UPLC-MS/MS Analysis of Methotrexate in Plasma and Serum for Clinical Research

The antifolate methotrexate is a widely used drug for the treatment of various malignancies. Traditionally, enzyme and fluorescence based immunoassay tests are used to analyze methotrexate concentrations, but they are known for their cross-reactivity to the inactive metabolite DAMPA. Immunoassay techniques can overestimate methotrexate concentrations post-glucarpidase administration due to elevated levels of DAMPA, emphasizing the importance of using a selective detection method when performing clinical trials of high dose methotrexate. Here we describe a clinical research method using deproteination of plasma or serum samples with methotrexate-²H₃ internal standard in methanol. Isocratic separation was achieved within five minutes using an ACQUITY UPLC HSS C₁₈ SB Column on an ACQUITY UPLC I-Class System followed by detection on a Xevo TQD Mass Spectrometer.

Benefits

• Analytical selectivity afforded by highly selective mass detection
• Wide (400 fold) dynamic range
• Simple, inexpensive sample preparation using low sample volumes
• Speed of analysis

The antifolate methotrexate is a widely used drug for the treatment of various malignancies. Traditionally, enzyme and fluorescence based immunoassay tests are used to analyze methotrexate concentrations, but they are known for their cross-reactivity to the inactive metabolite 2,4-Diamino-N¹⁰-methylpteroic acid (DAMPA). Off-label use of high dose methotrexate (>500 mg/m²) for patients with CNS lymphoma¹ is usually accompanied with the administration of glucarpidase (Voraxaze) or leucovorin rescue therapy to reduce methotrexate concentrations to below 1 mmol/L. Immunoassay techniques can overestimate methotrexate concentrations post-glucarpidase administration due to elevated levels of DAMPA, emphasizing the importance of using a selective detection method when performing clinical trials of high dose methotrexate.

Here we describe a clinical research method using deproteination of plasma or serum samples with methotrexate-²H₃ internal standard in methanol. Isocratic separation was achieved within five minutes using an ACQUITY UPLC HSS C₁₈ SB Column (2.1 x 30mm, 1.8 μm) on an ACQUITY UPLC I-Class System followed by detection on a Xevo TQD Mass Spectrometer (Figure 1).

Sample preparation

Methotrexate certified reference solution and its stable labeled internal standard (²H₃) were purchased from Cerilliant (Round Rock, TX, USA). Calibrators were prepared in a surrogate matrix of pooled plasma purchased from Sera Laboratories International Ltd (West Sussex, UK). The calibration range for methotrexate was 0.025–10 μmol/L (0.01–4.5 μg/mL). QC materials were prepared using this same pooled plasma at 0.1, 2.5, 10, and 75 μmol/L (0.045, 1.1, 4.5, and 34 μg/mL). The two highest concentrations of QC samples were diluted tenfold in plasma, prior to analysis, to demonstrate acceptability of a dilution protocol.

Sample extraction

To 50 μL of sample, 250 μL of 0.1 μmol/L internal standard (ISTD) in methanol was added as a precipitation solution, vortex mixed for 30 seconds, and centrifuged for 2 minutes at 16,100 g. 50 μL of supernatant was added to 950 μL water to prepare the final extract for analysis on a conditioned UPLC-MS/MS system.

Data management

MassLynx Software v4.1 with TargetLynx Application Manager

Under the chromatographic conditions mentioned previously, methotrexate is separated chromatographically from its principal metabolites 7-hydroxymethotrexate and DAMPA. Figure 2 shows a typical lowest concentration calibrator and metabolites.

No system carryover was observed following analysis of plasma samples with methotrexate levels of up to 100 μmol/L.

Analytical sensitivity investigations indicate that quantification (<20% RSD) at 0.0025 μmol/L for methotrexate is achievable.

Total precision was determined by extracting and quantifying five replicates of four concentrations of QC material over five separate days (n=25). Repeatability was assessed by analyzing five replicates at each QC level. Table 2 presents the results of these experiments, where total precision and repeatability at the four concentrations assessed was ≤5.5% RSD for methotrexate. Samples of 10 μmol/L and 75 μmol/L concentration were diluted tenfold prior to analysis to enable quantification.

