*3.2. Cytotoxicity of Mitoxantrone in the Presence of Quercetin*

To further confirm the effect of quercetin on the reversal of BCRP-mediated chemoresistance in HeLa cells, we examined the cytotoxicity (i.e., anticancer activity) of mitoxantrone in the absence and presence (1 or 100 µM) of quercetin. In this study, CCK-8 was used for the examination of mitoxantrone-associated cytotoxicity. As shown in Figure 3, mitoxantrone displayed concentration-dependent cytotoxicity in HeLa cells, which was further boosted in the presence of 1 µM Ko143, a stereotypical BCRP inhibitor. Likewise, the presence of 1 or 100 µM quercetin effectively enhanced the cytotoxicity associated with mitoxantrone as the IC<sup>50</sup> decreased to 19.3% (1.13 µM) or 8.2% (0.478 µM), respectively, which differed from that observed with mitoxantrone alone (5.83 µM; Figure 3A). In addition, the cytotoxicity of quercetin alone without mitoxantrone was also examined. Treatment with 100 µM quercetin alone led to no significant changes in cell viability in comparison with the control (0.1% DMSO), demonstrating that the increased cytotoxicity observed in mitoxantrone-treated cells was not likely associated with the toxicity of quercetin (Supplementary Figure S2). μ μ μ μ μ μ

μ μ **Figure 3.** Effect of co-incubation of mitoxantrone (MX) with (**A**) quercetin (1 or 100 µM) and (**B**) Ko143 (1 µM) on the cell viability of HeLa cells. The Cell Counting Kit-8 (CCK-8) assay was used to determine the cytotoxicity associated with the cellular accumulation of MX after 24 h of incubation. Asterisks indicate statistical differences (\* *p* < 0.05; \*\* *p* < 0.01; and \*\*\* *p* < 0.001) from the control group (i.e., without the quercetin or Ko143) according to one-way ANOVA, followed by Tukey's post hoc test. Data are presented as the mean ± SD of quintuplicate runs.

#### *3.3. Bi-Directional Transport Study in MDCKII*/*BCRP Cells*

μ

μ μ μ μ μ We performed bi-directional transport studies in MDCKII cells expressing human BCRP (MDCKII/BCRP) to investigate the in vitro inhibitory potency of quercetin against BCRP in a concentration-dependent manner. Co-incubation with quercetin increased the Papp, A-to-B of prazosin (Figure 4A) while simultaneously decreasing the Papp, B-to-A (Figure 4B) with an increasing concentration of quercetin, leading to a concentration-dependent decrease in the overall ER (Figure 4C). Additionally, the functional expression of the efflux transporter in MDCKII/BCRP cells was also confirmed in this study, with an ER of 5.4 for prazosin (the stereotypical substrate of BCRP [27,35,36]), which decreased to 0.9 in the presence of the known inhibitor Ko143 (Figure 5C). Notably, the inhibitory effect of 10 µM quercetin on the B-to-A transport and efflux ratio was comparable to 1 µM Ko143 (Figure 5; *p* > 0.05). At quercetin concentrations higher than 10 µM, the ERs were less than 1.2, indicating the nearly complete inhibition of prazosin efflux (the complete inhibition of efflux would theoretically result in an ER of ~1, Figure 5). Kinetic analysis of the transport process yielded an estimated IC<sup>50</sup> value of 4.22 µM for quercetin. Assuming the mechanism of inhibition to be competitive, the inhibitory constant (Ki) value was then estimated to be 3.91 µM using the K<sup>m</sup> value of 128 µM [27] for prazosin.

μ **Figure 4.** Bi-directional transport of prazosin in BCRP-overexpressing Madin-Darby Canine Kidney-II (MDCKII/BCRP) cells under various concentrations of quercetin (0.1–300 µM). (**A**) Apical-to-basolateral apparent permeability coefficient (Papp, A-to-B) and (**B**) basolateral-to-apical apparent permeability coefficient (Papp, B-to-A) of prazosin. (**C**) The percentage of the control efflux ratio (%ER, compared to the value without inhibitor) is shown together with the best-fit values generated from the nonlinear regression analysis based on Equation (2). Asterisks indicate statistical differences (\* *p* < 0.05; \*\* *p* < 0.01; and \*\*\* *p* < 0.001) from the control (i.e., without quercetin) according to one-way ANOVA, followed by Tukey's post hoc test. Data are presented as the mean ± SD of triplicate runs. Data are presented as the mean ± SD of triplicate runs. μ

