**4. Discussion**

Pharmacokinetic drug interactions involving drug absorption should be considered for optimum drug therapy, apart from the drug interactions attributed to the oxidative metabolism via the CYP-450 system of di fferent isozymes [40]. Ostensibly harmless natural products—such as juices, fruits, vegetables and herbal products in the form of ayurvedic medicine—have been reported in several studies to potentially cause many drug interactions a ffecting drug absorption mediated by transporters [41,42]. For example, emodin—a potential antineoplastic drug and a major component of the *Rhamnus, Rumex, Aloe*, *Rheum* and *Cassia* species—has been reported to be a possible *P*-gp inducer [26] or an inhibitor [18].

This study evaluated the e ffects of *R. acetosa* extract on the drug transporters discussed above, as well as its potential for drug interactions, while presenting a clear view of the interactions of emodin with the transporter *P*-gp. The major six anthraquinones present in *R. acetosa* were shown in our previous study [27]. A prior cytotoxicity assay was performed to establish the working range for the extract suitable for optimal viability of the cells during the experiment. Afterwards, the e ffects of these six anthraquinones on *P*-gp were demonstrated individually with an MDR assay kit using Caco-2 cells. Verapamil, being an inhibitor of *P*-gp, served as a positive control. Only groups treated with chrysophanol-8- *O*-β-<sup>d</sup>-glucoside and emodin showed higher fluorescence intensity than the control group, with average values of 121.4% ± 2.3% and 147.2% ± 12.4%, respectively, suggesting *P*-gp inhibition. This result is consistent with those obtained in a study by Min et al. [18], in which emodin was shown to inhibit *P*-gp. On the other hand, the results from the *P*-gp inhibition test of *R. acetosa* extract sugges<sup>t</sup> no significant inhibition of the e fflux transporter, as opposed to the emodin and chrysophanol-8- *O*-β-<sup>d</sup>-glucoside, which in contrast showed significant inhibition of the *P*-gp transporter when treated individually. A possible explanation is that the emodin content may not be high enough to exert its inhibitory e ffect in the extract. Chemical contents of herbal plant extracts can vary depending on various factors such as climate, harvesting seasons and extraction solvent. The probability of inhibition of *P*-gp by *R. acetosa* extract cannot be ruled out.

OATP1A2—the uptake transporter used in our in vitro test—is widely expressed in the intestines and serves as a major uptake mechanism for fexofenadine [43,44]. Sometimes, a substrate of *<sup>P</sup>*-gp—such as this study's selected model drug, fexofenadine—can also be a substrate for the OATP uptake transporter [43,44], making it necessary to di fferentiate between the contributions of *P*-gp and OATP to potential drug interactions and those of other simultaneously administered drugs that could a ffect these transporters. Therefore, our in vitro studies were also performed with HEK293 cells transfected with the polypeptide transporter OATP1A2. *R. acetosa* extract was found to inhibit the uptake of fexofenadine through in vitro studies. In other words, the uptake of fexofenadine by OATP1A2 into cells declined when *R. acetosa* extract was used as a co-treatment. This result suggests that *R. acetosa* extract can a ffect the absorption of fexofenadine through the inhibition of OATP1A2.

A pharmacokinetic study was designed to verify the results of our in vitro study in view of the observed herbal extract's drug interactions at the uptake transporter for fexofenadine in rats. Rat model is considered unsuitable to predict metabolic drug interaction in human [45]. However, there is a correlation in drug intestinal permeability with both carrier-mediated absorption and passive

di ffusion mechanisms between rat and human [46]. Because the property of our selected model drug, fexofenadine, has little metabolism, it is reasonable to use the rat model for predicting the intervention of extract on absorption. All rats were divided into 3 groups: an emodin administration group, an *R. acetosa* administration group and a control group. Eleven milligrams per kilogram of emodin, 2 g/kg of *R. acetosa* extract and 0.5% CMC as a control was administered orally to each group. Fexofenadine at the dose of 10 mg/kg was given orally to each group after 30 min. The results showed a smaller AUC of fexofenadine (132.1 ± 50.3 ng·h/mL) in the *R. acetosa* group in comparison to that of the control group, in which the AUC was 222.0 ± 92.1 ng·h/mL. These results sugges<sup>t</sup> decreased absorption of fexofenadine in the rats treated with *R. acetosa* extract. In other words, the gu<sup>t</sup> uptake transporter OATP1A2, which is responsible for fexofenadine absorption, was inhibited, as predicted by the in vitro results. Moreover, the alteration on the solubility of fexofenadine was also observed by *R. acetosa* extract through the physicochemical interaction study. The FT-IR spectra results sugges<sup>t</sup> that there is no functional group interaction between fexofenadine and the component of *R. acetosa* extract. The fexofenadine solubility in SIF changed from 1.03 ± 0.04 mg/mL to 0.83 ± 0.10 mg/mL after mixing with the extract. It means that the solubility alteration could also be the reason for the decreased fexofenadine AUC by *R. acetosa* extract because fexofenadine HCl is Biopharmaceutics Classification System (BCS) class 3 drug with high solubility and low permeability. Drug interactions due to changes in solubility can be avoided by adjusting the administration time. *R. acetosa* extract contains many kinds of compounds, not only anthraquinones, but also flavonoids and polysaccharides [15]. They have also the possibility of interference with the drug absorption through the intervention to the transporters [47,48]. Particularly, one of the flavonoids of *R. acetosa* extract, epicatechin-3- *O*-gallate [49], also has an inhibitory e ffect on the OATP1A2 [50]. Moreover, there was the possibility that *R. acetosa* extract may change the gastric emptying time [51,52] and the pH in the gastro-intestinal tract when coadministered with the fexofenadine. The e ffects of anthraquinones on OATP have been rarely reported. Further studies are needed to elucidate the components in *R. acetosa* extract responsible for inhibition of fexofenadine absorption. Meanwhile, emodin increased the AUC for fexofenadine, possibly via the inhibitory e ffect on an e fflux transporter of fexofenadine, *P*-gp [32], the e ffect of which on the uptake transporter of fexofenadine has ye<sup>t</sup> to be fully understood.

Given the evidence from both in vitro and in vivo studies, *R. acetosa* extract should be used with caution when substrates of the drug transporters or poorly water-soluble drugs are prescribed.
