Human epidermal growth factor receptor-2-positive breast cancer: does estrogen receptor status define two distinct subtypes?

Abstract

Background

Human epidermal growth factor receptor-2 (HER2) overexpression occurs in ∼20% of breast cancers and has historically been associated with decreased survival. Despite substantial improvements in clinical outcomes, particularly with the emergence of HER2-targeted therapy, a substantial minority of patients still relapses, and progression is inevitable in metastatic disease. Accumulating data indicate that HER2-positive disease is itself a heterogeneous entity.

Methods and results

In this article, we qualitatively review the data supporting the classification of HER2-positive disease as at least two separate entities, distinguished by estrogen receptor (ER) status. We summarize differences in clinical outcomes, including response to neoadjuvant therapy, timing and patterns of dissemination, efficacy of therapy in the metastatic setting and survival outcomes.

Conclusions

The collective data are sufficiently strong at this point to propose that ER status defines two distinct subtypes within HER2-positive breast cancer, and we highlight the implications of this knowledge in future research, including understanding of the basic biology of HER2-positive breast cancer and the design of future clinical trials.

introduction

Breast cancer is a heterogeneous disease, with substantial genotypic and phenotypic diversity [1]. In clinical practice, four main breast cancer subtypes drive treatment decisions: estrogen receptor (ER)-positive and human epidermal growth factor receptor-2 (HER2)-negative with a low or intermediate differentiation grade; ER-positive and HER2-negative with a high differentiation grade, HER2-positive and triple-negative breast disease (ER, progesterone receptor-negative and HER2-negative).

HER2 protein overexpression or gene amplification occurs in ∼20% of primary breast carcinomas and is associated with decreased survival [2]. During the last decade, there has been a remarkable progress in the management of HER2-positive breast cancers, with anti-HER2 regimens emerging as the cornerstone of treatment of this subtype of breast cancers.

Initially, an inverse association was described between HER2 positivity and the presence of ER; nonetheless, the review of baseline tumor characteristics among several studies indicates that ∼50% of the patients with HER2-positive tumors are also ER-positive [3, 4].

Studies of a wide variety of regimens, including trastuzumab and other HER2-targeted agents for early-stage and metastatic disease have shown benefit across ER status in patients with HER2-positive disease. It is becoming clear that in HER2-positive disease, the impact of HER2-directed therapy may differ by ER status. This leads to the question: does ER status define two distinct subtypes in HER2, recapitulating the effect of ER in breast cancer as a whole?

In this article, we summarize the available data regarding differences in tumor biology and clinical outcomes by ER status in HER2-positive breast cancer.

In some situations, the data are conflicting and firm conclusions cannot be drawn. We further recognize that cross-trial comparisons can have substantial limitations. In addition, we acknowledge a number of additional limitations including the variability of regimens used, varying lengths of follow-up time (e.g. if the median follow-up time does not exceed 5 years, the late events will not be detected, particularly in ER-positive disease) or simply small sample sizes with extremely limited power to detect important clinical differences.

Another important source of difficulty is the definition and assessment of hormone receptor (HR) and HER2 positivity. Recent reports highlighted this controversy, showing that the use of reverse transcriptase polymerase chain reaction for HER2 assessment was not perfectly correlated with the standard methodologies, such as fluorescence in situ hybridization or immunohistochemistry. Moreover, there is marked variability, in the used thresholds for calling ER and PR positivity and in the definition of categories [e.g. the inclusion (or not) of progesterone receptor status in the definition of HR-positive breast cancer] [5–10].

Despite these limitations, we consider that the collective data are sufficiently strong to propose that ER status defines two distinct subtypes within HER2-positive breast cancer, and we highlight the implications to future research.

molecular characterization of HER2 disease: stratification by ER

gene expression signatures in breast cancer

The phenotypic diversity of tumors is accompanied by genotypic diversity that can be captured by gene expression analysis [1]. Each subtype is defined based on an ‘intrinsic’ gene list that translates to clinically distinct tumor subtypes and prognosis [1, 11–13].

