Partial protein binding of uracil and thymine affects accurate dihydropyrimidine dehydrogenase (DPD) phenotyping

Fluorouracil is among the most used antimetabolite drugs for the chemotherapeutic treatment of various types of gastrointestinal malignancies. Dihydropyrimidine dehydrogenase (DPYD) genotyping prior to fluorouracil treatment is considered standard practice in most European countries. Yet, current pre-therapeutic DPYD genotyping procedures do not identify all dihydropyrimidine dehydrogenase (DPD)-deficient patients. Alternatively, DPD activity can be estimated by determining the DPD phenotype by quantification of plasma concentrations of the endogenous uracil and thymine concentrations and their respective metabolites dihydrouracil (DHU) and dihydrothymine (DHT). Liquid chromatography - mass spectrometry (LC-MS) detection is currently considered as the most adequate method for quantification of low-molecular weight molecules, although the sample preparation method is highly critical for analytical outcome. It was hypothesized that during protein precipitation, the recovery of the molecule of interest highly depends on the choice of precipitation agent and the extent of protein binding in plasma. In this work, the effect of protein precipitation using acetonitrile (ACN) compared to strong acid perchloric acid (PCA) on the recovery of uracil, thymine, DHU and DHT is demonstrated. Upon the analysis of plasma samples, PCA precipitation showed higher concentrations of uracil and thymine as compared to ACN precipitation. Using ultrafiltration, it was shown that uracil and thymine are significantly (60–65 %) bound to proteins compared to DHU and DHT. This shows that before harmonized cut-off levels of DPD phenotyping can be applied in clinical practice, the analytical methodology requires extensive further optimization. •Uracil and Thymine are partially bound to proteins determined by ultrafiltration.•Recovery differences can be related to the extent of molecule protein binding.•PCA precipitation provided higher concentrations compared to ACN.•The use of cut-off levels requires extensive further optimization and harmonization.•DPD phenotyping is hampered by a lack of method standardization.