using lysis buffer (50 mM Tris–HCl (pH 7.5), 1% TX-100, 150 mM NaCl, 50 mM β-glycerophosphonate, 50 mM NaF, 1 mM Na3VO4) and a cocktail of protease inhibitors. Lysates were incubated for 20 min on ice, centrifuged at 13,000 rpm at 4°C for 10 min, and clear supernatant was collected. Protein concentration was measured using bicinchoninic acid reagent (Pierce). Total cell lysates were run on SDS-PAGE and western blots were probed with various antibodies. Primary antibodies used for the western blotting include EGFR (clone 13) (BD transduction laboratories), HER2 antibody (Ab8) (Neomarkers), HER3 antibody (clone 2F12) (Neomarkers), Akt antibody (C67E7) (cell signaling), Phospho-Akt (ser 473) (587F12) (cell signaling), p44/42 MAP kinase (137F5) (cell signaling), Phospho-p44/42 MAP kinase (Thr202/Tyr204) (cell signaling), Phosphotyrosine (4G10 clone) antibody (Upstate). Horseradish peroxidase-linked IgG (Pierce) was used as secondary antibody. For co-immunoprecipitation, 100–400 μg of total lysates were mixed with 2–4 μg of antibody and incubated on ice for 1 h before adding 40–70 μl of protein A/G sepharose beads (SantaCruz Biotechnology). The lysate mixture was incubated overnight at 4°C on a rocking platform. Immunocomplexes were pelleted by centrifugation, washed five times with lysis buffer, and resuspended in 70 μl of SDS-PAGE sample buffer. Primary antibodies used for co-immunoprecipitation include HER3 (Santacruz biotechnology), EGFR (Ab12) (Neomarkers), and HER2 (Ab17) (Neomarkers).

EGFR phosphorylation, HER2 phosphorylation, HER3 phosphorylation, EGFR/HER2 heterodimerization, EGFR/EGFR homdimerization, HER2/HER2 homodimerization, and HER3/PI3K multiplex VeraTag assay

Primary antibodies used for the Veratag assays include EGFR (Ab11) (Labvision), HER2 (Ab5) (labvision), HER2 (Ab4) (Labvision), HER3 (1B4C3) (Santa Cruz), PI3K 100α (Millipore), and Phosphotyrosine (PT100) antibody (cell signaling). Fluorescent reporter “tags” (pro14, pro10, pro2, pro99) were synthesized and purified according to protocol described before (US patent 7,105,308). Total cell lysates were prepared and protein concentration was determined as described above. A series of seven 2-fold serial dilution of the protein lysates in total of 30 μl were prepared and transferred to a 96-well 2 μm membrane filter plate (Millipore). Subsequently, 5 μl of EGFR/HER2 antibody cocktail containing a mixture of biotin-HER2 (Ab5) (1 μg/ml), Pro14VeraTag-HER2 (Ab4) (0.1 μg/ml), Pro10VeraTag-EGFR (Ab11) (0.2 μg/ml), and Pro2VeraTag- phosphotyrosine PT100 (0.1 μg/ml) antibodies was added. In this multiplex assay, we simultaneously determine EGFR/HER2 heterodimerization and HER2 phosphorylation. The mixture was incubated in dark at room temperature for 45 min on a shaking platform. A 5 μl of streptavidin-conjugated beads containing pthalocyanin were added to the mixture and mixture was further incubated for 45 min. Immunocomplexes were pelleted by filtration, washed once with wash buffer 1 (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, 0.5% TX-100) and twice with wash buffer 2 (685 μM NaCl, 13.5 μM KCl, 50 μM Na2HPO4, 10 μM KH2PO4). Illumination buffer containing 3 pmol/l fluorescein and two CE internal markers (MF and ML) in buffer 3 (1370 μM NaCl, 27 μM KCl, 100 μM Na2HPO4, 20 μM KH2PO4) was added. The bound reporter VeraTags were released at room temperature by photo-activated cleavage with 670 nm light using a customized LED illuminator. The sample containing the released reporter VeraTags was collected, separated, and detected on ABI3100 capillary electrophoresis (22 cm capillary array) (Applied Biosystems) instrument under CE injections conditions of 6 kV for 80 s at 30°C. EGFR/EGFR homodimerization VeraTag assay was done as described above except EGFR/EGFR antibody cocktail containing a mixture of biotin-EGFR (Ab11) (0.3 μg/ml), pro10Veratag-EGFR (Ab11) (0.3 μg/ml), pro1Veratag-EGFR (Ab5) (0.3 μg/ml), and pro2Veratag-phosphotyrosine PT100 (0.3 μg/ml) was used. In this multiplex assay, we simultaneously determine total EGFR, EGFR/EGFR homodimerization, and EGFR phosphorylation. HER2/HER2 homodimerization assay was done as described above except HER2/HER2 antibody cocktail containing a mixture of biotin-HER2 (Ab5) (0.3 μg/ml), and pro10Veratag (Ab5) (0.3 μg/ml) was used. In this assay, we determine levels of HER2/HER2 homodimerization. HER3/PI3K protein–protein complex assay was also done as described above except HER3/PI3K antibody cocktail containing a mixture of biotin-HER3 (1B4C3) (2 μg/ml), pro99Veratag-HER3 (0.1 μg/ml), pro7Veratag-PI3K (0.1 μg/ml), and pro2Veratag phosphotyrosine PT100 (0.1 μg/ml) was used in the assay. In this multiplex assay we simultaneously determined total HER3, HER3 phosphorylation, and HER3/PI3K complex.