Of note, in studies that initially defined the intrinsic subtypes, based on the measurement of messenger RNA, there was segregation by ER before HER2, suggesting that ER status is the most important discriminator of breast cancers and ER divides breast tumors into two major groups: ER-positive (luminal A and B) and ER-negative subtypes (normal like, HER2-enriched, basal and claudin-low) [1, 11–16].The luminal ER-positive tumors are characterized by relatively high expression of many genes also expressed by normal luminal epithelial cells [1]. The expression of the proliferation cluster is the most prominent difference between luminal A and B subgroups [16]. The luminal A group has the highest expression of genes that are characteristic of the ER cluster and low expression of the proliferation markers [11]. Luminal B has a low-to-moderate expression of E- related genes, variable expression of the HER2 associated genes, higher expression of the proliferation markers and moderate expression of some genes shared with the basal-like subtype [11, 12]. Luminal B tumors appear to be far more heterogeneous than those characterized as luminal A [11].

Overall, all ER-negative subtypes are highly proliferative [16] and HER2-associated genes play an important role in their segregation.

The basal subtype seems to have gene expression similarities with the basal epithelial cells of the normal mammary gland, having high levels of cytokeratins 5 and 17 [12]. The claudin-low subtype is characterized by higher expression of epithelial–mesenchymal transition genes [15]. Finally, the HER2-enriched subtype has a more similar gene expression profile to the one present in progenitor and stem cell-like cells [17].

Could ER-positive/HER2-positive disease have a different cell of origin from ER-negative/HER2-positive disease? This argument is plausible given the observation that ER-positive/HER2-positive disease has high expression of genes expressed by normal luminal epithelial cells and ER-negative/HER2-positive disease has high expression of genes expressed by progenitor, stem cell-like cells and eventually basal cells [1, 15, 17].

Although the classical histological markers do not fully recapitulate the intrinsic subtypes, most of the clinically ER-positive/HER2-positive cancers tend to fall in the luminal subtypes and ER-negative/HER2-positive in the HER2-enriched subtype, clearly two different entities [18–20] (Figure 1).

Figure 1.

Hierarchical clustering of invasive breast cancers. Clustering orders the cancers according to the greatest similarity of gene expression. The top color bar indicates the immunohistochemistry results, ‘blue’ is positive, ‘green’ is negative and ‘light blue’ is low positive. In the figure below, each column represents a different tumor sample and each row represents a different gene. The expression scale is relative. The degree of expression is normalized to the mean, ‘white’ represents mean, overexpression is represented by ‘red’, and underexpression is represented by ‘blue’ (courtesy of Andrea Richardson) [91].

Nevertheless, it is important to stress that this segregation of HER2 by ER may misclassify a substantial proportion of patients. For example, a combined dataset that included 106 patients that belong to HER2-enriched subtype comprised 51% ER-negative/HER2-positive, 15% ER-positive/HER2-positive and even 34% HER2-negative analyzed by classical procedures [20, 21]. Therefore, the information provided by the clustering subtypes complements and expands the information provided by classical clinical-pathologic markers.

Marchio et al. using a microarray-based assay compared the genomic profiles of ER-positive and ER-negative/HER2-positive tumors. They concluded that HR status does not determine the overall genetic profile of HER2-positive breast cancers. Nevertheless, specific genetic aberrations were characteristic of subgroups of HER2 breast cancers defined by ER. For example, in this small dataset, TOP2A was substantially more prevalently amplified in ER-positive than in ER-negative [22]. Similarly, in a genome-wide copy number variants/gene expression profiling study, an amplicon with TOP2A defined a subgroup of ER-positive/HER2-positive breast cancers [23].

