*3.3. Method Validation: Recovery, Matrix E*ff*ect, and Stability*

≤ ≤ ≤ As shown in Table 2, we assessed the recovery and matrix effect of the method for REP and CEL at the four different QC levels and for IS at 50 ng/mL. The mean recovery of REP and CEL was observed to be 98.5–104% with CV values of ≤2.57%. There were no significant differences in recovery values among the four different QC levels (*p* = 0.066 for REP and 0.502 for CEL), indicating concentration-independent recovery for both drugs. The mean matrix effect for REP and CEL was observed to be 92.1–102% with CV values of ≤5.25%. The stability was assessed under various handling and storage conditions relevant to this HPLC-FL method. Bench-top stability, autosampler stability, freeze–thaw stability, and long-term stability were determined for REP and CEL at the four different QC levels. The extent of bias in the concentration was within ±15% of the corresponding nominal value, while the remaining fraction of REP and CEL was observed to be 89.2–106% with CV values of ≤6.04%, as shown in Table 3. These data clearly indicate that the sample preparation procedure employed in the bioanalytical method proposed herein offered sufficient extraction recovery with minimal matrix effect, and that REP and CEL remained stable under several conditions related to the present bioanalytical procedures.


**Table 2.** Recovery and matrix effect of REP, CEL, and IS in rat plasma (*n* = 5).

**Table 3.** Stability (%) of REP and CEL in rat plasma (*n* = 5).


<sup>a</sup> Room temperature for 3 h. <sup>b</sup> 10 ◦C for 1 day in the autosampler. <sup>c</sup> Three freezing and thawing cycles. <sup>d</sup> −20 ◦C for 30 days.

### *3.4. Pharmacokinetic Drug Interaction Studies*

Rats received oral REP (0.4 mg/kg) and CEL (2 mg/kg) either alone or in combination. Then, plasma concentration versus time profiles of REP and CEL were evaluated as shown in Figure 3. The relevant pharmacokinetic parameters are listed in Table 4. The oral doses used were selected based on previous rat pharmacokinetic studies on REP or CEL [34–36]. After oral dosing of REP, plasma REP levels increased for 20 to 45 min and then declined in a multi-exponential fashion. The AUClast, AUCinf, and *t*1/<sup>2</sup> of REP were not significantly changed by concurrent administration with oral CEL, as shown in Table 4. After oral administration of CEL, its plasma concentration profiles markedly fluctuated during the whole period of blood collection (480 min). Thus, the AUCinf and *t*1/<sup>2</sup> of CEL could not be determined in this study because there was no discernible linear terminal phase observed in the plasma concentration versus time curves of CEL. The multiple peaks in the plasma concentration profiles of CEL may be caused by slow and variable gastrointestinal absorption, which warrants further investigation. Notably, the AUClast of CEL was significantly higher (by 76.2%) after co-administration of CEL and REP than after administration of CEL alone (*p* = 0.0213). Because CEL is eliminated primarily by extensive metabolism [15], the increased systemic exposure of oral CEL could be attributable to a reduction of hepatic first-pass and/or systemic metabolism of CEL caused by concurrent administration of REP.

**Figure 3.** Plasma concentration versus time profiles of REP (**A**) and CEL (**B**) after oral administration of 0.4 mg/kg REP and 2 mg/kg CEL either alone (closed circle) or in combination (open circle) to rats. The circles and vertical bars represent means and standard deviations, respectively (*n* = 5).

**Table 4.** Pharmacokinetic parameters of REP and CEL in rats after oral administration of 0.4 mg/kg REP and 2 mg/kg CEL either alone or in combination (*n* = 5).


\* Significantly different from the single group (*p* < 0.05).

