|Home | About | Journals | Submit | Contact Us | Français|
Notorious for its poor prognosis and aggressive nature, triple-negative breast cancer (TNBC) is a heterogeneous disease entity. The nature of its biological specificity, which is similar to basal-like cancers, tumors arising in BRCA1 mutation carriers, and claudin-low cancers, is currently being explored in hopes of finding the targets for novel biologics and chemotherapeutic agents. In this review, we aim to give a broad overview of the disease’s nomenclature and epidemiology, as well as the basic mechanisms of emerging targeted therapies and their performance in clinical trials to date.
Breast cancer classification is in constant evolution, as advances in DNA and RNA microarrays as well as immunohistochemical (IHC) staining allow researchers to define the molecular heterogeneity of different disease subtypes and to guide the selection of appropriate treatment. With routine clinical testing for the expression of HER-2/neu in all breast cancer cases and a significantly improved survival rate by trastuzumab in women with HER-2/neu positive disease,1–5 a subtype—triple-negative breast cancer (TNBC)—has been recognized and garnered recent attention. The lack of HER-2/neu coupled with the absence of estrogen-receptors (ER) and progesterone-receptors (PR) defines triple-negative breast cancer. Without these targets, women with TNBC do not benefit from hormonal therapy or trastuzumab, and are left with chemotherapy as their only option. TNBC is a disease that accounts for approximately 7%–20% of all breast cancers6–12 and is known for its aggressive nature and poor prognosis. Traditional chemotherapy drugs may benefit some of these patients but the relapse rate is high and the survival rate continues to lag behind other subtypes. The biological specificity of TNBC however, may be exploited in the development of novel targeted therapy.
One of the difficulties in addressing TNBC is the heterogeneity of the disease entity. As a result, various terminologies have been used to describe the disease and associated biologies. TNBC is a clinical term, characterized by the lack of expression of ER, PR, and HER-2/neu in a subgroup of breast cancer cases. Perou et al defined five molecular subtypes (luminal A and B, HER-2/neu positive, normal breast-tissue like, and basal-like) in their microarray-based expression profiling study.13 Basal-like breast cancer, which expresses genes usually found in the basal cells of the normal breast, has since become an area of research interest.14 While TNBC is clearly defined by the absence of three marker expressions, there is no universally accepted profile of basal-like breast cancer.15 Nielsen et al compared transcriptomic and IHC profiles, concluding that a panel that was negative for ER and HER-2/neu, and positive for CK 5/6, and epidermal growth factor receptor (HER-1 or EGFR) could accurately identify basal-like carcinomas.16 Korsching et al included the presence of cytokeratins 14 and 17 in the definition.6 Others have proposed that some basal-like tumors may be positive for ER and/or HER-2/neu amplification.15,17–22
Though some studies claim basal-like tumors and TNBC may be considered synonymous,7,23–25 it has been shown repeatedly that though there is significant overlap between the two, they are not identical.9,26–28 When examining basal-like breast cancers, Bertucci et al found that 77% were TNBC and 23% were not.26 Similarly, when Livasy et al and Kandel et al performed IHC testing of a panel of characteristic markers of basal-like tumors on a group of TNBC tumors, they found that only about 85% were basal-like.27,28 What has become clear is that basal-like carcinoma and TNBC are neither exclusive nor synonymous diseases. Both represent heterogeneous types of breast cancer and further classification studies are underway.
The relationship between basal-like/TNBC, and BRCA-1-related disease is also of great relevance. While probing into the genetic machinery of basal-like disease, it became clear to researchers that the tumors arising preferentially in carriers of the BRCA-1 mutation, especially those diagnosed before the age of 50, bore transcriptomic and IHC profiles that were strikingly similar.29–31 BRCA-1 mutations lead to derangements in repair pathways of double-stranded DNA breaks and though a patient may lack the BRCA-1 somatic mutation, sporadically arising basal-like cancers often display a dysfunctional BRCA-1 pathway.32,33 Both share certain histological features (eg, central necrosis, lymphocytic infiltrate, and genomic instability34 as well as mutations in p53,22,31,35 which disrupt apoptosis and are associated with a poor prognosis.36 As high as 75% of tumors in BRCA-1 carriers are reported to be TNBC, basal-like, or both.15,37 Studies taking a converse approach, looking at patients with TNBC but without a significant familial breast cancer risk, found that 11%–29% of that population under 50 are BRCA-1 mutation carriers.19,38 It has been suggested recently that for women under the age of 50 who are diagnosed with TNBC, BRCA mutation testing is a cost-effective strategy and should be integrated into genetic testing guidelines.39
While distinct from basal-like cancers, claudin-low tumors are triple-negative and are thus considered another subtype of triple-negative disease. This recently discovered claudin-low subtype40,41 takes its name from its low expression of the claudin genes. It lacks epithelial cell junction proteins including E-cadherin, and is marked by intense immune cell infiltrate, stem-cell-like features, and epithelial-mesenchymal transition features. A study looking at tumor-initiating cells in tumor subtypes suggested that claudin-low tumors are enriched with stem cells, presenting the possibility of linking tumor initiating cells with stem cells.42 Tumors arresting at various points in differentiation would have different characteristics and it is thought that claudin-low tumors arrest at the step preceding that of basal-like phenotypes, making it the most primitive cell of cancer cells.43
At the other end of the spectrum lies the normal breast-like group, which can often be mistaken for normal breast tissue. Some studies have questioned the existence of this subtype, though it may be a question of pathological rigor.44 It is a heterogeneous subgroup that includes tumors with high stromal content, those with high lymphocytic infiltration, and those with tumors of low malignant cell content.45 From breast tumors that resemble primitive stem cells to those that closely mimic normal tissue, attempts at characterizing TNBC have only reaffirmed a true heterogeneity exists within the subtype.