biomarkers in breast cancer

Results of biomarker analysis of the NeOAdjuvant Herceptin (NOAH) and the NeoSphere trials indicated that ER-positive/HER2-positive and ER-negative/HER2-positive breast cancer are driven by distinct biologic mechanisms [24, 25]. For example, in the NeoSphere trial, a panel of biomarkers including HER1, HER2, HER3, IGF1R, PTEN, pAKT, amphiregulin, betacellulin, transforming growth factor-α and PIK3CA mutations showed a qualitatively different expression distribution in the ER-negative and ER-positive subgroups [25]. Previously, Saal et al. [26] had shown a correlation of PIK3CA mutations with ER-positive tumors and PTEN loss with ER-negative tumors. Unpublished data from the M.D. Anderson Cancer Centre showed that the prevalence of PIK3CA mutation within HER2 disease was 24%. Within this population, 43% were ER-negative and 57% were ER-positive (personal communication of Ana Maria Gonzalez-Angulo). Interestingly, in the former biomarker analysis of the NeoSphere trial, the frequency of PIK3CA mutations in this HER2-positive disease population was similar in ER subgroups, although we cannot rule out the existence of group-specific PIK3CA mutations [25]. Additional data will be required to more fully define the relationship between alterations in the PI3K pathway and ER status within HER2-positive tumors.

impact of androgen receptor on the biology of breast cancer

Overall, 70% of breast cancers express androgen receptor (AR) [27–29]. The presence of AR varies by subtype: more than 90% of luminal A, ∼70–90% of luminal B, ∼10–30% of triple-negative and ∼60% of HER2-positive [27, 28]. Within HER2-positive breast cancer, AR expression is found in the majority of both ER-positive and ER-negative tumors [30]. Additionally, expression of AR appears to be related with a different impact on tumor biology according to the ER receptor.

In ER-positive tumors, including those that are also HER2-positive, AR has been independently associated with good prognosis; leading to the inhibition of proliferation [31, 32].

In ER-negative tumors, AR expression is associated with HER2 overexpression. AR is expressed in the majority of ER-negative/HER2-positive tumors, compared with 10–30% of triple-negative tumors [27, 28]. Furthermore, in ER-negative tumors, AR expression may actually enhance tumor proliferation [32–34], via WNT and HER2 signaling pathways through up-regulation of WNT7B and HER3 [34].

These findings raise the hypothesis that AR antagonists may represent a novel approach to target the HER3 pathway, particularly among patients with AR-positive, ER-negative and HER2-positive breast cancers [34]. In this context, it is also of interest that the pathological complete response (pCR) rate to the combination of pertuzumab and trastuzumab in the NeoSphere study was substantially higher among patients with ER-negative, HER2-positive tumors [35].

ER/HER2 crosstalk in breast cancer

Bidirectional crosstalk between ER and HER2 has been suggested clinically and described in preclinical models. In some studies, HER2 overexpression is associated with resistance to endocrine therapy [36, 37], and similarly, ER pathways have been postulated as means of escape to HER2-directed therapy [35, 38–44].

On the one hand, nongenomic actions of ER are associated with an increase in phosphorylated levels of HER2 and activating cellular kinases such as PI3K [45]. On the other hand, HER2 interferes with the ER pathway through a variety of HER2 downstream kinases (e.g. MAPKs and AKT) that phosphorylate both ER and ER coactivators (e.g. AIB1) leading to ligand-independent ER activation and ER-dependent transcription stimulation [46, 47]. Additionally, it has been shown that tamoxifen behaves as an estrogen agonist in breast cancer cells that overexpress HER2 and AIB1 [37]. Moreover, the inhibition of HER2 activity appears to increase some of the tamoxifen inhibitory effect in cell proliferation, and in ER-positive/HER2-positive xenograft breast tumors, a combination of several HER inhibitors, together with endocrine therapy, either tamoxifen or estrogen deprivation, was more effective in inducing apoptosis and slowing proliferation than each individual drug [48]. Some, though not all, clinical data also support a role of ER/HER2 crosstalk in endocrine resistance, with retrospective studies showing lower endocrine therapy response when HER2 is present and with some clinical trials demonstrating the benefit of blocking both the ER and the HER2 pathway. For example, in the Tandem trial, patients treated in the first-line advanced setting with anastrozole and trastuzumab experienced substantially longer median progression-free survival (4.8 versus 2.4 months, P = 0.002) than patients treated with anastrozole in monotherapy; similarly, in the III trial of lapatinib with letrozole versus letrozole patients treated with the combination also substantially had longer progression-free survival (8.2 versus 3.0 months, P = 0.019) [49–51].

