Introduction
The human epidermal growth factor receptor 2 (HER2, ErbB2, or HER2/neu) is a member of the HER receptor tyrosine kinase (RTK) family, which includes three other members: epidermal growth factor receptor (EGFR or HER1), HER3, and HER4. Homo- and hetero-dimerization of ligand-bound HER receptors results in activation of multiple pathways, including the p44/42 mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K) pathways, which regulate cell proliferation and apoptosis [
1‐
3]. HER2, the preferred heterodimerization partner of the other HER receptors, does not have a ligand and is activated by overexpression and homodimerization, or by ligand-mediated stimulation of another HER receptor through heterodimerization. Approximately 20% of human breast cancers are HER2-amplified, and overexpression correlates with aggressive tumor behavior and poor patient outcome [
4].
To date, two distinct HER2-targeting agents, trastuzumab (T) and lapatinib (L), have been FDA-approved, and both have proven efficacy in the clinical setting [
5‐
8]. Trastuzumab is a humanized monoclonal antibody that binds to the extracellular domain of HER2, disrupting HER signaling and inducing antibody-dependent cell-mediated cytotoxicity (ADCC) [
9,
10]. Lapatinib, a small-molecule EGFR/HER2 dual tyrosine kinase inhibitor (TKI), antagonizes the kinase activity of these receptors, inhibiting phosphorylation of their substrates and downstream signaling [
11,
12]. Despite their proven clinical benefit,
de novo and acquired resistance to both L and T is common [
13,
14].
The HER signaling system has been described as a complex, robust, and redundant biological network, modulated by positive and negative feedback circuits [
2]. These features, which protect the system from various perturbations, can also play a key role in resistance to drugs targeting this pathway. As such, multiple escape mechanisms to circumvent inhibition of the HER system have been reported to cause resistance [
15,
16], including compensatory activation of the HER network [
17‐
19] or activation of other redundant survival pathways in the cell [
20,
21]. Therefore, multi-targeted therapies might be the optimal approach to prevent resistance in some patients.
Multiple levels of crosstalk between estrogen receptor (ER) and HER2 have been identified [
20,
21]. Our laboratory has previously shown that HER2 overexpression contributes to
de novo and acquired resistance in various endocrine therapies [
22,
23]. Similarly, in the clinical setting, gene amplification of HER2 is associated with resistance to endocrine therapy [
24‐
26]. Conversely, anecdotal observations from the clinic showed up-regulation of ER following treatment with trastuzumab in several patients with HER2-positive tumors [
27‐
29]. Likewise, a retrospective study suggested a greater benefit of lapatinib in those patients with HER2-amplified tumors that are ER- and PR-negative, compared with hormone receptor positive patients [
30]. An ER-positive/HER2-positive breast cancer cell line, BT474, has been reported to acquire resistance to lapatinib
in vitro by up-regulating ER [
20,
21]. However, it is not yet fully established if this up-regulation of ER expression and/or activity can function as an escape mechanism to cause resistance to HER2 targeted therapy in other cell lines or in human breast cancer.
We and others previously hypothesized that a common mechanism of resistance to single agent anti-HER2 therapy is the incomplete blockade of the HER pathway and its multiple potential homo- and heterodimer pairs. We then reported that combination regimens including L + T were superior to single agent therapy and were capable of eradicating most HER2-positive xenografts
in vivo [
24,
31]. However, some tumors still developed acquired resistance. In addition, we also showed that optimal antitumor effect in one cell line, MCF7-HER2, required endocrine therapy to block ER.
To further study the mechanisms of resistance to HER2-targeted therapies, we developed a panel of over 10 different HER2-positive human breast cancer cell lines de novo or acquired resistant to T, L, or L + T. We find that while de novo and acquired resistance to T is associated with reactivation of the HER2 pathway, resistance to L or L + T is due to alternative signaling through the ER pathway, providing clues to strategies to improve HER2-targeted therapies in the clinic.
