The purpose of this study was to classify breast carcinomas based on variations in gene expression patterns derived from cDNA microarrays and to correlate tumor characteristics to clinical outcome. A total of 85 cDNA microarray experiments representing 78 cancers, three fibroadenomas, and four normal breast tissues were analyzed by hierarchical clustering. As reported previously, the cancers could be classified into a basal epithelial-like group, an ERBB2-overexpressing group and a normal breast-like group based on variations in gene expression. A novel finding was that the previously characterized luminal epithelial/estrogen receptor-positive group could be divided into at least two subgroups, each with a distinctive expression profile. These subtypes proved to be reasonably robust by clustering using two different gene sets: first, a set of 456 cDNA clones previously selected to reflect intrinsic properties of the tumors and, second, a gene set that highly correlated with patient outcome. Survival analyses on a subcohort of patients with locally advanced breast cancer uniformly treated in a prospective study showed significantly different outcomes for the patients belonging to the various groups, including a poor prognosis for the basal-like subtype and a significant difference in outcome for the two estrogen receptor-positive groups.
TP53 status [mutations, immunostaining, and loss of heterozygosity (LOH)], expression of c-erbB-2, bcl-2, and histological grading were correlated to the response to doxorubicin monotherapy (14 mg/m2) administered weekly to 90 patients with locally advanced breast cancer. Mutations in the TP53 gene, in particular those affecting or disrupting the loop domains L2 or L3 of the p53 protein, were associated with lack of response to chemotherapy (P = 0.063 for all mutations and P = 0.008 for mutations affecting L2/L3, respectively). Similarly, expression of c-erbB-2 (P = 0.041), a high histological grade (P = 0.023), and lack of expression of bcl-2 (P = 0.018) all predicted chemoresistance. No statistically significant association between either p53 immunostaining or TP53 LOH and response to therapy was recorded, despite the finding that both were associated with TP53 mutation status (p53 immunostaining, P < 0.001; LOH, P = 0.021). Lack of immunostaining for p53 despite mutation of the TP53 gene was particularly seen in tumors harboring nonsense mutations or deletions/splices (7 of 10 negative for staining compared with 4 of 16 with missense mutations). TP53 mutations (total/affecting L2/L3 domains) were associated with expression of c-erbB-2 (P < 0.001 for both), high histological grade (P = 0.001 and P = 0.025), and bcl-2 negativity (P = 0.003 and P = 0.002). TP53 mutations, histological grade, and expression of bcl-2 (but not LOH or c-erbB-2 expression) all predicted for relapse-free as well as breast cancer-specific survival in univariate analysis (Ps between <0.0001 and 0.0155), but only tumor grade was found to be predictive in multivariate analysis (P = 0.01 and P = 0.0007, respectively). Our data are consistent with the hypothesis that certain TP53 mutations predict for resistance to doxorubicin in breast cancer patients. However, the observation that the majority of patients with TP53 mutations affecting or disrupting the L2/L3 domains with LOH in addition (n = 12) obtained a partial response (n = 4) or stabilization of disease (n = 5) during chemotherapy suggests redundant mechanisms to compensate for loss of p53 function. Our findings are consistent with the hypothesis that other defects may act in concert with loss of p53 function, causing resistance to doxorubicin in breast cancers.