An isotope dilution-liquid chromatography-tandem mass spectrometry-based candidate reference measurement procedure for the quantification of mycophenolic acid glucuronide in human serum and plasma

Chemicals and reagents

The standard MPAG (C₂₃H₂₈O₁₂, molecular weight=496.5 g/mol, conversion factor to molar unit [µmol/L]=2.0; CAS 31528-44-6, Cat. No. C2565, Lot JA-ALS-21-093-P3) was bought from Alsachim (Illkirch, France). MPA (CAS 24280-93-1) was purchased from LGC (Dr. Ehrenstorfer) (Luckenwalde, Germany). [²H₃]-MPAG (CAS 31528-44-6 (unlabeled), Cat. No. M831522, Lot 14-YSW-5-1) was bought from Toronto Research Chemicals (Toronto, Canada).

qNMR for determination of the purity of the standard materials

Glucuronides, known for their extreme hygroscopicity, pose significant challenges when it comes to accurate weighing on an ultra-microbalance [18], [19], [20]. To circumvent this issue, we equilibrated the analyte open air with regular stirring, which facilitated the material’s saturation with moisture and subsequently enabled precise weighing [18], 19]. Subsequently, we developed the qNMR methodology for this saturated material. A single-pulse ¹H{¹³C}NMR was utilized for the quantitation (δ=6.30 ppm; 1H; methyl-3,5-dinitrobenzoate as SI traceable qNMR ISTD; pyridine-d₅ as solvent) with an inter-scan delay of 70 s (Supplementary material 2 Figures 1 and 2). The identity of this resonance was confirmed by 2D-TOCSY, 2D-gHSQCAD and 2D-gHMBCAD pulse sequences (Supplementary material 2 Figures 3 and 4).

The choice of pyridine-d₅ as the solvent was deliberate, given its aromatic solvent-induced separation (ASIS) effects that enable optimal resolution of the quantitative resonance. This meticulous selection was crucial for verifying the analyte as the O-aryl glucuronide as opposed to the O-acyl glucuronide, a distinction that was substantiated by the ³J_CH correlation observed in the 2D-gHMBCAD spectrum. Furthermore, this specific resonance allows for a clear differentiation between MPAG and MPA, establishing it as the sole identifiable and quantifiable signal for MPAG, particularly in scenarios where the pure analyte is unavailable. Comprehensive details on the NMR acquisition and processing parameters are documented in Supplementary Material 2 (Table 1, Figures 1–4).

Utilizing these parameters, six individual experiments, each involving separate weighings, yield a final absolute content value of 93.7±0.4 % (k=1).

Additionally, we also performed quantum chemical calculations to ascertain the conformer distribution of MPAG in solution because the Boltzmann average of these structures constitute the NMR spectrum. Computational calculations to determine the most reactive chemical bond towards de-protonation was also performed owing to the labile nature of the methylene protons in the internal lactone group. Finally, DFT single-point calculations were performed at the wB97X-V/def2-TZVP level of theory and all the information is presented in Supplementary Material 2. These theoretical analyses contributed in the determination of ideal quantitative resonance for both analytes.

Preparation of calibrators and quality control samples

Two calibrator stock solutions were prepared which were further used for the preparation of spike solutions and the final matrix-based calibrator levels. To obtain concentrations of 120 mg/mL, approximately 600 mg of MPAG was weighed directly into a volumetric flask on an analytical balance and dissolved in 5 mL of DMSO.

The concentration of the stock solutions was precisely calculated based on the purity of the reference material (93.7±0.4 % determined by qNMR) and the exact amount weighed. The stock solutions were subsequently diluted with DMSO to obtain working solutions with concentrations of 3.00 mg/mL. Stock and working solutions were used for further preparation of the eight calibrator spike solutions in DMSO. Final matrix-based calibrators, uniformly distributed from 0.750 to 600 μg/mL, were prepared by a 1+99 dilution (v+v) into analyte-free human serum. The conversion factor from µg/mL to µmol/L is 2.0.

Four levels of matrix-based quality control levels were prepared using a third independent stock solution and resulting spike solutions, following the same approach as described for the calibrator levels. The concentrations for the control levels were established at four crucial points: above the lower limit of the measuring interval (LLMI) (2.00 μg/mL), at the lower margin of the expected steady state MPAG drug level range (40.0 μg/mL), at the upper margin of expected steady state drug level range (200 μg/mL), and above any expected drug level range (400 μg/mL).

