Triple negative tumours: a critical review
J S Reis-Filho, Molecular Pathology Laboratory, The Breakthrough Breast Cancer Research Centre, Institute of Cancer Research, 237 Fulham Road, London SW3 6JB, UK. e-mail: email@example.com
Breast cancer is a heterogeneous disease that encompasses several distinct entities with remarkably different biological characteristics and clinical behaviour. Currently, breast cancer patients are managed according to algorithms based on a constellation of clinical and histopathological parameters in conjunction with assessment of hormone receptor (oestrogen and progesterone receptor) status and HER2 overexpression/gene amplification. Although effective tailored therapies have been developed for patients with hormone receptor-positive or HER2+ disease, chemotherapy is the only modality of systemic therapy for patients with breast cancers lacking the expression of these markers (triple-negative cancers). Recent microarray expression profiling analyses have demonstrated that breast cancers can be systematically characterized into biologically and clinically meaningful groups. These studies have led to the re-discovery of basal-like breast cancers, which preferentially show a triple-negative phenotype. Both triple-negative and basal-like cancers preferentially affect young and African-American women, are of high histological grade and have more aggressive clinical behaviour. Furthermore, a significant overlap between the biological and clinical characteristics of sporadic triple-negative and basal-like cancers and breast carcinomas arising in BRCA1 mutation carriers has been repeatedly demonstrated. In this review, we critically address the characteristics of basal-like and triple-negative cancers, their similarities and differences, their response to chemotherapy as well as strategies for the development of novel therapeutic targets for these aggressive types of breast cancer. In addition, the possible mechanisms are discussed leading to BRCA1 pathway dysfunction in sporadic triple-negative and basal-like cancers and animal models for these tumour types.
epidermal growth factor receptor
poly ADP-ribose polymerase
Breast cancer is a heterogeneous disease, encompassing a number of distinct biological entities that are associated with specific morphological and immunohistochemical features and clinical behaviour.1–3 For many decades, invasive breast carcinomas were only classified according to histological type, grade, and expression of hormone receptors.1–3 More recently, following the success of the trastuzumab adjuvant clinical trials, characterization of HER2 expression has become an integral part of the pathological work-up for breast cancer patients. From an oncologist’s point of view, breast cancer patients fall into three main groups: (i) those with hormone receptor-positive tumours who are managed with a number of oestrogen receptor (ER)-targeted therapy options ± chemotherapy; (ii) those with HER2+ tumours, who will, in addition, receive HER2-directed therapy with trastuzumab or, in some situations, Lapatinib; and (iii) those with hormone receptor [ER and progesterone receptor (PR)]-negative and HER2− breast cancers, for whom chemotherapy is the only modality of systemic therapy available.
Concurrently with the development of trastuzumab as a targeted therapy for breast cancer patients, the first few results from genome-wide microarray analysis began to be reported. The class discovery expression profile studies pioneered by the Stanford group4–6 have demonstrated that the morphological heterogeneity of breast cancer can be recapitulated and systematically classified at the transcriptomic level.7,8 These studies have shown that the expression profiles of breast cancer display a systematic variation and allow classification of breast cancer into five main groups, two of them ER+ (luminal A and B) and three ER− groups [normal breast-like, ERBB2 (also known as HER2) and ‘basal-like’].4,5,7,9 In those, and in subsequent studies, it has been shown that the basal-like group is enriched for tumours that lack expression of hormone receptors and of HER2 and has a more aggressive clinical behaviour,5,8,10,11 a distinctive metastatic pattern12,13 and a poor prognosis despite responding to conventional neoadjuvant and adjuvant chemotherapy regimens.9,14
Based on the above it is clear that the interest in triple-negative cancers stems from (i) the lack of tailored therapies for this group of breast cancer patients and (ii) overlap with the profiles of basal-like cancers. However, there are practical questions that need to be addressed, including the definition of basal-like cancers, the characteristics of triple-negative disease, and whether basal-like phenotype equates with ‘triple negativity’ (i.e. lack of ER, PR and HER2 expression).
What is a basal-like breast carcinoma?