Materials and methods
Cell lines and reagents
The human breast cancer cell line BT474 was obtained from AstraZeneca (Cheshire, UK) [
24]. UACC-812, AU-565, and HCC-1569 cell lines were purchased from the American Type Culture Collection (Manassas, VA, USA). MDA-MB-361, MDA-MB-453, HCC-1954, ZR75-30, SKBR-3, and HCC-202 cell lines were obtained from Dr. Joe Gray (Berkeley Lab, Berkeley, CA, USA) [
32]. SUM-190 and SUM-225 cells were obtained from Dr. Stephen Ethier (Wayne State University, Detroit, MI, USA). MCF7-HER2 cells were established as previously described [
33]. BT474, UACC-812, MDA-MB-361, and MDA-MB-453 cell lines were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 1% penicillin-streptomycin-glutamine (PSG). AU-565, HCC-1569, HCC-1954, ZR75-30, and HCC-202 cells were cultured in RPMI 1640 with 10% heat-inactivated FBS and 1% PSG. SKBR3 cells were grown in McCoy's 5A with 10% heat-inactivated FBS and 1% PSG. SUM-190 cells were maintained in Ham's F12 media with 5 μg/ml insulin, 1 μg/ml hydrocortisone, 5 mM ethanolamine, 10 mM HEPES, 5 μg/ml transferrin, 10 nM triiodothyronine, 50 nM sodium selenite, and 0.5 g/l bovine serum albumin (BSA). SUM-225 cells were grown in Ham's F12 media with 5% heat-inactivated FBS, 1% PSG, 5 μg/ml insulin, and 1 μg/ml hydrocortisone. Cell lines resistant (R) to HER2-targeted therapy were generated by long term culture of the cells in their original media with increasing concentrations of trastuzumab (1 to 50 μg/ml), lapatinib (0.1 to 1 μM), or both. For cells showing no growth inhibition, the treatment duration was at least three months, while responsive cells were cultured with their respective treatments until growth resumed. The time to the development of resistant growth varied from 3 to 12 months.
Trastuzumab (Herceptin) was acquired from Genentech (San Francisco, CA, USA) and dissolved in sterile distilled water. Lapatinib (Tykerb) was obtained from GlaxoSmithKline (US headquarters in Research Triangle Park, NC, USA) and prepared with dimethyl sulfoxide (DMSO). Fulvestrant (Faslodex) was obtained from AstraZeneca and prepared with ethanol.
Cell growth assay
A total of 5,000 cells/well of the parental or resistant cell lines, cultured with their individual treatments, were plated in 96-well plates 24 hours before beginning respective additional treatments, which consisted of 10 μg/ml trastuzumab, 1 μM lapatinib, the combination of trastuzumab with lapatinib, or 10-7 M fulvestrant. Cell growth was assessed at different time points (zero, three, six, and nine days). Cell cultures were fixed with 4% glutaraldehyde and stained with 0.05% methylene blue. The dye was subsequently extracted with 3% HCl and absorbance measured at 655 nm. Growth fold change was determined by ((O.D. 655 nm at six days/O.D. 655 nm at zero days) Treatment)/((O.D. 655 nm at six days/O.D. 655 nm at zero days) Control). Growth curve and growth fold change experiments were executed in quadruplicate.
Immunohistochemistry (IHC)
Cells were fixed in 10% neutral buffered formalin prior to processing and paraffin embedding. Blocks were then organized into a 3-mm core tissue array and IHC was performed on 3-micron sections from these arrays [
24]. Briefly, after deparaffinization, sections were subjected to epitope retrieval in tris-HCl buffer (pH 9.0) and then blocked in 3% hydrogen peroxide for 10 minutes. Slides were incubated with primary antibody to ER (Vector Labs, Burlingame, CA, USA), PR (Dako Cytomation, Carpinteria, CA, USA), or phospho-HER2-Tyr877 (Cell Signaling Technology, Beverly, MA, USA), for one hour. Immunodetection was performed with the EnVision+ System (Dako Cytomation, Carpinteria, CA, USA).