Data analysis

The identification and quantification of the VeraTag assay fluorescent peaks from a CE electropherogram were done using developed software as described [17]. The CE fluorescence signal or peak area of each VeraTag reporter tag is calculated as the peak height integrated over the peak elution time. A titration curve using a protein dilution series was curve fitted to determine the concentration of the analyte per mg of the input protein. The results represent slope ±SE expressed as relative peak area (RPA) per mg of input protein. The statistical significance of different slopes was generated by linear regression analysis using graphpad prism software.

Results

Co-expression of EGFR and HER2 in SKBR3 cell line

The SKBR3 cell line contains high levels of HER2 and low levels of EGFR. To investigate the role of EGFR expression on trastuzumab resistance, we established EGFR expressing SKBR3 cell clones. The SKBR3 cell line was transfected with a pURE-EGFR vector encoding wild-type EGFR under the control of cytomegalovirus (CMV) promoter. Following G418 antibiotic selection, stable clones with varying levels of EGFR expression were established. SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5), respectively, exhibited moderate and high levels of EGFR expression, whereas HER2 and HER3 levels remained unaltered (Fig. 1a). To determine the receptor expression on the cell surface, flow cytometry analysis was performed. EGFR, HER2, and HER3 levels in SKBR3, SKBR3/EGFR (clone 4), and SKBR3/EGFR (clone 5) were consistent with the western blots (data not shown). Lastly, estrogen receptor (ER) α expression level in parental and EGFR over-expressing SKBR3 cell lines were not altered (Fig. 1b).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig1_HTML.gif
Fig. 1

a, b Immunoblot analysis from cell extracts using antibodies against EGFR, HER2, HER3, and ERα in SKBR3, SKBR3/EGFR (clone 4), and SKBR3/EGFR (clone 5) cells. SKBR3, SKBR3/EGFR (clone 4), and SKBR3/EGFR (clone 5) cells were grown and lysed to prepare cell extracts. Immunoblot analysis was done to detect total EGFR, HER2, HER3, and ERα. β-actin was used as a loading control

Over-expression of EGFR and trastuzumab induced growth inhibition

We studied the growth of parental and EGFR over-expressing SKBR3 cells in the absence and presence of trastuzumab. We selected a dose of trastuzumab ranging from 10 to 100 μg/ml because clinical trial data in humans showed trough concentration of ~20 μg/ml and peak plasma concentration of 185 μg/ml [19]. Growth was reduced at 10 μg/ml trastuzumab dose in both parental and SKBR3/EGFR (clone 4) cell lines (Fig. 2a) with parental and SKBR3/EGFR (clone 4), respectively, showing ~30 and 25% growth inhibition. In contrast, growth was minimally reduced by trastuzumab in the SKBR3/EGFR (clone 5) cell line that expressed high levels of EGFR and HER2 (Fig. 2a). Additionally, the trastuzumab-resistant cells maintained the phenotype on matrigel (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig2_HTML.gif
Fig. 2