Basal-like breast carcinomas were so named because the neoplastic cells of this tumour type consistently express genes usually found in normal basal/myoepithelial cells of the breast,15–18 including high-molecular-weight ‘basal’ cytokeratins (CK; CK5/6, CK14 and CK17), vimentin, p-cadherin, αB crystallin, fascin and caveolins 1 and 2.5,10,11,19–23 Basal-like breast carcinomas, as defined by gene expression microarray analysis, account for up to 15% of all breast cancers. The tumours often affect younger patients, frequently lack expression of hormone receptors and HER2,4,5,10,21 show either p53 immunohistochemical expression or TP53 gene mutations5,24 in up to 85% of cases, display exceedingly high levels of proliferation-related genes4–6,21,25 and express epidermal growth factor receptor (EGFR) in >60% of cases.10,26,27 Although a subgroup of breast carcinomas, characterized by expression of high-molecular-weight CKs, lack of ER, PR and HER2 expression but expressing EGFR, was described >10 years ago,15,28–33 basal-like breast carcinomas gained widespread interest only after their rediscovery by microarray-based expression profile analysis.4,5,34
Morphologically, basal-like breast carcinomas are characterized by high histological grade, high mitotic index, the presence of central necrotic zones, pushing borders and conspicuous lymphocytic infiltrate.35–39 Moreover, the presence of metaplastic elements27,35–37 and medullary/atypical medullary features36,37,40 are significantly more prevalent in basal-like breast carcinomas than in other types of breast cancer. In fact, recent studies have demonstrated that >90% of metaplastic breast carcinomas27 as well as the majority of medullary carcinomas40,41 consistently show a basal-like phenotype.
Basal-like cancers, as defined by microarrays or by immunohistochemical surrogates, have been shown to have a more aggressive clinical behaviour.10,42,43 In fact, some studies have demonstrated that expression of basal keratins is a prognostic factor independent of tumour size, grade and lymph node status.42 However, when compared with either ER−, non-basal-like cancers44 or with grade-matched non-basal-like cancers,12 carcinomas with a basal-like phenotype do not seem to be associated with a poorer outcome. In addition, the pattern of metastatic spread of tumours with a basal-like phenotype seems to be different: they are reported to disseminate to axillary nodes and bones less frequently12,38 and to favour a haematogenous spread,12,13,38,45 with a peculiar proclivity to develop metastatic deposits in the brain and lungs.
What is a triple-negative breast cancer?
Triple-negative cancers account for 10–17% of all breast carcinomas,14,46–52 depending on the thresholds used to define ER and PR positivity and the methods for HER2 assessment. The main characteristics of triple-negative cancers that have emerged from the literature illustrate the similarities between basal-like and triple-negative tumours, including the fact that they more frequently affect younger patients (<50 years),46,47,49,50 are more prevalent in African-American women,50–52 often present as interval cancers and are significantly more aggressive than tumours pertaining to other molecular subgroups.14,46,47,49–51 This aggressiveness is best illustrated by the fact that the peak risk of recurrence is between the first and third years and the majority of deaths occur in the first 5 years, following therapy.46,49 On the other hand, differences in outcome between triple-negative cancers and tumours with other phenotypes are reduced at 10 years of follow-up. Interestingly, patients with basal-like12 or triple-negative cancers46,51 have a significantly shorter survival following the first metastatic event when compared with those with non-basal-like/non-triple-negative controls.