μ μ μ μ μ μ μ μ **Figure 5.** Effect of 10 µM quercetin or 1 µM Ko143 on the apparent permeability coefficient and efflux ratio of prazosin, a BCRP substrate, in MDCKII/BCRP cells. (**A**) Apical-to-basolateral apparent permeability coefficient (Papp, A-to-B), (**B**) basolateral-to-apical apparent permeability coefficient (Papp, B-to-A), and (**C**) efflux ratios of prazosin in the absence of inhibitor (i.e., the control) or in the presence of quercetin (10 µM) or Ko143 (the standard inhibitor of BCRP; 1 µM). Asterisks indicate statistical differences (\* *p* < 0.05; \*\* *p* < 0.01; and \*\*\* *p* < 0.001) from the control group (i.e., without the inhibitor) according to one-way ANOVA, followed by Tukey's post hoc test. Data are presented as the mean ± SD of triplicate runs.

## *3.4. Oral Pharmacokinetic Study in Rats with or without Quercetin*

To investigate the possible pharmacokinetic impact of quercetin as a BCRP inhibitor, we performed an oral pharmacokinetic study with sulfasalazine, a BCRP substrate, in rats. In this study, the change in the plasma concentration of sulfasalazine was used as an indicator of the in vivo interaction of BCRP with quercetin. To our knowledge, sulfasalazine has only limited interactions with other efflux transporters, including P-gp and MRP2 [34], whereas prazosin (the substrate used in the bi-directional transport study) is a dual substrate of P-gp and BCRP in vivo [37]. Thus, sulfasalazine is considered a relatively selective in vivo probe substrate of BCRP [25,26]. The mean plasma concentration–time profiles following the oral administration of 2 mg/kg sulfasalazine with or without pretreatment with 10 mg/kg quercetin in rats are shown in Figure 6. The pharmacokinetic parameters, as estimated using non-compartmental analysis, are summarized in Table 1. The plasma AUC8h of sulfasalazine with or without quercetin pretreatment was 44.5 ± 11.8 min·µg/mL and 25.7 ± 9.98 min·µg/mL, respectively; this value was higher by 1.8-fold in the quercetin pretreatment group than in the control group, but it

was not significantly different (*p* < 0.05, power < 0.8). More importantly, the Cmax was significantly higher by 1.5-fold (*p* < 0.05, power > 0.8) in the quercetin pretreatment group (179 ± 23.0 ng/mL) than in the control group (i.e., 122 ± 23.2 ng/mL), whereas there was no significant change in the elimination half-life (t1/2) of sulfasalazine. Collectively, these results suggest that pretreatment with quercetin led to the increased oral absorption of sulfasalazine in vivo.

∙μ

● ■ **Figure 6.** Temporal profiles of orally administered sulfasalazine (2 mg/kg) with or without the pre-administration of quercetin (10 mg/kg). Key: Control (•; without quercetin), quercetin pre-administration (). Asterisks indicate statistical differences from the control (i.e., without quercetin) according to a two-tailed/unpaired Student's t-test (\* *p* < 0.05, power > 0.8). Data are expressed as the mean ± SD of quadruplicate runs.

**Table 1.** Pharmacokinetic parameters of sulfasalazine after its oral administration (2 mg/kg dose) with and without pretreatment with quercetin (10 mg/kg) in rats. Data are expressed as the mean ± SD (*n* = 4 per group).


\* significantly different from the control (i.e., without the pre-administration of quercetin) (*p* < 0.05, power > 0.8).

#### **4. Discussion**

Increasing lines of evidence from animal and human studies regarding food–drug interactions have indicated that a wide range of flavonoids can interact with ABC transporters, thereby leading to overexposure or underexposure of clinically important substrate drugs [13]. However, the accurate prediction of such interactions has been found to be difficult owing to limited in vitro data. The objective of this study was to investigate the inhibitory potential of quercetin against BCRP in vitro and in vivo. This study, which integrated the in vitro and in vivo effects of quercetin, was indeed necessary because a thorough understanding of the pharmacokinetic influence of this flavonoid is needed because of its high dietary intake as well as the lack of clear corresponding pharmacokinetic data.