Over-expression of EGFR in SKBR3 cell line inhibits growth by trastuzumab. Sensitivity of cell lines was determined by dose–response analysis. a SKBR3 and SKBR3/EGFR cells were plated in triplicate at 5000 cells/well in a 96-well plate. After overnight incubation, medium was replaced with a fresh complete growth medium with 10% FBS containing indicated concentrations of trastuzumab. After 5 days of incubation, the medium was replaced with a fresh drug-free medium and viability was measured using CellTiter Glo Cell proliferation assay (Promega Corporation, Madison, WI, USA). Viability of the cells was calculated as the percentage of cells in trastuzumab-treated culture compared to untreated cells

si-RNA-mediated EGFR downregulation and trastuzumab-induced growth reduction

To further define the requirement for a high level of EGFR expression for trastuzumab resistance in SKBR3 and SKBR3/EGFR cell lines, we down-regulated EGFR by EGFR-specific si-RNAs and measured cell growth in the absence and presence of trastuzumab. A considerable amount of EGFR was downregulated in SKBR3/EGFR (clone 5) after 3 days post-EGFR siRNA transfection (Fig. 3a). Greater than 80% downregulation of EGFR was determined by quantitative VeraTag assay (data not shown). Three days post-EGFR siRNA transfection, cells were treated with trastuzumab for 5 days. The reduction of EGFR expression by siRNA transfection in the SKBR3/EGFR (clone 5), restored trastuzumab sensitivity while non-targeted control si-RNA knockdown transfections performed in parallel remained trastuzumab resistant (Fig. 3b). Conversely, parental SKBR3 cells showed no change in trastuzumab sensitivity after EGFR knockdown. In the presence of trastuzumab, growth of both parental SKBR3 and SKBR3/EGFR (clone 5) was reduced by 40% upon reduced EGFR expression by siRNA transfections.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig3_HTML.gif
Fig. 3

Trastuzumab-resistance in SKBR3/EGFR (clone5) is abolished by siRNA knockdown of over-expressed EGFR. a SKBR3 and SKBR3/EGFR (clone 5) cells were seeded in a 6-well plate. After overnight incubation, cells were transfected with either 4 nM EGFR-specific siRNA or non-targeted siRNA control. After 3 days of incubation, cells were lysed and soluble lysates were prepared. Western blot analysis was done to detect the total levels of EGFR. b SKBR3 and SKBR3/EGFR (clone 5) cells were seeded in triplicate in a 6-well plate. After overnight incubation, cells were transfected with either 4 nM EGFR-specific siRNA or non-targeted siRNA control. After 3 days of incubation, medium was removed and fresh drug-free medium or fresh medium containing 20 μg/ml trastuzumab was added. After 5 days of incubation, cell was trypsinized and counted with a coulter counter

EGFR and HER2 co-expression and tyrosine kinase inhibitor sensitivity

Lapatinib is a dual HER2/EGFR kinase inhibitor that is clinically active in HER2 positive patients that have progressed on trastuzumab therapy [20]. Erlotinib is an EGFR kinase inhibitor that is clinically active in non-small cell lung and pancreatic cancer [21, 22]. To evaluate whether EGFR-targeted inhibitors would be effective in the trastuzumab-resistant cell lines, we examined proliferation of SKBR3 and SKBR3/EGFR clonal cell lines in the absence and presence of these drugs. SKBR3, SKBR3/EGFR (clone 4), and SKBR3/EGFR (clone 5) cell lines were equally sensitive to lapatinib (Fig. 4a). In all cases, growth was inhibited by 95% at a 1 μM dose of lapatinib indicating that lapatinib sensitivity in HER2 over-expressing cell lines is not affected by EGFR over-expression. On the other hand, while SKBR3 parental showed ~20% growth inhibition in the presence of 1 μM erlotinib, SKBR3/EGFR (clone 4), and SKBR3/EGFR (clone 5), respectively, showed a statistically significant (P = 0.0127 for SKBR3/EGFR (clone 4) and P = 0.02 for SKBR3/EGFR (clone 5)) increase in growth inhibition (~30%) at the same dose (Fig. 4b). This indicates that erlotinib sensitivity in HER2 over-expressing cells might be modulated by EGFR expression levels.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig4_HTML.gif
Fig. 4