It has been well-documented that African-American women are overrepresented in the TNBC group.8,10,11,46–48 A population based study of the California Cancer Registry reported by Bauer et al10 showed that non-Hispanic black women accounted for 10% of all TNBC patients diagnosed and treated in California. They were twice as likely to be diagnosed with TNBC when compared with whites and the incidence of black women with TNBC was more than twice the incidence of black women with other types of breast cancer. Studies by Bauer et al and Carey et al8 also showed a worse 5-year survival rate for black women with late stage TNBC than for other ethnicities.
While most breast cancer cases are associated with increasing age, TNBC has a preferential occurrence in younger/pre-menopausal women.7,8,10,25,49–51 Phipps et al52 and Freedman et al47 found age and menopausal status trended to affect recurrence and survival but neither reached statistical significance in women with TNBC. Demographics, while useful in targeting at-risk populations, may not be particularly prognostic in women already diagnosed with TNBC.
Other clinical associations with TNBC patients have been ventured as well, such as increased parity, young age at first full-term pregnancy (AFFTP), elevated waist-to-hip ratio (WHR), gain of adiposity since childhood, and obesity. Though several studies52,53 found that increased parity did not correlate with TNBC, Millikan et al48 found that women with basal-like carcinoma were more likely to display increased parity in combination with a lack of breast-feeding when compared to women with luminal A disease. The data is also split on the significance of AFFTP: Millikan et al found an association between basal-like disease and younger AFFTP while Phipps et al did not.
Women with TNBC, if premenopausal, were also more likely to be obese when compared to women with other disease subtypes.54–56 Other studies have looked at more specific measurements like WHR and adiposity gain since childhood. Millikan et al48 found positive associations between basal-like breast cancer and an elevated WHR and a gain of adiposity since childhood. Slattery et al57 similarly found that weight gain since age 15 and an elevated WHR were both associated with an increased risk of ER-negative breast cancer. Metabolic syndrome has also been noted to be more prevalent in TNBC patients than those with non-TNBC disease.58 Whether or not any of these clinical associations have a causal effect on developing TNBC has yet to be elucidated.
Basal-like carcinomas and TNBC are most likely to be infiltrating ductal carcinoma of no special type (IDC-NST) 7,13 but metaplastic, atypical or typical medullary, and adenoid cystic cancer, histologies that are usually quite rare, are prevalent in TNBCs.24,28,59 Medullary cancer in particular has been observed to be a subtype that occurs with notable frequency within basal-like populations.11,26,60 Both typical medullary cancer and basal-like carcinomas have an increased rate of p53 mutations35 and share certain genomic alterations (eg, 1q, 8q and X losses), though other alterations are specific to medullary cancer. It may be appropriate to consider medullary breast cancers as an entity within the basal-like spectrum.60
Large tumor size,8,10,26 high histological grade (75%–100% are grade 3),18,28,61 and poor differentiation10 also mark basal-like tumors. EGFR overexpression, though technically not a basal-like breast carcinoma-specific marker, has been found to be present in 44%– 50% of samples16,26,62 and, as Nielsen et al suggested, has a strong enough correlation with basal-like disease to aid in its identification. It has been shown that most tumors that do express c-KIT also express basal-like cytokeratins.16 They often have a high Ki-67, a marker of poor prognosis even though it is associated with a greater chance of chemotherapy response,63 high mitotic index, and marked nuclear pleomorphism.8 High proliferative rate, central necrosis, a pushing border, frequent apoptotic cells, scant stromal content, and stromal lymphocytic response are also often noted.11,19,28,64 As Rakha et al suggested, histological characteristics such as tumor grade, histological subtype, and tumor architecture, in combination with other features such as patient age and tumor size, may aid in the understanding of clinically-identified TNBC.61
As a group, TNBC and basal-like disease is frequently thought of as having poor outcomes (eg, development of distant metastasis, shorter survival, and higher mortality rate) than other disease subtypes.8,10,16,22,31,35,51,65–67 There are, however, data suggesting that prognostic outcome should be discussed in terms of specific subgroups. For instance, lymphnode status may be one qualifier, though its significance has yet to be clearly defined. Carey et al’s study8 found the basal-like subgroup had the poorest breast-cancer specific survival amongst all tumor subtypes in both lymph node-negative and lymph node-positive patients. Van de Rijn et al however, found that in node-negative breast cancer, the expression of CK 15 and/or CK 5/6 was a negative prognostic factor independent of tumor size and tumor grade, though they had no predictive value in node-positive disease.65 Nielsen et al found the presence of basal cytokeratin was associated with poor outcome only in the node-positive group.16
The importance of specificity in terms of TNBC vs. basal-like discussion was highlighted with Liu et al’s study which found that tumors that simultaneously over-express HER-2/neu and basal markers had a significantly worse 5-year overall survival rate than basal-like breast tumors and might require different treatment strategy, suggesting that the poor outcomes associated with basal-like disease may be a function of a variety of factors.68
Subdivisions within the TNBC category by additional marker profiling exist as well. While some studies have shown that the poor prognosis of TNBC is conferred almost entirely by tumors with basal markers,66 Choi et al69 subdivided TNBC into basal-like (ER, PR, HER-2/neu negative, and EGFR and CK 5/6 positive) and quintuple-negative breast cancer (QNBC) (negative for ER, PR, HER-2-neu, CK 5/6, and EGFR). Within the TNBC group, the QNBC group had a worse overall survival (OS) than the basal-like tumors, emphasizing that definition and specificity of nomenclature is important when discussing survival data.
