Measurement of methotrexate in human cerebrospinal fluid using a chemiluminescence immunoassay intended for serum and plasma matrices

Abstract Background The concentration of MTX in blood is often measured quickly and easily by immunoassays. Thus, immunoassays may facilitate the easy determination of the concentration of MTX in the cerebrospinal fluid (CSF). In this study, we measured methotrexate (MTX) concentrations in the CSF using a high‐performance liquid chromatography (HPLC) method intended for analyzing CSF matrices and a chemiluminescence immunoassay (CLIA) method intended for assessing serum and plasma matrices and verified the differences in the results of the two methods. Methods HPLC analysis for MTX in the CSF was performed using a Prominence UFLC system with a C18 column. The HPLC method was validated in accordance with the 2018 FDA guideline. The CLIA method was performed using an ARCHITECT i1000SR system intended for serum and plasma matrices. A total of 47 CSF samples (14 clinical and 33 spiked specimens) were analyzed using the two methods. Results The HPLC method passed the validation criteria. The concentration of MTX in the same sample, determined using the HPLC and CLIA methods, differed proportionally; the percent difference in the concentrations averaged −23.0% (95% confidence interval: −36.9% to −9.1%) as revealed by the Bland‐Altman plot. The relationship between the measured values, evaluated using the Passing‐Bablok regression, was as follows: HPLC = 1.205 × CLIA – 0.024. Conclusion The equation deduced in this study can be used to correct the concentration of MTX measured using the CLIA method.


| INTRODUC TI ON
Methotrexate (MTX) is a cytotoxic drug of the antifolate type, which is used to treat certain types of hematological cancers, solid tumors, and rheumatoid arthritis. MTX, at high doses, is used clinically as a cytotoxic drug for the treatment of solid tumors and leukemia. 1,2 Regular monitoring of MTX concentrations in the serum allows early detection of abnormal clearance so that measures, such as adjustment of the dose of leucovorin and enhancement of hydration, can be opted to prevent excessive toxicity. 3,4 If extracorporeal excretion is slower than the standard, enhancement of the normal cell rescue by the administration of leucovorin should be considered. Methotrexate is administered intrathecally as a measure against the recurrence of acute lymphocytic leukemia in the central nervous system to improve the therapeutic results.
The intraventricular administration of anticancer drugs has the advantage of maintaining high drug concentrations in the cerebrospinal fluid (CSF) and minimizing systemic side effects. However, sustained exposure to high concentrations of anticancer drugs is dangerous, warranting measures to avoid it. Methotrexate is cytotoxic at concentrations of more than 1 µM. 5 It has been reported that although MTX concentrations in the CSF do not exceed 1 µM after intravenous administration, they can be higher immediately after intrathecal administration. 6 Intraventricular drug clearance is not uniform in all patients. Therefore, to safely administer the anticancer drug by intrathecal injection, evaluation of local pharmacokinetics after administration is required.
The concentration of MTX in blood is often measured by immunoassays. The results can be obtained easily and quickly using this method, thereby greatly contributing to the management of delayed extracorporeal excretion. For the immunoassay of MTX concentrations in blood, fully automatic measuring devices are widely used, most of which use serum or plasma as the target matrix. Separation analyses methods, such as capillary electrophoresis 7 and LC-MS/MS, 8 are mainly used to measure MTX concentrations in the CSF. Although the separation analyses methods are highly versatile, they require advanced analytical techniques, unlike the immunoassay methods. Therefore, it is not very feasible to measure the concentration of drugs using separation analyses in daily clinical practice.
The use of immunoassay as an alternative to the separation analyses methods can make the routine determination of the concentration of MTX in the CSF easy. Herein, we focused on the use of chemiluminescence immunoassay (CLIA) for this purpose.
The CLIA method is now widely used to measure the concentration of MTX in blood. [9][10][11] In the present study, we aimed to construct an environment for measuring the concentration of MTX in the CSF using a CLIA method following past reports. 12,13 We measured the concentration of MTX in CSF using a high-performance liquid chromatography (HPLC) method, intended for CSF, and a CLIA method, intended for serum and plasma matrices. We also determined the differences in the results obtained using the two methods. We believe that our results will be significant for employing the CLIA method to determine the MTX concentration in the CSF.

| Instrumentation
HPLC analysis was performed on a Prominence UFLC system (Shimadzu). The Analyst software (LCsolution, Shimadzu) was used to acquire and process the data. The CLIA method was performed on an ARCHITECT i1000SR system (Abbott).
The mobile phase consisted of sodium acetate buffer (50 mM, pH 3.6):acetonitrile (77:13, v/v). The flow rate was 0.9 mL/min. The detector wavelength was set at 307 nm.

| Preparation of stock and working solutions
MTX stock solution (100 mM) and the internal standard (IS) (137U, 10 mM) were prepared in 0.01 M NaOH and HPLC-grade water, respectively, and stored at −20°C. The concentration of the IS working solution was 100 µM.

| Preparation of calibration and quality control samples
For drug-free CSF, the sample that remained after CSF testing was

| Patient enrollment, sample collection, and preparation
The leftover CSF test samples collected from patients receiving MTX intrathecally were used as clinical specimens. Spiked specimens were prepared by the addition of properly diluted stock solution to drug-free CSF.
For the solid phase extraction procedure, 14

| Method validation for HPLC
In accordance with the standard guideline 15 for method validation, selectivity, LLOQ, carry-over, linearity, accuracy, precision, dilution integrity, and stability were evaluated.

| Selectivity and LLOQ
Methotrexate-and IS-free CSF samples were used to evaluate the selectivity. The lowest concentration on the calibration curve was regarded as the LLOQ. For MTX, interfering peak areas were <20% of the peak area of the LLOQ. For LLOQ samples, the mean accuracy was within ±20% of the nominal value and the coefficient of variation (CV) value did not exceed 20%.