a Lapatinib sensitivity of SKBR3 and SKBR3/EGFR cell lines. b Erlotinib sensitivity of SKBR3 and SKBR3/EGFR cell lines. Sensitivity of cell lines was determined by dose–response analysis. SKBR3 and SKBR3/EGFR cells were plated in triplicate at 5000 cells/well in a 96-well plate. After overnight incubation, medium was replaced with a fresh complete growth medium with 10% FBS containing indicated concentrations of lapatinib or erlotinib. After 5 days of incubation, the medium was replaced with a fresh drug-free medium and viability was measured using CellTiter Glo Cell proliferation assay (Promega Corporation, Madison, WI, USA). Viability of the cells was calculated as the percentage of cells in drug-treated culture compared to untreated cells

Biochemical characterization of EGFR and HER2 over-expressing SKBR3 cell lines

Basal EGFR, HER2, and HER3 phosphorylation levels were determined in EGFR and HER2 co-expressing SKBR3 cell lines. Each receptor was immunoprecipitated and phosphorylation was evaluated by western blot analysis using a phosphotyrosine antibody. In comparison to the parental SKBR3 cell line, SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) cell lines showed significantly higher levels of EGFR phosphorylation (Fig. 5a) whereas HER2 and HER3 phosphorylation levels remained unchanged. Basal level of HER3 phosphorylation in SKBR3 cells is consistent with the recent data [23]. The level of EGFR/HER2 and HER2/HER3 heterodimer complexes in the clonal cell lines was evaluated by co-immunopreciptiation and western immunoblotting. The SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) cell lines that expressed higher levels of EGFR than the SKBR3 parental cells exhibited, higher levels of basal EGFR/HER2 complexes (Fig. 5b). HER2/HER3 heterodimer levels were not affected by the over-expression of EGFR in SKBR3 cell lines (Fig. 5b).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig5_HTML.gif
Fig. 5

Immunoblot analysis of immunoprecipitated EGFR, HER2, and HER3. a, b SKBR3 and SKBR3/EGFR cells were grown to subconfluence. Cells were lysed and soluble lysates were incubated with EGFR, HER2, or HER3 antibodies. The pulldown immunocomplexes on the beads were washed and eluted proteins were separated on SDS-PAGE. The proteins were transferred on nitrocellulose and immunoblotted with indicated antibodies. c Subconfluent SKBR3 and SKBR3/EGFR cells were treated overnight with or without lapatinib (0.1 μM) or trastuzumab (20 μg/ml) and cell lysates were prepared. Western blot analysis was done with indicated antibodies. β-actin was used as a loading control

HER family receptor expression and the status of downstream MAPK and Akt phosphorylation in the presence of lapatinib or trastuzumab were evaluated in the SKBR3 cell lines. While HER2 expression was slightly increased by lapatinib, there was a slight decrease in HER2 in the presence of trastuzumab in the parental SKBR3, SKBR3/EGFR (clone 4), and SKBR3/EGFR (clone 5) cell lines (Fig. 5c). However, lapatinib effectively reduced both P-MAPK and P-Akt levels (Fig. 5c). Interestingly, the SKBR3/EGFR (clone 5) cell line was particularly sensitive to lapatinib based on P-Akt inhibition.

