|Home | About | Journals | Submit | Contact Us | Français|
Lapatinib is a dual inhibitor of epidermal growth factor receptor (EGFR) and human epidermal growth factor receptor 2 (HER2) with activity in HER2-amplified metastatic breast cancer (MBC). Its role in non–HER2-amplified MBC remains unclear. EGF30001, a phase III trial of lapatinib and paclitaxel versus paclitaxel and placebo, demonstrated lapatinib does not significantly benefit HER2-negative or HER2-unselected patients with MBC. Published data support interactions between steroid hormone and peptide growth factor signaling. We hypothesized that molecular subgroups may exist within EGF30001 that would benefit from lapatinib.
A blinded, retrospective biomarker evaluation was performed using immunohistochemistry to semiquantitate estrogen (ER), progesterone (PR), and EGFR expression. HER2 amplification was determined by fluorescent in situ hybridization. Effects of these biomarkers on event-free survival (EFS) were examined in patients with available tissue (n = 493).
Lapatinib improved median EFS in HER2-amplified, ER- or PR-positive MBC (n = 36; 5.7 v 4.5 months; P = .351); benefit was greater and statistically significant in HER2-amplified, ER-negative, PR-negative MBC (n = 42; 8.3 v 5.0 months; P = .007). In HER2-negative, ER-positive MBC, median EFS improvement varied by degree of PR expression (H-score): no benefit if PR-strong (n = 133; 9.3 v 7.3 months; P = .373), benefit if PR-weak (n = 50; 7.3 v 2.4 months; P = .026), and potential antagonism if PR-negative (n = 40; 3.7 v 7.2 months; P = .004). No benefit was seen in triple-negative MBC (n = 131; median EFS, 4.6 v 4.8 months; P = .255). EGFR expression was not correlated with benefit from lapatinib.
Although subgroups are small, these analyses support the hypothesis that semiquantitative determination of hormone receptor status may be a surrogate for EGFR and/or HER2 dependency.
The epidermal growth factor receptor (EGFR) family has been pursued as a therapeutic target in breast cancer.1 Human EGFR 2 (HER2) amplification occurs in 20% to 25% of women and is associated with a poor prognosis.2,3 Elevated EGFR expression has been correlated with a poor prognosis in breast cancer specimens and in laboratory models.4 Whereas clinical data with trastuzumab and lapatinib have validated HER2 as a therapeutic target in breast cancer, studies using agents aimed at disrupting EGFR signaling have not shown meaningful activity in breast cancer.5,6 This raises the question of whether EGFR plays a role in breast cancer pathogenesis or whether the correct patients with EGFR-dependent disease are being selected. The latter is important, as EGFR mutations and amplification were only retrospectively found to be associated with response to EGFR inhibitors in non–small-cell lung cancer.7–9
Lapatinib is a small-molecule inhibitor of the EGFR and HER2 tyrosine kinases. Preclinical studies using human breast cancer cell lines have shown that HER2 amplification is predictive of response, whereas EGFR expression is not.10–12 Clinical activity of lapatinib in advanced HER2-overexpressing breast cancer was confirmed in open-label13 and randomized clinical trials.14,15 A randomized, placebo-controlled, phase III study (EGF30001) evaluated the role of lapatinib in combination with paclitaxel in first-line HER2-negative or -unknown metastatic breast cancer.16 Of the 579 women enrolled, 493 had tissue available for centralized fluorescent in situ hybridization (FISH) and immunohistochemistry (IHC). After central review, 86 patients had evidence of HER2 amplification, and 407 patients had HER2-nonamplified disease. The primary end point was time to progression (TTP). The intent-to-treat population did not show benefit in TTP when lapatinib was added to paclitaxel; however, the distinct subgroup of HER2-amplified patients receiving lapatinib with paclitaxel demonstrated clear clinical benefit.16 This supports the concept that patient selection and predictive markers are critical in determining certain subgroups of patients who are more likely to respond to treatment. It was hypothesized that a subgroup within the HER2-nonamplified cohort may derive benefit from the addition of lapatinib.
