Healthcare rationing---are additional criteria needed for assessing evidence based clinical practice guidelines?

ABSTRACT Alternative polyadenylation (APA) is a post-transcriptional mechanism that regulates gene expression. In human cancer, shortening of the 3ʹ-untranslated region (3ʹ-UTR) through APA is widespread affecting thousands of genes[1]. We previously identified that 3ʹ-UTR shortening (3ʹUS) disrupts the competing-endogenous RNA (ceRNA) network (3ʹUS-ceRNA effect) to promote breast cancer[2]. As different breast cancer subtypes are associated with different molecular mechanisms[3], we identified distinct 3ʹUS profiles of different breast cancer subtypes in this work, calling for the characterization of subtype-specific 3ʹUS-ceRNA effect in the ceRNA network. A quantitative challenge is that different sample sizes available for the different breast cancer subtypes can result in a systematic bias on size and topology of the constructed ceRNA networks. We addressed the bias by normalizing the networks in two-way, first between and second within the subtypes. Using the two-way network normalization, we built comparable ceRNA networks for estrogen receptor negative (ER-) and positive (ER+) subtype breast tumor samples of different size. Functional enrichment analyses associated subtype-specific 3ʹUS-ceRNA effect with ER-’s aggressive phenotype[4] and unique growth mechanism[5]. Especially, for ER-specific growth mechanism, subtype-specific 3ʹUS-ceRNA effect disrupts ceRNA crosstalk of housekeeping genes, which help maintain similar ceRNA network topology for ER- and ER+ normal samples. As ER-specific 3ʹUS-ceRNA effect is associated with ER-’s pathological features, aggressive phenotype and unique growth mechanism, our study provides new insights into the interactive mechanism of 3ʹUS and ceRNA for ER-specific cancer progression.