From a pathologist’s point of view, the differences between triple-negative and non-triple-negative breast cancers are not surprising, given that the majority of triple-negative cancers are of histological grade 3.46,48 Although basal-like and triple-negative cancers are predominantly of high histological grade,46,48,49 up to 10% of triple-negative tumours have been shown to be of grade 1 in one study.46 At the time of drafting this review, no systematic analysis of the morphological characteristics of triple-negative cancers compared with grade-matched non-triple-negative carcinomas is on record. However, based on the analysis of basal-like breast cancers27,35,37,40,53,54 and in these authors’ experience,27,35,53 the majority of triple-negative cancers are high-grade invasive ductal carcinomas of no special type, metaplastic carcinomas and medullary cancers. Although some salivary gland-like tumours of the breast are known to have a triple-negative phenotype,54 their inclusion in this group is arguable, given that they have distinct biological features and clinical behaviour.55
There are conflicting results on the prevalence of lymph node metastasis at the time of diagnosis in patients with triple-negative cancers; whereas in one study there was a higher prevalence of lymph node metastasis in triple-negative cancers compared with controls,46 others have found no difference47,48 or an inverse association between triple-negative phenotype and lymph node metastasis.49 Interestingly, it has been described that, unlike non-triple-negative cancers, no correlation between tumour size and presence of lymph node metastasis was observed in the triple-negative group.46 A similar dissociation between tumour size and prevalence of lymph node metastasis at diagnosis has been identified by Foulkes et al.56 in tumours arising in BRCA1 germ-line mutation carriers.
Triple-negative and basal-like breast cancers: synonyms?
Despite the great interest in basal-like cancers, there is still no internationally accepted definition of these tumours. From a scientific perspective, microarray-based expression profiling analysis remains the ‘gold standard’ for the identification of basal-like breast cancers.57 However, for the foreseeable future, this technology is unlikely to be introduced in the diagnostic armamentarium for breast cancer patients, and results of microarray-based expression profiling using RNA extracted from formalin-fixed archival samples are suboptimal.58 Therefore, several attempts to define an immunohistochemical surrogate for basal-like cancers have been described. Although many ‘immunohistochemical signatures’ have been described, little information about their specificity and sensitivity for the identification of basal-like cancers as defined by microarray analysis has been provided. The best example to date is the panel proposed by Nielsen et al.,10 where basal-like cancers are defined as those lacking both ER and HER2 expression and expressing CK5/6 and EGFR. This panel has a specificity of 100% and a sensitivity of 76% for the identification of basal-like cancers.
Given that by microarray-based expression analysis, basal-like cancers are preferentially negative for ER and PR and lack HER2 expression, some have claimed that ‘the basal-like category of tumours is composed almost entirely of triple-negative breast cancers’.46 Although there is a great deal of overlap between triple-negative and basal-like breast cancers, this overlap is by no means complete. Analysis of ER, PR and HER2 status of breast cancers classified by microarray-based expression profiling analysis as pertaining to the basal-like subgroup has revealed that 15–54% of them express at least one of these markers.9,10,24,34,44
Although at first glance, it seems obvious that triple-negative cancers mainly encompass those of basal-like phenotype, careful analysis of microarray-based expression profiles suggests that the triple-negative group also encompasses another molecular subgroup of tumours, namely the so-called normal breast-like cancers. In the majority of studies these latter have been shown to cluster together with basal-like and HER2 tumours in the ER− arm.4–6,21,34,43 Importantly, normal breast-like cancers not only have a slightly better prognosis than basal-like breast cancers,5,6,34,43 but also appear not to respond to neoadjuvant chemotherapy in the same fashion as basal-like cancers do.9
In fact, there is indisputable evidence that the group of triple-negative cancers is heterogeneous and does not comprise a ‘single entity’. When expression of basal CKs and EGFR was investigated in separate cohorts of triple-negative tumours, only 56–84% expressed these markers.48,49 Patients with triple-negative cancers expressing a basal phenotype had a significantly shorter disease-free survival than those with triple-negative cancers lacking the expression of basal markers.48,49 Therefore, caution should be exercised not to equate a triple-negative phenotype with a basal-like profile. Triple-negative cancer is not a synonym for basal-like cancer.
