Breast Cancer Stem Cells Mutations: A New Understanding Of Intratumoral Heterogeneity.

25 Background: Breast cancer has long been recognized as a heterogeneous disease. This has profound implications for diagnosis, treatment and disease recurrence. Oncogenic mutations have been identified in breast cancer cells with stem-like and progenitor properties (BCSC). We have previously reported that BCSC mutations correlated with axillary lymph node metastases. This was even more significant when micrometastatic disease was included. Our hypothesis is that tumor heterogeneity extends to the genetics of BCSC, and that BCSC mutations are better predictors of lymph node status than whole tumor genetics.BCSC from fresh tissue specimens were matched to their whole tumor specimens. BCSC and whole tumor DNA were sent for PCR-based mutation analysis. Patient data was collected by chart review.Twenty-eight matched BCSC and whole tumor samples were analyzed. PI3K/Akt signaling mutations in PIK3CA, AKT1, HRAS, and MET were identified in BCSC from 10 tumors. In 4 of these, mutations were also identified in the corresponding whole tumor specimens. In 4 patients, mutations were identified in whole tumor samples only. Fourteen tumors had no mutations. Tumor stage, grade, receptor status, and age did not correlate with tumor or BCSC mutation status. In contrast to BCSC mutations, mutation status of the whole tumor did not correlate with micro or macro metastatic disease in the lymph node (p = 0.92).Mutations in BCSC are more predictive of lymph node metastases than mutations identified in the tumors. Thus, PI3K/Akt pathway mutations in tumor precursor cells may have a stronger influence on tumor metastatic potential than mutations identified in whole tumor samples. Whole tumors and BCSC populations demonstrate significant heterogeneity, as mutations identified in BCSC and tumors were not always concordant. Rare BCSC populations must be tested separately as they provide crucial prognostic and treatment information in conjunction with whole tumor genetic analyses.