The data supporting EGFR as a target in breast cancer has been difficult to reconcile with the lack of clinical activity of EGFR-targeted agents in this disease. Recent data from a retrospective study of more than 35,000 clinical samples suggested a relationship between peptide growth factor signaling and hormone receptor status,17 specifically, that progesterone receptor (PR) loss in estrogen receptor (ER) –positive disease was associated with higher expression of EGFR and/or HER2. In addition, a presurgical study demonstrated cell-cycle inhibition in ER-positive and PR-weak/-negative patients with breast cancer who were exposed to a short course of the EGFR inhibitor gefitinib.18 Other data suggested that breast cancers lacking ER, PR, and HER2 amplification (“triple-negative” breast cancer) have higher levels of EGFR and may be more sensitive to EGFR-targeted agents.19,20 A blinded, retrospective, semiquantitative analysis of ER, PR, HER2, and EGFR expression was performed to correlate these measurements with benefit from lapatinib.
The eligibility criteria and study design for EGF30001 have been previously reported.16 Briefly, 579 women with advanced breast cancer (stage III or IV) previously untreated in the metastatic setting were randomly assigned to receive paclitaxel 175 mg/m2 intravenously every 3 weeks with either oral lapatinib 1,500 mg daily or placebo. Patients had either HER2-negative (per the enrolling site) or HER2-unknown breast cancer.
Tumor specimens from pretreatment or archival tumor biopsies were available for HER2 status (n = 493), ER and PR (n = 455), and EGFR (n = 443) determination. Paraffin-embedded samples were cut to 4 μm, mounted on slides, deparaffinized, and rehydrated through a graded series of alcohol. Endogenous peroxidase was quenched with 3% H2O2 in phosphate-buffered saline (PBS). Antigen retrieval was performed for ER only by placing slides in 0.1 mol/L of citrate buffer in a steam bath for 1 hour and allowing them to cool for 20 minutes. Slides were blocked with 10% normal goat serum in PBS at room temperature for 1 hour, and then primary antibody was applied for 1 hour at room temperature. Both ER 1D5 (Immunotech; Marseilles, France) and PgR 636 (Dako; Carpinteria, CA) antibodies were used at 1:50 dilute in 10% normal goat serum. Slides were washed in PBS, followed by antimouse secondary (Dako Polymer Envision Plus, Dako; Carpinteria, CA) for 30 minutes at room temperature. The chromogen used was diaminobenzidine, and slides were counterstained with ethyl green, passed through butanol, and dehydrated in xylene before mounting. Negative controls consisted of eliminating primary antibody, and positive controls were known ER-/PR-positive cell lines that were pelleted, embedded, and cut. Known positive human tumors were used as well. Controls were included in each staining batch. In a blinded fashion, the samples were read and scored by light microscopy.
ER and PR were semi-quantitated using the H-score. Tissue was scored (H-score) based on the total percentage of positive cells and the intensity of the staining (1+, 2+, or 3+), where H = (% 1+ × 1) + (% 2+ × 2) + (% 3+ × 3). The sample was considered negative if H = 0 and positive if H was more than 0; positive samples were also categorized as weak if H = 1 to 50 and strong if H was more than 50. A minimum of 100 cells were evaluated in calculating the H-score.
FISH assays were performed in a blinded fashion using the HER2 PathVysion FISH assay (Abbott Laboratories, North Chicago, IL) as described elsewhere21,22 and as approved by the US Food and Drug Administration. Enumerations of red HER2 signals and green chromosome 17 centromeres in each of at least 20 interphase carcinoma cell nuclei were performed by a clinical laboratory scientist/technician and confirmed by a board-certified pathologist. Samples with a HER2:CEP17 ratio ≥ 2.0 were considered FISH positive.
EGFR IHC was performed using a commercially available immunohistochemical assay kit (PharmDX; Dako). A monoclonal mouse antibody was used in a peroxidase antiperoxidase immunohistochemical assay to demonstrate the EGFR protein product in tissue sections. The immunohistochemical technique involved incubation of the tissue sections with two different antibodies: a primary antibody specific for the EGFR protein (clone 2-18C9; Dako) and a goat antimouse immunoglobulin conjugated to a horseradish peroxidase–labeled dextran polymer (Envision+; Dako). The site of immunoprecipitates was identified by the use of a chromogen, diaminobenzidine, which was visualized microscopically. Subjective assessment of the amount of membrane staining (0, 1+, 2+, 3+) was performed according to the manufacturer's package insert. Again, a minimum of 100 cells were evaluated in calculating the IHC score.
