ABSTRACT: Breast cancer brain metastases (BCBMs) are the second most frequent secondary central nervous system metastases following those associated with non–small-cell lung cancer. It is increasingly evident that BCBM arises as a function of the biology of the primary tumor and the metastatic niche, which combine to create a unique microenvironment in the brain impacting both metastatic colonization and therapeutic response. Clinical outcomes are improving for BCBM patients as a result of modern combinatorial therapies, challenging the traditionally nihilistic approach to this patient subgroup. This review will focus on the breast cancer subtypes with the highest incidence of BCBM—human epidermal growth factor receptor 2 (HER2)-positive breast cancer, and triple-negative (estrogen receptor [ER]-negative, progesterone(Drug information on progesterone) receptor [PR]-negative, and HER2-negative) breast cancer (TNBC)—and will characterize differences in the clinical behavior of brain metastases that arise from these different subtypes. We will also highlight some of the recent preclinical studies that may shed light on the biological mechanisms and mediators underlying brain metastases. Finally, we will review published and current prospective trials of systemic therapies specifically for BCBM, including novel pathway-specific therapies.
The Changing Landscape of Breast Cancer Brain Metastases
The diagnosis of central nervous system (CNS) recurrence is a much dreaded outcome among breast cancer patients, and its incidence varies with disease stage and cancer subtype. While less common than bony, lung, or liver metastases, breast cancer brain metastases (BCBMs) are associated with the shortest survival time once diagnosed. BCBMs are also the second most frequent secondary CNS metastases following those associated with non–small-cell lung cancer. BCBMs are typically multifocal and intracerebral, and less commonly solitary and leptomeningeal.[2,3] It is increasingly evident that BCBM arises as a function of the biology of the primary tumor and the metastatic niche; the latter is comprised of the blood-brain barrier (BBB), pericytes, astrocytes, and glial cells, which combine to create a unique microenvironment in the brain that impacts both metastatic colonization and therapeutic response. Improvements in systemic therapy have altered the natural history of breast cancer, and BCBM occurs in a significant proportion of patients. However, patients with BCBM are excluded from many clinical trials, despite the urgent need to develop treatments for this critical challenge. In this article, we will compare the clinical behavior of BCBM associated with the various breast cancer subtypes, with a focus on the subtypes with the highest incidence of BCBM, human epidermal growth factor receptor 2 (HER2)-positive breast cancer and triple-negative (estrogen receptor [ER]-negative, progesterone receptor [PR]-negative, and HER2-negative) breast cancer (TNBC). We will also review therapies and research strategies currently in use and in development.
Differences between the molecular subtypes of breast cancer extend beyond prognosis and therapeutic response to include clinical behavior and patterns of metastatic spread (Table 1). In a study of 3726 patients with early-stage breast cancer diagnosed between 1986 and 1992, the 15-year cumulative incidence rates of brain metastasis were highest in the HER2-positive and TNBC subtypes and lowest in the luminal subtypes. Similar subtype distributions have also been seen in the metastatic setting, where the incidence of BCBM in the HER2-positive and TNBC subtypes is 15% to 44% and 25% to 46%, respectively, with ER negativity and a higher disease burden predicting for a higher BCBM risk in HER2-positive breast cancer.[1,2,4-6] In a separate analysis of 213 patients with BCBM, the median brain metastasis–free survival (BMFS), defined as the time from the diagnosis of extracranial metastases to the time of BCBM, was 34, 18, and 12 months in the ER-positive, HER2-positive, and TNBC subtypes, respectively. In the HER2-positive subgroup, ER co-expression resulted in a prolonged BMFS (26 vs 15 months without ER co-expression). In our series of patients with HER2-positive metastatic breast cancer diagnosed between 1999 and 2005, 8% of patients had BCBM at the time of first metastatic diagnosis, and this increased to 55% with BCBM by the time of death or last follow-up. The BMFS and median survival from the time of BCBM diagnosis were 1.3 and 1.5 years, respectively (Olson EM et al, manuscript in preparation). The variability in the reported median time to CNS progression following diagnosis of HER2-positive metastatic disease may be explained in part by the variable use of HER2-directed therapy in the different studies. Control of extracranial disease at the time of diagnosis of HER2-positive BCBM is approximately 50%, in contrast to TNBC, in which such control is uncommon. Consequently, the cause of death in patients with triple-negative BCBM is rarely due to progressive CNS disease alone, in contrast to HER2-positive BCBM, a setting in which up to 50% of patients die of progressive CNS disease. Concordant with this observation is the longer median overall survival from time of BCBM diagnosis that is seen in the HER2-positive subtype compared with the TNBC subtype (1 to 2 years vs 3 to 5 months).[2-5,8,9] Finally, in a multi-institutional retrospective analysis of patients with newly diagnosed CNS metastases from various primary sites, including 400 patients with breast cancer, univariate and multivariate analyses of prognostic factors associated with outcomes revealed that performance status and tumor subtype were the primary determinants of outcome in BCBM. In contrast, the number of CNS metastases and presence of extracranial metastases were not among the main determinants. Based on the results of their analysis, the authors put forth a disease-specific graded prognostic index for brain metastases, and subsequently refined it further for patients with BCBM; the median survival in the best-graded BCBM group in this index was approximately 25 months.
