Gene expression profiles of poor-prognosis primary breast cancer correlate with survival.

The extensive heterogeneity of breast cancer complicates the precise assessment of tumour aggressiveness, making therapeutic decisions difficult and treatments inappropriate in some cases. Consequently, the long-term metastasis-free survival rate of patients receiving adjuvant chemotherapy is only 60%. There is a genuine need to identify parameters that might accurately predict the effectiveness of this treatment for each patient. Using cDNA arrays, we profiled tumour samples from 55 women with poor-prognosis breast cancer treated with adjuvant anthracycline-based chemotherapy. Gene expression monitoring was applied to a set of about 1000 candidate cancer genes. Differences in expression profiles provided molecular evidence of the clinical heterogeneity of disease. First, we confirmed the capacity of a 23-gene predictor set, identified in a previous study, to distinguish between tumours associated with different survival. Second, using a refined gene set derived from the previous one, we distinguished, among the 55 clinically homogeneous tumours, three classes with significantly different clinical outcome: 5-year overall survival and metastasis-free survival rates were respectively 100% and 75% in the first class, 65% and 56% in the second and 40% and 20% in the third. This discrimination resulted from the differential expression of two clusters of genes encoding proteins with diverse functions, including the estrogen receptor (ER). Another finding was the identification of two ER-positive tumour subgroups with different survival. These results indicate that gene expression profiling can predict clinical outcome and lead to a more precise classification of breast tumours. Furthermore, the characterization of discriminator genes might accelerate the development of new specific and alternative therapies, allowing more rationally tailored treatments that are potentially more efficient and less toxic.

Genome-wide search for loss of heterozygosity in Burkitt lymphoma cell lines.

The molecular biological characteristics of Burkitt lymphoma (BL), in addition to the presence of the Epstein-Barr virus (EBV) in some forms, relies on well-characterized alterations, such as MYC translocations and TP53 inactivations. To ascertain the number and location of other genome alterations, we used 191 polymorphic markers in a genome-wide search for loss of heterozygosity (LOH) in 31 Burkitt lymphoma cell lines and their normal counterparts. We were able to distinguish two types of altered allelic patterns: a bona fide LOH profile, indicative of deletion (LOH), and a profile indicative of increased dosage (ID). The former type was most frequent at chromosome arm 17p, most likely indicating TP53 gene inactivation. Increased dosage at 1q was found almost exclusively in non-EBV cell lines (P < 0.00004) and correlated well with karyotypic abnormalities affecting region 1q21-25. Our results suggest that a gene important for BL pathogenesis is located in region 1q21-25 and that the activation of this gene mimics the effects of EBV.