The primary population was the EGF30001 intent-to-treat population (all randomly assigned patients who received one or more dose of study medication), and the primary end point was TTP (time from random assignment until disease progression or death because of disease under study). However, event-free survival (EFS; time from random assignment until disease progression or death due to any cause) was the clinical end point used for biomarker correlations in this analysis, as it is a more conservative end point. Kaplan Meier curves were generated for EFS and used to calculate median EFS. Estimates of treatment hazard ratios (HR) based on log-rank tests and 95% CIs were calculated. Treatment arms were compared using log-rank tests. The EFS analysis was repeated in the retrospectively defined subpopulations defined by hormone and/or EGFR and HER2 status. Interactions between hormone status and treatment arms were examined using Cox's proportional hazards regression to model EFS.
EGF30001 enrolled women with HER2-negative or unknown breast cancer. On central review, 86 patients were found to have HER2 amplification. Eighty-three were FISH positive, and three patients had 3+ IHC staining and no material available for FISH. The addition of lapatinib to paclitaxel significantly improved median EFS (8.1 v 5.0 months; P = .008; HR = 0.49; 95% CI, 0.3 to 0.8; Fig 2A). Benefit from lapatinib was also analyzed on the basis of hormone receptor (ER and PR) status. Whereas lapatinib showed some improvement in median EFS in HER2-amplified and either patients with ER- or PR-positive disease (n = 36; 5.7 v 4.5 months; P = .351; HR = 0.7; 95% CI, 0.34 to 1.45; Fig 2B), improvement was greater in hormone receptor–negative patients (n = 42; 8.3 v 5.0 months; P = .007; HR = 0.34; 95% CI, 0.17 to 0.7; Fig 2C). Interaction between treatment arm and hormone status (ER and PR negative, ER or PR positive) was not statistically significant (P = .170), but the test of the interaction term lacked power because of the limited sample size. There was no benefit from lapatinib in patients without HER2 amplification (n = 407) on the basis of that biomarker alone.
The role of HER family inhibitors in HER2-nonamplified breast cancer is not defined. We analyzed EFS in patients with HER2-nonamplified disease as a function of ER and PR expression, which were calculated using the H-score as described in Methods. ER and PR expression in the non–HER2-amplified, hormone receptor–positive population is shown in Figure 3A. There was no correlation between ER and PR in this cohort, demonstrating that the expression of one is not a function of the other (R2 = 0.031; Fig 3B). As hypothesized, we sought to determine whether PR expression in the ER-positive cohort was associated with benefit from lapatinib. Patients were subgrouped and blinded to clinical outcome by PR H-score into three predetermined groups: (1) H-score = 0, PR negative (n = 40); (2) H-score 1 to 50, PR weak (n = 50); and (3) H-score more than 50, PR strong (n = 133). These cutoffs were based on data from a presurgical study with gefitinib that identified an H-score in these ranges as being associated with response to an EGFR inhibitor.17 In HER2-nonamplified, ER-positive patients (n = 223), median EFS improvement with the addition of lapatinib varied based on PR H-score: no improvement if PR-strong (n = 133; 9.3 v 7.3 months; P = .373; HR = 0.83; 95% CI, 0.55 to 1.25; Fig 4A); improvement if PR-weak (n = 50; 7.3 v 2.4 months; P = .026; HR = 0.49; 95% CI, 0.25 to 0.92; Fig 4B); and potential antagonism if PR-negative (n = 40; 3.7 v 7.2 months; P = .004; HR = 3.58; 95% CI, 1.58 to 8.12; Fig 4C). Interaction tests between treatment arm and the three strata of PR status (PR-strong, PR-weak, PR-negative) were highly significant (P < .001).
Triple-negative breast cancer poses a unique clinical challenge, as there are no therapeutic agents tailored to this population.23 In EGF30001, 131 patients were categorized as triple-negative (HER2 FISH negative or IHC 0, 1, or 2+ when FISH was not available and H-score of 0 for ER and PR). There was no EFS improvement with the addition of lapatinib in this subgroup (4.6 v 4.8 months; P = .255; HR = 1.25; 95% CI, 0.85 to 1.83; Fig 5).
Table 1 shows the frequency of EGFR expression as measured by IHC. The majority (315 of 443; 71%) of patients did not have detectable expression (EGFR = 0). Both treatment arms were balanced for levels of EGFR expression; there were slightly more IHC 3+ patients in the lapatinib arm. Consistent with other EGFR-targeted agents,24 EGFR expression was not predictive of improved median EFS with lapatinib (EGFR = 0; n = 315; 4.6 v 4.8 months; P = .502; Fig 6A; EGFR = 1+, 2+, 3+; n = 128; 4.8 v 4.9 months; P = .968; Fig 6B).