Quantitative measurement of ErbB homodimers, heterodimers, and phosphorylation in EGFR and HER2 co-expressing cell lines by VeraTag assay

The VeraTag assay technology is an immuno-based detection method to quantify analytes or protein complexes as either soluble protein lysates or as formalin-fixed paraffin embedded tissues. The VeraTag technology has been applied for measurements of EGFR and HER2 expression and homodimerization in formalin-fixed paraffin embedded cell lines as well as breast cancer tissue samples [16, 17]. The method is outlined in Fig. 6a. Briefly, soluble cell lysates are incubated with a mixture of antibodies that are either conjugated to fluorescent VeraTag reporters or to biotin. Streptavidin-conjugated beads containing a photo-sensitizer are added to the mixture that is subsequently illuminated at 670 nM. Photo-activation generates a reactive singlet oxygen that cleaves thio-ether bonds in close proximity, releasing the fluorescent VeraTag reporter, whereupon VeraTags are collected, separated, and quantified by capillary electrophoresis (CE).
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig6_HTML.gif
Fig. 6

a Schematic of quantitative proximity-based VeraTag assay. b Schematic of quantitative proximity-based homodimer and heterodimer VeraTag assay

By altering the antibodies included in the VeraTag assay, heterodimerization and homodimerization can be measured. In both instances, one antibody is conjugated to a VeraTag reporter and the other directly or indirectly to a photo-sensitive cleavage agent. To measure homodimerization two identical antibodies that compete for the same epitope are used (Fig. 6b) while to measure heterodimerization, two different antibodies targeting distinct proteins are used (Fig. 6c). Since VeraTags reporter molecules have distinct mobilities and are readily differentiated from one another during capillary electrophoresis, assays can be multiplexed.

In order to assess the effect of drugs on ligand-induced signaling in cell lines, the status of HER family phosphorylation and dimerization was evaluated in the presence of growth factor alone, or in combination with various concentrations of drug. First, we quantified the total amount of EGFR, HER2, and HER3 receptors by VeraTag assay. SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5), respectively, showed ~9-fold and ~32-fold more EGFR as compared to parental SKBR3 while total HER2 and HER3 levels were not changed significantly (Table 1). This is consistent with the immunoblot data (Fig. 1a). Basal EGFR phosphorylation in SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) was also higher by ~6 and ~20 times, respectively, compared to the parental SKBR3 cells (Fig. 7a). In contrast to the higher constitutive EGFR phosphorylation, basal levels of HER2 phosphorylation were similar in parental SKBR3 and EGFR over-expressing SKBR3 cell lines (Fig. 7a). In contrast to our western blot data (Fig. 5a), we could not detect basal HER3 phosphorylation under these conditions. The reason for this anomaly is that in order to be in the linear range of the HER3 multiplex assay we could only use 20 μg or lower protein input in the assay. At higher protein input, we were able to detect basal HER3 phosphorylation (data not shown). In addition, it is important to note that a prerequisite of using two antibodies in close proximity in the VeraTag phosphorylation assay results in high specificity. It is rare to find two antibodies bind non-specifically to the same target in close proximity. All three cell lines could be efficiently ligand activated based upon increased either EGFR and HER2 phosphorylation in response to TGFα stimulation or HER3 phosphorylation in response to heregulin (HRG). TGFα ligand-induced EGFR and HER2 phosphorylation was much higher in SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) relative to the parental SKBR3 cells, whereas HRG ligand-induced HER3 phosphorylation was equivalent in all three cell lines (Fig. 7a). Lapatinib potently inhibited ligand-induced EGFR, HER2, and HER3 phosphorylation while trastuzumab was either not effective in inhibiting receptor phosphorylation (Fig. 7a) or resulted in a slight increase in HER2 phosphorylation which is consistent with previous studies [2325].
Table 1

Total receptor protein expression of EGFR, HER2, and HER3 in SKBR3 and SKBR3/EGFR cell lines by quantitative VeraTag assay

Receptor

RPA/mg protein

SKBR3

SKBR3/EGFR (clone 4)

SKBR3/EGFR (clone 5)

