Purpose: Not all breast cancers respond to lapatinib. A change in Ki67 after short-term exposure may elucidate a biomarker profile for responsive versus nonresponsive tumors.

Experimental Design: Women with primary breast cancer were randomized (3:1) to 10 to 14 days of preoperative lapatinib or placebo in a multicenter phase II trial (ISRCTN68509377). Biopsies pre-/posttreatment were analyzed for Ki67, apoptosis, HER2, EGFR, ER, PgR, pAKT, pERK, and stathmin by IHC. Further markers were measured by RT-PCR. Primary endpoint was change in Ki67. HER2+ was defined as 2+/3+ by IHC and FISH+.

Results: One hundred twenty-one patients (lapatinib, 94; placebo, 27) were randomized; of these, 21% were HER2+, 78% were HER2 nonamplified, 26% were EGFR+. Paired samples containing tumor were obtained for 98% (118 of 121). Ki67 fell significantly with lapatinib (−31%; P < 0.001), but not with placebo (−3%). Whereas Ki67 reduction with lapatinib was greatest in HER2+ breast cancer (−46%; P = 0.003), there was a significant Ki67 decrease in HER2 breast cancer (−27%; P = 0.017) with 14% of HER2 breast cancer demonstrating ≥50% Ki67 reduction with lapatinib. Among HER2+ patients, the only biomarker predictive of Ki67 response was the EGFR/HER4 ligand epiregulin (EREG) (rho = −0.7; P = 0.002). Among HER2 tumors, only HER3 mRNA levels were significantly associated with Ki67 response on multivariate analysis (P = 0.01). In HER2 breast cancer, HER2 and HER3 mRNA levels were highly correlated (rho = 0.67, P < 0.001), with all Ki67 responders having elevated HER3 and HER2 expression.

Conclusions: Lapatinib has antiproliferative effects in a subgroup of HER2 nonamplified tumors characterized by high HER3 expression. The possible role of high HER2:HER3 heterodimers in predicting response to lapatinib merits investigation in HER2 tumors. Clin Cancer Res; 21(13); 2932–40. ©2014 AACR.

See related commentary by Campbell and Moasser, p. 2886

This article is featured in Highlights of This Issue, p. 2881

Translational Relevance

The presurgical trial design offers a useful translational tool to investigate the in vivo antitumor effects of a novel agent by comparing biopsies obtained at diagnosis and at surgery after 2 weeks of treatment. In this case, we investigated, in a placebo-controlled trial, the effect of the HER1/2 tyrosine kinase inhibitor lapatinib on proliferation (Ki67) of primary breast cancers. The MAPLE presurgical trial demonstrated that the effects of lapatinib were not limited to HER2-amplified breast cancer, with a significant decrease in Ki67 in the HER2 nonamplified subset. Among HER2 tumors, high HER3 mRNA identified Ki67 responders, and there was a trend for combined increase in HER3 and HER2 to predict response. Whether high levels of HER2–HER3 heterodimers may identify HER2 patients who could benefit from lapatinib treatment merits further investigation.

Lapatinib is a tyrosine kinase inhibitor (TKI) of both the epidermal growth factor receptor (EGFR, HER1) and human epidermal growth factor receptor-2 (HER2), which is an FDA-approved and commonly used drug in patients with HER2-positive (HER2+) breast cancer (1–3). Targeting this drug to the most appropriate subset of breast cancer patients depends on a full understanding of the molecular determinants of response. Although HER2 overexpression [3+ by immunohistochemistry (IHC)] or amplification is an important predictor of response to lapatinib, it may not be the only requirement. EGFR (HER1) protein expression by IHC has not been shown to correlate with response to lapatinib (4), whereas the sensitizing EGFR gene mutations described in other tumor types (5) have not been reported in breast cancer. Either moderate levels of HER2 expression in the absence of gene amplification, the expression of other HER partners (i.e., EGFR, HER3), elevated levels of HER ligands, or activation of downstream effectors such as AKT and ERK could all be relevant to lapatinib responsiveness in breast cancer. As such, these additional biomarkers potentially could identify a subset of HER2 nonamplified tumors that might respond to an EGFR/HER2 TKI.

Changes in the marker of cell proliferation as determined by Ki67 following 2 weeks of treatment have been shown to predict recurrence-free survival in groups of patients with breast cancer treated with different hormonal agents (6, 7). Assessment of in vivo surrogate markers, such as Ki67, associated with comprehensive molecular profiling may elucidate the differential profile of treatment-sensitive versus treatment-nonresponsive tumors (8–11). Ten to 14 days is the usual time interval between the diagnosis of breast cancer and primary surgery, offering a convenient window of opportunity for a short treatment with an experimental agent.

We therefore conducted a short-term presurgical study to determine the Molecular Antiproliferative Predictors of Lapatinib's Effects in breast cancer (MAPLE) to (i) determine whether lapatinib induces antiproliferative effects in both HER2+ and HER2 (i.e., nonamplified) breast cancer and (ii) identify biomarkers of sensitivity or nonresponsiveness to lapatinib. This was a multicenter presurgical double-blind study, designed to randomize 120 women with previously untreated primary breast cancer to lapatinib (1,500 mg/d; n = 90) versus lapatinib-placebo (n = 30) for 10 to 14 days between diagnosis and the day before surgery. Comprehensive molecular profiling of biopsies taken before treatment and at surgery was performed in an effort to identify predictors of sensitivity or nonresponsiveness to the antiproliferative effects of lapatinib. The primary objective was change in Ki67 after 2 weeks of therapy. Treatment-induced changes in apoptosis were also measured. These in vivo markers of drug effect (Ki67 and apoptosis) were correlated to baseline and treatment-induced changes in candidate molecular biomarkers.

Patients and samples

Women younger than 80 years presenting with newly diagnosed breast cancer scheduled for primary surgery were eligible. Other eligibility criteria included normal findings on screening cardiac examination, tumor tissue available for research, and at least 10 days' interval between randomization and the date of surgery. No concurrent or recent (within the last 28 days) hormonal therapy was allowed. Relevant demographics, medical history, physical examination, full blood count, and biochemistry were obtained at baseline, time of surgery, and at 1-month follow-up when the patient was discharged from the study. Toxicities were graded according to the NCI Common Terminology Criteria for Adverse Events (CTCAE) version 3.0. Patients received lapatinib 1,500 mg orally once daily or placebo for 10 to 14 days until the day before surgery. Dose modifications and discontinuation were conducted as detailed in the protocol. Informed consent was obtained from all patients, and the study was approved by the South East England Main Research Ethics Committee (MREC) and the Medicine and Healthcare products Regulatory Agency (MHRA). The Institute of Cancer Research–Clinical Trials and Statistics Unit (ICR-CTSU) was responsible for the trial management of the study.

Patients were recruited from 10 sites in the United Kingdom. Patients were randomized in a 3:1 ratio to receive either lapatinib (N = 90) or lapatinib-placebo (N = 30) to complete 2 weeks of presurgical treatment. Randomization was performed with the use of computer-generated random permuted blocks stratified according to center. Randomization was performed via a telephone call to the ICR-CTSU and the treatment allocation was unknown to the patient; investigators and GlaxoSmithKline personnel were blinded to treatment allocation.

Up to two core-cut 14-gauge biopsies were obtained at both diagnosis and surgery. In some cases, only sections of the full excision specimen were available. The first core biopsy was placed immediately in formalin and fixed for 24 to 48 hours before processing to paraffin. The second optional core biopsy was incubated in RNA later at 4°C overnight before freezing at −20°C.

IHC and in situ hybridization

IHC was performed as previously described (12) on 4-μm sections from paraffin-embedded blocks with the following antibodies: Ki67 (MIB-1), HER2 (Herceptest) from Dako, ERα (6F11) and PgR from VECTOR labs, stathmin, pAKT (Ser473), and pERK (Tyr202/Tyr204) from Cell Signaling Technology, and EGFR from Zymed. In addition, HER2 gene amplification was assessed in all cases by FISH using the Vysis method (Abott Molecular Inc.). For apoptosis, the terminal deoxynucleotidyltransferase–mediated UTP end-labeling (TUNEL) method was used.