Internal standard solution

For the preparation of the MPAG ISTD stock solution, [²H₃]-MPAG was dissolved in the appropriate amount of DMSO to obtain a 1,000 μg/mL ISTD stock solution. The ISTD working solution was prepared by mixing 100 µL of ISTD stock solution with 3,900 µL Milli-Q water to receive a concentration of 25.0 μg/mL [²H₃]-MPAG.

Sample preparation

100 µL of the ISTD working solution was pipetted into a 2 mL tube (Eppendorf, Hamburg, Germany) and 50 µL of the sample specimen (calibrator/QC/native sample) was added. For protein precipitation, 1,000 µL precipitation solution (75 % methanol (v+v)). 10 µL of the supernatant was first mixed with 1,000 µL 45 % acetonitrile in mobile phase A (v+v) followed by another 1+9 (v+v) dilution directly in an HPLC vial (Wicom, Heppenheim, Germany) with 45 % acetonitrile in mobile phase A (v+v).

Liquid chromatography mass spectrometry

An Agilent 1290 Infinity II LC system (Santa Clara, California, USA), equipped with a binary pump, a vacuum degasser, an autosampler at 7 °C and a column compartment tempered to 40 °C, was used for chromatographic separation. Baseline separation of MPAG and MPA, MMF and acyl-MPAG were achieved using a Phenomenex Luna C18(2) column (100 × 3 mm, 3 μm, Phenomenex, Lane Cove, Australia) using a gradient elution over 9 min with a flow of 0.6 mL/min. The mobile phases consisted of 2 mM ammonium acetate in Milli-Q water with 0.1 % formic acid (A) and 2 mM ammonium acetate in methanol with 0.1 % formic acid (B). Contamination of the mass spectrometer was reduced by switching the eluent flow until 0.5 min and from 4.5 min to the waste.

MPAG was detected in selected reaction monitoring (SRM) mode using an AB Sciex Triple Quad 6500+ and Q-Trap 6500+ mass spectrometer operating in positive electrospray ionization mode. The quantifier ion transition (MPAG m/z 514.2 → 303.1) and the corresponding ISTD transition ([²H₃]-MPAG m/z 517.2 → 306.1) served as the basis for MPAG quantitation. Supplementary Material 1 includes detailed LC and MS methods.

System suitability test

To check the sensitivity and chromatographic performance of the system a system suitability test (SST) was performed prior to each analysis and data was only collected if the SST was passed. Using a MPAG stock solution (1 mg/mL dissolved in DMSO), two samples (SST1 and SST2) were prepared in mobile phase A with concentrations corresponding to processed calibrator levels 1 and 8, respectively.

To pass the SST, signal-to-noise ratio of the quantifier transition had to be ≥50 for SST1 and the retention time for SST1 and SST2 had to be within 3.0 ± 0.5 min. To assess possible technical carry-over effects, SST2 was injected followed by two solvent blanks. The analyte peak area observed in the first blank after the injection of SST2 had to be ≤20 % of the analyte peak area of SST1.

Calibration and structure of analytical series and data processing

The final calibration function was obtained by measuring the calibration levels 1 to 8 in increasing concentration at the beginning and the end of the analytical series. The analyte to ISTD area ratios (y) was plotted against the analyte concentration (x), resulting in a quadratic regression model (y=a₀ + a₁x + a₂x ² ). Data processing was performed using Analyst^® software (version 1.6.3 or higher) with the IntelliQuant algorithm. The peaks were integrated at a retention time of 3.0±0.5 min using a smoothing factor of five, a peak splitting factor of one, a noise percent of 80 %, and a base sub-window of 0.8 min.

Method validation

Assay validation and determination of measurement uncertainty were performed according to the Clinical & Laboratory Standard Institute (CLSI) Guidelines C62A “Liquid Chromatography-Mass Spectrometry Methods” [21], the International Conference on Harmonization (ICH) guidance document “Harmonised Tripartite Guideline Validation of Analytical Procedures: Text and Methodology Q2 (R1)” [22] and the Guide to the expression of uncertainty in measurement (GUM) [23].

