A high-throughput LC-MS/MS method suitable for population biomonitoring measures five serum folate vitamers and one oxidation product

Abstract

Small specimen volume and high sample throughput are key features needed for routine methods used for population biomonitoring. We modified our routine eight-probe solid phase extraction (SPE) LC-MS/MS method for the measurement of five folate vitamers [5-methyltetrahydrofolate (5-methylTHF), folic acid (FA), plus three minor forms: THF, 5-formylTHF, 5,10-methenylTHF] and one oxidation product of 5-methylTHF (MeFox) to require less serum volume (150 μL instead of 275 μL) by using 96-well SPE plates with 50 mg instead of 100 mg phenyl sorbent and to provide faster throughput by using a 96-probe SPE system. Total imprecision (10 days, two replicates/day) for three serum quality control pools was 2.8-3.6% for 5-methylTHF (19.5-51.1 nmol/L), 6.6-8.7% for FA (0.72-11.4 nmol/L), and ≤11.4% for the minor folate forms (<1-5 nmol/L). The mean (±SE) recoveries of folates spiked into serum (3 days, four levels, two replicates/level) were: 5-methylTHF, 99.4 ± 3.6%; FA, 100 ± 1.8%; minor folates, 91.7-108%. SPE extraction efficiencies were ≥85%, except for THF (78%). Limits of detection were ≤0.3 nmol/L. The new method correlated well with our routine method [n = 150, r = 0.99 for 5-methylTHF, FA, and total folate (tFOL, sum of folate forms)] and produced slightly higher tFOL (5.6%) and 5-methylTHF (7.3%) concentrations, likely due to the faster 96-probe SPE process (1 vs. 5 h), resulting in improved SPE efficiency and recovery compared to the eight-probe SPE method. With this improved LC-MS/MS method, 96 samples can be processed in ~2 h, and all relevant folate forms can be accurately measured using a small serum volume.

Fig. 1

Spiking recovery and SPE efficiency of folate forms added to serum for methods 2 (8-probe) and 3 (96-probe) The low serum QC pool was spiked with a calibrator mixture containing each folate form at four levels (two replicates per level, three runs; 2, 4, 10, and 100 nmol/L spike for 5-methylTHF; 1, 2, 10, and 50 nmol/L spike for all other folate forms) and also measured unspiked (two replicates per run, three runs) for endogenous folate concentrations. The mixture of internal standards was added at the same time as the spike in the spike recovery experiment, but after SPE was completed in the SPE efficiency experiment. The spiking recovery and SPE efficiency were calculated as the measured concentration difference between the spiked and unspiked sample divided by the nominal concentration of the spike. Bars represent the average from the four spiking levels and error bars represent the 95% confidence limit of the mean of all data (n = 24). Spike recovery results for each individual spiking level can be found in Table 2 (method 3) and Table S2 (method 2). 5-MethyTHF, 5-methyltetrahydrofolate; FA, folic acid; THF, tetrahydrofolate; 5-formylTHF, 5-formyltetrahydro-folate; 5,10-methenylTHF, 5,10-methenyltetrahydrofolate; MeFox, pyrazino-s-triazine derivative of 4α-hydroxy-5-methylTHF.

Fig. 2

Effect of specimen type and anticoagulant on folate forms Twelve matched serum and plasma samples were analyzed with the scaled down 96-probe SPE LC-MS/MS method. K₂ EDTA and Na heparin were spray dried anticoagulants, while Na citrate was a liquid (0.5 mL/5-mL vacutainer tube); folate results were multiplied by 1.1 to correct for this dilution. 5-MethyTHF, 5-methyltetrahydrofolate; FA, folic acid; MeFox, pyrazino-s-triazine derivative of 4α-hydroxy-5-methylTHF; tFOL, total folate (sum of all folate forms including MeFox); other folate forms were < LOD.

Fig. 3

Comparison of total folate results in serum samples obtained by different LC-MS/MS methods and by microbiologic assay Method comparison consisted of two separate aliquots of 150 pristine serum samples, one aliquot analyzed by microbiologic assay (MA) for tFOL_MA and the second aliquot analyzed by three different LC-MS/MS methods (method 1: routine 8-probe SPE; method 2: scaled down 8-probe SPE; method 3: scaled down 96-probe SPE) for folate forms that were summed up for tFOL_{without MeFox} because the MA does not respond to the biologically inactive MeFox. Panels A, C and E show weighted Deming regressions (because the SD increased over the range of folate concentrations) with tFOL by MA as the reference. Panels B, D and F show relative Bland-Altman plots (because the distribution of differences was not normal) with the tFOL by MA as the reference; “mean of all” is the mean of the two methods.