As previously reported, EGFR expression was highest in the triple-negative subgroup (Table 2).19,20 However, even when the analysis was restricted to triple-negative cases, detectable levels of EGFR were not associated with improved median EFS with the addition of lapatinib (paclitaxel plus lapatinib, n = 67; 5.2 v 4.2 months; P = .064; Fig 6C; paclitaxel plus placebo, n = 59; 4.3 v 4.9 months; P = .450; Fig 6D).
Patient selection is critical to the clinical success of molecularly targeted agents. As seen from the results of this study, there was no overall benefit from adding lapatinib to paclitaxel in a mixed population of patients with HER2-negative or -unknown breast cancer. Clearly, when HER2-amplified patients were determined by central review, a significant benefit from lapatinib in the first-line metastatic setting was observed.15 On the basis of a growing body of literature supporting ER and PR expression and its relationship to EGFR and HER2 signaling, it was hypothesized that measurement of these receptors may identify a subgroup of patients who may benefit from lapatinib.
Significant cross-talk between the HER and steroid receptor families has been described.25,26 HER2-driven breast cancers may be more aggressive and carry a worse prognosis than HER2-negative cancers, which are usually ER-positive tumors.27 In addition, elevated EGFR expression and HER2 amplification have been associated with a worse response to hormone blockade,17,25 and an inverse correlation has been shown between HER2 and ER and PR expression.28 Analysis of 86 patients with HER2-amplified disease and their response as a function of hormone receptor status demonstrated that patients with hormone receptor–negative tumors derived greater benefit from lapatinib than patients with either ER- or PR-positive tumors. The observation of a potential interaction between hormone receptor status and response (or resistance) to lapatinib has not been previously described. We were able to ask this question in the front-line setting because these patients were thought to be HER2 negative based on local assessment. Whether this relationship between hormone receptor status and response to lapatinib exists in women who have received prior trastuzumab is unknown. Although these subgroups are small, these findings raise the question of whether the complete loss of steroid receptors identifies a group of patients with greater dependence on HER family signaling. A correlation between HER2 status and hormone receptors exists, and studies have shown that patients with HER2-positive and hormone receptor–positive tumors are less dependent on hormone signaling and respond less well to hormone manipulation.25,29 Still, there is a need to identify those patients with HER2-amplified disease who respond to lapatinib, and this hypothesis will be studied in a large, placebo-controlled study of letrozole with or without lapatinib (EGF30008, NCT #00073528). Of note, contrasting recent data suggest that patients with normal HER2 status may benefit from adjuvant trastuzumab in early breast cancer.30 In EGF30001, there was no evidence of activity of lapatinib in the large unselected HER2-negative population when evaluated in the metastatic setting.
The largest patient group (n = 243) in this study had HER2-negative, hormone receptor–positive tumors. The majority of these patients were ER positive, with variable levels of PR (v only PR positive; n = 20). We hypothesized that PR loss in ER-positive patients may be associated with a benefit from lapatinib in this group of 223 patients. It was initially demonstrated that ER and PR expression were independent variables. Next cutoffs for PR-expression subgroups were blindly selected and designated as described in the Methods section. These designations, as previously experienced,18 were based on hypotheses that (1) having no detectable PR was biologically significant; (2) there is a unique patient subgroup identified by a lower H-score, as can be seen by the increased number of patients in this range; (3) this subgroup is more likely to benefit from EGFR inhibition, as previously seen.18 The small subgroup size limits this analysis, and conclusions may reflect sample size rather than true biology; however, the ER-positive, PR-weak group demonstrated a significant benefit from the addition of lapatinib. It is also in ER-positive, PR-negative patients that responses were seen to the EGFR small-molecule inhibitors gefitinib and erlotinib in phase II and presurgical studies,18,31–33 providing precedent for linking EGFR dependence to hormone receptor status. Laboratory studies have identified a role for RTK signaling via PI3-kinase in modifying PR expression. In HER2-negative breast cancer cells, insulin-like growth factor and the HER family ligands EGF and heregulin downregulated PR expression.34 Furthermore, a recent study has shown that HER family signaling through PI3-kinase and AKT can decrease PR expression.35 Together, these data suggest that in the absence of amplification, decreased PR expression may be the readout for EGFR activation and dependence.