EGFR

9 ± 0.7

77 ± 1.6

282 ± 7.5

HER2

2266 ± 31

2535 ± 114

2431 ± 30

HER3

36 ± 0.8

41 ± 0.8

39 ± 0.8

https://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig7a_HTML.gifhttps://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig7b_HTML.gifhttps://static-content.springer.com/image/art%3A10.1007%2Fs10549-009-0592-x/MediaObjects/10549_2009_592_Fig7c-d_HTML.gif
Fig. 7

a, b, c P-EGFR, P-HER2, P-HER3, EGFR/HER2 heterodimer, EGFR/EGFR homodimer, HER3/PI3K complex, and HER2/HER2 homodimer analysis by proximity-based VeraTag assay. SKBR3 and SKBR3/EGFR cells were grown to subconfluency, serum starved overnight, and stimulated with indicated doses of TGFα or heregulin. Cell lysates were prepared and subjected to proximity-based VeraTag assay as described in “Material and methods” section

Next, we evaluated EGFR/HER2, EGFR/EGFR, and HER2/HER2 dimerization by VeraTag assays. Basal EGFR/HER2 heterodimerization and EGFR/EGFR homodimerization were higher in SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) than in the SKBR3 parental cell line (Fig. 7b). EGFR homodimers were increased 20-fold in SKBR3/EGFR (clone 4) and 100-fold in SKBR3/EGFR (clone 5), relative to SKBR3 cells. EGFR/HER2 heterodimerization was also increased almost 4-fold in SKBR3/EGFR (clone 4) and 11-fold SKBR3/EGFR (clone 5) cell line. On the other hand, HER2 homodimerization in SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) decreased almost 2-fold as compared to SKBR3 parental cells (Fig. 7d). Interestingly, while there was increased basal EGFR/HER2 dimerization in SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5), basal HER2 phosphorylation was not changed (Fig. 7a, b).

TGFα-induced EGFR homodimerization and EGFR/HER2 heterodimerization in all SKBR3 cell lines. These ligand-dependent dimerization events were increased in the cell lines with higher EGFR expression. While ligand-induced EGFR homodimers and EGFR/HER2 heterodimers were slightly increased in the presence of low doses of lapatinib (10 nM), dimerization was reduced at higher concentrations of the drug (160 nM) (Fig. 7b). Finally, in order to assess the HER3/PI3K pathway, we quantified the heregulin-induced HER3/PI3K complex. As expected, heregulin leads to an increase in HER3/PI3K complex (Fig. 7c). Basal levels of HER3/PI3K complex in parental SKBR3 and SKBR3/EGFR (clone 4) were not detectable under lower protein input (20 μg or lower) conditions. We have been able to detect albeit lower level of basal HER3/PI3K complex using higher protein input (data not shown) which is consistent with the recent data [23]. Lapatinib, reduced heregulin-induced HER3/PI3K complex in all cell lines (Fig. 7c). Conversely, trastuzumab had little effect on the heregulin-induced HER3/PI3K complex (Fig. 7c). SKBR3/EGFR (Clone 5) showed albeit a small but detectable levels of basal HER3/PI3K complex.

Discussion

The expression of ER/PR verses over-expression of HER2 dictates clinical treatment of breast cancer patients. While both treatments have proven clinical efficacy, resistance emerges in both instances [12, 15]. Furthermore, several studies have demonstrated a complex interplay of ErbB family receptor switching that maintains critical downstream signaling for the progression of HER2 positive tumors and can result in trastuzumab resistance [7, 911]. In this study, we investigated the potential of EGFR to confer resistance to trastuzumab in an ER-negative HER2 positive SKBR3 cell line, by exogenously expressing EGFR. The data indicate that resistance toward trastuzumab in high HER2 over-expressing cells can be achieved only when a very high level of EGFR is co-expressed (in the absence of ligand). However, under these conditions, the trastuzumab-resistant phenotype can be directly attributed to EGFR expression as targeted siRNA reduction of EGFR reversed the trastuzumab-resistant effect. Furthermore, our data suggest that the EGFR-driven resistance is mediated by the formation of EGFR homodimers and EGFR-HER2 heterodimers that rescue activation of downstream signals including MAPK and AKT.