All scoring was performed on paired samples by two experienced technicians blinded to treatment allocation. The percentages of Ki67- and TUNEL-positive cells were scored among 1,000 and 3,000 malignant cells, respectively. HER2 IHC and HER2 FISH were scored as per 2007 American Society of Clinical Oncology (ASCO) guidelines (13): as 0, 1+, 2+, or 3+ for IHC, and as FISH+ if the ratio of HER2 signal to chromosome 17 was greater than 2.2. A tumor was classified as HER2+ if scored HER2 3+ by IHC, or 2+ by IHC and FISH+. All cases were assessed by FISH as well, regardless of the IHC score. EGFR was scored using intensity (levels 1–3) and percentage positivity: strong was defined as >20% cells staining level 3, weak anything less, and negative as no staining. Nuclear staining for ER, PgR, pAKT, and pERK was done by H score, as previously published, resulting in a score between 0 and 300 (12). ER and PgR positivity was defined by an H score of >1. Cytoplasmic scoring for stathmin, pAKT, and pERK was quantified by quickscore as previously validated (14) and results in a score between 0 and 18.

Real-time PCR

When sufficient material was available, RNA was extracted from two 10-μm sections of FFPE samples using the RecoverAll (Ambion) kit, cDNA was synthesized using SuperScript III reverse transcriptase (Invitrogen), and Taqman Gene Expression assays (PE Applied Biosystems) were used to measure RNA expression of HER2, HER3, NRG1, EREG, AREG, ESR1, EGFR, and three housekeeping genes (MRPL19, TBP, and TFRC). The results obtained were converted into relative concentrations using an in-run standard curve, and the observed relative concentrations were normalized to the mean of the three housekeeping genes.

Statistical considerations

Enrollment of 120 patients was required to demonstrate an overall decrease in Ki67 of 50% after 2 weeks of treatment. A 50% decrease in Ki67 with a standard error of 1.07 would imply a standardized difference of 0.65; therefore, 90 patients would give a power of more than 99% (two-sided P value, 0.0027) using a one-sample t test. A control group of 30 patients was required to obtain a sufficiently robust difference for the primary endpoint.

An interim analysis to examine safety and futility was planned for when 60 patients were evaluable for Ki67. At this point, emerging trial data were reviewed by the independent data monitoring committee, which could consider recommending early stopping of the trial if there was clear evidence that either treatment regimen was contraindicated (statistical significance level of P < 0.001).

The primary aim of this study was to determine the 2-week change in Ki67 in response to lapatinib treatment. The secondary objectives were (i) to investigate the relationship between changes in Ki67 and molecular markers at baseline, (ii) to investigate change in apoptosis, and if this was found to occur, examine its relationship to baseline biomarkers. The Wilcoxon signed-rank matched pairs test was used to compare expression levels in paired pre-and posttreatment samples, and the Spearman rank correlation was used to assess the relationship between change in Ki67 and baseline biomarker values and, more generally, correlation between two continuous factors. To examine whether baseline molecular markers predicted a greater than 50% Ki67 reduction in Ki67, logistic regression was also used, multivariable analysis being used to assess the simultaneous effect of several factors. Differences between means were compared using the Mann–Whitney test. All statistical tests were two tailed. Geometric means were used as the summary measure because many biomarkers demonstrated a distribution that is approximately log normal; to avoid taking the log of zero, a small negligible quantity was added to biomarker values, e.g., 0.1 was added to Ki67 percentage values. Statistical analyses were performed using Microsoft Excel 2010 and Stata 11.2 for Windows. For secondary analyses, a false discovery rate (FDR) of 5% was applied to 119 analyses that examined the ability of biomarkers to predict a decrease in Ki67 in lapatinib-treated patients; this gave a critical P value of 0.0008.

Patients and treatment

A total of 121 patients were randomized over a 3-year period from 2008 until the accrual target was reached in 2011, and included in the toxicity analysis (Fig. 1). The clinicopathologic characteristics were representative of an unselected patient population with early breast cancer and were balanced between the treatment groups (Table 1). The majority of patients were peri/postmenopausal (57%), with a median tumor size of 23 mm, 46% were grade 3, and 81% were ER+. Twenty-one percent (26 of 121) of patients had centrally confirmed HER2+ breast cancer and included 25 HER2 3+ tumors and one HER2 2+ and FISH+ tumor. All HER2 3+ were also found to be FISH+. The proportion of HER2+ breast cancer (21%) was slightly higher than the 15% expected from a population with early breast cancer, and may reflect the fact that trial entry required a palpable tumor, which may bias toward younger women with self-detected cancers over older women with screen-detected tumors; 26% of patients demonstrated detectable EGFR expression.

Figure 1.

Consort diagram for treatment disposition in all randomized patients.

Figure 1.

Consort diagram for treatment disposition in all randomized patients.

Close modal
Table 1.

Patient and tumor characteristics

CharacteristicOverall (N = 121)Lapatinib (n = 94)Placebo (n = 27)
Age, years 
 Median (range) 53 (29–78) 53 (29–78) 57 (38–75) 
Menopausal status 
 Premenopausal 44 (36%) 35 (37%) 9 (33%) 
 Perimenopausal 13 (11%) 9 (9%) 4 (15%) 
 Postmenopausal 56 (46%) 42 (46%) 13 (48%) 
 TAH/BSO 7 (6%) 6 (6%) 1 (4%) 
 UK 1 (1%) 2 (2%) 
ECOG performance status 
 0 110 (91%) 87 (94%) 23 (82%) 
 1 8 (7%) 5 (5%) 3 (11%) 
 2 1 (1%) 0 (0%) 1 (4%) 
 UK 2 (2%) 2 (2%) 0 (0%) 
Histology 
 Invasive ductal 97 (80%) 75 (80%) 22 (81%) 
 Invasive lobular 13 (11%) 10 (11%) 3 (11%) 
 Other invasive 9 (7%) 8 (8%) 1 (4%) 
 UK 2 (2%) 1 (1%) 1 (4%) 
Tumor size (N = data available)a N = 63 N = 48 N = 15 
 Median, mm (range) 23 (6–75) 23 (6–71) 22 (11–75) 
Nodal status 
 Positive 46 (38%) 32 (34%) 14 (52%) 
 Negative 69 (57%) 58 (62%) 11 (41%) 
 Unknown 6 (5%) 4 (4%) 2 (7%) 
Grade 
 1 13 (11%) 11 (12%) 2 (7%) 
 2 48 (40%) 37 (39%) 11 (41%) 
 3 56 (46%) 43 (46%) 13 (48%) 
 UK 4 (3%) 3 (3%) 1 (4%) 
Lymphovascular invasion 
 Absent 71(59%) 56(60%) 15(55%) 
 Present 38 (31%) 27 (29%) 11 (41%) 
 Unknown 12 (10%) 11 (11%) 1 (4%) 
Breast surgery 
 Lumpectomy 76 (63%) 60 (65%) 16 (57%) 
 Mastectomy 44 (36%) 34 (37%) 10 (36%) 
 UK 1 (1%) 0 (0%) 1 (4%) 
Tumor characteristics central review 
Hormone receptor statusb 
 ER+ 98 (81%) 76 (81%) 22 (81%) 
 ER 20 (17%) 15 (16%) 5 (19%) 
 UK 3 (2%) 3 (3%) 0 (0%) 
HER2 statusb 
 HER2+ 26 (21%) 19 (20%) 7 (26%) 
  HER2 3+/FISH+  25 (21%)  19 (20%)  6 (22%) 
  HER2 2+/FISH+  1 (1%)  0 (0%)  1 (4%) 
 HER2 92 (76%) 72 (77%) 20 (74%) 
  HER2 0/FISH  15 (12%)  12 (13%)  3 (11%) 
  HER2 1+/FISH  65 (54%)  50 (53%)  15 (56%) 
  HER2 2+/FISH  12 (10%)  10 (11%)  2 (7%) 
 UK 3 (2%) 3 (3%) 0 (0%) 
Hormone receptor statusb 
 ER+/PgR+ 81 (67%) 64 (68%) 17 (63%) 
 ER+/PgR 17 (14%) 12 (13%) 5 (19%) 
 ER/PgR 19 (16%) 19 (15%) 5 (19%) 
 ER/PgR+ 1 (1%) 1 (1%) 0 (0%) 
 UK 3 (2%) 1 (3%)  
EGFR statusb 
 EGFR+ 31 (26%) 23 (24%) 8 (30%) 
 EGFR 87 (72%) 68 (72%) 19 (70%) 
 UK 3 (2%) 3 (3%) 0 (0%) 
CharacteristicOverall (N = 121)Lapatinib (n = 94)Placebo (n = 27)
Age, years 
 Median (range) 53 (29–78) 53 (29–78) 57 (38–75) 
Menopausal status 
 Premenopausal 44 (36%) 35 (37%) 9 (33%) 
 Perimenopausal 13 (11%) 9 (9%) 4 (15%) 
 Postmenopausal 56 (46%) 42 (46%) 13 (48%) 
 TAH/BSO 7 (6%) 6 (6%) 1 (4%) 
 UK 1 (1%) 2 (2%) 
ECOG performance status 
 0 110 (91%) 87 (94%) 23 (82%) 
 1 8 (7%) 5 (5%) 3 (11%) 
 2 1 (1%) 0 (0%) 1 (4%) 
 UK 2 (2%) 2 (2%) 0 (0%) 
Histology 
 Invasive ductal 97 (80%) 75 (80%) 22 (81%) 
 Invasive lobular 13 (11%) 10 (11%) 3 (11%) 
 Other invasive 9 (7%) 8 (8%) 1 (4%) 
 UK 2 (2%) 1 (1%) 1 (4%) 
Tumor size (N = data available)a N = 63 N = 48 N = 15 
 Median, mm (range) 23 (6–75) 23 (6–71) 22 (11–75) 
Nodal status 
 Positive 46 (38%) 32 (34%) 14 (52%) 
 Negative 69 (57%) 58 (62%) 11 (41%) 
 Unknown 6 (5%) 4 (4%) 2 (7%) 
Grade 
 1 13 (11%) 11 (12%) 2 (7%) 
 2 48 (40%) 37 (39%) 11 (41%) 
 3 56 (46%) 43 (46%) 13 (48%) 
 UK 4 (3%) 3 (3%) 1 (4%) 
Lymphovascular invasion 
 Absent 71(59%) 56(60%) 15(55%) 
 Present 38 (31%) 27 (29%) 11 (41%) 
 Unknown 12 (10%) 11 (11%) 1 (4%) 
Breast surgery 
 Lumpectomy 76 (63%) 60 (65%) 16 (57%) 
 Mastectomy 44 (36%) 34 (37%) 10 (36%) 
 UK 1 (1%) 0 (0%) 1 (4%) 
Tumor characteristics central review 
Hormone receptor statusb 
 ER+ 98 (81%) 76 (81%) 22 (81%) 
 ER 20 (17%) 15 (16%) 5 (19%) 
 UK 3 (2%) 3 (3%) 0 (0%) 
HER2 statusb 
 HER2+ 26 (21%) 19 (20%) 7 (26%) 
  HER2 3+/FISH+  25 (21%)  19 (20%)  6 (22%) 
  HER2 2+/FISH+  1 (1%)  0 (0%)  1 (4%) 
 HER2 92 (76%) 72 (77%) 20 (74%) 
  HER2 0/FISH  15 (12%)  12 (13%)  3 (11%) 
  HER2 1+/FISH  65 (54%)  50 (53%)  15 (56%) 
  HER2 2+/FISH  12 (10%)  10 (11%)  2 (7%) 
 UK 3 (2%) 3 (3%) 0 (0%) 
Hormone receptor statusb 
 ER+/PgR+ 81 (67%) 64 (68%) 17 (63%) 
 ER+/PgR 17 (14%) 12 (13%) 5 (19%) 
 ER/PgR 19 (16%) 19 (15%) 5 (19%) 
 ER/PgR+ 1 (1%) 1 (1%) 0 (0%) 
 UK 3 (2%) 1 (3%)  
EGFR statusb 
 EGFR+ 31 (26%) 23 (24%) 8 (30%) 
 EGFR 87 (72%) 68 (72%) 19 (70%) 
 UK 3 (2%) 3 (3%) 0 (0%) 