More recently, it has been suggested that ER/HER2 crosstalk may be implicated in escape from HER2-directed therapy. Wang et al. showed that following treatment with lapatinib and trastuzumab, ER and its downstream products increased in four out of five ER-positive/HER2-positive cell lines. Furthermore, in the setting of HER2 blockage with lapatinib and trastuzumab, acquisition of resistance required the activation of the ER pathway, via Bcl2 family members [38].

neoadjuvant therapy in HER2 disease: stratification by ER

efficacy

In general, in the preoperative setting, ER-negative tumors are associated with a substantially higher pCR compared with ER-positive tumors, across a wide variety of chemotherapy regimens [52–58].

In HER2 disease treated with neoadjuvant chemotherapy, ER status continues to influence the likelihood of pCR. In a retrospective analysis by Guarneri et al., with most of patients receiving anthracycline- and taxane-based therapy, the rate of pCR was 15% in patients with HR-positive/HER2-positive tumors versus 29% in patients with HR-negative/HER2-positive tumors (P < 0.001). These results have been replicated in several other cohorts [53, 59], though it should be noted that some conflicting results have also been observed [60].

The addition of trastuzumab substantially increases pCR rates in HER2-positive disease. pCR rates appear numerically higher in ER-negative compared with ER-positive disease; however, most studies did not report any direct statistical comparison [35, 39–44] (supplementary Figure S1, available at Annals of Oncology online).

Results of trials with the combination of anti-HER2-targeted therapy with neoadjuvant chemotherapy reveal, again, the differences in this subgroup by ER, with higher pCR rates in the ER-negative group in the NeoSphere, Neo-ALTTO and CHER-LOB trials [35, 43, 44] (supplementary Figure S1, available at Annals of Oncology online). For example, in NeoSphere, the rate of pCR among patients treated with biological therapy alone (e.g. trastuzumab–pertuzumab) was 27% among patients with ER-negative tumors, compared with 6% among patients with ER-positive tumors. In the three studies, the rate of pCR in the ‘triple combination’ arms was 63%, 61% and 56% among patients with ER-negative tumors, compared with 26%, 42% and 46% among patients with ER-positive tumors, respectively [35, 43, 44]. The same relationship between ER status and likelihood of pCR is also seen in combinations with lapatinib or pertuzumab.

These results are consistent with the previous suggestion of ER as an escape mechanism of HER2 inhibition [38].

meaning of pCR

Achievement of a pCR appears to be associated with excellent outcomes across tumor subtypes [52, 53, 57]. What is less clear is the prognostic significance of a lack of pCR by ER status. Notably, in a retrospective study of Ring et al. [54], pCR had no prognostic significance in patients with ER-positive tumors. In the setting of HER2-positive disease, a very recent meta-analysis that included 6377 patients with operable or locally advanced nonmetastatic breast cancer showed that pCR strongly correlated with disease-free survival in ER-negative/HER2-positive breast cancer but not in ER-positive/HER2-positive breast cancer [61] (supplementary Figure S2, available at Annals of Oncology online). This finding may reflect fundamental underlying biologic differences by ER or the fact that patients with ER-positive tumors generally go on to receive additional active therapy in the form of endocrine therapy. Notably, despite a lower rate of pCR among patients with ER-positive/HER2-positive tumors across neoadjuvant trials, the prognosis of such patients does not appear worse than the prognosis of patients with ER-negative/HER2-positive tumors [62, 63].

meaning and implications for trial design

In the past, patients enrolled into adjuvant and neoadjuvant trials were selected primarily on the basis of stage. More recently, distinct trials have been developed for patients with distinct biologic subtypes of disease. Patients with HER2-positive disease have not been further separated by ER status with respect to clinical trial eligibility or design. Based on the available data, it appears that the next generation of HER2 clinical trials should consider the impact of ER status on outcome.

