|
Abstract INTRODUCTION:
Direct Analysis in Real Time (DART) is a soft ionization technique that produces excited species from gases such as helium, argon or nitrogen followed by a cascading effect of ionizing atmospheric/dopant molecules and biological samples to produce analyte ions. DART can be combined with tandem Mass Spectrometry for the rapid analysis of a large variety of samples by minimizing sample preparation, detection and analysis time.
OBJECTIVE:
In a proof-of-concept study for measuring selected antifungals in human serum, 5-fluorocytosine (5FC), itraconazole (ITR), hydroxyitraconazole (OH-ITR), isavuconazole (ISV), ketoconazole (KETO), fluconazole (FLU) and posaconazole (POSA) were tested as DART candidates on a Bruker EVOQ-DART TQ MS. DART results for some analytes were compared with existing LC-MS/MS results.
METHODS:
Analytes were optimized for unique MS/MS transitions (quant and qual) and collision energies (CE). DART cone and gas stream temperatures were optimized to 350oC and 375oC respectively for all analytes except FLU (275oC and 300oC). DART was operated in linear scan mode (0.5 mm/s, 5.0mm range) with helium as the carrier gas. Acquisition time was optimized to 12 sec. Monitored analyte transitions were: 5FC m/z 130.1/113.1 and 58.1 with CEs 14.8 and 31.6, respectively (5-FC -13C15N2 m/z 133.1/115.1, CE 14.8), ITR m/z 705.2/392.2 and 255.7 with both CEs 29.6 (ITR-d5 m/z 710.2/397.2, CE 29.6), OH-ITR m/z 721.2/408.2 and 256.2 with CEs 34.5 and 33.6, respectively (OH-ITR-d5 m/z 726.2/413.2, CE 34.5), ISV m/z 438.1/224.0 and 369.0 with CEs 10.9 and 18.8, respectively (ISV-d4 m/z 443.1/224.0, CE 10.9), KETO m/z 531.1/489.1 and 244.0 with CEs 26.6 and 35.5, respectively (KETO-d8 m/z 539.1/497.1, CE 26.6), FLU 306.9/238.0 and 121.0 with CEs 7.9 and 34.5, respectively (FLU-d4 m/z 310.9/242.1, CE 7.9), POSA m/z 701.3/614.6 and 344.1 with CEs 30.6 and 39.5, respectively (OH-ITR-d5 used as internal standard for POSA).
50 µL of three matrix matched calibrators, two quality controls per analyte, and patient samples were spiked with a 100 µL mixture of internal standards and subjected to protein precipitation with acetonitrile (for 5FC, ITR, OH-ITR, ISV, KETO and POSA) and samples mixed with 10 µL mixture of internal standards were subjected to liquid-liquid extraction using 90 µL ethyl acetate for FLU. The samples were vigorously mixed for 30 seconds and centrifuged at 4000 rpm for 5 mins. 3.5-5 µL of the supernatant was spotted on the QuickStrip HTS 96 well DART screen plate and dried at 40oC for 20 mins under nitrogen (40 psi). Precision and linearity across AMR were evaluated. Results for 5FC, ITR, OH-ITR, FLU, and POSA were compared with existing LC-MS/MS results.
RESULTS:
Lower and upper limits of measurement interval were 5.23 and 111.0 µg/mL (5FC), 0.13 and 2.98 µg/mL (ITR), 0.18 and 3.91 µg/mL (OH-ITR), 0.45 and 10.20 µg/mL (ISV), 0.40 and 8.84 µg/mL (KETO), 0.67 and 14.1 µg/mL (FLU), and 0.23 and 5.420 µg/mL (POSA) respectively. Within-run, between-run/day and total % CVs evaluated using two matrix matched quality controls for 5FC at 21.10 µg/mL were 2.86, 0.10 and 2.86% respectively (accuracy 99.8%), at 50.3 µg/mL were 2.37, 0.84 and 2.51% respectively (accuracy 99.8%). For ITR at 0.59 µg/mL % CVs were 2.98, 2.72, and 4.04% respectively (accuracy 98.5%), at 1.41 µg/mL were 2.64, 6.82 and 7.31%, respectively (accuracy 98.4%). For OH-ITR % CVs at 0.73 µg/mL were 6.14, 0.83, and 6.19% respectively (accuracy 101.2%), at 1.72 µg/mL were 2.21, 2.54 and 3.37%, respectively (accuracy 100.6%). For ISV % CVs at 1.99 µg/mL were 7.94, 6.00, and 9.95 % respectively (accuracy 111.0%), at 4.59 µg/mL were 10.10, 12.89 and 16.37%, respectively (accuracy 94%). For KETO % CVs at 1.73 µg/mL were 3.20, 0.14, and 3.20%, respectively (accuracy 94.1%), at 3.92 µg/mL were 2.02, 1.37 and 2.44%, respectively (accuracy 99.3%). For POSA % CVs at 0.92 µg/mL were 4.54, 2.69, and 5.28%, respectively (accuracy 100.8%), at 2.25 µg/mL were 1.38, 4.02, and 4.25%, respectively (accuracy 99.6%).
Good agreement was observed with existing LC-MS/MS methods for 5FC (slope = 1.083 , r = 0.949, bias = 3.87%, n = 27), ITR (slope = 1.112, r = 0.975, bias = 12.9%, n = 37), OH-ITR (slope = 0.96 , r = 0.944, bias = 1.94%, n = 38) and FLU (slope = 0.875, r = 0.979, bias = -3.49, n = 17).
POSA samples extracted by protein precipitation, when compared with LC-MS/MS, showed a high positive bias by DART-MS/MS, although the relationship was fairly linear (slope = 2.32, r = 0.845, bias = 79.9%, n = 38). Using a different MRM transition for POSA (701.3/683.1) and liquid-liquid extraction (LLE), better comparison was observed (slope = 0.896, r = 0.957, bias = 6.76% n = 30). Although the average accuracy % by LLE was 112%, a bias of > 40% was observed in 5 out of 30 samples suggesting that in some cases DART could be less specific than LC-MS/MS.
%CVs of replicates spotted on the DART sample plate were <15%. No carryover was observed between low and high samples. Internal standard contribution at LLOQ was < 20% for all analytes. All samples passed the ion-ratio tolerance maintained at ±30%.
CONCLUSION:
In this proof-of-concept study, the measurement of antifungals in human serum was demonstrated using protein precipitation and DART-MS/MS. The 12.9% bias observed for ITR could be attributed to calibration bias between DART-MS/MS and the compared LC-MS/MS method. The moderate linear association, with a systemic bias observed for posaconazole, later alleviated by liquid-liquid extraction, could be attributed to matrix effect. Greater selectivity in sample preparation methods could be evaluated for improving accuracy in DART measurements. Overall, the results demonstrate the suitability of DART-MS/MS as a fast alternate to conventional LC-MS/MS assays for some analytes, otherwise requiring extensive sample preparation steps, expensive HPLC columns, and long analysis times.
|
= Discovery stage. (53.14%, 2025)
= Translation stage. (22.33%, 2025)
= Clinically available. (24.53%, 2025)
