*2.4. Sample Preparation*

For deproteinization, 400 µL ice-cold ACN containing IS (50 ng/mL) was added to 120 µL plasma samples. The resultant mixture was vortex-mixed for 5 min, followed by centrifugation at 15,000× *g* for 5 min. Next, 400 µL supernatant was transferred to another microtube and dried by N<sup>2</sup> gas stream. For reconstitution, 60 µL mobile phase was added to the resultant residue, and after sufficient vortex-mixing, 20 µL finally prepared sample solution was injected to the HPLC system.

#### *2.5. Chromatographic Conditions*

In this study, we used a Shimadzu HPLC system (Shimadzu Co., Kyoto, Japan) equipped with a fluorescence detector (RF-20A), column oven (CTO-20A), autosampler (SIL-20AC), and pump (LC-20AT). A Kinetex C18 column (250 × 4.6 mm, 5 µm, 100 Å; Phenomenex, Torrance, CA, USA) protected by a C18 guard column (SecurityGuard HPLC Cartridge System; Phenomenex) at 40 ◦C was

used for chromatographic separation. Isocratic elution of mobile phase consisting of 10 mM phosphate buffer (pH 6.0) and ACN (53.6:46.4, *v*/*v*) was performed at a flow rate of 1 mL/min. The injection volume and total run time were 20 µL and 23 min, respectively. The fluorescence excitation and emission wavelengths for REP, CEL, and IS were 240 and 380 nm, respectively.

#### *2.6. Method Validation*

This new bioanalytical method for simultaneous determination of REP and CEL was validated based on the US-FDA guidelines [28]. The selectivity was assessed based on the comparison among chromatograms of REP, CEL, and IS in blank rat plasma; blank rat plasma spiked with REP, CEL, and IS; and rat plasma sample obtained from a pharmacokinetic study in rats. The presence of endogenous interferences at the acquisition windows of the analytes was examined.

The linearity was determined by the addition of increasing amounts of REP and CEL to a blank biological matrix. Calibration curves (*n* = 5) were constructed by plotting the peak area ratios of analytes to IS (*y*-axis) versus the concentration ratios of REP and CEL (10–2000 ng/mL) to IS (50 ng/mL) in plasma (*x*-axis), and linear regression analysis was conducted using the least squares method with a weighting factor of 1/*x* (*x* = concentration). The sensitivity was assessed based on LLOQ, defined as the lowest quantifiable concentration levels of REP and CEL in calibration curves (signal-to-noise [S/N] ratio of more than 5). REP and CEL peaks at the LLOQ level should be identifiable, reproducible, and discrete with acceptable accuracy (within 80–120%) and precision (<20%).

The precision and accuracy were estimated by comparison between the measured concentrations and their respective nominal concentrations in the QC samples, which were prepared as five separate sets on one day (intra-day) and five different days (inter-day). Precision was expressed as a coefficient of variation (CV) of the mean values of the measured concentration. Accuracy was expressed as a relative error between the measured and nominal concentrations. They were determined with plasma samples spiked with REP and CEL at the four different QC levels in five replicates.

The extraction recovery and matrix effect were determined by comparison among the analytical signals (peak area) obtained from (A) the extracted sample, (B) the post-extracted spiked sample (extracts of blanks spiked with the analyte post extraction), and (C) non-extracted neat sample (diluted stock solution). The recovery was calculated as 'A/B × 100', and the matrix effect was calculated as 'B/C × 100'. Five replicates were assessed at the four different QC levels.

The stability was assessed by comparison of the analytical signals (peak area) obtained from plasma samples exposed to various handling and storage conditions with those obtained from plasma samples. Bench-top stability was determined by exposing spiked plasma samples to room temperature for 180 min. Freeze–thaw stability was determined by exposing spiked plasma samples to freeze–thaw cycles (from −20 ◦C to room temperature) three times on consecutive days. Long-term stability was determined by storing spiked plasma samples at −20 ◦C for 30 days. Autosampler stability (post-preparative stability) was determined by exposing extracted plasma samples to 25 ◦C for 1 day in an autosampler. The stability was determined at the four different QC levels.

#### *2.7. In Vivo Pharmacokinetic Study in Rats*

Rats were fasted for 12 h prior to the pharmacokinetic experiment and then anesthetized with zoletil (intramuscular, 20 mg/kg). The femoral artery and vein of the rats were cannulated with a polyethylene tube (BD Medical; Franklin Lakes, NJ, USA) at 240 min prior to drug dosing. A single oral dose of REP alone (0.4 mg/kg), CEL alone (2 mg/kg), or REP and CEL at the same doses was administered to the rats (*n* = 5 per group). Drugs were dissolved in a vehicle that is a clear mixture of DMSO, ethanol, polyethylene glycol 400, and saline at a ratio of 1:5:30:64 (*v*/*v*/*v*/*v*). Approximately 300 µL aliquots of blood were collected in heparin pre-treated microcentrifuge tubes via the femoral artery at 0, 10, 20, 30, 45, 60, 90, 120, 180, 240, 360, and 480 min after the oral dosing. Following centrifugation of blood samples at 2000× *g* at 4 ◦C for 10 min, 120 µL aliquots of plasma were stored at −80 ◦C until HPLC analysis.