In some cases, it may be best to design distinct treatment protocols for patients with ER-positive/HER2-positive disease compared with ER-negative/HER2-positive disease; in other settings, patients can be enrolled to a common treatment protocol, but one that stratifies patients by ER status and/or is designed prospectively to explore the value of therapy by ER status. The recently completed TBCRC006 trial provides a good example of the understanding of the underlying differences in ER-positive/HER2-positive and ER-negative/HER2-positive tumors and the crosstalk between them. Patients on the TBCRC 006 study received lapatinib/trastuzumab plus concurrent endocrine therapy if ER-positive and lapatinib and trastuzumab alone if ER-negative. Although the pCR was higher in patients with ER-negative disease, a high proportion of ER-positive patients achieved a clinical response and what was characterized as a near-pCR, suggesting that there may be value in this combined targeted approach [51].

adjuvant therapy in HER2 disease: stratification by ER

efficacy

Patients with either ER-positive and ER-negative tumors seem to derive benefit from adjuvant polychemotherapy; however, the magnitude of absolute improvement can be different between subgroups: in the Early Breast Cancer Trialists' Collaborative Group overview analysis, women under age 50 with ER-positive tumors had a 5-year absolute risk reduction in recurrence of 8%, while patients with ER-poor tumors had a 13% 5-year absolute risk reduction in recurrence. Similar differences were seen in women ages 50–69 [64]. A retrospective analysis of three consecutive randomized clinical trials of Cancer and Leukemia Group B (CALGB) (CALGB 8541, 9344, 9741) also demonstrated higher benefit in the ER-negative, compared with ER-positive groups [65]. Nevertheless, not all studies reported the same predictive value of ER status in the adjuvant setting, particularly from the addition of taxanes [66]. In patients with HER2-positive disease paclitaxel (Taxol, Bristol-Myers Squibb, New York, USA) was associated with benefit in both ER-positive and ER-negative patients [67].

In HER2 disease, across each of adjuvant trastuzumab trials, similar benefit, defined by relative risk reduction, was seen to the addition of trastuzumab in both ER-positive and ER-negative subgroups [62, 63, 68, 69].

pattern of relapse

As is well appreciated, the pattern of recurrence differs between ER-positive and ER-negative tumors. For ER-positive breast cancer, relapses are relatively consistent over time and continue to occur after 15 years of follow-up. For ER-negative breast tumors, relapses tend to occur in the first 5 years, and in fact, the impact of adjuvant chemotherapy is seen mostly within the first 3 years [64]. Notably, ER appears to influence timing of relapse in HER2 disease as well [63, 67, 70].

In the same way, while disease-free survival appears to plateau within a few years in patients with ER-negative/HER2-positive cancers, patients with ER-positive/HER2-positive cancers are at continuing risk of recurrence even several years out from diagnosis. This pattern of relapse seems to be an intrinsic feature, occurring in patients treated alone or those treated with chemotherapy and trastuzumab combinations. For example, this difference in the hazard of relapse over time is apparent in both the CALGB9344/INT0148 and the HERA trials (supplementary Figure S3, available at Annals of Oncology online). Consistently, in a cohort of 3394 patients with HER2-positive disease of the national comprehensive cancer network (NCCN) cohort, the risk of death for patients with HER2-positive disease stratified by HR was not proportional over time. Patients with HR-negative tumors had a higher risk of death within 5 years of initial diagnosis; however, beyond 5 years no statistically significant differences in hazards of death between groups were found [63, 67, 71].

sites of recurrence

The pattern of metastatic spread differs by ER status. ER-negative tumors are associated with a substantially higher rate of tumor recurrence in the viscera, soft tissues and brain, while ER-positive status is associated with substantially higher rates of tumor recurrence involving the bone [72, 73].