μ μ Previously reported pharmacokinetic parameters of intravenous REP and CEL in rats are listed in Table S1. Because the blood-to-plasma concentration ratio (RB) was 0.61 for REP and 2.66 for CEL in rat blood (our in-house data), the blood CL (CLB) was determined to be 8.52 mL/min/kg for REP and 2.92 mL/min/kg for CEL (calculated as plasma CL/RB). Because the urinary excretion of the unchanged drug was reported to be negligible for both drugs, as shown in Table S1, it is plausible to assume that the CL<sup>B</sup> of REP and CEL could represent their hepatic clearance, which is far below the reported hepatic blood flow rate in rats (QH; ranging from 50 to 80 mL/min/kg) [37]. This indicates that REP and CEL are drugs with low hepatic extraction ratios of 0.037 to 0.170 (calculated as CLH/QH). Based on the well-stirred model, the CL<sup>H</sup> of a drug with a low hepatic extraction ratio primarily depends on its intrinsic metabolic clearance (CLint) and fraction unbound in blood (fB) [30]. Thus, the hepatic metabolism and plasma protein binding of the drugs were further investigated using an in vitro rat and human liver microsomes and plasma. As shown in Figure 4, dose-response curves for the inhibitory effect of REP on the metabolism of CEL were constructed in RLM and HLM. REP significantly inhibited the metabolic reaction of CEL with IC<sup>50</sup> values of 16.1 ± 4.5 µM in RLM and 14.4 ± 0.6 µM in HLM. However, the metabolism of REP in RLM and HLM was not significantly altered by CEL (data not shown). Additionally, protein binding interactions between the two drugs in rat and human plasma were assessed. As shown in Figure 5, there were no significant differences in fractions of unbound drugs either alone or in combination, suggesting the minimal possibility of protein binding-based interactions between the two drugs.

**110 120**

**Figure 4.** Dose-response curves for the inhibitory effect of REP on metabolic reactions of CEL in RLM (**A**) and HLM (**B**). The circles and vertical bars represent the means and standard deviations, respectively (*n* = 4).

**REP with CEL CEL with REP Figure 5.** Fraction of unbound REP (**A**) and CEL (**B**) either alone or in combination in rat and human plasma (*n* = 3).

μ μ μ μ Since oral AUC is calculated as F × D/CL (F, oral bioavailability; D, dose; CL, total clearance), an increase in F and/or decrease in CL result in an increase in AUC. Moreover, hepatic metabolism is the major elimination route for both REP and CEL that are drugs with a low hepatic extraction ratio. Thus, the inhibition of hepatic metabolism of the two drugs can reduce their hepatic first-pass effect (increase in F) and hepatic systemic clearance (decrease in CL), consequently leading to an increase in AUC. In our present in vitro metabolism study in RLM and HLM, as shown in Figure 4, the metabolism of REP was not significantly changed by CEL, while the metabolism of CEL was inhibited by REP with a mean IC<sup>50</sup> of 16.1 µM in RLM and 14.4 µM in HLM. Assuming that the in vivo concentration levels of REP in the rat liver after oral dosing are high enough to inhibit the metabolism of CEL, the increased oral systemic exposure of CEL by concurrent administration of REP (AUClast in Table 4) could be attributable to a reduction of hepatic first-pass effect and/or hepatic systemic clearance of CEL caused by the inhibitory activity of REP on the hepatic metabolism of CEL.

The present rat study highlighted the possibility for metabolism-based interactions between CEL and REP in clinical settings. As shown in Figure 4, REP significantly inhibited the metabolic reaction of CEL with comparable IC<sup>50</sup> values between RLM and HLM. The *C*max of REP administered orally to rats was reported to be 297 ± 103 ng/mL (dose: 0.4 mg/kg) in this study, as shown in Table 4, and 105.1 ± 30 ng/mL (dose: 0.5 mg/kg) in a previous study [35], which are roughly comparable to the reported *C*max of 65.8 ± 30.1 ng/mL (dose: 4 mg; converted to 0.41 mg/kg in rats based on the

human equivalent dose concept proposed by the FDA) in humans (FDA drug label information). Moreover, there were no significant differences in the unbound fraction of REP between rat and human plasma, as shown in Figure 5. Based on these findings, it is plausible that the increased systemic exposure of CEL by co-administration of REP in the present rat pharmacokinetic study could have some clinical relevance, depending on species differences in hepatic distribution profiles of REP between rats and humans.

Additionally, there have been no reported studies on the relationships between the AUC and toxicity of CEL. However, the FDA drug label of CEL indicates that the steady-state AUC of CEL is increased about 40% and 180% by mild (Child-Pugh Class A) and moderate (Child-Pugh Class B) hepatic impairment, respectively. Thus, the daily recommended dose of CEL should be reduced by approximately 50% in patients with moderate (Child-Pugh Class B) hepatic impairment. In our present study, the AUClast of CEL in rats was observed to be 189 (ranging from 113 to 285) µg·min/mL after administration of CEL alone and 333 (ranging from 458) µg·min/mL after administration of CEL and REP in combination. If these rat data could be extrapolated to humans, it is plausible that careful toxicity monitoring and/or dose modification may be needed for the combined dose of CEL and REP in clinical practice. Despite intrinsic limitations associated with nonclinical studies, our present in vivo rat and in vitro HLM data warrant further clinical study on drug interactions between REP and CEL.