Abbreviation: TAH/BSO, total abdominal hysterectomy/bilateral salpingo-oophorectomy.

aTumor measurements were not required.

bCentral review.

Two weeks of lapatinib treatment was well tolerated and never resulted in any delay to scheduled surgery. The only toxicities that were significantly increased in the lapatinib versus placebo group were the occurrence of a rash (grade 3 rash was 6% vs. 0%; grade 1–2 rash was 64% vs. 12%; P < 0.001) and grade 1–2 diarrhea (56% vs. 8%; P = 0.001; Supplementary Table S1). The only other grade 3 toxicity was infection in 1 patient.

Sample collection and baseline Ki67

Paired pre- and posttreatment samples containing sufficient tumor cells for analysis were collected in 98% (118 of 121) of randomized patients. In the case of 3 patients, the posttreatment sample contained no viable cancer cells, and all three tumors were HER2+ from patients randomized to lapatinib. Therefore, 118 patients were included in the biomarker analyses (lapatinib, n = 91, or placebo, n = 27), and 100% (118 of 118) paired samples were of sufficient quality and cellularity (>1,000 tumor cells visualized) for analysis of the primary endpoint, Ki67. Of 118 patients available for primary endpoint analysis, three tumors were too small to allow more than two sections. Ki67 and HER2 staining was therefore available on 118 cases, while comprehensive IHC profiling was conducted on 115. The protocol requested that surgeons obtain posttreatment samples as core biopsies in the operating room and that they be fixed immediately, because extreme loss of staining of pAKT and pERK can occur during routine fixation of resection specimens (12). In 72 of 118 (61%) cases, surgeons were able to obtain core-needle biopsies (CNB) at the time of surgery; for the remaining 46 cases, the posttreatment sample was a block cut from the surgical excision specimen (ES) in pathology. Analyses of pERK and pAKT were confined to CNB samples.

In terms of the baseline proliferation levels in the tumor, as expected, Ki67 levels were significantly higher among ER versus ER+ tumors (P = 0.015), among PgR versus PgR+ tumors (P < 0.0001), among HER2+ versus HER2 tumors (P = 0.03), and among EGFR+ versus EGFR (P = 0.001) tumors (Supplementary Fig. S1). In absolute terms, the difference in baseline Ki67 levels between HER2+ and HER2 tumors was quite small, even though it was significant at the P < 0.05 level (mean Ki67 = 31% versus 20% for HER2+ versus HER2 tumors, respectively).

Change in Ki67 and apoptosis following 2 weeks of lapatinib

In the lapatinib-treated group as a whole (n = 91), there was a significant 31% reduction in Ki67 (95% CI, −43.6% to −16.3%). Importantly, there was no significant Ki67 change in the placebo group (−2.8%; 95% CI, −16.3% to +13%). Although the Ki67 reduction was greatest among the lapatinib-treated HER2+ subset (n = 19) with a 46% reduction (95% CI, −62.6% to −23%; Fig. 2A), there was also a significant 27% reduction in Ki67 among the HER2 lapatinib-treated tumors (n = 72; 95% CI, −41.9% to −7.6%; Fig. 2A and Table 2). Within this group, the proportion of Ki67 responders was higher among HER2 mRNA and HER3 mRNA–high patients (both above median) than among HER2 mRNA or HER3 mRNA–low (either below median): 6 of 22 (27%) versus 1 of 43 (2%; Supplementary Fig. S2).

Figure 2.

Change in Ki67 with lapatinib according to growth factor receptor status. A, mean pre- and posttreatment Ki67 score in HER2+ (n = 19) and HER2 nonamplified (n = 72) lapatinib-treated cases; B, pre- and posttreatment Ki67 scores in individual lapatinib-treated cases, with those demonstrating a Ki67 response defined by a greater than 50% decrease (shown in yellow).

Figure 2.

Change in Ki67 with lapatinib according to growth factor receptor status. A, mean pre- and posttreatment Ki67 score in HER2+ (n = 19) and HER2 nonamplified (n = 72) lapatinib-treated cases; B, pre- and posttreatment Ki67 scores in individual lapatinib-treated cases, with those demonstrating a Ki67 response defined by a greater than 50% decrease (shown in yellow).

Close modal
Table 2.