| Carry-over and linearity
The carry-over effects of MTX and IS were evaluated by testing the response of a blank CSF sample immediately following the highest concentration of the calibration sample. The peak area of the blank CSF sample was <20% of the peak area of the LLOQ sample for MTX and 5% for the IS. The ordinary least squares method was used to fit the peak area ratio vs. the analyte concentration for linearity.

| Accuracy and precision
To evaluate the accuracy and precision, five replicates of LLOQ samples and QC samples at three levels were analyzed. The mean accuracy was within ±15% of the nominal values for the QC samples, except for the LLOQ, which was within ±20% of the nominal value.
The CV value did not exceed 15% for the QC samples and 20% for the LLOQ samples.

| Dilution integrity and stability
To evaluate the dilution integrity for the clinical samples, a 10-fold

| CLIA
The CLIA was performed on the ARCHITECT i1000SR system (Abbott). The recommended matrices included serum and plasma.
The sample volume was 60 µl. The calibration range was 0.040-1.500 µM. Samples with MTX concentrations >1.500 µM were diluted 20-fold with the ARCHITECT i1000SR system mechanically.
Calibration (ARCHITECT Methotrexate Calibrators, Abbott) and QC samples (ARCHITECT Methotrexate Controls, Abbott) were routinely tested according to the manufacturer's instructions before the analysis of the samples.

| Linearity
Linearity of MTX detection in CSF was assessed using samples pre-

| Precision
Four samples with MTX concentrations of 0.070, 0.450, 1.000, and 10.000 µM were prepared by the addition of properly diluted stock solution to drug-free CSF. To evaluate the precision, five replicates of samples at each level were analyzed.

| Application and comparison
A total of 47 CSF samples (including 14 clinical and 33 spiked specimens) were analyzed using the two methods. For HPLC, samples with MTX concentrations >1.50 µM were diluted 10-fold with the blank CSF. For CLIA, samples with MTX concentrations >1.500 µM were diluted 20-fold with the ARCHITECT i1000SR system mechanically.
A Bland-Altman plot was used to evaluate the agreement between HPLC and CLIA for the analysis of MTX in human CSF. 16,17 The equation and the correlation coefficient describing the relationship between the two methods were evaluated using the Passing-Bablok regression 18 and Pearson's product-moment correlation analysis.

| Statistical analysis
R v.3.5.1 (www.r-proje ct.org) was used for statistical analyses. Results with a p-value of <.05 were regarded as statistically significant.

| LLOQ and selectivity
The typical chromatograms obtained using the HPLC method are shown in Figure 1. The interfering peaks were not observed in the blank CSF sample ( Figure 1A) at the retention times for MTX and IS ( Figure 1B). The IS did not affect the measurement of MTX. For LLOQ samples, the mean relative error was 15.8% and the CV was 1.97%.

| Carry-over and linearity
No carry-over was observed. A typical calibration curve is shown in

| Accuracy and precision
The results for the accuracy and precision of the HPLC method are presented in Table 1. For the three concentrations of the QC

| Dilution integrity and stability
For diluted samples, the mean relative error was 4.4% and the CV was 1.88%. The results of the stability tests are shown in Table 2.
Because the mean relative error for each condition and level was <15%, MTX was stable in CSF under all the tested conditions at concentrations of 0.20, 0.60, and 1.20 µM.

| Linearity
The linearity of MTX detection in CSF by CLIA was excellent in the verified concentration range (0.1-15.0 µM); the correlation coefficient (r) was 0.997.

| Method comparison
The ranges of MTX measured using the CLIA and HPLC methods    Figure 4B).

| D ISCUSS I ON
In this study, we established an HPLC method for measuring the concentration of MTX in the CSF and successfully verified the con- can be converted into non-analyte targets. If a drug is to be analyzed, these non-analyte targets may be detected. Furthermore, when the sample to be analyzed is directly supplied to the reaction system, there is a possibility that the reaction is affected by the sample components. Although immunoassays are less versatile than separation analyses methods, they are very much applicable for clinical use, such as in commercial reagent kits. In this study, we measured the concentrations of MTX in different matrix samples using an immunoassay for which the intended matrix is serum or plasma. The concentration of MTX in the CSF sample measured using the CLIA method was slightly different from that measured using the HPLC method. The concentration determined using the HPLC method developed for the quantification of the MTX concentration in the CSF can be considered the actual value.
We found that the concentration of MTX in the CSF could not be accurately measured by the CLIA method intended for serum or plasma matrices. Because of the cross-reactivity, it is easily considered that the results obtained using the CLIA method would be higher than those obtained using the HPLC method. In addition, there is a concern that protein in the sample could influence drug quantification by the immune reaction. The protein binding rate of MTX in blood is about 50%. 20 The documentation for the reagent kit for the CLIA method states that the measured value may be low because of high albumin content in the sample. However, the amount of protein in CSF is significantly lower than that in blood because of the presence of the blood-CSF barrier. Hence, we believe MTX quantification in the CSF sample is not easily affected by protein. However, the results obtained using the CLIA method were actually lower than those obtained using the HPLC method.
In immunoassays, the pH and ionic concentration of the sample affects the reactivity. 21 In this study, we verified the accuracy of measuring MTX concentrations in the CSF using the CLIA method and deduced an equation that can correct the values with respect to the actual ones determined using the HPLC method. The evaluation of the existing immunoassays other than the CLIA method in the same manner may lead to further adoption of the measurement of the MTX concentration in the CSF in routine clinical practice. We believe such evaluations will contribute to providing safer and more personalized treatment.

ACK N OWLED G M ENTS
We would like to thank Editage (www.edita ge.com) for English language editing.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author, NY, upon reasonable request.