Kennecke et al. analyzed a cohort of 3726 patients, 6% HR-positive/HER2-positive and 7.1% HR-negative/HER2-positive. Among patients who received adjuvant therapy, patients with HR-negative/HER2-positve tumors had a lower frequency of bone metastases. Additionally, a higher rate and frequency of brain involvement in HR-negative/HER2-positive disease compared with HR-positive/HER2-positive disease [70]. Concordant with this, in a retrospective cohort of 296 patients with HER2 disease, there was a nonsignificant trend to higher frequency of bone metastasis in patients with ER-positive/HER2-positive tumors than in the ER-negative/HER2-positive disease (52.9% versus 41.4%, P = 0.072). On the other hand, there was, as well, a nonsignificant trend to less pleural and central nervous system metastasis in ER-positive/HER2-positive tumors compared with ER-negative/HER2-positive (8% versus 16.2%, P = 0.08 and 31.8% versus 43.5%, P = 0.09) [73]. In an analysis of patterns of recurrence of the NCCN cohort, HR-positive/HER2-positive patients were also more likely to experience recurrence in the bone [odds ratio for first recurrence after adjusting for age, stage and adjuvant trastuzumab therapy = 1.89, 95% confidence interval (CI): 1.20–2.94] [71].

These data seem to recapitulate the seed and soil theory of Paget [74] and add evidence to the role of ER in HER2 heterogeneity. The mechanisms underlying this site specificity are under active investigation. Nonoverlapping organ-specific signatures of recurrence in breast cancer have previously been recognized, including those for the bone, lung and brain. Some of these organ-specific signatures may also correlate with specific breast cancer subtype signatures [75–78].

metastatic HER2 disease: stratification by ER

efficacy of anti-HER2 therapies chemotherapy combinations

The median overall survival appears to be longer for patients with metastatic HR-positive/HER2-positive tumors compared with HR-negative/HER2-positive tumors in most of the studies. Nevertheless, it is important to highlight that even in this setting, the risk of death may be not proportional over time by HR status [71, 79].

Despite data in the neoadjuvant setting indicating a higher pCR rate in patients with ER-negative/HER2-positive tumors, whether there is any differential benefit in the metastatic setting is less clear (Table 1). In trials testing trastuzumab monotherapy or combinations of trastuzumab with chemotherapy, clinical benefit was observed regardless of HR status [79–82].

Table 1.

Results of objective response rate in phase III/II trials of trastuzumab in monotherapy or trastuzumab chemotherapy combinations in metastatic disease [79–82]

Study/author	Treatment	ER + n, %	Objective response rate
			ER+ (%)	ER− (%)
Slamon et al.'s studya (Brufsky et al. [79])	Trastuzumab + chemotherapy (AC or T)	269, 45	58	51
	Chemotherapy (AC or T)		30	32
Vogel et al.'s studya (Brufsky et al. [79])	Trastuzumab		33	34
CobLeigh et al.'s studya (Brufsky et al. [79])	Trastuzumab		21	18
Schilling et al.	Trastuzumab + vinorelbine	25, 69	44	43
Bayo-Calero et al.a	Trastuzumab + vinorelbine	10, 19	52	65
Wardley et al.	Trastuzumab + docetaxel + capecitabine	50, 45	20	25
	Trastuzumab + docetaxel	39, 35	5	24

No display of statistical tests for difference, because for most of the studies they were not published.

^aResults related with hormone receptor (HR) status (HR-positive: ER and/or progesterone receptor-positive, HR-negative: ER and progesterone-negative).

ER, estrogen receptor; AC, doxorubicin and cyclophosphamide; T, paclitaxel.

In registHER, a prospective observational study of ∼1000 patients with metastatic HER2-positive breast cancer treated with trastuzumab-based chemotherapy regimens, HR negativity was associated with a shorter progression-free survival (median progression-free survival: 8.3 months versus 11.2 months, for HR-negative versus HR-positive, P = 0.010, respectively) [83].