Triple-negative tumours: clinical implications
Triple-negative cancer constitutes one of the most challenging groups of breast cancers. As mentioned above, the only systemic therapy currently available for patients with such cancers is chemotherapy. Furthermore, although triple-negative tumours, in a way akin to basal-like breast cancers,9 show high rates of objective response to neoadjuvant anthracycline plus taxane neoadjuvant chemotherapy,14 patients with triple-negative cancers that have not evolved to pathological complete response still show a significantly poorer prognosis than those with tumours pertaining to other molecular subgroups.14
The similarities between basal-like and triple-negative cancers53,59,60 (see below) with breast cancers occurring in BRCA1 mutation carriers point to data indicating BRCA1 genotype-specific sensitivity and resistance to common chemotherapy drugs in model systems. These indicate that cells and murine tumours with BRCA1 mutations may be less sensitive to taxanes.61–63 It is therefore not surprising that triple-negative cancers appear not to show increased sensitivity to taxanes51 compared with non-triple-negative cancers, despite the fact that these tumours demonstrate high proliferation rates4–6,21,25 and a high prevalence of TP53 gene mutation,5,24 which have been shown to be predictors of response to taxanes.51,64 Given the evidence from randomized clinical trials for the significant role of taxanes in breast cancer, data from prospective randomized trials are required before the widespread use of taxanes in triple-negative breast cancer is called into question. However, as a result of the sensitivity in cellular models and in tumours with a dysfunctional BRCA1 pathway to cross-linking agents (e.g. platinum salts),59,60,63,65–67 clinical trials are now testing the efficacy of carboplatin, cisplatin or docetaxel in the management of patients with triple-negative cancers in advanced disease scenarios (Table 1).
Table 1. Current clinical trials testing the efficacy of platinum salts for the management of triple-negative cancers
|A phase II study of cisplatin as first-line therapy for triple-negative metastatic breast cancer|
|Cisplatin||Phase II |
|Overall response rate of TN tumours to cisplatin as first-line therapy||1 (single group assignment, and historical control)|| 39||June 2007||Metastatic breast cancer||Assess PFS, clinical benefit rate and OS. To assess safety and toxicity of the drug|
|Triple-negative trial (TNT)|
|Carboplatin versus Docetaxel||Phase III|
|Efficacy of carboplatin compared with docetaxel||2 (parallel assignment, active control)||350–450||September 2007||Metastatic or locally advanced recurrent breast cancer||Assess response according to RECIST criteria (PFS and OS)|
Basal-like and triple-negative tumours have been shown to express EGFR in up to 66% of cases,10,26,36,48,49,60,68,69 but EGFR activating gene mutations are remarkably rare.68,70 In contrast, EGFR gene amplification has been shown in up to 25% of cases of metaplastic breast cancers,26,68 a subgroup of tumours that consistently show a triple-negative/basal-like phenotype. Given that increased copy numbers of the EGFR gene, in the form of amplification and chromosome 7 polysomy, have been shown to be a predictor of response to anti-EGFR targeted therapies,71–73 clinical trials testing the efficacy of humanized anti-EGFR monoclonal antibodies and EGFR tyrosine kinase inhibitors are current underway in patients with triple-negative breast cancer (Table 2). In addition, c-kit has been shown to be preferentially expressed in tumours lacking hormone receptors and HER2 expression.10 As a result, it has been hypothesized that imatinib mesylate could be used as a tailored therapy for triple-negative cancers, based on its inhibition of c-kit. There is, however, little evidence of mutation in the KIT gene (a strong predictor of response to imatinib in gastrointestinal stromal tumours74) in breast cancer.75 The efficacy of KIT/PDGFRα pathway inhibition in the management of triple-negative cancers remains to be determined. More recently, dasatinib, a src pathway inhibitor, has been shown in preclinical studies to be effective in breast cancer cell lines with a triple-negative phenotype.76,77 Furthermore, patients who responded to this tyrosine kinase inhibitor had cancers preferentially lacking ER, PR and HER2 and expressing CK5 and CK17.77 Although the actual targets of dasatinib in triple-negative breast cancer are yet to be validated, a six-gene predictor including EPHA2, CAV1, CAV2, ANXA1 PTRF and IGFBP2 has recently been developed.77 We23,78 and others79,80 have demonstrated that caveolins 1 and 2, which are substrates for SRC family kinases,81 are preferentially expressed in basal-like and triple-negative cancers. Interestingly, caveolins 1 and 2 have also been shown to predict response to platinum salts.82
Table 2. Current clinical trials testing targeted therapies for the management of patients with triple-negative breast cancer
|Cetuximab and cisplatin in the treatment of ‘triple-negative’ metastatic breast cancer (BALI-1)|