Biomarker changes with lapatinib treatment

NPretreatmentNSurgeryNChange (%)P
A. Protein biomarker 
Ki67 (%) 91 22.8 (19–27.4) 91 15.7 (12–20.4) 91 −31.3 (−43.6 to −16.3) 0.0004 
 HER2 72 20.7 (16.7–25.8) 72 15.2 (11–20.9) 72 −26.7 (−41.9 to −7.6) 0.0173 
 HER2+ 19 32.7 (24.4–43.8) 19 17.6 (11.7–26.3) 19 −46.3 (−62.6 to −23) 0.0029 
HER2 (FISH) 88 1.255 (1.21–1.3) 88 1.22 (1.2–1.27) 88 −0.03 (−0.08 to 0) 0.05 
 HER2 70 1.215 (1.2–1.25) 70 1.2 (1.13–1.22) 70 −0.02 (−0.06 to 0) 0.12 
 HER2+ 18 7.1 (4.5–9.2) 18 7.1 (4.7–9.5) 18 −0.28 (−0.9 to 0.1) 0.16 
EGFR (% pos)a 91 25.3% 91 26.4% 91 1.1% up /0% down 
 HER2 72 22.2% 72 23.6% 72 1.4% up /0% down 
 HER2+ 19 36.8% 19 36.8% 19 0% up /0% down 
pERK nuclearb 57 172 (165–179) 57 132 (107–162) 57 −23.2 (−37.4 to −5.9) 0.0011 
 HER2 47 170 (163–177) 47 128 (100–164) 47 −24.6 (−41.1 to −3.5) 0.0077 
 HER2+ 10 181 (163–201) 10 151 (121–188) 10 −16.6 (−29 to −2) 0.037 
pERK cytoplasmicb 57 7 (5.7–8.6) 57 5.1 (3.8–6.8) 57 −27.5 (−43.6 to −6.7) 0.029 
 HER2 47 6.8 (5.4–8.5) 47 4.9 (3.4–6.9) 47 −28.2 (−46.6 to −3.6) 0.06 
 HER2+ 10 8 (4.5–14.3) 10 6.1 (3.9–9.6) 10 −23.7 (−52 to 21.4) 0.22 
pAKT nuclearb 57 105 (91–122) 57 99 (86–115) 57 −5.7 (−19 to 9.9) 0.24 
 HER2 47 99 (84–117) 47 95 (80–112) 47 −4.5 (−20.5 to 14.8) 0.49 
 HER2+ 10 140 (116–168) 10 124 (96–160) 10 −11.1 (−25.5 to 6.1) 0.2 
pAKT cytoplasmicb 57 3.7 (2.5–5.4) 57 3.2 (2.3–4.7) 57 −11.4 (−34.2 to 19.4) 0.036 
 HER2 47 3.1 (2–4.9) 47 3.1 (2–4.6) 47 −1.2 (−28.3 to 36.2) 0.208 
 HER2+ 10 8.1 (5.9–11.1) 10 4.3 (1.6–11.9) 10 −46.9 (−77.2 to 23.4) 0.044 
Stathmin 91 8.1 (7.2–9) 91 6.1 (5.4–6.9) 91 −24.3 (−31.3 to −16.6) P < 0.001 
 HER2 72 7.9 (7–8.9) 72 6 (5.2–6.8) 72 −24.6 (−32.4 to −15.8) P < 0.001 
 HER2+ 19 8.7 (6.5–11.7) 19 6.7 (5.2–8.6) 19 −23.3 (−38.6 to −4.4) 0.014 
ER 91 56.9 (37.3–86.8) 91 51.1 (33–79.2) 91 −10.1 (−26.7 to 10.2) 0.13 
 HER2 72 71.9 (46.7–110.6) 72 60.9 (37.6–98.7) 72 −15.2 (−30.8 to 3.9) 0.04 
 HER2+ 19 23.5 (7.1–77.9) 19 26.3 (9.1–76.2) 19 12 (−41 to 112.7) 0.66 
PgR 91 28.1 (17.4–45.4) 91 27.5 (16.9–44.9) 91 −2.1 (−16.4 to 14.6) 0.78 
 HER2 72 39.2 (23.3–65.9) 72 36.3 (21.4–61.5) 72 −7.4 (−20 to 7.3) 0.28 
 HER2+ 19 8 (2.7–23.9) 19 9.6 (2.9–31.9) 19 20.5 (−30 to 107.2) 0.17 
Apoptosis (%)c 74 2.5 (2.1–2.8) 74 2.3 (2–2.6) 74 −8.4 (−18.2 to 2.6) 0.11 
 HER2 61 2.2 (1.9–2.6) 61 2.1 (1.8–2.5) 61 −5.5 (−16.8 to 7.5) 0.35 
 HER2+ 13 4 (3.4–4.8) 13 3.2 (2.5–4) 13 −21 (−38.9 to 2.2) 0.1 
B. mRNA biomarker 
HER2 mRNA 79 0.6 (0.5–0.8) 79 0.6 (0.5–0.8) 79 0 (−11.2 to 12.6) 0.9 
 HER2 62 0.4 (0.3–0.4) 62 0.4 (0.3–0.4) 62 2.8 (−6 to 12.3) 0.81 
 HER2+ 17 4.3 (2.8–6.8) 17 3.9 (2.3–6.6) 17 −9.4 (−44.2 to 47) 0.65 
HER3 mRNA 79 1.6 (1.4–1.9) 79 1.7 (1.4–2) 79 2.1 (−9 to 14.5) 0.83 
 HER2 62 1.7 (1.5–2.1) 62 1.7 (1.5–2) 62 −0.8 (−9.5 to 8.6) 0.94 
 HER2+ 17 1.3 (1–1.9) 17 1.5 (0.8–2.7) 17 13.3 (−27.9 to 78.1) 0.83 
Amphiregulin mRNA 79 0.3 (0.2–0.5) 79 0.4 (0.3–0.7) 79 49 (14.9 to 93.2) 0.0056 
 HER2 62 0.3 (0.2–0.5) 62 0.4 (0.3–0.7) 62 49 (10.1 to 101.6) 0.018 
 HER2+ 17 0.3 (0.1–1) 17 0.5 (0.2–1.3) 17 49 (−13.9 to 157.9) 0.14 
Neuregulin mRNA 61 5.7 (4–8.1) 61 8.3 (6.1–11.2) 61 46 (5.9 to 101.3) 0.07 
 HER2 49 6.2 (4.1–9.4) 49 8.9 (6.2–12.8) 49 43.6 (0.4 to 105.4) 0.07 
 HER2+ 12 3.9 (1.8–8.6) 12 6.1 (4–9.4) 12 56.3 (−33.2 to 266) 0.58 
Epiregulin mRNA 79 0.5 (0.3–0.7) 79 0.6 (0.4–0.9) 79 31.7 (−12 to 97.1) 0.36 
 HER2 62 0.4 (0.3–0.7) 62 0.6 (0.4–1) 62 39 (−13.9 to 124.4) 0.32 
 HER2+ 17 0.5 (0.2–1.5) 17 0.6 (0.2–1.6) 17 8.4 (−49 to 130.2) 0.98 
ESR1 mRNA 70 1 (0.7–1.6) 70 1 (0.6–1.5) 70 −3.1 (−18.4 to 15.1) 0.76 
 HER2 55 1.2 (0.7–2.1) 55 1.1 (0.7–1.9) 55 −8.6 (−24.7 to 11) 0.39 
 HER2+ 15 0.5 (0.2–1.2) 15 0.6 (0.3–1.1) 15 20.1 (−19.2 to 78.4) 0.39 
EGFR mRNA 70 1.7 (1.3–2.3) 70 2.2 (1.7–3) 70 32.1 (5 to 66.2) 0.02 
 HER2 55 1.6 (1.1–2.2) 55 2 (1.4–2.7) 55 22.8 (−6.8 to 61.9) 0.2 
 HER2+ 15 2.1 (1.2–3.8) 15 3.6 (2.4–5.6) 15 72.5 (18.7 to 150.7) 0.015 
NPretreatmentNSurgeryNChange (%)P
A. Protein biomarker 
Ki67 (%) 91 22.8 (19–27.4) 91 15.7 (12–20.4) 91 −31.3 (−43.6 to −16.3) 0.0004 
 HER2 72 20.7 (16.7–25.8) 72 15.2 (11–20.9) 72 −26.7 (−41.9 to −7.6) 0.0173 
 HER2+ 19 32.7 (24.4–43.8) 19 17.6 (11.7–26.3) 19 −46.3 (−62.6 to −23) 0.0029 
HER2 (FISH) 88 1.255 (1.21–1.3) 88 1.22 (1.2–1.27) 88 −0.03 (−0.08 to 0) 0.05 
 HER2 70 1.215 (1.2–1.25) 70 1.2 (1.13–1.22) 70 −0.02 (−0.06 to 0) 0.12 
 HER2+ 18 7.1 (4.5–9.2) 18 7.1 (4.7–9.5) 18 −0.28 (−0.9 to 0.1) 0.16 
EGFR (% pos)a 91 25.3% 91 26.4% 91 1.1% up /0% down 
 HER2 72 22.2% 72 23.6% 72 1.4% up /0% down 
 HER2+ 19 36.8% 19 36.8% 19 0% up /0% down 
pERK nuclearb 57 172 (165–179) 57 132 (107–162) 57 −23.2 (−37.4 to −5.9) 0.0011 
 HER2 47 170 (163–177) 47 128 (100–164) 47 −24.6 (−41.1 to −3.5) 0.0077 
 HER2+ 10 181 (163–201) 10 151 (121–188) 10 −16.6 (−29 to −2) 0.037 
pERK cytoplasmicb 57 7 (5.7–8.6) 57 5.1 (3.8–6.8) 57 −27.5 (−43.6 to −6.7) 0.029 
 HER2 47 6.8 (5.4–8.5) 47 4.9 (3.4–6.9) 47 −28.2 (−46.6 to −3.6) 0.06 
 HER2+ 10 8 (4.5–14.3) 10 6.1 (3.9–9.6) 10 −23.7 (−52 to 21.4) 0.22 
pAKT nuclearb 57 105 (91–122) 57 99 (86–115) 57 −5.7 (−19 to 9.9) 0.24 
 HER2 47 99 (84–117) 47 95 (80–112) 47 −4.5 (−20.5 to 14.8) 0.49 
 HER2+ 10 140 (116–168) 10 124 (96–160) 10 −11.1 (−25.5 to 6.1) 0.2 
pAKT cytoplasmicb 57 3.7 (2.5–5.4) 57 3.2 (2.3–4.7) 57 −11.4 (−34.2 to 19.4) 0.036 
 HER2 47 3.1 (2–4.9) 47 3.1 (2–4.6) 47 −1.2 (−28.3 to 36.2) 0.208 
 HER2+ 10 8.1 (5.9–11.1) 10 4.3 (1.6–11.9) 10 −46.9 (−77.2 to 23.4) 0.044 
Stathmin 91 8.1 (7.2–9) 91 6.1 (5.4–6.9) 91 −24.3 (−31.3 to −16.6) P < 0.001 
 HER2 72 7.9 (7–8.9) 72 6 (5.2–6.8) 72 −24.6 (−32.4 to −15.8) P < 0.001 
 HER2+ 19 8.7 (6.5–11.7) 19 6.7 (5.2–8.6) 19 −23.3 (−38.6 to −4.4) 0.014 
ER 91 56.9 (37.3–86.8) 91 51.1 (33–79.2) 91 −10.1 (−26.7 to 10.2) 0.13 
 HER2 72 71.9 (46.7–110.6) 72 60.9 (37.6–98.7) 72 −15.2 (−30.8 to 3.9) 0.04 
 HER2+ 19 23.5 (7.1–77.9) 19 26.3 (9.1–76.2) 19 12 (−41 to 112.7) 0.66 
PgR 91 28.1 (17.4–45.4) 91 27.5 (16.9–44.9) 91 −2.1 (−16.4 to 14.6) 0.78 
 HER2 72 39.2 (23.3–65.9) 72 36.3 (21.4–61.5) 72 −7.4 (−20 to 7.3) 0.28 
 HER2+ 19 8 (2.7–23.9) 19 9.6 (2.9–31.9) 19 20.5 (−30 to 107.2) 0.17 
Apoptosis (%)c 74 2.5 (2.1–2.8) 74 2.3 (2–2.6) 74 −8.4 (−18.2 to 2.6) 0.11 
 HER2 61 2.2 (1.9–2.6) 61 2.1 (1.8–2.5) 61 −5.5 (−16.8 to 7.5) 0.35 
 HER2+ 13 4 (3.4–4.8) 13 3.2 (2.5–4) 13 −21 (−38.9 to 2.2) 0.1 
B. mRNA biomarker 
HER2 mRNA 79 0.6 (0.5–0.8) 79 0.6 (0.5–0.8) 79 0 (−11.2 to 12.6) 0.9 
 HER2 62 0.4 (0.3–0.4) 62 0.4 (0.3–0.4) 62 2.8 (−6 to 12.3) 0.81 
 HER2+ 17 4.3 (2.8–6.8) 17 3.9 (2.3–6.6) 17 −9.4 (−44.2 to 47) 0.65 
HER3 mRNA 79 1.6 (1.4–1.9) 79 1.7 (1.4–2) 79 2.1 (−9 to 14.5) 0.83 
 HER2 62 1.7 (1.5–2.1) 62 1.7 (1.5–2) 62 −0.8 (−9.5 to 8.6) 0.94 
 HER2+ 17 1.3 (1–1.9) 17 1.5 (0.8–2.7) 17 13.3 (−27.9 to 78.1) 0.83 
Amphiregulin mRNA 79 0.3 (0.2–0.5) 79 0.4 (0.3–0.7) 79 49 (14.9 to 93.2) 0.0056 
 HER2 62 0.3 (0.2–0.5) 62 0.4 (0.3–0.7) 62 49 (10.1 to 101.6) 0.018 
 HER2+ 17 0.3 (0.1–1) 17 0.5 (0.2–1.3) 17 49 (−13.9 to 157.9) 0.14 
Neuregulin mRNA 61 5.7 (4–8.1) 61 8.3 (6.1–11.2) 61 46 (5.9 to 101.3) 0.07 
 HER2 49 6.2 (4.1–9.4) 49 8.9 (6.2–12.8) 49 43.6 (0.4 to 105.4) 0.07 
 HER2+ 12 3.9 (1.8–8.6) 12 6.1 (4–9.4) 12 56.3 (−33.2 to 266) 0.58 
Epiregulin mRNA 79 0.5 (0.3–0.7) 79 0.6 (0.4–0.9) 79 31.7 (−12 to 97.1) 0.36 
 HER2 62 0.4 (0.3–0.7) 62 0.6 (0.4–1) 62 39 (−13.9 to 124.4) 0.32 
 HER2+ 17 0.5 (0.2–1.5) 17 0.6 (0.2–1.6) 17 8.4 (−49 to 130.2) 0.98 
ESR1 mRNA 70 1 (0.7–1.6) 70 1 (0.6–1.5) 70 −3.1 (−18.4 to 15.1) 0.76 
 HER2 55 1.2 (0.7–2.1) 55 1.1 (0.7–1.9) 55 −8.6 (−24.7 to 11) 0.39 
 HER2+ 15 0.5 (0.2–1.2) 15 0.6 (0.3–1.1) 15 20.1 (−19.2 to 78.4) 0.39 
EGFR mRNA 70 1.7 (1.3–2.3) 70 2.2 (1.7–3) 70 32.1 (5 to 66.2) 0.02 
 HER2 55 1.6 (1.1–2.2) 55 2 (1.4–2.7) 55 22.8 (−6.8 to 61.9) 0.2 
 HER2+ 15 2.1 (1.2–3.8) 15 3.6 (2.4–5.6) 15 72.5 (18.7 to 150.7) 0.015 