Immunoblotting assay
Cells were lysed in buffer consisting of 10% Triton X100, 50 mM Hepes (pH 7.4), 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 100 mM NaF, 10 mM NaPPi, 10% glycerol, 1 mM Na3VO4, and 1X protease inhibitor cocktail (Roche Molecular Biochemicals, Indianapolis, IN, USA). Protein lysates were collected and microcentrifuged at 14,000 g for 10 minutes at 4°C. Cell supernatants were aliquoted and stored at -80°C. Protein concentration was measured using the Bio-Rad Protein Assay kit (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer's directions. Equivalent amounts of protein (25 μg) from each sample were separated under denaturing conditions by electrophoresis on polyacrylamide gels containing sodium dodecyl sulfate (SDS-PAGE) and transferred by electroblotting onto nitrocellulose membranes (Invitrogen, Carlsbad, CA, USA). The blots were first stained with Ponceau S to confirm uniform loading and transfer, followed by immunoblotting with the specific primary antibodies according to the manufacturer's instructions. Briefly, blots were blocked with appropriate blocking buffer and then reacted at 4°C with primary antibodies at dilutions as per the manufacturer's directions overnight. Primary antibodies were: phospho-EGFR-Tyr1173 (Epitomics, Burlingame, CA, USA), EGFR, phospho-HER2- Tyr877, phospho-HER2-Tyr1221, HER2, phospho-HER3-Tyr1289, phospho-AKT-Thr308, phospho-AKT-Ser374, AKT, phospho-p44/42 MAPK- Thr202/Tyr204, p44/42 MAPK, β-actin, insulin-like growth factor-I receptor (IGF1R), cleaved PARP, caveolin-1 (Cav-1), Bik (all from Cell Signaling Technology), phospho-HER2-Tyr1248, HER3 (from Millipore, Billerica, MA, USA), ERα (Lab Vision, Fremont, CA, USA), progesterone receptor (PR), Cyclin-D1, and Bcl2 (from Santa Cruz Biotechnology, Santa Cruz, CA, USA). Blots were then incubated with a horseradish peroxidase-linked or a fluorescently-labeled secondary antibody for one hour, after which the labeled proteins were visualized by chemiluminescence or by the Odyssey Infrared Imaging System (LI-COR Biosciences, Inc., Lincoln, NE, USA). Gels were produced at least three independent times. For HER quantitation, protein levels of three independent samples from each resistant cell line were quantified with the Odyssey Infrared Imaging System and normalized to β-actin (protein levels/actin levels).
Quantitative reverse transcription-polymerase chain reaction (qRT-PCR)
Total RNA was extracted using the RNeasy Mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer's directions. For ER and PR analysis, the cDNA of each sample was generated by Superscript II reverse transcriptase and random hexamers (Invitrogen). Real time quantitative PCR (qRT-PCR) was performed using SYBR Green PCR Master Mix (ABI, Carlsbad, CA, USA), with human β-actin acting as an endogenous control. For analysis of HER ligands and receptors, gene expression was quantified using 100 ng of total RNA and Taqman One-Step Universal Master Mix in each qRT-PCR reaction, as described previously [
19]. Normalization of EGFR family receptor and ligand gene expression was performed using the house-keeping gene
HP1BP3 (heterochromatin protein 1, binding protein 3). All qRT-PCR reactions were performed in triplicate in a standard 96-well plate format with the ABI 7500 Real-Time qPCR System. Fold changes in mRNA expression were determined by the 2-ΔΔCt method. Target primer and probe sequences are available in supplemental material (Additional file
1).
Xenograft studies
UACC-812 (ER-positive/HER2 amplified) cells were maintained as described in the "Cell lines and reagents" section. Animal care was in accordance with institutional guidelines. UACC-812 (ER-positive/HER2 amplified) xenografts were established in ovariectomized five- to six-week-old athymic mice (Harlan Sprague Dawley, Madison, WI, USA) supplemented with estrogen pellets by inoculating 5 × 10
6 cells subcutaneously as described previously [
24]. When tumors reached the size of 150 to 200 mm
3 (two to four weeks), mice bearing the UACC-812 xenografts were randomly allocated to eight treatment groups, including continued estrogen (E2), E2 plus trastuzumab, E2 plus lapatinib, E2 plus the combination regimen (L + T), estrogen deprivation alone (ED) by removal of the estrogen pellets, ED plus trastuzumab, ED plus lapatinib, and ED plus the combination regimen. Each treatment group contained a minimum of 12 mice. Tumor volumes were measured weekly as previously described [
24]. Each tumor analyzed was from a different mouse.
