**4. Discussion**

This study focused on the in vitro and in vivo inhibitory effects of SPN on human CYPs, especially CYP2C9. We screened the inhibitory effects of SPN on the major human CYP isoforms (CYP1A2, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6, CYP2E1, and CYP3A4) in human liver microsomes. Of the nine tested CYP isoforms, SPN exerted the strongest inhibitory effect on CYP2C9 activity, with the lowest IC50 value of 0.966 ± 0.149 μM (Table 2; Figure 2). In addition to CYP2C9, SPN mildly inhibited several CYP enzymes, with potency ranked in the order CYP2C8 > CYP2C19; the IC50 values were 13.6 ± 3.15 μM and 16.8 ± 3.21 μM, respectively (Table 2; Figure 2). Although the IC50 values could not been calculated, SPN also appears to weakly inhibit CYP2D6 and CYP3A4; the residual enzyme activities at the highest tested concentration (50 μM) were 53.9 ± 3.53% and 53.3 ± 4.00%, respectively (Figure 2). No apparent inhibition of the other CYPs (CYP1A2, CYP2A6, CYP2B6, and CYP2E1) was observed (Figure 2). SPN also strongly inhibited other CYP2C9-catalyzed diclofenac 4-hydroxylation and losartan oxidation activities (Figure 3). The inhibition mechanisms of SPN on CYP2C9-catalyzed tolbutamide 4-hydroxylation and diclofenac 4-hydroxylation activities were both competitive, with *K*i values of 0.503 ± 0.0383 μM and 0.587 ± 0.0470 μM, respectively. Pre-incubation of SPN for 30 min with human liver microsomes and an NADPH-generating system did not alter the inhibition potencies against the nine CYPs, suggesting that SPN is not a time-dependent inactivator.

The reversible inhibition of SPN-mediated CYP3A4 activity was less consistent with the published literature. Li et al. [16] reported that among 44 tested flavonoids, SPN inhibited CYP3A4-catalyzed bufalin 5-hydroxylation activity with a *K*i value of 2.17 ± 0.29 μM. They only focused on the in vitro inhibitory potentials of several flavonoids against CYP3A4 activity. To the best of our knowledge, to date, bufalin has not been used as the in vitro probe substrate for the CYP3A4 activity, and the reference material of 5-hydroxybufalin is not commercially available. Because of the presence of several binding regions within the CYP3A4 active site, multiple probe substrates are often used for in vitro CYP3A4-mediated drug–drug interaction studies, including midazolam, nifedipine, and testosterone [34]. In that study, when other CYP3A4 substrates were tested, the ranges of IC50 values by SPN were reported to be 5.62–38.4 μM [16]. Additionally, we examined the inhibitory effect on another CYP3A4-catalyzed testosterone 6β-hydroxylation and found that SPN also inhibited the activity with an IC50 value of 31.5 ± 4.79 μM, which showed a higher percentage inhibition compared to midazolam (data not shown). Altogether, the in vitro CYP3A4 inhibition by SPN seemed to be substrate-specific.

Generally, alterations in the activities of hepatic CYPs through in vitro inhibition or induction represent the major mechanisms underlying pharmacokinetic drug–drug interactions [11–13]. It has been estimated that CYP2C9 is responsible for the metabolic clearance of up to 15–20% of all drugs undergoing phase I metabolism, including clinically important drugs such as *S*-warfarin, phenytoin, tolbutamide, losartan, and several anti-inflammatory drugs [23,35]. Considering that SPN is a potent CYP2C9 inhibitor in vitro, there may be potential for herb–drug interactions between SPN and CYP2C9 substrates after concomitant oral administration.

Using the in vitro reversible inhibition results, a clinical drug–drug interaction risk was initially predicted by the basic static model approach, as recommended by the EMA [36] and US FDA [37] with calculating the R1 value (R1 = 1 + [Imax,u/*K*i,u]), which representing the predicted AUC ratio in the presence or absence of inhibitor. Where, Imax,u (Cmax,u) is maximal free plasma concentration of the inhibitor and *<sup>K</sup>*i,u is the unbound in vitro inhibition constant. However, little information is ye<sup>t</sup> to be reported on the Cmax values of SPN after oral administration of SPN. As stated in the Introduction, from our previous study, the Cmax of SPN was reported to be 13.1 ng/mL in rats after oral dosing of 12.9 mg/kg SPN in rats [8]. Thus, we investigated whether SPN affects the pharmacokinetics of diclofenac and 4-hydroxydiclofenac, produced by hepatic CYP2C9 enzyme, in rats. In the group that received co-administration of SPN (75 mg/kg), the Cmax of SPN was found to be 33.7 ± 14.8 ng/mL (0.0732 ± 0.0321 μM) at 60–75 min (Figure 6C). These results sugges<sup>t</sup> that SPN has low oral bioavailability. The calculated values of Imax, u and *<sup>K</sup>*i,u for SPN used in this study were 0.0420 ± 0.0184 nM and 3.39 ± 0.258 nM (3.95 ± 0.316 nM for diclofenac 4-hydroxylation), respectively. Considering these values, the R1 value of SPN for the inhibition of CYP2C9 in vitro was calculated as 1.0124 (*K*i, u for tolbutamide 4-hydroxylation) or 1.0106 (Ki, u for diclofenac 4-hydroxylation) which are both below the EMA and US FDA cut-off criteria of R1, 1.02 [36,37], indicating that the potential for clinically relevant drug interaction-mediated CYP2C9 inhibition by SPN may be low and no clinical interaction studies are warranted. In our results, also no significant differences were observed in any of the other pharmacokinetic parameters of diclofenac and 4-hydroxydiclofenac in rats in the absence or presence of oral co-administration of SPN at a dose of 75 mg/kg (Table 3). Furthermore, the molar metabolic conversion ratio, expressed as AUC4-hydroxydiclofenac/AUCdiclofenac, which indicated a causal factor for the evaluation of the capacities of CYP2C9 activity in vivo, did not show significant differences (0.799 ± 0.167 versus 0.904 ± 0.0534) in both groups (Table 3).

