Simultaneous quantification method for 5-FU, uracil, and tegafur using UPLC-MS/MS and clinical application in monitoring UFT/LV combination therapy after hepatectomy

Combination therapy of tegafur/uracil (UFT) and leucovorin (LV) is widely used to treat colorectal cancers. Although this therapy has a significant therapeutic effect, severe adverse effects occur frequently. Therapeutic drug monitoring (TDM) may help to prevent adverse effects. A useful assay that can quantitate plasma levels of 5-FU, uracil, and tegafur simultaneously for TDM has been desired, but such a method is not currently available. In this study, we aimed to develop a sensitive method for simultaneous quantification of 5-FU, uracil, and tegafur in human plasma using ultra-performance liquid chromatography coupled to tandem mass spectrometry (UPLC-MS/MS). After preparing plasma samples by protein precipitation and liquid extraction, 5-FU, uracil, and tegafur were analyzed by UPLC-MS/MS in negative electrospray ionization mode. Validation was performed according to US Food and Drugs Administration guidance. The calibration curves were linear over concentration ranges of 2–500 ng/mL for 5-FU, 20–5000 ng/mL for uracil, and 200–50,000 ng/mL for tegafur. The corresponding average recovery rates were 79.9, 80.9, and 87.8%. The method provides accuracy within 11.6% and precision below 13.3% for all three analytes. Matrix effects of 5-FU, uracil, and tegafur were higher than 43.5, 84.9, and 100.2%, respectively. This assay was successfully applied to assess the time courses of plasma 5-FU, uracil, and tegafur concentrations in two patients with colorectal liver metastasis who received UFT/LV therapy after hepatectomy. In conclusion, we succeeded to develop a sensitive and robust UPLC-MS/MS method for simultaneous quantification of 5-FU, uracil, and tegafur in human plasma. This method is potentially useful for TDM in patients receiving UFT/LV combination therapy.

Plasma sample preparation. Two hundred µL of blank plasma was added to 100 µL of each calibration or QC solution and vortexed with 100 µL of the internal standard (IS) mixture (5-FU-13 C, 15 N 2 ; 200 ng/mL, uracil-15 N 2 ; 2000 ng/mL, and tegafur-13 C, 15 N 2 ; 5000 ng/mL). The patient sample (200 µL) was also vortexed with 100 µL of the IS mixture and 100 µL of 50% acetonitrile aqueous solution containing 2 mM ammonium formate for volume adjustment. After adding 500 µL of methanol for protein precipitation, all samples were vortexed for 1 min. After centrifugation at 15,000 rpm at 10 °C for 10 min, 750 µL of the supernatant was carefully transferred to a new polypropylene tube and evaporated with nitrogen (N 2 ) gas. After evaporation, 1800 µL of ethyl acetate was added and liquid-liquid extraction was performed for 10 min. After centrifugation at 15,000 rpm at 10 °C for 10 min, 1440 µL of the supernatant was transferred to a new polypropylene tube and evaporated with N 2 gas. The residue was redissolved with 100 µL of 27.5% acetonitrile aqueous solution containing 2 mM ammonium formate. Mass spectrometry. The conditions for mass spectrometry were as described previously by Tanaka et al. 21 .
Samples were ionized using the following parameters: electrospray voltage was 0.5 kV, cone voltage was 20 V, source temperature was 150 °C, cone gas (N 2 ) flow rate was 150 L/hour, desolvation gas (N 2 ) flow rate was 1000 L/hour, and desolvation temperature was 500 °C. The mass spectrometer was optimized automatically to 5-FU, uracil, tegafur, and each internal standard, using the MassLynx V4.1 system software package (Waters) and Intel-liStart standard optimization procedures (Waters). Multiple reaction monitoring (MRM) analysis was carried out using argon as collision gas for collision-induced dissociation (CID Validation of the analytical method. Analytical validation was performed according to the guidelines published by the US Food and Drug Administration (FDA) 23 . All analytical parameters concerning validation were determined according to previous reports [19][20][21][22] . Three validation batches, each containing eight calibration samples and 24 QC samples at different concentrations (LLOQ, QC A, B, and C; each in sextuplicate), were analyzed. Accuracy was determined as the ratio of mean measured concentration to the nominal concentration. Precision was defined by the ratio of the standard deviation to the mean concentration measured. Accuracy and precision were calculated for each analytical batch (within-batch) and for three validation batches (batchto-batch). Selectivity was evaluated by comparing chromatograms of blank plasma samples obtained from six healthy volunteers without the addition of the internal standard mixture. The baseline signals at the expected analyte retention times were evaluated for interfering peaks. The recovery rate from plasma was determined by comparing the peak areas obtained from QC samples A, B or C with the respective peak area obtained from blank plasma spiked at the corresponding QC level after extraction (representing 100% of the analyte amount in an identical matrix) in triplicate determination. The potential matrix effect of plasma components affecting the measurement of analytes was calculated by comparing the peak area of blank plasma spiked at QC levels A, B or C after extraction with the respective peak area of matrix-free LC eluent containing the same amount of analyte in triplicate determination. The freeze-thaw stability of analytes was tested using QC samples B and C after undergoing three freeze-and-thaw cycles, and accuracy was calculated. Stability in the autosampler was tested by performing repeated analysis on QC samples B and C after being placed in the autosampler at 5 °C for 24 h. Long-term stability has already been reported 24,25 .
Incurred sample reanalysis. Incurred sample reanalysis (ISR) was conducted to verify the reliability of the patient sample measurements. All the patient samples were reanalyzed. According to the FDA guidance 23 , the percentage difference of the results between the original measurement and the reanalyzed measurement was determined with the following equation: % Difference = (Reanalyzed -Original)/Mean × 100%.
Calculations. Calibration curves were obtained using the standard samples and analyte-specific MRM quantifier transitions. Peak area ratios of 5-FU, uracil and tegafur to the corresponding internal standards were calculated, and weighted linear regression (1/x) was carried out for each analytical batch using the TargetLynx V4.1 software package (Waters). Method linearity was validated by performing simple linear regression between the nominal concentrations and back-calculated concentrations of the calibration samples using the corresponding calibration curve.

