Objectives: Cytotoxic nucleosides (gemcitabine, cytarabine…) are used for the treatment of various malignancies. Their activity is dependent on the interaction with several proteins and enzymes of nucleotide metabolism. It has for a long time been hypothesized that the clinical activity of nucleoside analogues can be predicted by studying corresponding genes or gene products in clinical samples.
Methods: In this short review, I will present old and new published data from our group and others about the prediction of activity of these drugs concentrating on gene-candidate approaches, and discuss biological and technical limitations of these.
Results: A large number of studies have been conducted in various clinical settings (drugs, disease, patient cohort…) evaluating DNA, mRNA or protein-related markers. Although some individual parameters and associations thereof have been validated, only a very few numbers have been implemented in pretreatment evaluations of patients.
Conclusion: There is still much to do in the field of outcome-prediction with nucleoside analogues. The use of multiparametric methods could increase the success rate but at the cost of a poorer understanding of molecular mechanisms.
Keywords: Gemcitabine; cytarabine; cytidine deaminase; deoxycytidine kinase; hENT1; ribonucleotide reductase.