In contrast, a phase III trial of first-line lapatinib and paclitaxel versus paclitaxel demonstrated that the addition of lapatinib to paclitaxel improved event-free survival in the subset of patients with HER2-positive tumors (n = 86), with numerically larger differences in HR-negative disease [8.3 versus 5.0 months; P = 0.007; hazard ratio (HR) = 0.34; 95% CI: 0.17–0.7] compared with HR-positive disease (5.7 versus 4.5 months; P = 0.351; HR = 0.7; 95% CI: 0.34–1.45). The interaction was not significant (P = 0.170), though the study was underpowered to examine this question [84]. Furthermore, results must be interpreted cautiously given its post hoc nature.

There are no definitive data on the results of different strategies used in patients treated beyond trastuzumab progression. The German breast group 26/Breast international group 03–05 study that compared capecitabine and trastuzumab with capecitabine in monotherapy showed that progression-free survival was superior in the combination therapy. Time to progression and overall survival were independently associated with HR status. Nevertheless, this was not consistent with other prospective trials and retrospective analysis [85–87].

In trials that have included pertuzumab, all subgroups seem to derive benefit for the addition of pertuzumab, independently of ER status. Results of CLEOPATRA, a phase III trial where 808 patients with previously untreated metastatic breast cancer received either pertuzumab–trastuzumab and docetaxel (Taxotere, Sanofis, Paris, France) chemotherapy or trastuzumab and docetaxel, showed an HR of progression-free survival of 0.72 (CI: 0.55–0.95) in the HR-positive subgroup and 0.55 (CI:0.42–0.72) in the HR-negative subgroup [88, 89].

Interestingly, in a phase I–II study of HER2-positive metastatic breast cancer treated with trastuzumab and gefitinib, time to progression was higher in the HR-positive group [90]. While these results must be interpreted cautiously and as hypothesis-generating only, they do raise the possibility that ER status (as a marker of signaling networks within a tumor) may have more relevance in the context of combinations of targeted therapies, as opposed to chemotherapy-based combinations.

conclusions

More than two decades after HER2, overexpression was recognized as a critical determinant of breast cancer biology and over a decade following the introduction of HER2-directed therapy, we still do not have a complete picture of the diversity of HER2 breast cancers. The heterogeneity of HER2-positive tumors is manifested in differences in response to a wide variety of therapies, patterns of recurrence and survival. The challenge in this setting is to define what underlying biological factors actually contribute to the observed differences, and how such factors may be targeted in an effort to improve patient outcomes.

Despite some conflicting results reported on the significance of ER status in HER2 disease, there appear to be clinically relevant differences in the response to neoadjuvant therapy, predictive value of pCR, timing and sites of relapse and risk of death over time.

In the years ahead, it will be critical to translate the clinical knowledge to basic biology, with the ultimate goal of improving patient outcomes. The clinical observations suggesting differences within HER2-positive tumors by ER status are consonant with gene expression profiling data, indicating that these tumors do segregate differently across molecular subtypes. It is possible that the same pathway(s) may play distinct roles in ER-positive /HER2-positive, when compared with ER-negative/HER2-positive tumors, and an understanding of these differences may allow for more directed development of novel targeted approaches.

As we move forward, it will be critical to continue translational efforts bridging the clinic and laboratory science. Access to tumor tissue is urgently needed to better characterize signaling networks and resistance mechanisms with a sufficient sample size to move beyond the broad tumor subtypes and to tailor individual therapies to those patients most likely to benefit.

funding

Additional support for this work has been provided by the Breast Cancer Research Foundation (no grant number assigned). IV-L is supported by Fundacao da Ciencia e Tecnologia (HMSP-ICS/0004/2011) and by the Scholars in Clinical Science Program of Harvard Catalyst | The Harvard Clinical and Translational Science Center (Award No. UL1 RR025758 and contributions from Harvard University and its affiliated academic health-care centers). The content of this work is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University and its affiliated academic health-care centers, or the National Institutes of Health.

disclosures

EPW: Research Funding: Genentech. NUL: Research Funding: Genentech, GlaxoSmithKline, Infinity, Boehringer Ingelheim; Consultant or Advisory role: GlaxoSmithKline, Novartis. IV-L: has declared no conflicts of interest.