Results
Mass spectrometric and chromatographic characteristics. Figure 1 shows the precursor ion and product ion mass spectra of 5-FU, uracil, and tegafur.  uracil, and (c) tegafur. No interfering peaks due to plasma matrix were observed in plasma samples obtained from healthy volunteers. A good peak shape was also obtained in LLOQ. LLOQ, lower limit of quantification; QC, quality control. www.nature.com/scientificreports/ for 3 QCs was between 3.65 and 4.80% for 5-FU, between − 1.37 and 0.82% for uracil, and between 0.48 and 3.07% for tegafur. Batch-to-batch precision for 3 QCs was less than 9.27% for 5-FU, less than 8.85% for uracil, and less than 9.72% for tegafur. The extraction recovery rate and matrix effect were evaluated by triplicate determination of 3 QCs for each compound. The average recovery rates (mean ± SD) of 5-FU, uracil, and tegafur were 79.9 ± 11.2, 80.9 ± 5.9, and 87.8 ± 10.6%, respectively. Matrix effects of 5-FU, uracil, and tegafur ranged from 43.5 to 69.7, 84.9 to 102.2, and 100.2 to 119.8%, respectively. There was no difference in matrix effect among all QC levels. The freeze-and-thaw stability was tested within the validation process using three freeze-and-thaw cycles at QC B and C levels. No significant changes in measured concentrations were observed (accuracy ranged between − 0.2 and 9.8% for 5-FU, between − 5.7 and − 3.6% for uracil, and between 2.2% and 20.7% for tegafur). Stability in the autosampler at 5 °C for 24 h was evaluated based on the accuracy for QC B and C samples. The accuracy for 5-FU, uracil, and tegafur measurements ranged from − 5.7 to − 2.6%, − 8.2 to − 4.3% and − 10.5 to − 4.6%, respectively. Since the range of accuracy fell within 15%, autosampler stability was acceptable.