NOTE: Pretreatment and posttreatment (surgical) protein (A) and mRNA expression biomarker (B) values for lapatinib-treated HER2 nonamplified and HER2+ tumors, together with percentage change and P values showing the statistical significance of the change. All protein biomarkers were assessed by IHC and scored as detailed in Materials and Methods. All mRNA biomarkers were assessed by RT-PCR and expressed as relative concentrations normalized to housekeeping genes. Geometric means and 95% confidence intervals are shown, except for HER2 FISH (median, CI) and EGFR (% of tumors positive). Only cases with quality RNA available on pre- and posttreatment samples were included.

aEGFR+ included weak or strong staining.

bOnly 57 pairs were evaluable for phosphorylated markers, because only pairs containing a posttreatment core biopsy were included.

cOnly 74 pairs were evaluable for apoptosis, e.g., more than 3,000 tumor cells.

As shown in Fig. 2B, the individual paired Ki67 changes show significant interpatient variability. Defining Ki67 responders or nonresponders as cases showing a >50% or <50% decrease in Ki67, respectively, there was a greater proportion of Ki67 responders (yellow lines in Fig. 2B) among the HER2+ subset (7 of 19; 37%); however, 14% (10 of 72) of HER2 tumors demonstrated >50% reduction in Ki67 (Fig. 2B).