Traceability to SI units

Traceability to the SI unit of mass (kilogram) and the SI unit of amount of substance, has been established by the utilization of methyl-3,5-dinitrobenzoate as the qNMR ISTD which is directly traceable to NIST PS1 (primary qNMR standard). Since Planck’s constant is now the main parameter for defining kilogram and Avogadro’s constant for mole, qNMR methodology is best suited to fulfill both the traceability chains. Utilizing the afore mentioned parameters, six individual experiments (Supplementary Material 2 Figure 1 and 2; Table 1) involving six individual weighings, yield a final absolute content value of 93.7±0.4 % (k=1).

Figure 1:

LC-MS/MS read outs of spiked analyte-free human serum sample containing 15.0 μg/mL of each mycophenolic acid glucuronide (peak 1), mycophenolic acid (peak 2), mycophenolic acid acyl glucuronide (peak 3), and mycophenolate mofetil (peak 4), SRM ion traces of (A) mycophenolic acid glucuronide, (B) mycophenolic acid, (C) mycophenolic acid acyl glucuronide, (D) mycophenolate mofetil.

Selectivity

Baseline separation of MPAG and MPA, acyl-MPAG, and MMF was achieved using a Phenomenex Luna C18(2) column resulting in a chromatographic resolution ≥2.7 for all tested interferences (Figure 1). Moreover, no signals were detected at the expected retention time of MPAG (3.0±0.5 min) when sample pools of analyte-free human serum, TDM-free serum, and Li-heparin plasma were measured (Figure 2). Furthermore, the ISTD showed no residue of the unlabeled analyte and no interferences were observed within the retention time window of MPAG, making it suitable for use in this RMP.

Figure 2:

Mycophenolic acid glucuronide LC-MS/MS derived analytical readouts. (A) Chromatogram of a matrix blank showing the analyte SRM ion trace (left) and the ISTD SRM ion trace (right). (B) Chromatogram of the lowest calibrator level with a concentration of 0.750 μg/mL mycophenolic acid glucuronide spiked in analyte-free human serum; analyte (left) and ISTD (right). (C) Patient sample pool (n≥5) with a concentration of 58.3 μg/mL; analyte (left) and ISTD (right). ISTD, internal standard; LC-MS/MS, liquid chromatography-tandem mass spectrometry.

Matrix effect

A post-column infusion experiment revealed no significant shift in the ionization field at the expected retention time in analyte-free human serum, TDM-free serum, and plasma matrices (K₂-EDTA, K₃-EDTA, and Li-heparin). Additionally, MEs were investigated by comparing mean slopes and coefficients of determination of calibrations in different matrices (analyte-free human serum, TDM-free serum, Li-heparin plasma) and calibration in neat solution (45 % acetonitrile in mobile phase A (v+v)). Mean slopes were found to be 0.0121 (95 % CI 0.0117 to 0.0124) for the analyte-free human serum, 0.0122 (95 % CI 0.0118 to 0.0125) for the neat solution, 0.0121 (95 % CI 0.0116 to 0.0126) for the TDM-free serum and 0.0123 (95 % CI 0.0116 to 0.0129) for the Li-heparin plasma. 95 % CIs are overlapping and therefore they are not significantly different from each other, and no ME is present. In addition, the coefficients of determination were ≥0.999. Matrix samples (n=6, sample preparations), when calibrated against calibrators in neat, showed a relative bias for all matrices and levels ranging from −1.8 to 4.0 %.

Furthermore, relative MEs were evaluated based on absolute peak areas. Analyte and ISTD peak areas, as well as area ratios from spiked matrix samples, were compared to those from spiked neat samples. In all tested matrices, the MEs for the analyte ranged from 98 to 106 %, while ISTD values ranged from 100 to 103 %. At the lowest concentration level, the MEs for the analyte ranged from 103 to 111 %, while the ISTD values ranged from 100 to 102 % (Table 1), indicating the possibility of MEs. This variability results from two out of five preparations showing increased analyte peak areas while ISTD peak areas remained constant. However, since all concentration levels were prepared from the same matrix pool, the presence of significant MEs can be excluded. Furthermore, all other measurements were unaffected, and the ISTD effectively corrected for any potential variations in the mass spectrometry signals in the remaining preparations.