The markers that each study uses to define “basal-like” are also of critical importance. A study by Fulford et al used only CK14 staining to identify basal-like tumors among a sample set of grade III IDC-NST tumors. The authors found that in the five years following diagnosis, those grade III IDC-NST tumors had similar relapse-free survival and OS regardless of CK14 expression, but those expressing CK14 had a better prognosis after 5 years. They suggested that two subgroups may exist within basal carcinomas: one exhibiting early relapse and aggressive clinical course and a separate group that despite the traditionally poor prognostic indicators do not relapse.64 Banerjee et al’s study, which also looked at grade III carcinomas, screened for CK 5/6, CK 14, and CK 17, qualifying a case as basal-like if any one of these markers was found positive. Here, though women with basal-like disease had shorter disease-free survival (DFS) and OS, basal-like status, as an independent prognostic variable, did not reach significance in multivariable analysis.70
A unique pattern of relapse has been observed amongst TNBC: in the first two years following diagnosis, there is a rapid rise in rate of relapse, with a peak within three years, followed by a rapid decline over the next five, and a very low risk of subsequent recurrence.25 The location of relapse also requires some discussion. Whether specifically local-regional relapse (LRR) is higher for basal-like disease than other subtypes is of some debate, with some studies reporting high rates of LRR71 and others failing to find a significantly increased risk of isolated LRR after breast-conserving surgery.47 The pattern of metastatic relapse has been examined in a number of studies, and lung and soft-tissue relapse has been found to be more common than bone relapse or lymph-node metastases.64,71–74 There is also a greater risk of brain metastases, which, along with lung metastases, has been associated with a poorer prognosis.64,73,75,76 Because many studies did not find a relationship between an increase in tumor size and an increase in node-positivity in TNBC disease and because this phenomenon has also been shown to be present in BRCA-associated cancers, it has been hypothesized that basal-like disease may have a hematogenous pattern of spread.72,77
When triple-negative breast cancer is diagnosed in young women, African-American women, women of Jewish descent, and women with a high-risk family history of breast and/or ovarian cancer, BRCA testing should be included as part of the pretreatment assessment. For those who test positive for the BRCA1 and BRCA2 mutation are frequently advised to undergo bilateral mastectomy, especially if they are young. Other than this subset of patients, the considerations for choosing loco-regional treatment for TNBC are the same as for other infiltrating ductal cancers. Breast conservation surgery with postoperative radiation remains to be the choice of local therapy for women with T1 and some T2 TNBCs. Mastectomy is reserved for women with multicentric disease or with persistently involved margins after re-excision. Women with large TNBC may still be candidates for breast conservation surgery as studies such as ours78 have demonstrated the extreme sensitivity of TNBC to neoadjuvant chemotherapy and the significant size reduction of the tumor following neoadjuvant treatment. Mastectomy in general renders radiation unnecessary unless the tumor is 5 cm or larger, margins are involved, or there is nodal metastasis, but Tseng et al suggested that adjuvant radiation in all patients with metaplastic breast cancer may lead to improved overall survival.79
Voduc et al80 suggested that basal-like breast cancer and HER-2 positive breast cancer have the worst 10 year loco-regional survival rate when compared with other molecular subtypes of breast cancer after breast conservation surgery. This finding raises concerns about breast conserving surgery for women with TNBC. However, the same study showed that the 10-year loco-regional recurrence rate after mastectomy was also the highest among basal-like TNBC and HER-2 positive breast cancer. Therefore, the poor relapse-free survival rate observed in these women is more likely to be the result of the biology of TNBC and less likely to be dictated by the type of surgery. In our own analysis of TNBC treatment at UCLA, we found that treatment factors such as lumpectomy, radiation, and negative surgical margins were associated with significantly better relapse-free survival in women with TNBC. Though LRR rate may be higher7,81 and time to recurrence may be shorter in TNBC patients,25 we believe that lumpectomy followed by postoperative adjuvant radiation is an excellent local treatment for many with this disease subtype,82 and we put a strong emphasis on clean surgical margins regardless of the type of surgery chosen.
Though new targeted biologic therapies show promise in many other subtypes of breast cancer, chemotherapy remains the only therapeutic option for patients with TNBC. TNBC’s superior sensitivity and responsiveness to chemotherapy has been well documented and while doxorubicin and taxanes are the classic choices, the most efficacious chemotherapeutic regimen has not yet been clearly established. Recent interest has focused on several classes of chemotherapeutic agents whose mechanisms of action target the unique molecular defects of TNBC.
It is well established that TNBC is prevalent among carriers of BRCA1 and BRCA2 mutations.83–85 The cancers of these women frequently have a defect in homologous recombinant DNA repair, which prevents the repair of double-stranded DNA breaks. A similar derangement has also been seen in sporadic TNBC. It is thought that DNA damaging agents, such as the platinum salts, which bind directly to and cross-link DNA, are likely to lead to an irreversible collapse of DNA repair and achieve the desirable therapeutic result.86 The expression of p63/p73 proteins expressed in about 33% of TNBC patients, might be a potential biomarker indicating platinum sensitivity of the tumor.87
Antitubulin agents can be divided into taxanes (paclitaxel and docetaxel) and non-taxane (vinca alkaloids, ixabepilone, eribulin) drugs. Both work through the stabilization of microtubules; by acting on the spindle, they block the metaphase-anaphase transition and ultimately lead to cell-cycle arrest and apoptosis.
