This discrepancy may be due to different subtypes of breast cancers and different percentages of samples from primary and metastatic breast tumors. Although CD44+/CD24- percentage was not associated with ER or HER2 expression, we observed an association between high CD44+/CD24- percentage and PR expression. This linkage was more prominent in samples from

recurrent and metastatic tumors with more than 25% CD44+/CD24- cells. In contrast, previous studies showed that the presence of CD44+/CD24- tumor cells was not associated with ER or PR status [20]. CD44+/CD24- cells have been observed in 63% of basal-like subtype (SR-HER2- basal-like) breast tumors.[20] Although we did not observe a significant difference in the proportion of CD44+/CD24- selleck kinase inhibitor cells in samples from tumors with and without basal-like features, we found that the CD44+/CD24- subpopulation was higher in samples of recurrent and metastatic tumors with basal-like features. Several studies have shown an association between CD44+/CD24-

cells and the metastasis of basal-like breast cancers. For example, the expression of several metastasis-associated genes was found to be higher in cells with than without the CD44+/CD24- phenotype, and only malignant cell lines with the CD44+/CD24- subpopulation were able to invade matrigel, indicating that CD44+/CD24- cancer cells are more metastatic than non-CD44+/CD24- cells [21, 22]. Importantly, a unique 186-gene invasiveness gene signature has been observed in CD44+/CD24- FHPI supplier malignant cells,[22] linking the presence

of CD44+/CD24- cells to distant metastasis although not to survival.[8, 23] We found that the time to tumor relapse (including recurrence and metastasis) was significantly shorter in patients with than without CD44+/CD24- tumor cells. Metastasis is a complex process involving invasion, intravasation, survival in the blood stream, extravasation and homing and proliferation at the sites of metastasis.[8, 24, 25] The poor prognosis of patients with Tolmetin primary tumors having higher levels of CD44+/CD24- cells, but whose metastatic cells had the CD44±/CD24+ phenotype,[26, 27] suggests that CD44+/CD24- tumor cells may be a transient mTOR inhibitor phenotype and that these cells have an intrinsic program to transition to a phenotype that enhances their heterotypic interaction and survival/proliferation in distant organs.[8] This hypothesis, however, cannot explain the difference in time to tumor relapse in patients with and without CD44+/CD24- cancer cells who had undergone surgical resection plus immunotherapy. Conclusion We observed variations in the prevalence of CD44+/CD24- tumor cells in breast tumors of different subtypes. This phenotype was highly prevalent in primary tumors with high PR expression and in secondary tumors.