Cancer research: from prognostic genes to therapeutic targets

In recent years, numerous studies have been performed to identify biomarkers that predict prognosis or response to chemother-Non-small cell lung cancer represents one example where expression profiling was used to successfully identify prog-nostic signatures (Wigle et al. assist in the decision whether patients should receive adjuvant chemotherapy after surgical resection. A recently published meta-analysis based on gene expression microar-ray data from five lung cancer cohorts (n=860 patients) identified 14 genes as significantly associated with survival (Botling et al., 2013). Using a multiplex real-time PCR-based assay followed by validation on diagnostic paraffin embedded patient tissue, two further studies established gene signatures based on the combination of a few genes that robustly predict prognosis in the clinically-important stage I non-small lung cancer subgroup (Kratz et al., 2012; Wistuba et al., 2013). Gene expression studies have also been performed in breast cancer (review: Schmidt et al., 2009; Hellwig et al., 2010), and most of the identified prognostic genes are proliferation-associated (Schmidt et al., 2008; Siggelkow et al., 2012) or indicate immune cell infiltration (Schmidt et al. However, redox factors (Cadenas et al., 2010), anti-apoptotic proteins (Petry et al., 2010), and cytoskeletal factors controlling mechanoreactivity and migration (Martin et al., 2012) have also been identified. A major challenge facing physicians when deciding on the best course of treatment for their patients is the suboptimal accuracy of prognostic markers (Schmidt et al., 2009). For example, in node-negative breast cancer only approximately 30 % of all patients will go on to develop metastasis (Cianfrocca and Goldstein, 2004). However, the majority of patients receive chemothera-py. Therefore, biomarkers are urgently needed that accurately predicts the 70 % of patients who do not require chemotherapy because they will never develop metastasis. Although numerous biomarkers are significantly associated with the risk of metastasis, the accuracy of prediction is not sufficient to convince physicians and patients to waive chemotherapy. Despite the unsatisfactory situation in predicting prognosis, many scientists have shifted their focus to understand whether the biomarkers identified in expression profiling can be used as therapeutic targets. Ideally, the best candidate genes would be expressed specifically in carcinomas and not in healthy tissues. A further advantage would be that these genes are expressed on the plasma membrane of tumor cells, such as ERBB2 (Brase et al., 2010) or Ep-CAM (Schmidt et al., 2011), making them the good targets for therapeutic antibodies. Such membranous