Lack of correlation between markers of breast cancer initiating cells

The existence of breast cancer-initiating cells was initially demonstrated by Al-Hajj and colleagues [1] using antigen expression, and subsequent studies have employed several methodologies to identify and isolate these cells. However, there are limited data describing whether similar cell populations are recognized by the different approaches.

Using breast cancer cell lines MCF7, MDAMB231 and MDAMB468, we have compared the antigen expression profile (CD44⁺CD24^-/low) against the side population and the ability to form tumour spheroids. Immunostaining on cells and xenografts was also performed to search for expression of potential stem cell markers.

Our data showed increased CD44⁺CD24^-/low population in both MCF7 and MDAMB468 spheroids, but growth advantage was only observed in sorted MDAMB468 CD44⁺CD24^-/low cells. In contrast, analysis of the antigen profile of the side population did not demonstrate any correlation and no growth advantage was found in sorted MCF7 and MDAMB468 cells. Immunostaining of MCF7-derived tumour xenografts showed two potential markers, p63 and sox2, in addition to CD44; both MDAMB231 and 468-derived xenograft expressed strong CD44, and the latter was also stained for p63 and aldehyde dehydrogenase (ALDH). In addition, comparison between the antibodies only demonstrated partial overlap between CD44 and p63/ALDH in MCF7 and MDAMB468 xenografts. Therefore, in MCF7/MDAMB468-like breast tumours, p63 and sox2/ALDH recognize different stem/progenitor cell populations, and the combination of CD44 and p63/ALDH further clarifies the boundary of these cells.

Our results indicate that each breast cancer is unique, and therefore tumour-initiating cell markers and methodologies should be applied specifically.