Since the late 1970s, clinicians have distinguished breast cancers types according to the presence or absence of certain receptors that sit on the surface of these tumor cells. Depending on the receptors found—namely, the estrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2)—a doctor can get a better sense of the prognosis and which treatments might work.
Now, a study published this week in the Journal of Clinical Investigation proposes a new conceptual framework that classifies breast cancers based on whether the cells possess receptors for other molecules, such as androgen and vitamin D. Under this system, current cancer subtypes would be stratified into one of four groups.
“I’m very excited about this paper,” says Jorge Reis-Filho, a pathologist at the Memorial Sloan Kettering Cancer Center in New York who was not involved in the research. He adds that the proposed classification system could point to new therapies for breast cancers previously categorized as unlikely to respond to treatment.
In the new study, researchers sought to explain some of the diversity observed in human breast tumors by obtaining a more detailed understanding of normal breast cell subtypes. “I approached this question like an evolutionary biologist trying to figure out the ancestry of a species,” says study co-author Tan Ince, a pathologist at the University of Miami Miller School of Medicine.
Ince and his team scanned samples of normal breast tissue for proteins expressed in a “bimodal” or on/off pattern—highly expressed in some cells but completely absent in others. The researchers focused in on the handful of proteins that displayed this pattern. They subsequently characterized 11 previously undescribed subtypes of luminal cells—one of the two major epithelial cell types found in mammary glands—that each expressed distinct combinations of these proteins.
Using the expression patterns of just three such proteins—the receptors for estrogen, androgen and vitamin D—the researchers grouped the 11 cell subtypes into four categories that were dubbed HR0, HR1, HR2 and HR3 based on how many of the hormone receptor types the cells contained.
Ultimately, the researchers then used their classification system to characterize more than 3,000 human breast tumors. To their surprise, Ince says, 95% of breast tumors displayed protein expression patterns that resembled one of the newly defined breast cell subtypes. Moreover, the categories that the tumors belonged to were associated with significantly different prognostic outcomes: for example, patients whose tumors belonged to the HR0 category had an almost threefold greater risk of death within the first five years of diagnosis compared with those who had HR3 cancers. The classification scheme also provided some insights into the treatment of each tumor type. For example, the authors showed that compounds that bind the vitamin D receptor effectively blocked the growth of cell lines derived from triple negative breast cancers—those that lack ER, PR and HER2 and therefore have few effective treatments.
“The current system is very well established—changing that is not something that is really in the grand scheme,” says Sandro Santagata, a pathologist at Brigham and Women’s Hospital and Harvard Medical School in Boston and lead author of the study. “We’re just hoping to add an additional layer of information.”
Lajos Pusztai, an oncologist at the Yale School of Medicine in New Haven, Connecticut, called the findings “thought provoking” and likely to motivate further research, but noted a few caveats that limit their clinical implication. “The new classification put forward by this paper includes markers that are not standardized,” he says. Furthermore, the prognostic ability of the new HR categories is not substantially better than the existing methods for predicting disease recurrence, he adds.
Other scientists agree that the classification system is still not ready for primetime. “Additional work is required to see how we can translate [this classification system] into something that we can use in the clinic,” says Reis-Filho. “The task of translating these findings into assays that we could run in pathology departments across the country is not a trivial one.”