Molecular and cellular heterogeneity in breast cancer: challenges for personalized medicine

Abstract

Breast cancer is noted for disparate clinical behaviors and patient outcomes, despite common histopathological features at diagnosis. Molecular pathogenesis studies suggest that breast cancer is a collection of diseases with variable molecular underpinnings that modulate therapeutic responses, disease-free intervals, and long-term survival. Traditional therapeutic strategies for individual patients are guided by the expression status of the estrogen and progesterone receptors (ER and PR) and human epidermal growth factor receptor 2 (HER2). Although such methods for clinical classification have utility in selection of targeted therapies, short-term patient responses and long-term survival remain difficult to predict. Molecular signatures of breast cancer based on complex gene expression patterns have utility in prediction of long-term patient outcomes, but are not yet used for guiding therapy. Examination of the correspondence between these methods for breast cancer classification reveals a lack of agreement affecting a significant percentage of cases. To realize true personalized breast cancer therapy, a more complete analysis and evaluation of the molecular characteristics of the disease in the individual patient is required, together with an understanding of the contributions of specific genetic and epigenetic alterations (and their combinations) to management of the patient. Here, we discuss the molecular and cellular heterogeneity of breast cancer, the impact of this heterogeneity on practical breast cancer classification, and the challenges for personalized breast cancer treatment.

Figure 1

Natural history of breast cancer development. Breast cancer develops from normal breast epithelial cells that evolve through atypical hyperplasia (and eventually dysplasia), DCIS, and invasive breast cancer. Multiple molecular alterations occur during this process, involving genetic and epigenetic alterations in precursor and neoplastic cells. Genetic predisposition can contribute to this process, but early molecular alterations (preceding DCIS) have not been well characterized. Original magnification, ×20.

Figure 2

Clinical classification of invasive breast cancer based on expression of ER, PR, and HER2. Representative examples of invasive breast cancers that correspond to the general clinical classifications are shown. Cancer histology is depicted using H&E staining; expression of ER, PR, and HER2 is visualized using immunohistochemistry. Breast cancers are generally classified as positive or negative for hormone receptors ER and PR and for HER2, resulting in four major clinical groupings: ER⁺/PR⁺/HER2⁻, ER⁺/PR⁺/HER2⁺, ER⁻/PR⁻/HER2⁺, and the triple-negative ER⁻/PR⁻/HER2⁻. Original magnification, ×40.

Figure 3

Correspondence between immunohistochemical and molecular classification of breast cancer. This analysis reflects data from a cohort of breast cancers for which complete immunohistochemical staining results for ER, PR, and HER2 were known (no missing values) and for which molecular classification had been performed based on gene expression patterns. The 381 breast cancers in this cohort were 205 ER⁺/PR⁺/HER2⁻, 33 ER⁺/PR⁺/HER2⁺, 34 ER⁻/PR⁻/HER2⁺, and 109 ER⁻/PR⁻/HER2⁻ based on immunohistochemical staining and 111 luminal A, 60 luminal B, 57 HER2⁺, 80 basal-like, 43 claudin-low, and 30 normal-like based on gene expression analysis.

Figure 4

Expression of ER among molecular subtypes of breast cancer. This analysis reflects data from a cohort of breast cancers for which ER status based on immunohistochemical staining was known and molecular classification had been performed based on gene expression patterns. The 804 breast cancers in this cohort were 548 ER⁺ and 256 ER⁻ breast cancers based on immunohistochemical staining and 195 luminal A, 162 luminal B, 144 HER2⁺, 138 basal-like, 89 claudin-low, and 76 normal-like based on gene expression analysis.

Figure 5

Expression of HER2 among molecular subtypes of breast cancer. This analysis reflects data from a cohort of breast cancers for which HER2 status based on immunohistochemical staining was known and molecular classification had been performed based on gene expression patterns. The 498 breast cancers in this cohort were 101 HER2⁺ and 397 HER2⁻ breast cancers based on immunohistochemical staining and 128 luminal A, 90 luminal B, 92 HER2⁺, 96 basal-like, 53 claudin-low, and 39 normal-like based on gene expression analysis.

Figure 6

Heterogeneity of HER2 expression in a single invasive breast cancer. Heterogeneity for HER2 positivity but uniform positivity for ER expression is observed in a single cancer in a single patient. A: H&E staining of an ER⁺/HER2⁺ region of an invasive breast cancer. B and C: This region of the cancer demonstrates strong (3⁺ positivity) HER2 expression (HER2⁺) by immunostaining (B) and uniform strong ER expression (C). D: H&E staining of an ER⁺/HER2⁻ region of the same invasive breast cancer. E and F: This region of the cancer demonstrates weak (1⁺ positivity) HER2 expression (HER2⁻) by immunostaining (E) and uniform strong ER expression (F). Original magnification: ×10 (A and D); ×40 (B, C, E, and F).