siRNA transfection
Pooled small-interfering RNA (siRNA) oligos targeting EGFR, HER2, HER3, ERα, and nontargeting siRNA were purchased (Dharmacon, Lafayette, CO, USA). Cells were transfected with siRNA by reverse transfection per the manufacturers' directions. Briefly, 5,000 cells/well were seeded into 96-well plates containing a pre-incubated mixture of pooled siRNA oligos at 50 nM final concentration and Lipofectamine RNAiMax (Invitrogen) diluted in Opti-MEM (Invitrogen). The appropriate cell-specific medium supplemented with the relevant, respective drugs was added 24 hours after transfection and the effect of siRNA was determined after an additional 48 hours. For parallel protein expression analysis, 2 Χ 105 cells/well were plated into six-well plates and subjected to the transfection protocol as above.
In vitro cell proliferation assay and apoptosis assay
The cell proliferation assay was performed using the Click-iT EdU (5-ethynyl-2'-deoxyuridine) Microplate Assay (Invitrogen) according to the manufacturer's directions. Following transfection with siRNA for 72 hours, cells were cultured with 10 μM EdU for 4 hours and the proliferation rate was analyzed by the Celigo Cytometer (Cyntellect, San Diego, CA, USA). Change in percent cell proliferation within parental and resistant derivatives was calculated as ((percentage of EdU-incorporating cells transfected with target siRNA/percentage of EdU-incorporating cells transfected with nontargeting siRNA) Χ 100). All measurements were performed in quadruplicate. Apoptosis assays were performed using the Annexin V-FITC Apoptosis Detection Kit (Abcam, Cambridge, MA, USA). Cells transfected with siRNA for 72 hours were incubated with Annexin V-FITC and DAPI for 30 minutes and apoptosis was analyzed by the Celigo Cytometer (Cyntellect, San Diego, CA, USA). Change in percent apoptosis was calculated as ((percentage of Annexin-V positive cells transfected with target siRNA/percentage of Annexin-V positive cells transfected with nontargeting siRNA) Χ 100). All measurements were performed in triplicate.
Statistical analysis
Experiments assessing proliferation and apoptosis of various cell-lines under various treatment conditions were analyzed using one-way ANOVA. Data were log-transformed to stabilize variances. Differences between groups were determined by multiple comparisons using contrasts, and the Sidak method for P-value adjustment. Growth curve and growth fold change data in vitro were analyzed similarly. Error bars on plots represent +/- standard error (SE).
Xenograft tumor growth curves were constructed using the mean tumor volume at each time point with error bars representing the standard error of the mean. Animals that died of other causes prior to the first animal developing a resistant tumor were not included in the calculation of tumor growth curves.
P-values for the xenograft studies were adjusted for multiple comparisons using the Hommel method to control for type I error when appropriate [
34]. Progression of the tumor was defined as: tumor size more than zero and at least two consecutive measurements with ≧10% increments in tumor size. Time to progression (PFS) is the day of the measurement on which the tumor qualifies as a progression.
Discussion
In this report we show that a dynamic transition between HER2 and ER activity plays a role in resistance to L-containing regimens, while sustained HER pathway activity is a prominent feature in TR cells. Our data suggest that ER-positive/HER2-positive cells, in general, exploit ER activity as a mechanism of de novo or acquired resistance to effective L-containing HER2-targeted regimens.