|Cetuximab + cisplatin or cisplatin alone||Phase II|
randomized, treatment and efficacy study
|Objective overall response||2 (parallel assignment, active control)||180||June 2007||Metastatic breast cancer||Assess time to disease progression (PFS) and overall survival (OS) and time to response|
|Phase II trial of cetuximab alone and in combination with carboplatin (LCCC 0403) NCT00492375||Cetuximab alone or in combination with carboplatin||Phase II|
|Objective overall response||2 (parallel assignment, active control)||100 patients||NS||Metastatic breast cancer||Assess PFS and OS|
Assess downstream effects of EGFR inhibitor
|Phase-II trial evaluating the efficacy and safety of EndoTAG-1 in triple receptor-negative breast cancer patients|
|EndoTAG-1 + paclitaxel or EndoTAG-1 monotherapy in comparison to paclitaxel (control group)||Phase II|
|4-month progression free survival (PFS)||23||135||January 2007||Metastatic or relapsed breast cancer||Median PFS time|
4-month survival rate Median overall survival time
Clinical benefit assessment via quality of life (QoL) scale assess 4-month PFS, survival rate and tumour response
|A Study of dasatinib (BMS-354825) in patients with advanced ‘triple-negative’ breast cancer|
|Dasatinib as second-line therapy||Phase II|
non-randomized, treatment and efficacy study
|Objective overall response||1 (single group assignment, uncontrolled)|| 45||December 2006||Recurrent, locally advanced or metastatic breast cancer||Assess disease control rate and proportion of free or progression, PFS distribution, and response duration|
|Phase II study with abraxane, bevacizumab and carboplatin in triple negative metastatic breast cancer NCT00479674||Abraxane, bevacizumab and carboplatin||Phase II|
|Safety and tolerability||1 (single group assignment, and historical control)||70 patients||May 2007||Metastatic breast cancer||Assess PFS|
|A phase II trial of bevacizumab and ABI-007 (abraxane) as second-line therapy in Her-2-negative, hormone receptor negative metastatic breast cancer|
|Bevacizumab and abraxane||Phase II|
|Effect of addition of bevacizumab to abraxane as second-line therapy||2 (single group assignment and active control)||NS||May 2007||Metastatic breast cancer||NS|
|Study of SU011248 (Sunitinib) versus chemotherapy for patients with previously treated triple receptor-negative breast cancer|
|SU011248 versus standard of care chemotherapy or metastatic breast cancer||Randomized phase II, open lable, active control, parallel assignment, safety/efficacy study phase II||Progression-free survival||2 (active control, parallel assignment) ||200||November 2006||Metastatic breast cancer||Objective response rate; duration of response; overall survival; 1-year survival; patient reported outcomes of health related quality of life and disease related symptoms assess PCR, safety and toxicity of this regimen|
Triple-negative/basal-like carcinomas and BRCA1 tumours: variations of the same theme?
There is increasingly more coherent evidence to suggest a link between the BRCA1 pathway and triple-negative and basal-like breast cancers.59,60 In fact, the vast majority of tumours arising in BRCA1 germ-line mutation carriers, in particular those diagnosed before 50 years of age, have morphological features similar to those described in basal-like cancers83,84 and they display a triple negative47 and basal-like phenotype as defined by immunohistochemistry36,85 or expression arrays.6
Although, even at the genetic level, sporadic basal-like cancers and tumours arising in BRCA1 mutation carriers show similar molecular genetic profiles,86–89 they differ in the lack of BRCA1 somatic mutations in sporadic basal-like cancers. Despite this lack of BRCA1 mutations, it has recently been demonstrated that the BRCA1 pathway may be dysfunctional in sporadic basal-like tumours;53,59,60 BRCA1 protein expression levels have been shown to be significantly lower in tumours of high histological grade, lacking ER and PR expression and of basal-like phenotype.22 We, and others, have hypothesized that this down-regulation would be mediated by epigenetic mechanisms, such as gene promoter methylation and/or transcriptional silencing of BRCA1. In fact, the BRCA1 gene promoter is methylated in >60% of medullary90,91 and metaplastic53 breast cancers of basal-like phenotype. However, despite the significantly lower levels of BRCA1 mRNA expression in sporadic basal-like cancers compared with grade matched controls,53 sporadic invasive ductal carcinomas both with and without a basal-like phenotype showed similarly low prevalence of BRCA1 gene promoter methylation.53,92 We therefore investigated alternative epigenetic mechanisms of BRCA1 pathway inactivation and found that sporadic invasive ductal carcinomas with a basal-like phenotype expressed ID4, a negative regulator of BRCA1,93,94 at significantly higher levels than grade-matched controls.53 This mechanism may account for the low levels of BRCA1 expression in sporadic basal-like carcinomas of ductal/no special type morphology.