The question of whether there is true antagonism in the ER-positive, PR-negative group needs further clarification and may be an artifact of the small subgroup size. Alternatively, this represents a true biologic effect of lapatinib that is unique to those cancers that have no detectable PR expression by immunohistochemistry as compared with those that have weak PR expression. In the neoadjuvant study performed with gefitinib, only six samples were ER positive and PR negative; five of six showed no molecular response to gefitinib, and one showed molecular inhibition.18
Despite data supporting EGFR as a therapeutic target in breast cancer, the predictive marker for those patients who will respond to EGFR inhibition remains elusive. To answer this important question, ER, PR, and EGFR were analyzed in the large cohort of cases with HER2-negative tumors. Like other studies, the highest levels of EGFR expression were in ER-, PR-, and HER2-negative patients (triple negatives). These patients have only chemotherapy available as a treatment option and are in need of new approaches. In our analysis, this group did not benefit from the addition of lapatinib to paclitaxel, suggesting that EGFR may not be a good target for this population. Ongoing studies with the EGFR monoclonal antibody cetuximab are in progress in combination with platinum salts.36,37 It is hoped these patients may respond to an antibody approach or benefit from the potential synergy between platinum salts and EGFR inhibition.38,39 Furthermore, it is possible that EGFR is still a potential target, but additional selection criteria (ie, more accurate biomarkers) are necessary. There was no group in our study in which EGFR measurements by IHC predicted benefit from the addition of lapatinib.
This study adds to the growing data describing a relationship between the HER family and steroid receptor signaling. The conclusions drawn are limited by the progressively small subgroups analyzed and, even though blinded, the study's retrospective nature. Still, it represents a biomarker analysis of the largest prospective trial of a HER2 or EGFR inhibitor (and the only dual-kinase inhibitor studied) in patients with non–HER2-amplified breast cancer. These data can be used to guide further biomarker assessments of ongoing lapatinib and EGFR inhibitor studies and, if validated in EGF30008, may ultimately guide patient selection in a prospective study.
We thank all patients and their treating physicians, Ivonne E. Villalobos, Angela Santiago, Roberta Guzman, Armen Gasparyan, Yanling Ma, Melinda Epstein, Anamaria Ioan, and Sherin Shirazi at University of Southern California, and Steve Ashton, Louise Downie, and Kevin Jackson at GlaxoSmithKline.
|Characteristic||Paclitaxel + Lapatinib (n = 251)||Paclitaxel + Placebo (n = 242)||Total (n = 493)|
|ECOG performance status, %|
|2||0||< 1||< 1|
|Unknown||1||< 1||< 1|
|Stage of disease, %|
|Disease sites, % viscera||63||65||64|
|No. of metastatic sites|
|0||0||< 1||< 1|
Abbreviation: ECOG, Eastern Cooperative Oncology Group.
Supported by grants from GlaxoSmithKline, and a National Institutes of Health Loan Repayment Program award (R.S.F.).
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical trial information can be found for the following: NCT00075270.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: Michael Arbushites, GlaxoSmithKline (C); Maria Koehler, GlaxoSmithKline (C); Cristina Oliva, GlaxoSmithKline (C); Lisa S. Williams, GlaxoSmithKline (C) Consultant or Advisory Role: Michael F. Press, Genentech (C), GlaxoSmithKline (C); Angelo Di Leo, GlaxoSmithKline (C) Stock Ownership: Michael Arbushites, GlaxoSmithKline; Maria Koehler, GlaxoSmithKline; Cristina Oliva, GlaxoSmithKline; Lisa S. Williams, GlaxoSmithKline Honoraria: Richard S. Finn, GlaxoSmithKline; Michael F. Press, Genentech, GlaxoSmithKline; Angelo Di Leo, GlaxoSmithKline Research Funding: None Expert Testimony: None Other Remuneration: None
Conception and design: Richard S. Finn, Michael F. Press, Judy Dering, Maria Koehler, Angelo Di Leo
Provision of study materials or patients: Richard S. Finn, Michael Arbushites, Maria Koehler, Cristina Oliva, Lisa S. Williams, Angelo Di Leo
Collection and assembly of data: Richard S. Finn, Michael F. Press, Judy Dering, Michael Arbushites, Maria Koehler, Cristina Oliva, Lisa S. Williams, Angelo Di Leo
Data analysis and interpretation: Richard S. Finn, Michael F. Press, Judy Dering
Manuscript writing: Richard S. Finn, Michael F. Press, Judy Dering, Michael Arbushites, Maria Koehler, Cristina Oliva, Lisa S. Williams, Angelo Di Leo
Final approval of manuscript: Richard S. Finn, Michael F. Press, Judy Dering, Michael Arbushites, Maria Koehler, Cristina Oliva, Lisa S. Williams, Angelo Di Leo