Four out of five ER-positive/HER2-positive cell lines in our panel showed up-regulation of ER signaling following treatment with combined L + T. However, only the MDA- MB-361 cell line, which showed the highest increase in ER activity upon L + T treatment, displayed a
de novo resistance phenotype. Therefore, ER in this particular cell line acts as the dominant and primary driver of growth even before anti-HER2 therapy is initiated. The other ER-positive lines were initially sensitive to L + T treatment, but later ER was used as an escape pathway to cause acquired resistance to L + T. Thus, in ER-positive/HER2-positive breast cancer cells, either ER or HER2 can function initially as the major promoter of proliferation and survival. Eventually, however, with sustained, effective HER2 inhibition with L or L + T in these cell lines, ER becomes the primary driver of cell survival resulting in resistance to L or L + T therapy. These findings are consistent with two recent neoadjuvant trials in HER2-positive patients, where chemotherapy was administered in addition to HER2-targeted therapy. These trials demonstrated significantly lower pathological complete response rates (pCR) in ER-positive/HER2-positive than in ER-negative/HER2-positive tumors [
39,
40]. However, neither of these trials included ER-targeted therapy. One of these trials, which combined T plus the HER2 dimerization inhibitor pertuzumab [
40], also included a group without chemotherapy. In this group, a 6% pCR rate was reported for the ER-positive tumors. A further recently reported neoadjuvant trial in patients with HER2-positive tumors, used L + T without chemotherapy but with combined endocrine therapy if the tumors were ER-positive [
41]. This trial, which included patients with larger tumors, reported a 21% pCR rate, a pCR greater than three times that reported in the trastuzumab plus pertuzumab trial. Although it is difficult to compare across trials, the lower response rate in the T plus pertuzumab trial could be due to the failure of this regimen to target EGFR, ER, or both. Collectively, these results suggest that targeting the ER and HER2 pathways simultaneously in ER-positive/HER2-positive tumors is essential for obtaining optimal benefit. The results from our UACC-812 xenograft model, together with our previous findings in the MCF7-HER2 and BT474 models [
24,
31], demonstrate the capability and superiority of the potent L + T regimen in combination with endocrine therapy in achieving complete tumor regression and preventing the onset of therapeutic resistance. Therefore, these data strongly suggest a potential role for this strategy in the clinic.
Unlike UACC-812 LR and LTR, which exhibit no HER pathway activity, BT474 LR (early stage) and LTR maintain AKT activity, even in the presence of reduced HER receptor activity. Previously, sustained PI3K/AKT activity in BT474 LR clones was suggested to be regulated by AXL, a membrane-bound receptor tyrosine kinase [
20]. In addition, ER has the ability to induce the expression of AXL, which could subsequently lead to activated AKT [
20]. However, in our early BT474 LR derivatives, AXL expression was unchanged. When treated with F, BT474 LR displayed evidence of ER degradation, but no substantial effect on AKT activity was observed. These results suggest that other unknown mechanisms may also be maintaining PI3K/AKT activity in these cells.
While ER activity was dominant in the LR and LTR derivates of our cultured models, we found that HER2 activity was crucial for resistance to T, as siRNA knocking down HER2 in our TR derivatives inhibited proliferation and also induced apoptosis. One of the mechanisms of action of T is to disrupt ligand-independent HER2-HER3 heterodimer signaling [
37]. UACC-812 and BT474 TR cells maintained high levels of EGFR and HER2 but showed decreased phosphorylated HER3, suggesting that T still manages to effectively disrupt HER2-HER3 heterodimer signaling in the resistant derivatives. Although it has been reported that EGFR and HER3 contribute to TR [
19,
42], our data demonstrate that HER2 is still required for growth in TR cells, while knockdown of EGFR or HER3 failed to elicit significant growth inhibition in BT474 TR. Importantly, the contribution of changes in antibody-dependent cell-mediated cytotoxicity (ADCC), thought to be one partial mechanism of action of T [
43], could not be studied in our
in vitro models. Therefore, in our culture studies, the observed inhibitory effect of T in comparison to L-containing regimens is related to the potency of this treatment directly on the HER signaling pathway. Collectively, we did show that TR derivatives are still dependent on the HER pathway and, therefore, remain sensitive to L, as previously reported [
44].
Of note, we did not observe up-regulation of ER expression or signaling in the LR and LTR derivatives of HER2-positive/ER-negative cell lines, in which the HER2 pathway remains suppressed. However, further investigation, both in vitro and in the clinical setting, is required to evaluate whether more prolonged exposure to these HER2-targeted therapies will reactivate the ER pathway.