Modelling basal-like breast carcinomas
Based on the fact that the majority of basal-like breast cancers show a dysfunctional BRCA1 pathway53,59,60 and harbour TP53 gene mutations,5,6,24 we have engineered the conditional mouse BLG-Cre;Brca1F22–24/F22–24;p53+/−, where Brca1 gene is inactivated in β-lactoglobin-expressing cells (i.e. luminal epithelial cells of the mouse mammary gland) and all cells of the animal have only one wild-type allele of p53.95 Consistent with our findings in human tumours, pathological analysis of the tumours arising in these mice has revealed that 78% lacked hormone receptors and HER2 and expressed basal markers (CK14 and/or EGFR) and 88% showed homologous metaplastic elements. This mouse model provides another line of evidence for the link between basal-like phenotype and BRCA1 pathway dysfunction and may prove useful for testing novel therapies for basal-like cancers.95 Interestingly, another Brca1F/F; p53F/F conditional mouse model has recently been described by Jonkers’ group.96 The morphological and phenotypic characteristics of the tumours arising in these Brca1F/F; p53F/F animals are remarkably similar to those observed in our study,95 despite the fact that Brca1 was inactivated in CK14+ cells of the mouse mammary gland.
Taken together, there is strong circumstantial evidence to suggest that BRCA1 pathway dysfunction is integral to the biology of a significant subgroup of triple-negative and basal-like breast carcinomas. Given that tumours that have a dysfunctional BRCA1 pathway show an exquisite sensitivity to cross-linking agents (e.g. platinum salts)63,66,67 and to inhibitors of the poly ADP-ribose polymerase (PARP) enzyme,97 these findings suggest new therapeutic strategies for the management of patients with basal-like breast cancers53,59,60,65 for testing in clinical trials. In fact, trials of both approaches are currently underway in women with breast cancer. The ‘Triple Negative Trial (TNT) Trial’ compares carboplatin with docetaxel in women with advanced sporadic triple-negative breast cancer (http://www.controlled-trials.com/cctspringview2/mrct/showTrial.html?mrid=260717&srch). The BRCA trial tests the same hypothesis in a more genetically defined advanced breast cancer population with germ-line BRCA1 or BRCA2 mutation (http://www.breakthrough.org.uk/researchcentre/clinical_trials/brca_trial/index.html). The preliminary results of PARP inhibitors phase I clinical trials that have included patients with BRCA1-deficient tumours have been encouraging,97 and a phase II clinical trial in BRCA1/BRCA2-associated breast cancer has recently begun (http://clinicaltrials.gov/ct/show/NCT00494234?order=1).
The term triple-negative breast cancer encompasses a heterogeneous group of tumours that show distinctive, but rather heterogeneous, pathological and clinical features. Although a significant overlap with basal-like carcinoma is observed, it seems clear that ‘triple negativity’ should not be used as a surrogate marker for basal-like cancers. Although triple-negative cancers are reported to respond to neoadjuvant chemotherapy,9,14 survival of patients with such tumours is still poor and their management may therefore require a more aggressive alternative intervention. Thus, the development of biologically informed systemic therapies and targeted therapies for triple-negative cancers is of paramount importance and may prove to be a Herculean task, only achievable by understanding the complexity of this heterogeneous group of tumours. We envisage that this can be achieved only through collaborative work and clinical trial design involving pathologists, oncologists and laboratory scientists alike.