SKBR3, SKBR3/EGFR (clone 4), and SKBR3/EGFR (clone 5) cell lines showed basal levels of EGFR/EGFR, EGFR/HER2, and HER2/HER2 preformed homodimers and heterodimers. Interestingly, as compared to parental SKBR3, over-expression of EGFR in SKBR3/EGFR (clone 4) and SKBR3/EGFR (clone 5) resulted in higher basal EGFR/HER2 dimerization while decreasing HER2/HER2 homodimerization. Similarly preformed homodimer and heterodimer structures have also been observed by bimolecular fluorescence complementation assay [26]. Similarly, EGFR was also demonstrated to preform homodimers on the cell surface independent of ligand binding [27].

Over-expression of EGFR in the SKBR3/EGFR (clone 5) caused a shift in drug susceptibility. Specifically, the SKBR3 HER2 over-expressing, ER-negative cells became resistant to trastuzumab and became slightly more sensitive to the EGFR tyrosine kinase inhibitor, erlotinib. The data is consistent with a study that showed that long-term trastuzumab exposure of trastuzumab-resistant cell lines induces increased EGFR expression and also increased de-novo sensitivity to the EGFR-targeted agents such as gefitinib or cetuximab [28]. Interestingly, the HER2/EGFR dual kinase inhibitor, lapatinib blocked proliferation of all cells lines equally, despite the differences in EGFR expression. This is in agreement with a recent study that lapatinib activity is not dependent on EGFR levels [29]. Despite any observable difference in lapatinib sensitivity between the different cell lines, based on proliferation assays, lapatinib was most effective at blocking AKT phosphorylation in the cells with the highest EGFR expression. Interestingly, MAPK was the least inhibited by lapatinib in these cells possibly suggesting compensatory signaling between the MAPK and AKT pathways [30]. Lapatinib also increased the levels of total HER2 while trastuzumab resulted in a moderate decrease in HER2 downregulation. This data is consistent with a recent study demonstrating accumulation of HER2 in the presence of lapatinib [31].

Trastuzumab sensitivity has also been linked to P-HER2 (Tyr 1248) phosphorylation [7]. However, in our study, no obvious alteration in HER2 phosphorylation was observed in the SKBR3/EGFR (clone 5) resistant cell line, suggesting that HER2 phosphorylation alone might not be sufficient to predict trastuzumab resistance. The level of ligand-induced EGFR/HER2 heterodimerization was slightly higher at lower doses of lapatinib while it was reduced at higher doses. Similar increase in EGFR/HER2 dimerization complex was recently observed when SKBR3 cells were incubated with lower dosage of lapatinib [31]. Quantitative analysis of EGFR/EGFR homodimers showed a substantial increase in homodimerization coupled with increased EGFR phosphorylation in trastuzumab resistance SKBR3/EGFR (clone 5) cell line suggesting that EGFR activation plays an important role in trastuzumab sensitivity.

HER2 positive breast cancer patients respond to trastuzumab [4, 5] but most of the responders eventually escape trastuzumab therapy and develop resistance. Our data suggests that EGFR activation may modulate trastuzumab sensitivity. VeraTag technology allows for the quantitative measurement of protein levels and interaction states with high sensitivity and specificity. In addition, two-antibody requirement against the same target in close proximity in the VeraTag assay further enhances the specificity of the phosphorylation and other assays. The technology may advance the potential to profile ErbB pathway activation in tumor samples. Overall, our results demonstrate the role of over-expressed EGFR in trastuzumab resistance in ER-negative cells and support further clinical investigation of the use of EGFR-targeted and trastuzumab-based therapies in breast cancer patients resistant to trastuzumab.

Acknowledgments

The authors wish to thank John Winslow and Youssouf Badal for their input on the proximity-based assays. Lili Chen helped in FACS analysis. Yining Shi and Sailaja Pidaparthi helped in the initial development of VeraTag assays. Hasan Tahir helped in the synthesis of the VeraTag reporters. Jeff Sperinde helped in setting up a method to calculate analyte concentration by slope analysis.

Copyright information

© Springer Science+Business Media, LLC. 2009