We found that HER3 expression levels increased upon commencement of HER2-targeted therapy, while HER2 phosphorylation was suppressed in most of our HER2-overexpressing models. Previous studies have indicated that AKT inhibition induces HER3 expression in HER2-positive cell lines [
45], and consistent with this, AKT activity is significantly inhibited by HER2-targeted therapy in the majority of the models examined. SKBR3 and SUM-190 cells, however, maintain AKT phosphorylation and still up-regulate HER3 expression, suggesting that additional mechanisms must also control HER3 expression.
Reactivated HER signaling did confer resistance to L in BT474 cells but only after the cells had experienced a period of ER dependency. In contrast, UACC-812 LR cells were driven by ER activity and maintained a fairly stable phenotype even after prolonged L treatment. In BT474 LR cells, however, a switch in dependence from the ER to the HER2 pathway was observed during the late phase of acquisition of LR. In this model, enhanced ER activity reduced cell death in LR cells at the early stage, acting as a transitional pathway. Following prolonged treatment with L, a significant compensatory rearrangement of HER receptor and ligand expression occurred, ultimately leading to up-regulated levels of HER2, HER3, and many HER ligands. Interestingly, doubling the dose of L inhibited the HER2-dependent BT474 LLR cells, but not the ER-dependent BT474 LR cells. A therapeutic strategy that applies high doses of L intermittently has been shown to more effectively inhibit tumor growth in mouse models with minimal toxicity [
46], a strategy that might be considered in the clinical setting. Another recent report suggests that up-regulated HER3 compensates for inhibition of L [
18]. Although HER3 knockdown has no effect on BT474 early stage LR, HER3 siRNA induced increased apoptosis in BT474 LLR, suggesting that HER3 could contribute to LR. Repeat biopsy of tumors from patients with LR tumors might be helpful in differentiating those tumors with a greater dependence on ER from those that remain dependent on the HER pathway, thus acting as a guide to further therapy.
Conclusions
The complexity and redundancy of the HER network requires more complete inhibition of the HER family of receptors with combination therapy. In cultured cells, treatment with L is more effective than T in achieving this inhibition, and the additive effect of the L + T combination achieves a more powerful blockade of the pathway than either therapy in isolation. In this study, we illustrate that TR derivatives show reactivation of the HER pathway as a mechanism of resistance. However, with a more complete HER2 blockade, resistance to L-containing regimens requires the activation of an alternative cell survival pathway. This is evident in ER-positive/HER2-positive cell lines, where up-regulation of the ER pathway occurs in order to create an escape survival pathway.
The findings of this study have several therapeutic implications: (i) A more potent HER pathway inhibitor, or a combination therapeutic strategy such as L + T, could improve the outcome of patients with HER2-positive breast cancer. Recent reports of clinical studies using L + T regimens support this idea. (ii) A combination of endocrine and anti-HER2 therapies given simultaneously might benefit ER-positive/HER2-positive patients, including those with tumors with low ER levels that clinically might be reported as ER-negative, especially if PR is still expressed. These ideas are currently being tested in clinical trials.
Competing interests
RS and CKO have received research grant funding and payments for participating in advisory panels from GlaxoSmithKline and AstraZeneca. JG and GLP are employees of Genentech and hold shares of Roche.
Authors' contributions
Y-CW contributed to concept design, established the resistant lines, carried out most of the molecular and functional studies, contributed to data analysis and interpretation, and drafted the manuscript. RS, CKO, and MFR contributed to concept design, data analysis and interpretation, and manuscript writing. GM, RG, RMW, XF, and MFB contributed to the establishment and characterization of the resistant lines and to the molecular and xenograft studies, and helped with data analysis. JG and GLP performed the qRT-PCR of the HER receptors and ligands and the data analysis. SGH contributed to concept design and performed the statistical analysis and data interpretation. GCC and NH contributed to data analysis and manuscript writing. All authors read and approved of the final manuscript.