**3. Results**

#### *3.1. Reversible Inhibition of (-)-Sophoranone toward the Nine CYP Isoforms in Human Liver Microsomes*

The inhibitory effects of SPN on the activities of nine CYP isozymes (CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4) in human liver microsomes are shown in Figure 2, and the IC50 values are listed in Table 2. The IC50 values for the positive controls used in the reversible inhibition studies were in an acceptable degree of accuracy with published values [12,19–21]. Of the P450 isoforms tested, SPN exerted the strongest inhibitory effect on CYP2C9-catalyzed tolbutamide hydroxylation, with an IC50 value of 0.966 ± 0.149 μM (Table 2). SPN showed weak inhibitory effects toward CYP2C8 and CYP2C19, with IC50 values of 13.6 ± 3.15 μM and 16.8 ± 3.21 μM, respectively. However, SPN had no apparent inhibitory effects toward the other CYPs tested (Table 2); the residual enzyme activities at the highest tested concentration (50 μM) were greater than 80%, except for CYP2D6 (53.9 ± 3.53%) and CYP3A4 (53.3 ± 4.00%) (Figure 2).

**Figure 2.** Inhibition curves of SPN on the nine major P450 activities in human liver microsomes using substrate cocktails including CYP1A2 for phenacetin *O*-deethylase (**A**), CYP2A6 for coumarin 7-hydroxylase (**B**), CYP2B6 for bupropion hydroxylase (**C**), CYP2C8 for rosiglitazone *p*-hydroxylase (**D**), CYP2C9 for tolbutamide 4-hydroxylase (**E**), CYP2C19 for omeprazole 5-hydroxylase (**F**), CYP2D6 for dextromethorphan *O*-demethylase (**G**), CYP2E1 for chlorzoxazone 6-hydroxylase (**H**), and CYP3A4 for midazolam 1-hydroxylase (**I**). The activity is expressed as a percentage of remaining activity compared with the control, no containing SPN. Data are the mean ± standard deviation of triplicate incubations. The dashed lines represent the best fit to the data with non-linear regression.

To determine whether the inhibitory effects of SPN on CYP2C9 was substrate specific, we examined the inhibitory effects on other CYP2C9-specific biotransformation pathways (i.e., diclofenac 4-hydroxylation and losartan oxidation) and found that SPN also markedly inhibited their activities, with IC50 values of 0.879 ± 0.0888 μM and 0.455 ± 0.0486 μM, respectively, (Figure 3).

**Figure 3.** Inhibition curves of SPN on the CYP2C9-catalyzed diclofenac 4-hydroxylation (**A**) and losartan oxidation (**B**) activities in human liver microsomes. Data are the mean ± standard deviation of triplicate incubations. The dashed lines represent the best fit to the data with non-linear regression.

#### *3.2. Determination of the Ki of (-)-Sophoranone for CYP2C9 Activity*

Based on the lowest IC50 value for CYP2C9, to characterize the type of reversible inhibition of CYP2C9 by SPN, enzyme kinetic experiments were performed in the presence of various concentrations of SPN and tolbutamide, or diclofenac. Otherwise, identical samples containing a known potent CYP2C9 inhibitor (sulfaphenazole), were included in the analysis. Representative Dixon plots of CYP2C9 inhibition by SPN and sulfaphenazole in human liver microsomes are shown in Figure 4, and the Ki values are summarized in Supplemental Table S1. Using a nonlinear regression analysis, SPN demonstrated competitive inhibition against CYP2C9-catalyzed tolbutamide hydroxylation or diclofenac hydroxylation, with calculated *K*i values of 0.503 ± 0.0383 μM and 0.587 ± 0.0470 μM (Figure 4A,B). Sulphafenazole competitively inhibited CYP2C9 with a *K*i value of 0.267 ± 0.0170 μM (Figure 4C), which was similar to a previously reported value [28].

**Figure 4.** Dixon plots to determine *K*i values of SPN on the CYP2C9 enzyme activity, using tolbutamide (**A**) or diclofenac (**B**) as substrates. The well-known inhibitor of CYP2C9, sulfaphenazole, is used as a positive control (**C**) using tolbutamide as a substrate. The concentrations of tolbutamide were determined 50 (•), 100 (-), and 150 (-) μM, respectively; diclofenac was used at 2 (•), 5 (-), and 10 (-) μM, respectively. *v* represents formation rate of 4-hydroxytolbutamide (nmol/min/mg protein) or 4-hydroxydiclofenac (pmol/min/mg protein). Data are the mean ± standard deviation of triplicate incubations. The dashed lines of SPN (**A**,**B**) and sulfaphenazole (**C**) all fit well to competitive inhibition types.

#### *3.3. Time-Dependent Inactivation of (-)-Sophoranone towards the Nine CYP Isoforms in Human Liver Microsomes*

The IC50 shift method incorporating a dilution is one of the most efficient and convenient methods for evaluating time-dependent inhibitory effects. A shift in IC50 to a lower value ("shift") with pre-incubation indicates time-dependent inactivation [29–31]. After 30 min pre-incubation of SPN with human liver microsomes in the presence of NADPH, no obvious shift in IC50 was observed for inhibition of the nine CYPs (Figure 5), suggesting that SPN is not a time-dependent inactivator for the nine CYPs.

**Figure 5.** Time-dependent inhibition curves of SPN on the nine major P450 activities in human liver microsomes using substrate cocktails after 30 min pre-incubation with the presence (•) or absence (-) of an NADPH-generating system. Data are the mean ± standard deviation of triplicate incubations.