The natural course of BCBM is thus strongly influenced by the biology of the primary tumor subtype. In HER2-positive disease, some patients experience extended survival relative to historical estimates due to effective targeted therapies in the extracranial setting; multiple lines of CNS-directed therapies are therefore needed, and there is an urgent need to develop therapies that are effective in the HER2-positive CNS. In contrast, patients with TNBC have relatively poorer outcomes, and both CNS and extracranial disease contribute to the shorter survival; thus, there is a need for better systemic treatments targeting all metastatic sites.
Biology of Breast Cancer Brain Metastases
The biological basis of BCBM is largely unknown. Distant metastases in breast cancer have been postulated as an early event, following which the disseminated tumor cells enter a state of proliferative dormancy. The period of metastatic latency is defined by the temporal gap between metastatic infiltration and the acquisition of colonization competency in a distant organ—a competency that may result from progressive malignant evolution of the
tumor and changes in the tumor microenvironment.
The characterization of the molecular subtypes of breast cancer was a major advance in our understanding of the heterogeneity of breast cancer biology. These studies of breast cancer heterogeneity have so far yielded little insight into the different patterns of development of distant metastases across subtypes. One conceptual model classifies genes that underlie the metastatic process into “metastases initiation genes,” which provide transformed cells with the ability to invade and enter the circulatory pathway, and “progression and virulence genes,” which determine the ability to infiltrate and colonize distant organs. Differences in the expression of these genes may underlie the different frequencies of and time to BCBM seen in the different breast cancer subtypes.
Recent breakthroughs in efforts to characterize organ-specific mediators of BCBM include results from gene-expression analyses of BCBM patient samples, which identified cyclooxygenase (COX)-2, epidermal growth factor receptor (EGFR) ligands, and ST6GALNAC5 (a sialyltransferase whose expression is normally restricted to the CNS), as mediators of cancer cell passage through the BBB. In contrast to COX-2 and EGFR, which are also linked to other organ metastases, aberrant expression of ST6GALNAC5 specifically mediated BCBM, potentially by providing a means of enhancing its adhesion to the CNS endothelium and its passage through the BBB. In another study comparing 47 brain metastases and 165 primary breast cancer specimens, the authors found that KISS1 protein, a known metastasis suppressor, was downregulated in metastases compared with primary tumors, and was a prognostic marker for increased risk of breast cancer progression. Finally, an understanding is also emerging of the role of the CNS stroma in the metastatic niche: it appears that reactive glial cells are recruited by highly proliferative CNS metastases, promoting metastatic cell colonization in vivo and supporting tumor cell growth in vitro.
A major hurdle in preclinical brain metastasis research has been the development of suitable models that accurately reflect the biology of BCBM in patients. A unique feature of the CNS is the presence of the BBB, a tight layer of endothelial cells and adipocyte foot processes that acts as a selective barrier to both tumor cell infiltration and the diffusion of systemic therapies, in contrast to the more porous fenestrated endothelia found in bone marrow sinusoids and the liver, which are more readily traversed by circulating tumor cells. The BBB is also characterized by drug efflux mechanisms; for example, preclinical studies in mice have identified a synergistic role between P-glycoprotein and breast cancer resistance protein in modulating the CNS penetration of lapatinib (Tykerb), a small molecule tyrosine kinase inhibitor, under steady-state conditions. There are limited data comparing the concentration of antineoplastic agents in the CNS and in tumor tissue, and there is considerable heterogeneity in the methods by which these data were obtained. A recent overview found CNS-to-blood ratios that, in CNS tumors, were lowest for carboplatin and temozolomide (Temodar), intermediate for paclitaxel and lapatinib, and highest for mitoxantrone(Drug information on mitoxantrone). Despite these findings, temozolomide(Drug information on temozolomide) is one of the most effective agents for the treatment of primary CNS tumors; therefore, penetration of the BBB is apparently not always necessary for CNS activity, and the therapeutic effect is also dependent on other properties of the drug and the inherent sensitivity of the tumor. While intrathecal drug administration may represent a potential direct passage into the CNS, this has not routinely been done in the setting of solid malignancies due to the need for reformulation and testing, and the concern that only superficial lesions would be exposed to sufficient drug levels.