Plasma concentrations of patients undergoing UFT/LV combination therapy. The time courses
of plasma concentrations in two patients receiving UFT/LV combination therapy measured by the UPLC-MS/ MS method developed in this study are shown in Fig. 3. The two patients underwent partial hepatectomy for CRLM before starting UFT/LV chemotherapy. One patient received UFT three times a day at a dose of 400 mg/ day (200-100-100), and the other received UFT three times a day at a dose of 500 mg/day (200-200-100). At day 1, several data points of 5-FU, uracil, and tegafur were lower than LLOQ because of the absorption lag. At day 8, however, all measured concentrations of 5-FU, uracil, and tegafur were within the calibration ranges. ISR was conducted to verify the patient sample concentrations' reliability by reanalyzing all the patient specimens used

Discussion
To the best of our knowledge, it is the first report of a method for simultaneous quantification of 5-FU, uracil, and tegafur. A method for simultaneous determination of tegafur and 5-FU was reported previously 17 . However, this method cannot be used to measure uracil together with 5-FU and tegafur simultaneously. Uracil inhibits the degradation of 5-FU by dihydropyrimidine dehydrogenase (DPD) in a competitive manner. High uracil levels are associated with higher toxicity as well as efficacy. To maintain an optimal balance between effectiveness and toxicity, tegafur and uracil are combined in a molar ratio of 1:4 in the preparation of UFT 26 . Besides, it is known that hepatectomy causes a reduction in DPD activity, leading to elevation of not only plasma 5-FU but also uracil level 27,28 . In other words, it is critically important to monitor these three compounds to accurately assess the therapeutic and adverse effects of this combination therapy. For these reasons, we developed a novel method to quantify 5-FU, uracil, and tegafur simultaneously in this study. We used a combination of protein precipitation and liquid extraction for sample preparation. Indeed, protein precipitation is a speedy and straightforward method for sample preparation. However, protein precipitation alone is insufficient to remove matrices that interfere with ionization. Moreover, considering routine use of the MS system, using a combination of sample preparation procedures significantly reduces the wear of the MS system. A combination of protein precipitation and liquid extraction is expected to improve the recovery rate and reduce the matrix effect. Furthermore, simultaneous and more selective detection of 5-FU and uracil was achieved by using tandem mass spectrometry instead of a UV detector system. However, in order to quantify 5-FU, uracil, and tegafur simultaneously, the conditions for sample preparation, MS system, and LC system in our method were set in favor of tegafur. As a result, the recovery rates of the highly polar compounds; 5-FU and uracil, were lower than that of tegafur. In a previously reported combined LC-UV and LC-MS/MS method for simultaneous determination of 5-FU and tegafur, the calibration range of 5-FU was 8-200 ng/mL, and that of tegafur was 800-20,000 ng/mL 17 . Despite using conditions favorable for measuring tegafur, our method has a more extensive calibration range and superior LLOQ for 5-FU compared to the previous method.
Severe adverse effects tend to occur during adjuvant chemotherapy given after liver resection compared to without liver resection. We previously reported that major hepatectomy enhanced the toxicity of 5-FU in a rat model 27 . The changes in drug metabolism after hepatectomy observed in the rat model probably also apply to humans. UFT/LV combination therapy is considered more tolerable and safer than oxaliplatin or irinotecanbased regimen. Although we consider UFT/LV the most recommended regimen for CRLM patients after hepatectomy, the impact of hepatectomy on the pharmacokinetics of UFT/LV remains unclear. Therefore, the optimal dose of UFT in this regimen cannot be determined. To provide more reliable therapy, TDM is a useful approach.
After our novel method was developed and validated, we examined the clinical application of this method to TDM for patients receiving UFT/LV therapy. Although several data points of each compound at day 1 were lower than LLOQ, all data points at day 8, when steady state was assumed to have reached, were within the calibration ranges. These results demonstrate that the method is clinically applicable for TDM of patients undergoing UFT/LV therapy.
Currently, the department of gastroenterological and pediatric surgery and the department of clinical pharmacy in Oita University Hospital are jointly conducting phase I clinical study (UMIN 000,021,146) designed to elucidate the effect of major hepatectomy, as reflected by the efficacy and toxicity of UFT/LV combination therapy. In the future, we shall evaluate the impact of major hepatectomy on drug metabolism in patients on UFT/LV combination therapy.