#### *3.4. Caco-2 Cell Permeability of (-)-Sophoranone*

A bi-directional permeability assay using Caco-2 monolayer cells was performed to predict the intestinal absorption of SPN. SPN showed very low permeability in both directions (from A-to-B and B-to-A). The calculated Papp values of SPN from A-to-B were (0.115 ± 0.0369) × 10−<sup>6</sup> cm/s and (0.172 ± 0.0488) × 10–<sup>6</sup> cm/s at 10 μM and 50 μM, respectively, (*n* = 3, each). These results indicated that SPN is poorly absorbed in vivo. The Papp values from B-to-A were (0.101 ± 0.00444) × 10−<sup>6</sup> cm/s at 10 μM (*n* = 3) and (0.152 ± 0.0353) × 10−<sup>6</sup> cm/s at 50 μM (*n* = 3). SPN was not a substrate for efflux transporters, that is, P-gp and BCRP, as the efflux ratio (B-to-A/A-to-B) is less than 2. The Papp of propranolol, a reference high permeable compound, from A-to-B and B-to-A were (26.8 ± 3.31) × 10−<sup>6</sup> cm/s and (21.5 ± 2.19) × 10−<sup>6</sup> cm/s, respectively, (*n* = 3, each), similar to the reported values [24,25].

#### *3.5. E*ff*ects of (-)-Sophoranone on the Pharmacokinetics of Diclofenac in Rats*

We conducted pharmacokinetic studies to investigate the e ffects of SPN on the pharmacokinetics of diclofenac in rats. Findings in the literature on the dried *S. tonkinensis* herbs indicate that a recommended daily dose for an adult human with the body weight of 60 kg were to be 6–10 g [32], which correlated to the equivalent dose ranges in rats, 0.620–1.03 g/kg [33]. He et al. [2] reported that the average contents of SPN in various *S. tonkinensis* samples were found to be approximately 2.53 mg/g (0.0253%). Reflecting this content, the dosage in rats, 0.620–1.03 g/kg of the dried herb, might be consistent with 15.7–26.1 mg/kg in terms of SPN. Thus, in this study, the SPN dose of 75 mg/kg was used in rats, which is approximately 2.87- to 4.87-fold greater than the recommended human dose.

The mean plasma concentration-time profiles of diclofenac and 4-hydroxydiclofenac after oral administration of diclofenac (2 mg/kg) in the absence or presence of oral co-administration of SPN (75 mg/kg) in rats are illustrated in Figure 6, and the relevant pharmacokinetic parameters are shown in Table 3. The plasma levels of diclofenac and 4-hydroxydiclofenac were similar in both groups (Figure 6A,B). Likewise, no significant di fferences were observed in any other pharmacokinetic parameter of diclofenac and 4-hydroxydiclofenac (Table 3). The in vivo marker for CYP2C9 activity, expressed as the molar AUC ratio of 4-hydroxydiclofenac to diclofenac, was not significant (0.799 ± 0.167 versus 0.904 ± 0.0534; *p* value of 0.215) in the presence or absence of SPN (Table 3). In the treatment group with co-administration of SPN, the Cmax of SPN was found to be 33.7 ± 14.8 ng/mL (0.0732 ± 0.0321 μM) at approximately 60–75 min post-dose (Figure 6C). Given the *K*i values of SPN on CYP2C9 activity (0.503 ± 0.0383 μM for tolbutamide hydroxylation and 0.587 ± 0.0470 μM for diclofenac hydroxylation), the plasma concentrations of SPN are too low to inhibit CYP2C9-mediated metabolism of diclofenac in vivo. Overall, the co-administration of SPN did not alter the pharmacokinetics of diclofenac and 4-hydroxydiclofenac.

**Figure 6.** Mean plasma concentrations of diclofenac ( **A**) and 4-hydroxydiclofenac (**B**) after oral administration of diclofenac at a dose of 2 mg/kg without (•, *n* = 6) or with ( -, *n* = 6) oral dosing of SPN (75 mg/kg) to rats. Mean plasma concentrations of SPN ( **C**) after co-administration of SPN (75 mg/kg) and diclofenac (2 mg/kg) to rats (, *n* = 6). Vertical bars mean standard deviation.


**Table 3.** Mean (±standard deviations) pharmacokinetic parameters of diclofenac and 4-hydroxydiclofenac after oral administration of diclofenac at a dose of 2 mg/kg without or with oral administration of SPN (75 mg/kg) to rats.

No significant differences were observed in all pharmacokinetic parameters of diclofenac and 4-hydroxydiclofenac. a Total area under the plasma concentration–time curve from time zero to time last sampling time; b total area under the plasma concentration–time curve from time zero to infinity; c terminal half-life; d peak plasma concentration; e time to reach Cmax. Median (ranges); f the metabolic conversion ratio, AUC<sup>∞</sup>,4-hydroxydiclofenac/AUC<sup>∞</sup>,diclofenac, was calculated based on a molar basis.

#### *3.6. Determination of the Unbound Fraction of (-)-Sophoranone in Plasma and Human Liver Microsomes*

SPN was extensively bound to plasma proteins, regardless of species. The free fractions (%) of SPN at 10 and 50 μM in human plasma were 0.0457 ± 0.00612% and 0.0927 ± 0.0400%, respectively, (*n* = 3, each). Similarly, when 10 and 50 μM SPN were added to the rat plasma, the free fractions were 0.0380 ± 0.0102% and 0.0531 ± 0.0149%, respectively, (*n* = 3, each). After adding 10 and 50 μM SPN to rat and human plasma, free fractions remained relatively unchanged, suggesting that SPN has no binding saturation in plasma.

SPN also exhibited marked non-specific bindings to human liver microsomes, although to a lesser extent than those in human plasma. The unbound fractions of SPN at 10 and 50 μM were calculated to be 0.621 ± 0.0405% and 0.724 ± 0.170%, respectively (*n* = 3, each), at a microsomal protein concentration of 0.1 mg/mL.