Ixabepilone, a semi-synthetic antineoplastic agent derived from the natural epithilones,88 was designed to have a low susceptibility to mechanisms causing drug resistance,89 holding a theoretical advantage over taxanes by bypassing drug efflux pumps and binding to beta-tubulin in a different manner than taxanes.89–93 Ixabepilone-sensitivity may be correlated with the tumor expression of high beta-III tubulin (a type of tubulin highly expressed in TNBC, basal-like, and HER2+ tumors, and a marker of taxane-resistance)93 and inversely related to ER expression levels.89 Both ixabepilone and eribulin, new non-taxane microtubule dynamics inhibitors, may also have an important role in the treatment of metastatic disease, especially in patients with anthracycline/taxane-resistant metastatic disease.94,95
Agents of this class are a promising targeted therapeutic for TNBC. PARP is an enzyme recruited by either single-stranded or double-stranded DNA breaks (SSB or DSB) for base-excisional repair. Its zinc finger domain binds to the SSB and cleaves off NAD+, which in turn causes the attachment of multiple ADP-ribose units and unwinding of the damaged DNA for repair. Because of the depletion of NAD+, tumor necrosis is frequently seen in tumors with PARP overaction such as TNBC/basal-like breast cancers. The overactive PARP can also increase the release of apoptosis-inducing factor from mitochondria and cause cell death and necrosis.96 Most PARP inhibitors mimic NAD+, thus blocking the binding of NAD+ to the PARP enzyme and inhibiting base-excision repair.
In tumor cells with BRCA1 and BRCA2 deficiencies, the repair of DSB is impaired through deranged homologous recombination repair pathways. Further blockage by PARP1 inhibitors induce SSBs, stalled replication forks, and persistent DSBs ultimately lead to cell-cycle arrest and apoptosis.97 Augmented cell death caused by the repair block of both SSB and DSB is known as synthetic lethality.
Beyond its role in base-excision repair of DNA damage, PARP has also been implicated in other vital functions for cancer growth, such as tumor angiogenesis through the modulation of tumor-released hypoxia-inducible factor and vascular endothelial growth factor.98,99
Given the BRCA1 pathway dysfunction also seen in sporadic TNBC, PARP inhibitors should theoretically be effective not only in the tumors of carriers with BRCA mutations but also in sporadic TNBC as well. Currently, clinical studies are investigating the efficacy of PARP inhibitors in both patient populations while bench research is delving into the mechanisms of tumor growth suppression and predictive markers of response to PARP-inhibitor treatment.
Shown to be elevated in TNBC, vascular endothelial growth factor (VEGF), a key mediator of angiogenesis, may play an important role in the progression of TNBC given this disease subtype’s penchant for high proliferation.100 VEGF stimulated the proliferation and migration of epithelial cells, inhibits apoptosis of endothelial tissue, increases vascular permeability and vasodilation. Bevacizumab (Avastin), the best known anti-angiogenic agent, is a humanized monoclonal antibody (mAb) that binds to VEGF and prevents it from interacting with vascular endothelial cells.101,102 Bevacizumab was shown to have added value when combined with chemotherapy in patients with hormone receptor (HR) negative breast cancer, although as a group the benefits and toxicities of anti-angiogenesis drugs in breast cancer treatment has not been clearly established.
Although EGFR/HER1 is not a specific marker for basal-like breast cancer, its over-expression has been found in 44%–78%16,62 of these tumors and may be an important prognostic marker in long-term survival.51 Similarly, over-expression of EGFR is also found in TNBC23,26,103 and there may be an inverse relationship between estrogen receptor expression and EGFR amplification.62 TNBC cell growth and survival may be supported by signaling via EGFR over-expression and increased ligand levels.
Expression of TIMP-2, an endogenous inhibitor for several ADAM (a disintegrin and metalloproteinase) and matrix metalloproteinase (MMP) family members, inhibits Erb-B ligand and receptor shedding by the tumor and tumor suppression in vivo. In many human tumors, reduced TIMP-3 expression correlated with disease suppression.104 These results suggest ADAM inhibitors INCB7839 (an inhibitor of ADAM 10 and ADAM 17) and TMI-002, an inhibitor specific for ADAM 17, may suppress the downstream signaling from all EGFR family members. Drugs have been developed to target both the extra-cellular domain of EGFR (monoclonal antibodies) and the intracellular domain (tyrosine kinase inhibitors). Clinical trials evaluating cetuximab, a humanized anti-EGFR IgG1 antibody, panitunumab, a full human anti-EGFR antibody, gefitinib and erlotinib, both small molecule tyrosine kinase inhibitors in TNBC are encouraging.
C-src, the cellular homolog of the viral oncogene v-src, is a non-receptor signaling kinase that works downstream of multiple growth factors including platelet-derived growth factor receptor (PDGFR), EGFR, IGF−1. It plays an important role in cancer cell proliferation and invasion through multiple pathways.
