*2.8. LC-MS*/*MS Analysis*

#### 2.8.1. In Vitro Samples

Metabolites of nine P450-selective substrates were analyzed using a Shimadzu Nexera X2 UPLC system coupled to an LCMS-8050 triple quadruple mass spectrometer (Shimadzu Corporation, Kyoto, Japan) equipped with an electrospray ionization interface as previously described with a slight modification [17,18]. Separation was performed on a reversed-phase column (Luna C18, 50 mm × 2.0 mm i.d.; 3 μm particle size; Phenomenex, Torrance, CA, USA) maintained at 40 ◦C. The mobile phase consisted of distilled water containing 0.1% formic acid (A) and acetonitrile containing 0.1% formic acid (B), with a flow rate of 0.5 mL/min. The gradient elution program used was as follows: (1) Mobile phase A was set to 95% at 0 min, (2) a linear gradient was run to 5% in 2.6 min, and (3) a linear gradient was run to 95% in 3.0 min and re-equilibrated for 2 min. The total run time was 5 min. The optimized compound-dependent parameters of the metabolites of the nine P450-selective substrates and the internal standard are listed in Table 1. Three-day validations were performed to confirm the e ffectiveness of the LC-MS/MS system for simultaneous determination of the nine P450-selective substrate metabolites at the respective ranges of 0.01–10 μM in blank microsomal incubation mixtures. We found that the precision (≤12.1%) and accuracy (95.4–110.2%) values were within acceptable ranges. Supplemental Figure S1 shows the representative LC-MS/MS chromatograms of a human liver microsomal incubation sample containing nine P450-selective metabolites and an internal standard.

The auto-optimized mass transitions were *m*/*z* 312 > 231 and *m*/*z* 437 > 207.1 for quantification of 4-hydroxydiclofenac and losartan carboxylic acid, respectively. HPLC conditions were the same as those in the cocktail assay.

#### 2.8.2. In Vivo Samples

The plasma concentrations of diclofenac and 4-hydroxydiclofenac were determined by a previously reported LC-MS/MS method [26] with some modifications. Briefly, 50 μL aliquots of plasma were extracted with 300 μL aliquots of acetonitrile containing chlorpropamide (internal standard), followed by LC-MS/MS (Shimadzu Corporation). Chromatographic separation was performed on a Phenomenex Luna C18 column (100 × 2.00 mm; 3.0 μm). The isocratic mobile phase consisted of 0.1% formic acid in distilled water (A) and 0.1% formic acid in acetonitrile (B) (45:55, *v*/*v*), with a flow rate of 0.3 mL/min. The transitions were *m*/*z* 296.0 > 214.0 for diclofenac, *m*/*z* 312 > 231 for 4-hydroxydiclofenac, and *m*/*z* 277 > 111 for the internal standard. The data acquisition was computed using LabSolutions LCMS Ver.5.6 (Shimadzu Corporation). The calibration curves for diclofenac and 4-hydroxydiclofenac were linear (*r* ≥ 0.996) over the concentration range of 20–5000 ng/mL.

The LC-MS/MS condition for the determination of SPN in plasma was the same with a previously reported method [27]. The calibration curve for SPN was linear (*r* ≥ 0.995) over the concentration range of 1–250 ng/mL.

#### *2.9. Analysis of Inhibition Kinetics and Pharmacokinetic Parameters*

The IC50 values were calculated via nonlinear least-squares regression analysis from logarithmic plots of inhibitor concentration versus percentage of activity remaining after inhibition, using SigmaPlot (ver. 14.0; Systat Software Inc, Chicago, IL, USA). The *K*i values were determined from the equations for a single substrate single inhibitor model and the software available in the SigmaPlot Enzyme Kinetics module. Competitive, non-competitive, uncompetitive, or mixed inhibition models were evaluated and ranked according to the best fit based on Akaike Information Criterion (AIC) values. For visual inspection, the data were presented as Dixon plots.

Pharmacokinetic parameters were calculated by a non-compartmental analysis using WinNonlin Professional software (version 5.2, Pharsight Corp., Mountain View, CA, USA) that used the total area under the plasma concentration–time curve from time zero to infinity (AUC∞) or the last measured time (AUCt). The logarithmic trapezoidal rule was used during the declining plasma level phase and the linear trapezoidal rule was used for the rising plasma-level phase. The peak plasma concentration (Cmax) and time to reach Cmax (Tmax) were read directly from the experimental data. Statistically significant differences were recognized at *p* < 0.05.
