Abstract 2225: First double validation of transcription infidelity biomarkers for breast cancer diagnosis

Abstract

Abstract Comparison of cancer versus normal transcriptome identified single base variations occurring at increasing rates in cancer (Brulliard et al PNAS 2007). These variations are not explained by somatic mutations, polymorphisms nor alternate splicing but are caused by errors made by RNA polymerase II while copying DNA i.e. transcription infidelity (TI). Omission of single base is the TI event that most dramatically increases in cancer. RNAs carrying these frame shifts translate into proteins with distinctive carboxy terminal AA sequences. Molecular characterisation of such variant that can not be encoded by translation of the 6 phases of human or mice genome led to its detection in mice and human cancer cell lines (LLC1, B16F10, NCI-H23 and HT29). TI proteins contain specific epitopes of IgG that are part of innate humoral immunity. Injections of cancer cells to syngenic mice elicit a shift from innate to adaptive humoral response thereby establishing a causal relationship between over production of TI proteins and humoral immune response to cancer. These pre-clinical models provide for the first time a quantitative measure of cancer induced changes in IgG with reactivity specific of TI proteins and no reactivity with proteins encoded by RNA without frame shift. Using samples collected in 2 independent centres and that included age matched [median 53 y range 30 to 70 y] controls (n=227) and breast cancers (n=285), a combination 24 TI biomarkers yielded detection of breast cancers with 95 % sensitivity and 97 % specificity. Area under receiver operating curve was 0.98. These collections contain all stages of the disease with 84 % ductal, 10 % lobular and 6 % other histologic types. Test performances were 94 %, 100 % and 88% for the each type respectively. Triple negative patients represented 14 % of breast cancers and 95 % had positive testing. There were no differences in performances between the 2 recruitment centres and no detectable interferences caused by associated non malignant pathological conditions. Two men with breast cancer not included in training set were correctly classified. We propose that a novel mechanism -TI- increases in cancer contributing to very high heterogeneity of cancer transcriptome and proteome. TI proteins are recognized as danger signals contributing to cancer immunogenicity. The first clinical application yielded high performances blood based breast cancer diagnosis. Prospective large scale multi centres cases controls study applying standardized blood sampling procedures and optimized analytical conditions and including controls, breast cancer, as well as lung cancer patients is ongoing. Preliminary analysis indicates that sensitivity and specificity > 98 % will be achieved for both types of cancer. Further breast cancer test does not cross react with lung cancer and vice versa. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 2225. doi:10.1158/1538-7445.AM2011-2225