Dasatinib is an orally active small molecule inhibitor of both scr and abl proteins. In vitro studies show that dasatinib inhibits growth of “basal-like/triple-negative” breast cancer cell lines both as a single-agent, and also in combination with chemotherapy (namely 5’-5’-DFU or cisplatin).105
The serine-threonine kinase mammalian target of rapamycin (mTOR) promotes protein translation, angiogenesis, proliferation, migration, and metabolism.106 mTOR has two complexes, mTOR complexes 1 and 2 (mTORC1 and mTORC2). The mTORC1 consists of mTOR, mammalian LST8 (mLST8), proline-rich Akt substrate 40 (PRAS 40) and raptor.107 Release of PRAS 40 leads to mTORC1 activation and phosphorylation of eukaryotic initiation factor 4E-binding protein (4E-BP1) and S6 kinase 1 (S6K1). Activation of 4E-BP1 enhances cell proliferation, survival and angiogenesis.108 Phosphorylation of S6K1 leads to many important cellular functions including activation of insulin receptor substrate 1 (IRS-1), eukaryotic initiation factor 4B, cellular apoptosis, eukaryotic elongation factor-2/kinase/mTOR, and glycogen synthase kinase 3.109 Both 4E-BP1 and S6K1 have been associated with cellular transformation and poor prognosis of cancer patients.108,110 The other mTOR complex, mTORC2, consists of mTOR, SIN1, and mLST8, PRR, and rector.111–115 This complex has been shown to activate Akt phosphorylation and has been implicated in cellular migration and apoptosis.111,116
Inhibiting mTOR’s mediated PI3K/Akt signaling pathway abolishes cellular proliferative responses and causes cell cycle arrest. As PI3K/Akt overactivity has been identified in a number of breast cancers,117 rapamycin and its analogs temsirolimus, everolimus, and deforolimus, are undergoing clinical evaluation in TNBC treatment.
Insulin-like growth factor I receptor belongs to a class of tyrosine kinase receptors that contribute to proliferative control, apoptosis, angiogenesis and tumor invasion.118 Expressed in 29%–36% of all TNBC tumors119 has been implicated in the activation of the PI3 K/Akt proliferative pathway in breast cancer.120,122 Preclinical studies in TNBC tumor grafts treated with anti-IGF-IR/InsR dual TKI and chemotherapy have demonstrated complete tumor regression.123 Drugs targeting IGF-1R are of two types: monoclonal antibodies specific for IGF-1R (eg, cixutumumab, ganitumab, figitumumab) and TKIs (linsitinib, XL-228). Drugs of both types are being investigated in treating TNBC.
Preclinical in-vitro studies demonstrated that androgens can induce proliferative changes in breast cancer cell lines and promote tumorigenesis in animal models by androgen receptor stimulation.124 Doane and colleagues examined MDA-MB-453, a cell line with the same biomarker phenotype as TNBC and found that androgen enhanced growth of this cells line was ER-independent and AR-dependent.125 10%–35% of TNBC express androgen receptors,126,127 and it has been suggested that a subset of TNBC cases may benefit from the addition of androgen blockade to their therapy.128 Bicalutamide, a nonsteroidal competitive androgen inhibitor, is used in the treatment of advanced prostate cancer, but until recently, its anticancer effects were not tested in women.
Hsp 90 is a chaperone protein that is widely expressed in breast cancer. It stabilizes client oncogenic proteins and contributes to the survival of tumor cells. In a preclinical study, Caldas-Lopes and colleagues demonstrated that the Hsp 90 inhibitor PU-HTI suppressed TNBC xenograft growth in vivo, showing both partial tumor regression and complete response.129 In vitro, Hsp 90 inhibition has been shown to 1) down-regulate members of the Ras/Raf/MAPK pathway and G 2-M phase to suppress Hsp 90 dependent tumor proliferation, 2) degrade the activated Akt and Bcl-XL, thus inducing apoptosis, and 3) inhibit the activated NF-KB, Akt, ERK2, Tyk2, and PKC, therefore reducing the invasive potential of TNBC. Their findings suggest that Hsp 90 may be an effective and pluripotent target for TNBC therapy.
Until recently, due to a lack of a specific target, systemic treatment options for TNBC were limited to cytotoxic chemotherapy. TNBC, when compared with other phenotypes, were found to have a more favorable outcome after chemotherapy.14 Shorter OS and disease-free intervals have been seen in patients who did not receive adjuvant chemotherapy.9 In addition, TNBC patients are known to have a greater pathologic complete response (pCR) rate when compared with non-TNBC patients.130 But does chemoresponsiveness lead to better overall survival? The NSABP B-18 and B-27 trials, which looked at a combination of neoadjuvant and adjuvant regimens of doxorubicin and cyclophosphamide (AC) with or without docetaxel, found that patients who achieved a pCR continued to have superior DFS and OS when compared with patients who did not.131 However, there exists what is known as the “triple-negative paradox”: while TNBC may be more chemosensitive, the poor prognosis associated with the disease can be explained by the high relapse rate in those patients who are unable to achieve a pCR.132
Many studies examining the timing of chemotherapy in the treatment of breast cancer have found that neoadjuvant therapy is equivalent to adjuvant therapy in OS and disease-free survival. A meta-analysis of nine randomized studies by Mauri et al however, found that neoadjuvant therapy was associated with an increased risk of loco-regional recurrence in patients treated with radiation therapy without surgery.133 As this meta-analysis lacked a subset for TNBC patients, further investigation into the issue of neoadjuvant versus adjuvant therapy for TNBC patients is warranted.
