Heterogeneity in breast cancer (original) (raw)

Breast cancer is one of the few tumor types in which molecular classification has successfully been used for the design of individualized therapies, leading to significant improvements in disease-specific survival (5). Based on comprehensive gene expression profiling, breast tumors are classified into at least three major subtypes: luminal, human epidermal growth factor receptor 2+ (HER2+), and basal like (6, 7). Each of these subtypes has different risk factors for incidence, response to treatment, risk of disease progression, and preferential organ sites of metastases. Luminal tumors are positive for estrogen and progesterone receptors, and the majority respond well to hormonal interventions, whereas HER2+ tumors have amplification and overexpression of the ERBB2 oncogene and can be effectively controlled with a diverse array of anti-HER2 therapies. Basal-like tumors in general lack hormone receptors and HER2; thus, the majority of these tumors are also called triple-negative breast cancer (TNBC). Currently there is no molecular-based targeted therapy for TNBC, and unfortunately only approximately 20% of these tumors respond well to standard chemotherapy. Thus, developing improved treatments for TNBC is one of the highest priorities of current breast cancer research. Numerous agents are in various phases of clinical development, including several different poly(ADP-ribose) polymerase inhibitors, JAK kinase, and EGFR inhibitors as well as “revived” classical chemotherapeutic agents such as platinum salts. However, thus far none of these shows promise for treating all TNBCs, a finding that perhaps is not surprising given that several recent studies have described that even this relatively small class of breast tumors can be further divided into five or six subclasses, each with its own molecular features and unique sensitivity to therapeutic agents (8–10).

Several hypotheses have been proposed to explain the origin of intertumor heterogeneity in breast cancer, including subtype-specific tumor cell–of–origin and transforming events (11). Accordingly, luminal and HER2+ tumors may originate from luminal lineage-committed progenitors, whereas basal-like cases arise from less differentiated stem cell–like cells. However, gene expression patterns and experimental evidence in model systems imply that luminal progenitors may also serve as precursors to basal-like tumors following genetic or epigenetic event(s) that switch cellular phenotypes (12–14). For example, loss of BRCA1 or PTEN in luminal epithelial cells leads to loss of luminal differentiation, and the oncogenic transformation of these cells results in the formation of basal-like tumors (15). However, because we know that not all ER+/HER2+, or basal-like tumors, are the same, it is likely that there are multiple ways to develop each of these tumor types.

Defining the cell-of-origin and evolutionary pathway of a breast cancer in humans is a nearly impossible task, as we are rarely able to diagnose tumors at their earliest stages and follow their molecular evolution. Currently three main approaches have been used to trace the evolutionary history of human cancer. One method is to analyze tumors at the single-cell level for phenotypic traits and somatic genetic alterations, based on the assumption that some cells may be relics of the tumor’s past, and that their frequency within a tumor may reveal the probable steps of the tumor’s evolution (16). Another approach analyzes a large collection of tumors at different progression stages for molecular changes, and based on the frequency at which these are detected at a specific progression step, their probable order of events can be assembled (17, 18). Thus far these approaches have been used to map colon (19) and pancreatic cancer (20) evolution, but similar studies have not been conducted in breast tumors. Of course both of these methods are based on certain assumptions, such as frequency of somatic genetic alterations as they relate to their order, though this may be in question given a recent study describing the acquisition of several somatic genetic changes by tumor cells due to a single catastrophic rearrangement of chromosomes (21). Defining the cell-of-origin of cancer was thought to be relevant only for risk prediction and chemoprevention studies. However, if the cell-of-origin has a major influence on the evolutionary path of a tumor as it relates to the identity and frequency of acquired transforming events, then its characterization would be important in better understanding breast tumor subtypes as well.