As shown in the waterfall plots for individual tumor responses (Fig. 3A and B), some of the tumors demonstrating the greatest decrease in Ki67 were neither HER2+ nor EGFR+. There was no significant change in apoptosis with lapatinib in either HER2+ or HER2 tumors (Table 2).

Figure 3.

Waterfall plot of percent Ki67 change among lapatinib-treated cases according to HER2 (A) or EGFR (B) expression.

Figure 3.

Waterfall plot of percent Ki67 change among lapatinib-treated cases according to HER2 (A) or EGFR (B) expression.

Close modal

Biomarker changes with 2 weeks of lapatinib

As shown in Table 2, there were no significant changes in EGFR or HER2 protein expression with lapatinib treatment. The effect of lapatinib on signaling downstream from those receptors was examined by measuring changes in pAKT and pERK by IHC (nuclear and cytoplasmic staining). Given the artifactual loss of phosphorylated proteins in excision versus core biopsies (12), only paired samples containing a posttreatment core biopsy were included in the analyses of pAKT and pERK changes (n = 57). Phosphorylated ERK was significantly downregulated with lapatinib treatment; in particular, nuclear pERK was reduced by 25% (P = 0.008) in HER2 cases (Table 2). A significant reduction in cytoplasmic pAKT was observed in the HER2+ subset (47% reduction; P = 0.044). Stathmin has been suggested as a surrogate marker of PI3K–AKT pathway activation in gene expression assays that can be reproducibly measured by IHC (15). Stathmin was significantly reduced by 23% and 25% in HER2+ and HER2 lapatinib-treated tumors (P = 0.014 and P < 0.001, respectively; Table 2). There was no significant change in pERK, pAKT, or stathmin expression in paired placebo-treated cases (Supplementary Table S2); some reductions were observed, for example, stathmin was downregulated by 25% in the placebo group, a reduction similar to the lapatinib group. However, the numbers were small and the 95% CI too wide to be significant. In contrast, although there was a statistically significant 15% reduction in ER expression among HER2 lapatinib-treated cases, it was not considered relevant, as a similar trend was observed with placebo (Supplementary Table S2).

Changes in gene expression levels (mRNA) of EGFR and HER2 and ErbB ligands (neuregulin, amphiregulin, and epiregulin) were measured. Amphiregulin expression was significantly upregulated (P = 0.02) with lapatinib treatment in the HER2 subset only (Table 2), but not with placebo treatment, although the magnitude of the effect seen with placebo was similar but did not reach statistical significance due to the small numbers in this group (Supplementary Table S2).

Predictive biomarkers in HER2 (nonamplified) lapatinib-treated tumors

Because some of the tumors showing the greatest Ki67 decrease were neither HER2+ nor EGFR+, a multivariable analysis was undertaken, testing baseline molecular markers for their ability to identify Ki67 responders (≥50% Ki67 decrease) in HER2 nonamplified tumors. Five baseline pretreatment parameters were significant at P < 0.05 and were considered in the analysis, namely protein expression of ER, PgR, pERK (cytoplasmic), and RNA expression levels of HER2, and of HER3 (Supplementary Table S3). The analysis showed that only HER3 mRNA levels were independently significant, and HER2 lost significance once HER3 levels were allowed for. Moreover, there was a highly significant correlation between HER2 and HER3 mRNA expression (rho = +0.67, P < 0.001) in the HER2 nonamplified tumors (Fig. 4), and responders demonstrated high expression levels (above the median) for both HER2 and HER3 mRNA. The only biomarker change between baseline and 2 weeks that was significantly associated with Ki67 reduction was stathmin (rho = +0.44, P = 0.001; Supplementary Table S4).

Figure 4.

Correlations between expression of HER3 and HER2 mRNA among HER2 nonamplified lapatinib-treated cases. Tumors with a Ki67 response defined as a greater than 50% decrease are shown in blue.

Figure 4.

Correlations between expression of HER3 and HER2 mRNA among HER2 nonamplified lapatinib-treated cases. Tumors with a Ki67 response defined as a greater than 50% decrease are shown in blue.

Close modal

Predictive biomarkers among HER2+ lapatinib-treated tumors

A similar analysis was conducted among HER2+ cases and showed that the only baseline marker significantly associated with Ki67 response was epiregulin, where high EREG mRNA levels predicted a Ki67 reduction (rho = −0.70; P = 0.002; Supplementary Fig. S3). In contrast with HER2 nonamplified cases, stathmin change with treatment did not correlate with Ki67 change in HER2+ tumors. However, mRNA expression of HER2, HER3, AREG, and EREG increased nonsignificantly with Ki67 response (rhoHER2 = −0.4, P = 0.1; rhoHER3 = −0.5, P = 0.05; rhoAREG = −0.5, P = 0.06; rhoEREG = −0.5, P = 0.05; Supplementary Table S4).

There is a need for novel trial designs more specifically adapted to the molecular era of targeted cancer medicine. Conventional adjuvant trials in early breast cancer with large numbers of unselected patients and traditional clinical endpoints, such as progression-free survival (PFS) or overall survival (OS), require large numbers, with prolonged follow-up, and may not be appropriate for taking forward targeted therapies aimed at increasingly small subsets of patients. Presurgical trials using surrogate markers of in vivo drug activity, such as the cell proliferation marker Ki67, have emerged as a useful setting in which to investigate biologic predictors of response (9). In addition, because such short-term (10–14 days) treatment exposure is of no proven clinical efficacy and study participation does not extend normal waiting time to surgery, an unselected patient population can be enrolled quickly in this clinical setting in an attempt to establish which tumors may exhibit the desired antiproliferative response. Furthermore, the randomization to a placebo is both permissible and acts as a valuable control for the biologic efficacy observed with the experimental therapy.

In the MAPLE study, we have confirmed that a short-term presurgical trial design in early breast cancer using an EGFR/HER2 TKI with a primary biologic antiproliferative endpoint is both feasible and safe, and that patients are willing to contribute sequential research biopsies. Our target patient recruitment was achieved in a reasonable time frame, with paired samples of adequate quality and cellularity obtained in 98% of patients. Because the primary endpoint was change in Ki67 and biomarker assessments were conducted as soon as paired samples were received, results were available shortly after recruitment was completed, thus illustrating how the presurgical trial design can provide a valuable tool to screen new agents in a fast and cost-effective manner.

Short-term lapatinib treatment was associated with a significant Ki67 downregulation. Importantly, there was no evidence of a change in Ki67 in the placebo control group, indicating that no protocol-related interventions affected the primary endpoint. The trial included 26 patients with a HER2+ tumor; 25 were HER2 3+ by IHC and were found to also demonstrate HER2 amplification by FISH. The only patient with a HER2 2+/FISH+ tumor was randomized to placebo. Although Ki67 reduction was greatest among HER2+ tumors (Fig. 2A), lapatinib also resulted in a statistically significant 27% Ki67 downregulation among HER2 nonamplified tumors. Although lapatinib is currently used exclusively in HER2+ disease, MAPLE has shown that lapatinib may also have antiproliferative activity in a subset of HER2 nonamplified breast cancers. Although Ki67 has not been validated as a surrogate endpoint in ER disease, the changes in Ki67 observed in patients with HER2 breast cancer enrolled in MAPLE nonetheless indicate that this group is not entirely biologically refractory to lapatinib and are consistent with a beneficial effect of the drug. The challenge is, therefore, characterizing the sensitive subset in MAPLE as defined by an antiproliferative response. When considering individual cases (Fig. 2B), there was marked interpatient variability in Ki67 change, with 14% of HER2 nonamplified lapatinib-treated tumors showing a >50% reduction in Ki67. We therefore investigated whether baseline biomarkers could identify these “Ki67 responders” (i.e., with a >50% Ki67 decrease), and demonstrated that among HER2 tumors, the only independently significant baseline predictor for Ki67 response in a multivariable analysis was elevated levels of HER3 mRNA. These data are consistent with a recently reported smaller non–placebo-controlled study in 31 patients with HER2 primary breast cancer where a similar >50% decrease in Ki67 was reported in four (14%) tumors, all of which expressed HER3 (16).