To explain the lack of in vitro–in vivo correlation, we assessed two factors that could limit the accuracy of in vitro models in predicting metabolic drug interactions in vivo, which were SPN's degree of plasma protein binding and its permeability in Caco-2 cells. We found that SPN was extensively bound in both human and rat plasma proteins (>99.9%) with a mean unbound fraction value of 0.0574% in the range of 10 and 50 μM. Thus, taking the plasma protein binding of SPN into account, the unbound

maximum concentrations of SPN in plasma might be 0.0420 ± 0.0184 nM, which is much lower than the unbound *K*i values of SPN in vitro. Some drugs that indicate in vitro–in vivo discrepancy because of high plasma protein bindings have been reported [38–40]. Tolfenamic acid strongly inhibited CYP1A2 in vitro but not in vivo because of high plasma protein binding (99.7%) [38]. Montelukast is a very potent inhibitor of CYP2C8 in vitro with *K*i values ranging from 0.0092–0.15 μM [41]. However, in humans, montelukast has had no e ffect on the pharmacokinetics of the CYP2C8 substrates, pioglitazone [39] and rosiglitazone [40]. The high degree of protein binding of montelukast in plasma (>99.7%) is similar to that of tolfenamic acid and explicitly explains the lack of its in vivo interaction, irrespective of its strong inhibitor potency in vitro. The Caco-2 cell model is widely used to predict the absorption across the intestinal barrier, and a good correlation between its oral absorption in humans and its apparent permeability (Papp) across the Caco-2 cell barrier has been shown [24,25]. A recent study has provided some updated guidelines on how permeability values might correlate with human oral absorption: Low permeability (0–20% absorbed) is correlated to Papp values < 1–2 × 10−<sup>6</sup> cm/s; moderate permeability (20–80% absorbed) to Papp values < 2–10 × 10−<sup>6</sup> cm/s; and high permeability (80–100% absorbed) to Papp values > 10 × 10−<sup>6</sup> cm/s [42]. Propranolol had >90% human absorption and exhibited high permeability with a Papp value of (26.8 ± 3.31) × 10−<sup>6</sup> cm/s in our assay. SPN exhibited a very low permeability with mean Papp values of 0.115 × 10−<sup>6</sup> cm/s (0.429% of propranolol Papp) and 0.172 × 10−<sup>6</sup> cm/s (0.642% of propranolol Papp) at 10 and 50 μM, respectively, indicating that it is poorly absorbed in vivo. SPN was not a substrate for e fflux transporters, that is, P-gp and BCRP, as the e fflux ratio (B-to-A/A-to-B) is less than 2.

Overall, SPN is a potent inhibitor of CYP2C9 in vitro but not in vivo. This apparent discrepancy is due to the extensive plasma protein binding and very low permeability of SPN, which resulted in poor oral absorption. These approaches could help in making more reliable in vitro–in vivo extrapolations about the risk of in vivo inhibition potential. In conclusion, these findings have provided useful information on the safe and e ffective use of SPN in clinical practice.

**Supplementary Materials:** The following is available online at http://www.mdpi.com/1999-4923/12/4/328/s1, Figure S1: Typical LC-MS/MS chromatograms for the formed metabolites of the nine CYP-specific probe in human liver microsomes and their internal standard in the positive electrospray ionization mode. Table S1: *K*i values and inhibition types for CYP2C9 by SPN and sulfaphenazole in human liver microsomes (*n* = 3).

**Author Contributions:** Conceptualization, Y.F.Z., S.H.B., and S.K.B.; methodology, Z.H., S.U.C., C.B.L., and D.K.; software, Z.H. and S.U.C.; validation, Y.F.Z., Z.H., S.U.C., C.B.L., S.J.J., and H.J.S.; formal analysis, Y.F.Z., S.U.C., and S.J.J.; investigation, Z.H., C.B.L., D.K., and H.Y.; resources, D.K. and H.Y.; data curation, Y.F.Z., S.H.B., and S.K.B.; writing—original draft preparation, Y.F.Z. and S.H.B.; writing—review and editing, S.H.B. and S.K.B.; visualization, S.J.J. and H.J.S.; supervision, S.K.B.; project administration, S.K.B.; funding acquisition, S.K.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by Industrial Core Technology Development Program funded by the Ministry of Trade, Industry and Energy (No. 10063475) and Technology Program for establishing biocide safety managemen<sup>t</sup> funded by the Ministry of Environment (No. 1485016231) of the Korean government, and Research Fund of The Catholic University of Korea (2018).

**Conflicts of Interest:** The authors declare no conflict of interest. Soo Hyeon Bae is the employee of the company Q-fitter, Inc. Doyun Kim, Hunseung Yoo are the employees of the company SK Chemicals. The companies had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

**Previous Presentation of Information:** A portion of the information in this manuscript has been presented as a poster in a conference previously: Soo Hyun Jang, Yu Fen Zheng, Chae Bin Lee, Soo Kyung Bae. 2019. Cytochrome P450 2C9 inhibition by sophoranone in human liver microsomes. 15th Euro-Global Summit on Toxicology and Applied Pharmacology, Berlin, Germany.