The method was shown to be linear over the range of 0.0175–13.0 μmol/L when different ratios of high and low concentration pools of methotrexate were combined and analyzed.

A tenfold dilution was performed on spiked plasma samples (20 and 75 μmol/L), with a mean accuracy of 102.1% (n=3) relative to the expected concentration.

Matrix effects were evaluated at low (0.1 μmol/L) and high (7.5 μmol/L) methotrexate concentrations in plasma and serum samples (n=7). The matrix factor range was 0.94 to 1.40 for serum samples compared with 3.12 to 3.77 for the plasma samples. The use of methotrexate internal standard adjusted response compensated for this enhancement with matrix factor ranges of 1.01 to 1.11 and 1.01 to 1.12 for serum and plasma samples, respectively.

Equivalence between plasma and serum matrix samples was determined by fortifying six individual matched plasma and serum samples with 0.1 μmol/L, 2.5 μmol/L, 10 μmol/L, and 75 μmol/L of methotrexate. Mean % differences between plasma and serum in the range -1.2% to 6.5% were observed.

Potential interference from endogenous compounds (albumin, bilirubin, cholesterol, triglycerides, and uric acid) and the exogenous material intralipid (20% emulsion) spiked at high concentrations was assessed by determining the recovery of methotrexate (n=3) from low and high pooled plasma samples (0.1 μmol/L and 1.0 μmol/L). Recovery ranged from 90.4–102.9%, and when assessing metabolites 7-hydroxymethotrexate and DAMPA at 5 μmol/L and 50 μmol/L the recovery range was 98.8–103.3%. A substance was deemed to interfere if a recovery range of 90–110% was exceeded.

An initial assessment of method accuracy was made by analyzing 12 External Quality Assurance (WEQAS, Cardiff, UK and LGC, Teddington, UK) serum samples spanning a concentration range from 0.057–3.80 μmol/L. Samples were analyzed over five separate days. Mean % deviations from absolute reference values ranged from -7.4% to +0.4%, with an overall mean of -5.7%.

Methotrexate measurements were made on a plasma series drawn from an individual 27 hours pre- to 147 hours post-glucarpidase administration. Methotrexate measurements by immunoassay differed to UPLC-MS/MS values by 36–99% <45 hours post-glucarpidase; the magnitude of the difference was proportional to the DAMPA peak area. Between 34–57 hours post-glucarpidase, the UPLC-MS/MS method detected an increase in methotrexate concentration which was not observed by immunoassay. Figure 3 presents a comparison of data.

An analytically sensitive and selective clinical research method has been developed for the analysis of plasma and serum methotrexate.

Using only 50 μL of sample with fast and inexpensive sample preparation, this analytical method provides sufficient analytical sensitivity to analyze low levels of methotrexate (0.025 μmol/L). Methotrexate is separated from its principal metabolites both chromatographically (UPLC) and selectively (mass detection using multiple mass monitoring), enabling accurate and precise quantification of methotrexate across the investigated concentration range.

Applying the UPLC-MS/MS method to research samples, post-glucarpidase administration shows a methotrexate ‘rebound’ phenomenon, not detected by immunoassay. This highlights the importance of using this selective detection method when performing clinical trials of high dose methotrexate in combination with glucarpidase rescue therapy to further understanding the pharmacokinetics of methotrexate.

1. Ahmed and Hasan, J Cancer Sci Ther. 2013;5(3):106–12.

Acknowledgements

Christopher Chaloner and Lesley Tetlow and their colleagues at the Department of Clinical Biochemistry, Royal Manchester Children’s Hospital, Manchester, UK, are thanked for the provision of anonymized plasma samples for this research. Additionally, the scientific contribution of Christopher Chaloner is acknowledged.

LGC (Teddington, UK) who kindly provided EQA samples for accuracy testing.