There has already been some evidence to suggest that HER2 inhibition may have a role in HER2-low breast cancer. Roughly 10% of patients enrolled in a trial of adjuvant chemotherapy alone or with trastuzumab revealed that some of the randomized patients were in fact neither HER2 3+ nor FISH amplified (17). These HER2 patients enrolled in the phase III National Surgical Adjuvant Breast and Bowel Project (NSABP) B-31 trial had significantly better disease-free survival with a relative risk of 0.40 (P = 0.026) when given trastuzumab after completing treatment for early breast cancer. These findings have provided the rationale for a large ongoing clinical trial, B-47. In addition, data from the Cancer and Leukemia Group B neoadjuvant study investigating the benefit of dual HER2 targeting suggest that intrinsic subtype is a strong driver of response to HER2-targeted therapy. Therefore, although HER2 overexpression or amplification may be an important predictor of response to HER2 targeted therapies, it may not be an obligate predictor (18).

In addition, there is a biologic rationale for expecting that in the setting of modest HER2 protein expression in nonamplified tumors, other molecular factors could determine an antiproliferative response to an intracellular TKI, such as levels of other HER family receptors that may heterodimerize with HER2 or levels of HER family ligands (19, 20). HER2 activation requires receptor dimerization; HER2 has no known ligand, and activation occurs either via ligand-independent HER2 homodimerization associated with high levels of receptor membrane expression such as that found in HER2 gene–amplified tumors or, alternatively, via ligand-dependent receptor heterodimerization (i.e., EGFR:HER2, or HER2:HER3; ref. 21). In contrast with the antibody trastuzumab that recognizes an epitope on the extracellular domain that is only exposed in the case of HER2 homodimers (21, 22), the TKI lapatinib inhibits the intracellular domains of EGFR and HER2, and as such may also target various HER2-containing heterodimers. HER3 has a docking site for the p85 subunit of PI3K and acts as the preferred heterodimerization partner for HER2. HER2–HER3 heterodimers are increasingly identified as key drivers of the PI3K cell survival pathway (23–26). Activation of HER3 and downstream PI3K has been shown to predict resistance to trastuzumab but sensitivity to lapatinib in vitro (27). In MAPLE, we have demonstrated that the HER2 nonamplified tumors showing a Ki67 response with lapatinib tended to show elevated levels of both HER3 and HER2 (Fig. 4), raising the possibility that a subset of HER2 breast cancers with significant HER2:HER3 dimerization and activated PI3K signaling may be sensitive to the antiproliferative effects of lapatinib. Novel imaging technologies such as FRET/FLIM (Förster resonance energy transfer/fluorescence lifetime imaging) that can detect the physical association between these two proteins could be a useful technology to investigate the relevance of various HER family heterodimers and lapatinib responsiveness (28, 29).

Whereas phosphorylated HER3 and AKT are useful readouts of the PI3K pathway in vitro, pAKT and pHER3 have proven more challenging to detect in patient samples, and baseline pAKT did not predict response to lapatinib in MAPLE. The chromatin-associated protein stathmin has been identified as a surrogate measure of PI3K signaling (15). In MAPLE, stathmin was significantly downregulated by lapatinib treatment in the HER2 subset, whereas pAKT was not. However, a reduction in stathmin of similar magnitude was also observed in the placebo group, but was not statistically significant due to the small numbers in this group, indicating that further work is needed to confirm the effect of lapatinib on stathmin in the presurgical setting. More recent data indicate that stathmin is strongly related to proliferation, and this is likely to be detrimental to its use as a surrogate of PI3K signaling.

Not all HER2+ patients benefit from lapatinib, and in MAPLE only 37% of HER2+ tumors demonstrated a >50% Ki67 decrease, suggesting that additional biomarkers may improve the selection of patients with HER2+ breast cancer for lapatinib therapy. Among HER2+ cases, baseline epiregulin emerged as a significant predictor of response. Epiregulin is a ligand for EGFR and HER4, and high epiregulin may identify HER2+ tumors demonstrating high levels of HER2–EGFR or HER2–HER4 heterodimers that may preferentially benefit from lapatinib.

In conclusion, the short presurgical trial design is a feasible, safe, and useful setting to investigate candidate predictive biomarkers for antiproliferative response to novel therapeutic agents. Lapatinib demonstrated an antiproliferative response in a subset of HER2 nonamplified primary breast tumors. Among these HER2 tumors, high HER3 mRNA identified a subset of Ki67 responders, and there was a trend for combined increase in HER3 and HER2 mRNA expression to predict this antiproliferative response. It would, however, be necessary for further neoadjuvant trials to validate the use of HER2 TKIs in HER2 patients before this is taken any further, with high HER2 and HER3 expression being examined as indicators of response. Among HER2+ tumors, high levels of the EGFR/HER4 ligand EREG identified the Ki67 responders. Although expression of HER family receptors, their ligands, or HER receptor hetero- or homodimerization status may merit further investigation, greater characterization of the responding HER2 breast tumors is needed before clinical use can occur.

S.R.D. Johnston reports receiving speakers bureau honoraria from GlaxoSmithKline, and is a consultant/advisory board member for Novartis and Roche. J.M. Bliss reports receiving a commercial research grant from GlaxoSmithKline. M. Dowsett reports receiving a commercial research grant from GlaxoSmithKline and Roche and is a consultant/advisory board member for Roche/Genentech. No potential conflicts of interest were disclosed by the other authors.

Conception and design: A. Leary, S.R.D. Johnston, R. A'Hern, J.M. Bliss, G. Coombes, I. Smith, M. Dowsett

Development of methodology: A. Leary, S.R.D. Johnston, R. A'Hern, R. Sahoo, S. Detre, C. Harper-Wynne

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A. Leary, A. Evans, S.R.D. Johnston, R. Sahoo, B.P. Haynes, C. Harper-Wynne, N. Bundred, G. Coombes, I. Smith

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A. Leary, A. Evans, S.R.D. Johnston, R. A'Hern, J.M. Bliss, B.P. Haynes, M. Dowsett

Writing, review, and/or revision of the manuscript: A. Leary, A. Evans, S.R.D. Johnston, R. A'Hern, J.M. Bliss, N. Bundred, G. Coombes, I. Smith, M. Dowsett

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Hills, G. Coombes

Study supervision: A. Leary, A. Evans, S.R.D. Johnston, J.M. Bliss, C. Harper-Wynne, N. Bundred, G. Coombes, M. Dowsett

Other (PI and lead recruiter): A. Evans

The authors are grateful for research grant support for the clinical trial management and analysis and laboratory studies from GlaxoSmithKline Oncology, and for research support from the NIHR Biomedical Research Centre at The Royal Marsden NHS Foundation Trust.

This study was supported by grant CSMD CR007226. The MAPLE trial received research grant support for the clinical trial management, analysis, and laboratory studies from GlaxoSmithKline Oncology and had research support from the NIHR Biomedical Research Centre at The Royal Marsden NHS Foundation Trust. The clinical trial was cosponsored by The Institute of Cancer Research and The Royal Marsden NHS Foundation Trust, and endorsed by the Clinical Trials Awards and Advisory Committee of Cancer Research UK (CRUK E/06/039). ICR-CTSU receives core funding from Cancer Research UK, which supplemented the project-specific research funding. The trial was supported at participating sites by the NIHR Clinical Research Network.