The specific scheduling of chemotherapy may also be important in treating TNBC.130,134 Dose-dense (in which intertreatment intervals are shortened) and/or metronomic scheduling (chronic, low-dose administration of therapy) have been shown not only to improve progression-free survival (PFS), but also increase pCR;135–137 this in turn could mean significantly greater OS, whereby weekly or bi-weekly AC and paclitaxel may greatly benefit TNBC patients. Dose intensification may also improve event-free survival and overall survival in TNBC patients with multiple positive nodes.138
Though trials have yet to demonstrate a clear increase in DFS and OS with neoadjuvant chemotherapy, there is still a clinical advantage given the availability of tissue and the ability to correlate potential biomarkers with pathologic response. More experimental neoadjuvant regimens including platinum salts paired with a taxane and excluding the use of anthracyclines, have shown to achieve high pCR rates in TNBC but choice of drug in this setting has yet to be established.139,140
In the adjuvant setting, anthracyclines and taxanes remain the standard of care for TNBC patients with operable, node-positive breast cancer.141–143 Relative anthracycline sensitivity and taxane-resistance among TNBC patients may hinge on BRCA-1 function. The loss of BRCA-1 is associated with sensitivity to DNAdamaging chemotherapy as well as resistance to spindle poisons, such as taxanes and vinca alkaloids.144 This is relevant not only for carriers of the BRCA-1 mutation but for patients with sporadically-occurring TNBC whose tumors have DNA repair defects similar to BRCA-1 associated tumors; in this population, it has been demonstrated that anthracycline sensitivity and taxane-resistance may be predicted by a BRCA-1 associated expression signature.145 A recent study showed that the classical regimen of cyclophosphamide, methotrexate, and fluorouracil (CMF) had a greater benefit in node-negative TNBC patients than in patients with hormone-receptor positive or HER-2 positive/hormone-receptor negative disease, suggesting CMF may be a good choice for adjuvant therapy in certain populations.146 Currently, there is no standard first line agent to recommend for use in metastatic disease.
The use of platinum salts in the neoadjuvant setting is promising, as TNBC patients undergoing regimens containing platinum salts with or without other agents showed pathological complete response rates ranging from 15%–83%.78,83,84,132 The best partner agents for platinum salts in the adjuvant setting has yet to be determined; regimens combining platinum salts with epirubicin, adriamycin, taxol, and taxotere all showed high pCR rates in TNBC patients.78,147,148 Pairing neoadjuvant cisplatin with bevacizumab did show 15% complete pathologic response in TNBC patients, though toxicity limited completion of therapy in about 10% of patients.149 The tumor response to platinum-based drugs in metastatic TNBC is also being evaluated.85 Mature data from prospective randomized controlled trials, such as NCT00532727, a phase III randomized trial comparing carboplatin and docetaxel as first-line treatment in metastatic and recurrent TNBC, and CALGB 40603, which is testing neoadjuvant carboplatin and taxane therapy in stage II and III TNBC, are not yet available (CALGB NCT00861705). While the role of this class of drug in treating patients with TNBC is being actively pursued, routine use of platinum-containing regimens in patients with early-stage TNBC is not recommended.
The taxanes include paclitaxel and docetaxel and has proven effective in all breast cancer types in both the neoadjuvant and adjuvant setting.78,85 TNBC has shown to have a better response to taxane-containing regimens than to chemotherapy without taxanes142 and to have a significantly better response rate to neoadjuvant taxane treatment.85,150,151 Whether they prove more effective in TNBC patients in the adjuvant setting than other breast cancer subtypes is questionable. Subset analysis from the BCIRG001 trial (docetaxel, doxorubicin, and cyclophosphamide vs. fluorouracil, doxorubicin and cyclophosphamide) found that the benefits of the docetaxel-containing regimen were independent of hormone receptor status.142 Similarly equivocal results between hormone-receptor subgroups were obtained in the NSABP B28 trial, which looked at doxorubicin and cyclophosphamide with or without paclitaxel.152
Its antitumor activity in TNBC has been demonstrated both when used as monotherapy or in combination with capecitabine. When administered as monotherapy, ixabepilone induced a higher pCR in TNBC groups (26%–28%) when compared to non-TNBC patients or to the overall study patient population (15%–18%).89,153–155
Several phase II and III trials have also looked at ixabepilone’s efficacy when paired with capecitabine, a second-line therapy widely used in anthracycline and taxane-resistant disease. Analysis of pooled data from these trials found that overall response rate (ORR) (31 vs. 15%) and PFS (4.2 vs. 1.7 months) were improved in TNBC patients who received combination therapy as opposed to those who received single-agent capecitabine.156
Ongoing trials are examining ixabepilone activity in combination with sunitinib (as first-line therapy in TNBC patients), cetuximab (in metastatic TNBC patients), and in direct comparison to docetaxel and paclitaxel-containing regimens.153 Ixabepilone has been shown to have a manageable safety profile, with neutropenia, sensory neuropathy, fatigue, arthralgias, myalgias, and stomatitis as its main side effects.157
Preclinical data on the mechanisms of PARP inhibitors have led to early phase clinical trials in the targeted treatment of BRCA-deficient breast cancer and TNBC. This class of drug includes olaparib (AZD2281, KU-0059436), iniparib (BSI-201), and veliparib (ABT888). The following PARP inhibitors are being studied in various phases of clinical trials (Table 1).