Linearity

The linearity of the method was proven by analyzing calibration curves extended by ±20 % at the lower and higher end (0.600 and 720 μg/mL). Residuals were randomly distributed in a quadratic model. The mean coefficient of determination was 0.999 (95 % CI from 0.998 to 1.000). The samples 1 to 11 were evaluated, demonstrating a linear dependence with a coefficient of determination of 1.000. The relative deviation ranged from −3.8 to 2.7 % and the CV was found to be ≤3.8 %.

Lower limit of measuring interval

The LLMI was determined by using spiked samples at the lowest calibrator level (0.750 μg/mL). Relative bias showed a deviation of −3.3 % and the CV was 3.2 % (n=4, preparations).

Precision and accuracy

Precision was evaluated through a five-day validation experiment. The repeatability CV for both serum and plasma samples ranged from 0.8 to 1.0 %, except at the lowest concentration level where it was ≤4.0 % (Table 2). The intermediate precision CV for all concentrations and matrices remained consistently below 1.5 %, except for the lowest concentration, which showed CVs of 3.7 and 4.6 %, respectively.

The mean bias was evaluated using three preparations of each operator (n=6). The relative mean bias ranged from −0.9 to 3.2 % for spiked analyte-free human serum samples and from −0.3 to 2.6 % for Li-heparin plasma samples. Dilution integrity was proven for high concentrated samples above the measuring range with a bias ranging from 2.6 to 3.2 % (Table 3). The corresponding CIs showed that the results were distributed around the nominal concentration.

Stability

Samples spiked in analyte-free human serum were stable at 7 °C over a period of 6 days with a mean recovery of 99 %. MPAG spiked in analyte-free human serum stored at −20 °C were found to be stable over a period of 27 days with a mean recovery of 101 %.

Equivalence of results between independent laboratories

A method comparison study containing 101 native anonymized patient samples was performed between two independent laboratories. From these samples one sample was highlighted as an outlier using the LORELIA (local reliability) outlier test [28] and therefore excluded from the evaluation.

Passing-Bablok regression analysis showed a good agreement between the two laboratories and resulted in a regression equation with a slope of 1.02 (95 % CI 1.01 to 1.03) and an intercept of 0.03 (95 % CI –0.24 to 0.25). Moreover, the Pearson correlation coefficient was 0.999. The Bland-Altman analysis showed a mean bias of 1.8 % with a (95 % CI 1.4 to 2.2) (Figure 3B). The data scatter was independent of the analyte concentration, ranging from −1.9 to 5.6 % (lower limit CI interval from −2.6 to −1.3 %, upper limit CI interval 4.9–6.2 %).

Figure 3:

Results from the patient sample-based mycophenolic acid glucuronide comparison study performed between two independent laboratories. (A) Passing-Bablok regression plot including the Pearson regression analysis for the method comparison study of the RMP (n=100 patients) between the independent laboratories (site 1 Dr. Risch Ostschweiz AG and site 2 Roche Diagnostics GmbH). Passing-Bablok regression analysis resulted in a regression equation with a slope of 1.02 (95 % CI 1.01 to 1.03) and an intercept of 0.03 (95 % CI –0.24 to 0.25). The Pearson correlation value was ≥0.999. (B) Bland-Altman plot resulted in an inter-laboratory measurement bias of 1.8 % (95 % CI interval from 1.4 to 2.2 %) and a 2S interval of the relative difference of 3.7 % (lower limit CI interval from −2.6 to −1.3 %, upper limit CI interval from 4.9 to 6.2 %).

Precision performance evaluation within the second laboratory showed CVs for repeatability and intermediate precision of ≤2.7 % and ≤4.4 %, respectively. These data indicate that the proposed RMP for MPAG is transferable between laboratories, emphasizing the robustness of the protocol design.

Uncertainty of results

The estimation of uncertainty of the calibrators was done as a type B evaluation. In the precision experiment all other aspects, e.g., calibration, sample preparation, measurement and evaluation of the sample result, were evaluated as type A uncertainty. Measurement uncertainties (k=1) for single measurements of spiked analyte-free human serum samples ranged from 1.2 to 3.8 % and was reduced to 0.9–1.6 % for target value assignment (n=6, three measurements on two days) (Tables 4 and 5). Consequently, the expanded measurement uncertainties (k=2) ranged from 2.4 to 7.7 % for single measurements and from 1.8 to 3.3 % for target value assignment.