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

1.
Blackwell
KL
,
Burstein
HJ
,
Storniolo
AM
,
Rugo
HS
,
Sledge
G
,
Aktan
G
, et al
Overall survival benefit with lapatinib in combination with trastuzumab for patients with human epidermal growth factor receptor 2-positive metastatic breast cancer: final results from the EGF104900 Study
.
J Clin Oncol
2012
;
30
:
2585
92
.
2.
Cameron
D
,
Casey
M
,
Press
M
,
Lindquist
D
,
Pienkowski
T
,
Romieu
CG
, et al
A phase III randomized comparison of lapatinib plus capecitabine versus capecitabine alone in women with advanced breast cancer that has progressed on trastuzumab: updated efficacy and biomarker analyses
.
Breast Cancer Res Treat
2008
;
112
:
533
43
.
3.
Johnston
S
,
Pippen
J
 Jr
,
Pivot
X
,
Lichinitser
M
,
Sadeghi
S
,
Dieras
V
, et al
Lapatinib combined with letrozole versus letrozole and placebo as first-line therapy for postmenopausal hormone receptor-positive metastatic breast cancer
.
J Clin Oncol
2009
;
27
:
5538
46
.
4.
Johnston
S
,
Trudeau
M
,
Kaufman
B
,
Boussen
H
,
Blackwell
K
,
LoRusso
P
, et al
Phase II study of predictive biomarker profiles for response targeting human epidermal growth factor receptor 2 (HER-2) in advanced inflammatory breast cancer with lapatinib monotherapy
.
J Clin Oncol
2008
;
26
:
1066
72
.
5.
Lynch
TJ
,
Bell
DW
,
Sordella
R
,
Gurubhagavatula
S
,
Okimoto
RA
,
Brannigan
BW
, et al
Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib
.
N Engl J Med
2004
;
350
:
2129
39
.
6.
Dowsett
M
,
Ebbs
SR
,
Dixon
JM
,
Skene
A
,
Griffith
C
,
Boeddinghaus
I
, et al
Biomarker changes during neoadjuvant anastrozole, tamoxifen, or the combination: influence of hormonal status and HER-2 in breast cancer–a study from the IMPACT trialists
.
J Clin Oncol
2005
;
23
:
2477
92
.
7.
Dowsett
M
,
Smith
IE
,
Ebbs
SR
,
Dixon
JM
,
Skene
A
,
Griffith
C
, et al
Short-term changes in Ki-67 during neoadjuvant treatment of primary breast cancer with anastrozole or tamoxifen alone or combined correlate with recurrence-free survival
.
Clin Cancer Res
2005
;
11
:
951s
8s
.
8.
The Australian and New Zealand Breast Cancer Trials Group, Clinical Oncological Society of Australia
. 
A randomized trial in postmenopausal patients with advanced breast cancer comparing endocrine and cytotoxic therapy given sequentially or in combination
.
J Clin Oncol
1986
;
4
:
186
93
.
9.
Dowsett
M
,
Nielsen
TO
,
A'Hern
R
,
Bartlett
J
,
Coombes
RC
,
Cuzick
J
, et al
Assessment of Ki67 in breast cancer: recommendations from the International Ki67 in Breast Cancer working group
.
J Natl Cancer Inst
2011
;
103
:
1656
64
.
10.
Guix
M
,
Granja Nde
M
,
Meszoely
I
,
Adkins
TB
,
Wieman
BM
,
Frierson
KE
, et al
Short preoperative treatment with erlotinib inhibits tumor cell proliferation in hormone receptor-positive breast cancers
.
J Clin Oncol
2008
;
26
:
897
906
.
11.
Martin
LA
,
Davies
GL
,
Weigel
MT
,
Betambeau
N
,
Hills
MJ
,
Salter
J
, et al
Pre-surgical study of the biological effects of the selective cyclo-oxygenase-2 inhibitor celecoxib in patients with primary breast cancer
.
Breast Cancer Res Treat
2010
;
123
:
829
36
.
12.
Pinhel
IF
,
Macneill
FA
,
Hills
MJ
,
Salter
J
,
Detre
S
,
A'Hern
R
, et al
Extreme loss of immunoreactive p-Akt and p-Erk1/2 during routine fixation of primary breast cancer
.
Breast Cancer Res
2010
;
12
:
R76
.
13.
Wolff
AC
,
Hammond
ME
,
Schwartz
JN
,
Hagerty
KL
,
Allred
DC
,
Cote
RJ
, et al
American Society of Clinical Oncology/College of American Pathologists guideline recommendations for human epidermal growth factor receptor 2 testing in breast cancer
.
Arch Pathol Lab Med
2007
;
131
:
18
43
.
14.
Detre
S
,
Saclani Jotti
G
,
Dowsett
M
. 
A “quickscore” method for immunohistochemical semiquantitation: validation for oestrogen receptor in breast carcinomas
.
J Clin Pathol
1995
;
48
:
876
8
.
15.
Saal
LH
,
Johansson
P
,
Holm
K
,
Gruvberger-Saal
SK
,
She
QB
,
Maurer
M
, et al
Poor prognosis in carcinoma is associated with a gene expression signature of aberrant PTEN tumor suppressor pathway activity
.
Proc Natl Acad Sci U S A
2007
;
104
:
7564
9
.
16.
Coombes
RC
,
Tat
T
,
Miller
ML
,
Reise
JA
,
Mansi
JL
,
Hadjiminas
DJ
, et al
An open-label study of lapatinib in women with HER-2-negative early breast cancer: the lapatinib pre-surgical study (LPS study)
.
Ann Oncol
2013
;
24
:
924
30
.
17.
Paik
S
,
Kim
C
,
Jeong
J
,
Geyer
CE
,
Romond
EH
,
Mejia-Mejia
O
, et al
ASCO annual meeting proceedings: benefit from adjuvant trastuzumab may not be confined to patients with IHC 3+ and/or FISH-positive tumors: Central testing results from NSABP B-31, Part I
.
J Clin Oncol
2007
;
25
:
511
.
18.
Carey
LA
,
Berry
DA
,
Ollila
D
,
Lyndsay
H
,
Krop
IE
,
Weckstein
D
, et al
Clinical and translational results of CALGB 40601: A neoadjuvant phase III trial of weekly paclitaxel and trastuzumab with or without lapatinib for HER2-positive breast cancer
.
J Clin Oncol
2013
;
31
(
suppl, abstr 500
).
19.
McIntyre
E
,
Blackburn
E
,
Brown
PJ
,
Johnson
CG
,
Gullick
WJ
. 
The complete family of epidermal growth factor receptors and their ligands are co-ordinately expressed in breast cancer
.
Breast Cancer Res Treat
2010
;
122
:
105
10
.
20.
Hayes
NV
,
Gullick
WJ
. 
The neuregulin family of genes and their multiple splice variants in breast cancer
.
J Mammary Gland Biol Neoplasia
2008
;
13
:
205
14
.
21.
Hynes
NE
,
Lane
HA
. 
ERBB receptors and cancer: the complexity of targeted inhibitors
.
Nat Rev Cancer
2005
;
5
:
341
54
.
22.
Ghosh
R
,
Narasanna
A
,
Wang
SE
,
Liu
S
,
Chakrabarty
A
,
Balko
JM
, et al
Trastuzumab has preferential activity against breast cancers driven by HER2 homodimers
.
Cancer Res
2011
;
71
:
1871
82
.
23.
Stern
DF
. 
ERBB3/HER3 and ERBB2/HER2 duet in mammary development and breast cancer
.
J Mammary Gland Biol Neoplasia
2008
;
13
:
215
23
.
24.
Liu
J
,
Nethery
D
,
Kern
JA
. 
Neuregulin-1 induces branching morphogenesis in the developing lung through a P13K signal pathway
.
Exp Lung Res
2004
;
30
:
465
78
.
25.
Lazrek
Y
,
Dubreuil
O
,
Garambois
V
,
Gaborit
N
,
Larbouret
C
,
Le Clorennec
C
, et al
Anti-HER3 domain 1 and 3 antibodies reduce tumor growth by hindering HER2/HER3 dimerization and AKT-induced MDM2, XIAP, and FoxO1 phosphorylation
.
Neoplasia
2013
;
15
:
335
47
.
26.
Balz
LM
,
Bartkowiak
K
,
Andreas
A
,
Pantel
K
,
Niggemann
B
,
Zanker
KS
, et al
The interplay of HER2/HER3/PI3K and EGFR/HER2/PLC-gamma1 signalling in breast cancer cell migration and dissemination
.
J Pathol
2012
;
227
:
234
44
.
27.
O'Brien
NA
,
Browne
BC
,
Chow
L
,
Wang
Y
,
Ginther
C
,
Arboleda
J
, et al
Activated phosphoinositide 3-kinase/AKT signaling confers resistance to trastuzumab but not lapatinib
.
Mol Cancer Ther
2010
;
9
:
1489
502
.
28.
Waterhouse
BR
,
Gijsen
M
,
Barber
PR
,
Tullis
ID
,
Vojnovic
B
,
Kong
A
. 
Assessment of EGFR/HER2 dimerization by FRET-FLIM utilizing Alexa-conjugated secondary antibodies in relation to targeted therapies in cancers
.
Oncotarget
2011
;
2
:
728
36
.
29.
DeFazio-Eli
L
,
Strommen
K
,
Dao-Pick
T
,
Parry
G
,
Goodman
L
,
Winslow
J
. 
Quantitative assays for the measurement of HER1-HER2 heterodimerization and phosphorylation in cell lines and breast tumors: applications for diagnostics and targeted drug mechanism of action
.
Breast Cancer Res
2011
;
13
:
R44
.