Olaparib, an oral PARP 1 and PARP2 inhibitor, is active in BRCA-deficient ovarian and breast cancers. In phase I and II studies, single agent olaparib has shown antitumor activity in BRCA-mutation carriers with refractory and/or advanced disease. A greater partial response rate to olaparib has been demonstrated in TNBC patients than in non-TNBC patients (54% vs. 29% respectively).158 Toxicities observed were primarily grade 1 and 2 and were similar to those observed with conventional chemotherapy (fatigue, nausea, vomiting, anemia).159
Encouraging as these results are, it still remains unclear if olaparib is effective outside the BRCA-associated cancer. Canadian study 20, a phase II trial looking at four cohorts of patients with advanced breast or ovarian disease, closed the arm of sporadic TNBC patients as no response to olaparib treatment was seen.160
The efficacy of olaparib in combination with conventional chemotherapy agents has yet to be determined. Concerning toxicity patterns (mainly grade 2–4 neutropenia) resulted when the drug was paired with paclitaxel in the treatment of metastatic TNBC.161 Given pre-clinical data that PARP1 inhibition may potentiate the effects of platinum compounds,162 olaparib is now being tested in combination with carboplatin and cisplatin in TNBC. Safety data from these trials will be important in determining olaparib’s therapeutic place in TNBC.
Iniparib (BSI 201) is a PARP 1 inhibitor administered intravenously. The addition of iniparib to gemcitabine and carboplatin in a phase II study in metastatic TNBC prolonged the median overall survival from 7.7 months to 12.3 month, translating to a 43% reduction in the risk of death (HR = 0.57, P = 0.01). Median PFS in the iniparib group was 5.9 months compared to 3.6 months for the chemotherapy group (HR = 0.59, P = 0.01). No significant difference in adverse events was seen between the groups.163
These promising results paved the way for a phase III study to evaluate OS and PFS in metastatic TNBC (NCT00938652). 519 women with metastatic TNBC were randomized to receive chemotherapy (gemcitabine and carboplatin) with or without iniparib. The study admittedly failed to meet the pre-specified criteria for significance for its co-primary endpoints of OS and PFS. There was, however, an improvement in OS and PFS for patients treated in the second and third-line. The safety analysis indicated that the addition of iniparib did not add to the toxicity profile of gemcitabine and carboplatin [JC, Sanofi-Aventis press release, January 2011]. The use of iniparib in TNBC is currently being tested in the neoadjuvant setting (NCT00813956, NCT01204125), and in the treatment of brain metastases (NCT01173497).
Veliparib (ABT888), an oral PARP 1 and PARP 2 inhibitor, is also being investigated. It has shown to be well tolerated in combination with metronomic cyclophosphamide and to have activity in TNBC.164 Veliparib with temozolomide, an agent found to be synergistic in breast cancer xenograft models, was shown to have activity in patients with metastatic breast cancer.165 Though the preliminary data from this phase II trial did not include a TNBC subgroup analysis, full accrual and final efficacy results are pending.
Whether all TNBC patients will benefit from PARP-inhibitors or if only a portion of TNBC patients, such as BRCA-deficient tumors, will have clinical improvement beyond chemotherapy alone remains to be seen.166 The clinical utility of PARP inhibitors may become better realized if predictive biomarkers can be identified.158
Numerous studies have examined bevacizumab as treatment for metastatic disease and subset analyses suggest that TNBC may have an increased sensitivity to anti-angiogenic agents. Multiple studies looking at the addition of bevacizumab to different chemotherapy agents have shown an increase in PFS in TNBC patients.167,168 Several multicenter randomized trials, including the Cancer and Leukemia Group B (CALGB) 40503 and National Surgical Adjuvant Breast and Bowel Project (NSABP)-B40 studies, hope to gather more data on the effect of bevacizumab on TNBC. However, as Greenberg and Rugo pointed out, all trials to date have used PFS as an endpoint and an improvement in OS has yet to be shown.100 In late 2010, the FDA began the process to remove breast cancer as an indication from the Avastin label not only due to a lack of efficacy, but safety. A 2011 meta-analysis in JAMA highlighted the dangers of the drug, finding that compared with chemotherapy alone, the addition of bevacizumab was associated with an increase risk of fatal adverse events (FAEs), the most common being hemorrhage (23.5%), neutropenia (12.2%), and gastrointestinal tract perforation (7.1%).169 While there were differences in relative risk across tumors types and between drug doses and combinations, it warned of the possible increased risk of FAEs, especially when pairing bevacizumab with taxanes or platinum drugs.
Cetuximab as a single agent appears to have low activity in metastatic TNBC and so recent research has focused on finding the right therapeutic partner for this monoclonal antibody. Cetuximab combined with the platinum salts has seen encouraging results. Carey et al’s study showed little response in the cetuximab alone-group, but patients who received cetuximab with carboplatin had an 18% response rate (CR and PR) and 27% saw clinical benefit (PR or SD > 6 months).170 The BALI-I trial demonstrated a response rate of 20% in the cetuximab plus cisplatin arm, nearly doubled the response rate of cisplatin alone. Overall survival data is still forthcoming and though it failed to reach its primary endpoint (a response rate of more than 20% in the combination arm), the findings reinforce the idea that anti-EGFR agents may have an important role to play in TNBC.171
Adding cetuximab to irinotecan and carboplatin resulted in an increased in ORR in the TNBC subset of O’Shaughnessy’s phase II trial conducted in patients with metastatic disease. A drawback, however, was that the primary toxicity of the irinotecan/carboplatin combination (diarrhea) was exacerbated in patients who received cetuximab.172
Preliminary results of a phase I/II trial of cetuximab in combination with either paclitaxel or docetaxel demonstrated a response (defined as a clinical response, decreased tumor markers, or a decrease in size of metastases) in 9 of 11 patients. The observed toxicity in this combination was the cumulative expected toxicity of each individual agent.173
Patients who suffer infusion reactions (bronchospasm, stridor, urticaria, hypotension, and cardiac arrest) to cetuximab, a chimeric monoclonal antibody, may be treated by panitumumab, a fully human anti-EGFR monoclonal antibody.174 This new agent is currently under clinical investigation in the setting of metastatic TNBC (NCT01009983).
