Tailored therapies based upon tumor subtype biology

Breast cancer is a spectrum of diseases comprised of different tumor subtypes, each with a distinct biology and clinical behavior. To capture this diversify, we characterized the variation in gene expression across human breast tumors using DNA microarrays and identified at least five distinct tumor subtypes that are statistically significant predictors of patient overall survival [1]. Recently, we further validated these findings using a training set of 102 tumors, which was used to derive a new 'intrinsic gene set'. This gene set was then validated using a true test set of 311 tumors compiled from three different microarray studies. Our analyses demonstrate that common patterns of gene expression can be identified across different microarray platforms, that the breast tumor 'intrinsic' subtypes are reproducible across different datasets, and that this classification was a significant predictor of outcomes after correcting for standard clinical parameters such as estrogen receptor (ER), grade and node status [2].

The biology of the 'intrinsic' subtypes is rich and extensive, and many of these expression features suggest distinct therapies. The 'intrinsic' subtypes include at least two types of ER-negative tumors (Basal-like and HER2⁺/ER^-) and at least two types of ER-positive tumors (Luminal A and Luminal B). Basal-like tumors typically show low expression of HER2 and ER, and these tumors exhibit high expression of genes characteristic of the basal epithelial cell layer, including expression of keratin 5, keratin 6, keratin 17 and four Kallikrein genes (KLK5-KLK8). The Basal-like tumors pose a challenge from the treatment perspective because they lack ER and HER2. However, we have recently shown that most are HER1-positive and/or c-KIT-positive [3], and we have initiated a clinical trial to evaluate the efficacy of HER1-inhibitors in preselected Basal-like tumor patients.

HER2-positive (i.e. gene amplified) tumors fall into at least two distinct expression groups: those that are ER-negative and typically cluster near the Basal-like tumors (HER2⁺/ER^-), and those that are ER-positive and cluster with tumors of luminal cell origin. These findings suggest that both types of HER2⁺ patients should receive transtuzumab, but that the ER⁺/HER2⁺ may gain a benefit from hormone therapy.

Finally, the Luminal subtype A and Luminal subtype B tumors express ER, GATA3, and genes regulated by both ER and GATA3. Compared with Luminal B tumors, Luminal A tumors express higher levels of ER, BCL2 and GATA3, and they show more favorable patient outcomes. Luminal B tumors more often express HER1, HER2 and/or cyclin E₁, and they show worse outcomes. Our data, when coupled with data from others [4], suggests that Luminal A patients are likely to benefit from hormone therapy and are not likely to benefit from chemotherapy, while the opposite may be true of Luminal B patients. Experiments to answer these questions in Luminal patients are underway and will be discussed.