TKIs showed early promise in pre-clinical studies, demonstrating efficacy in treating anti-hormone resistant breast cancer.175 In theory these drugs should be very effective in TNBC, given that the proliferation of these tumors seemed to be EGFR-dependent.176 But clinical studies have not supported the hypothesis; single-agent TKI studies were not impressive in the heavily pre-treated metastatic population nor in the ER(−), EGFR-overexpressing population.177 Instead, the TKIs seemed to be more effective in ER(+), tamoxifen-resistant patients even though the EGFR expression in their tumors tends to be low-tomoderate.178
However, like cetuximab, the key to effective TKI use probably lies in treatment combinations. Gefitinib paired with carboplatin and docetaxel has been shown to be synergistic and enhance response in TNBC cells.179 Inhibitors of ADAM (enzymes involved in the activation of EGFR ligands) may also be potential partners for TKIs in TNBC treatment. Studies testing gefitinib with TMI-002 (a compound that specifically inhibits ADAM-17 in breast cancer cell lines), did not see any additional benefit when the two agents were administered simultaneously; gefitinib treatment administered 72 hours after the ADAM inhibitor, however, was more effective, though the difference did not reach statistical significance. An un-named inhibitor of both ADAM 10 and ADAM 17 has been found to reduce cell growth by 91% in pre-clinical studies and has also been shown to reduce TNBC’s migratory ability [EM, EORTC-NCI-AACR press release, November 2010].
Dasatinib and sunitinib have been tested mostly in patient populations that have been heavily pre-treated. A Phase II trial of single-agent dasatinib in patients with locally advanced or metastatic TNBC and prior anthracycline and/or taxane therapy found only modest activity (clinical benefit rate of 9.3%).180 Candidate genomic markers for dasatinib therapy selection have been identified in breast cancer patients181 and are currently being tested for clinical utility (NCT00780676).
Sunitinib, a TKI that targets the VEGF-associated TK, has been found to elicit response in TNBC patients. One phase II trial given to patients with metastatic disease who had previously been treated with an anthracycline and taxane, found an overall response rate (ORR) of 15% in the TNBC subset.182 But, like bevacizumab, this drug is increasingly thought to be ineffective in breast cancer.183,184 This lack of efficacy may be due to drug’s short half-life and the fact that optimal biologic and therapeutic dosing has yet to be defined. Judicious patient selection may also play a key role in maximizing the efficacy of these anti-angiogenic agents.
Preclinical studies demonstrate that mTOR inhibitors used alone are cytostatic in most tumor types and may clinically stabilize disease.185 Data from clinical studies looking at single-agent everolimus have not been impressive. A phase II study comparing daily dosing with weekly dosing of single-agent everolimus in patients with recurrent/metastatic breast cancer found a low response rate, with no biologic correlates of response despite trends favoring benefit in ER-positive and HER-2 negative breast cancer. However the fairly modest drug-related toxicities encourage drug-combination studies.117,186 Two trials are currently investigating the use of everolimus in the treatment of TNBC (NCT01272141 is looking at the combination of lapatinib and everolimus in locally advanced or metastatic TNBC, and NCT00827567 is examining the use of single agent everolimus in metastatic TNBC). As with all other targeted therapy, markers of mTOR treatment response will prove paramount in patient selection. Patients with cancer showing decreased PTEN, activated PI3K activation, or high p-mTOR have been reported to benefit the most from this class of drugs187–190 but work in this area must progress.
Multiple phase I studies have found multiple humanized mAbs and TKIs to be safe and tolerable in patients with solid tumors.191–194 Data from a tissue study by Witkiewicz and colleagues demonstrated that IGF-1R is overexpressed and amplified in 29% of their TNBC samples.195 High IGF-1R expression was significantly correlated with negative lymph nodes and, in patients younger than 55 years of age, with longer survival. The IGF-1R/insulin receptor tyrosine kinase domain inhibitor BMS-754807 has demonstrated activity in TNBC123 and ongoing trials are evaluating the efficacy of this class of targeted treatment in breast cancer.
NCT00468715 is an ongoing study evaluating the use of bicalutamide, an anti-androgen agent used for treatment of prostate cancer in the treatment of HR-negative, AR-positive breast cancer. Bicalutamide has been well-tolerated in this population and preliminary analysis has demonstrated disease stabilization in ER/PR negative, AR positive with AR inhibition.196
Clinical studies are evaluating Hsp 90 inhibitor AUY922 and IPI-504, but only in ER and HER2 positive disease (NCT0181613 and NCT01081600). Whether agents of this class will prove effective in vivo and in TNBC specifically, remains to be seen.
The tumor biology of TNBC, basal-like breast cancer, BRCA-mutated machinery, and claudin-low disease is both specific and diverse. While conventional chemotherapeutic regimens can be successful in treating women with TNBC and basal-like disease, it is clear that this pool of diseases is heterogeneous in nature and must be further sub-categorized. Emerging therapies aimed at damaging DNA, angiogenic players, tubulin structures, mTOR, IGF-1R, AR, and HSP 90 show promise in early stage studies, but their clinical performance has yet to be definitively proven. No doubt much of the work to come must focus on generating more specific terminology in order to identify the optimal patient population for each treatment.
This manuscript has been read and approved by all authors. This paper is unique and is not under consideration by any other publication and has not been published elsewhere. The authors and peer reviewers of this paper report no conflicts of interest. The authors confirm that they have permission to reproduce any copyrighted material.