Here, we demonstrated that the presence of quercetin can effectively enhance the cellular accumulation and associated cytotoxicity of mitoxantrone in HeLa cells (Figures 2 and 3), consistent with previous reports [38]. In the current study, the efficacy of quercetin as a BCRP inhibitor was quantitively demonstrated via a significant reduction in the IC<sup>50</sup> of mitoxantrone even in the presence of quercetin at a concentration as low as 1 µM (i.e., it decreased to 19.3% of the control value; from 5.83 to 1.13 µM). When the concentration of quercetin increased to 100 µM, the IC<sup>50</sup> of mitoxantrone was further decreased (i.e., to 8.23% of the control value; 0.48 µM), similar to that in the presence of Ko143 (i.e., 0.62 µM), a stereotypical BCRP inhibitor. In addition, pharmacokinetically relevant parameters were obtained in a bi-directional transport study using MDCKII/BCRP cells, where the IC<sup>50</sup> values of quercetin for the inhibition of BCRP-mediated efflux were estimated to be 4.22 µM. Assuming the mechanism of inhibition to be competitive, the IC<sup>50</sup> value was further transformed to a Ki value of 3.91 µM, using the K<sup>m</sup> value of 128 µM [27] for prazosin. The values obtained in the bi-directional transport studies were comparable to those previously observed in MCF-7/MX and MDCKII/BCRP cells using Hoechst 33342 accumulation (IC<sup>50</sup> values of 7.6 and 6.9 µM, respectively) [21]. In both assays, it was shown that while quercetin is a less potent inhibitor compared to Ko143, it can show a similar inhibitory effect compared to 1 µM Ko143 in higher concentrations (Figures 3 and 5).

The US Food and Drug Administration recommends that orally administered compounds with an [Igut] value (the maximal gastrointestinal concentration; defined as the dose divided by 250 mL) divided by the Ki value greater than 10 be evaluated for potential in vivo interactions [36]. For quercetin, the estimated [Igut] value (662 µM, assuming a dietary quercetin intake of 50 mg/day) or even the estimated intestinal concentration (86.2 µM, when the intestinal fluid volume is assumed to be 1.92 L [39]) divided by the Ki (3.91 µM) value is far greater than 10. Thus, although the bioavailability of quercetin is somewhat low [40] and the daily dietary intake reportedly results in sub-micromolar concentrations in circulation [3], the substantially higher concentration in the gut is likely to result in the inhibition of intestinal BCRP and thereby an increase in the intestinal absorption of BCRP transporter substrates.

Consequently, the in vivo inhibitory potency of quercetin was further assessed to clarify its interaction with intestinal BCRP. In this study, the pharmacokinetic profile of orally administered sulfasalazine was used as an indicator of any alterations in intestinal BCRP activity. While sulfasalazine has been reported to be effluxed by P-gp and MRP2 to a low extent, previous studies have consistently demonstrated that the intestinal absorption of the compound following its oral administration was essentially unaffected in P-gp- or MRP2-knockout rats in contrast to the significantly higher AUC8h and Cmax values observed in BCRP-knockout rats [24], strongly suggesting that sulfasalazine is a good probe for observing intestinal BCRP activity. In this study, higher AUC8h and Cmax values of sulfasalazine (1.8-fold (*p* < 0.05, power < 0.8) and 1.5-fold (*p* < 0.05, power > 0.8), respectively) were observed in the presence of 10 mg/kg quercetin than in its absence (Table 1). The increased absorption in the presence of quercetin is clearly significant, but the degree is somewhat lower than that expected considering the approximately 20-fold increase observed in knockout rats [24] and, especially, the low Ki value of the flavonoid obtained in the current study. One possible reason for this discrepancy might be the rapid conjugation of quercetin to quercetin-3-glucuronide that occurs in the small intestine [20]. Once quercetin enters the intestinal cells by passive diffusion or uptake by the uptake transporters, it is subjected to glucuronidation by a UDP-glucuronosyltransferase present in both rat and human intestines [41–43], which results in the rapid clearance of quercetin from sites adjacent to the efflux transporter. Indeed, the oral bioavailability of quercetin was only 5.3% and the Cmax value was the sub-micromolar range (i.e., 0.21 µg/mL) following 10 mg/kg oral administration to rats [44]. Another possibility that might result in relatively limited alterations in sulfasalazine absorption is the involvement of OATP2B1 in the intestinal absorption of sulfasalazine [25]. Sulfasalazine is a high-affinity substrate of OATP2B1 [25,45], whereas quercetin has been reported to be an inhibitor of OATP2B1 [46]. Therefore, the relatively low increase in sulfasalazine exposure in the presence of quercetin may be attributed to complex interactions between the simultaneous inhibition of the efflux by BCRP and the uptake by OATP2B1. In addition, considering that we only observed a single dosing of quercetin on the sulfasalazine pharmacokinetics, further studies regarding multiple dosing of quercetin are likely needed.

In a previous study by Zhang et al., an apparent discrepancy between the in vitro and in vivo inhibition of BCRP by the flavonoids chrysin and 7,8-benzoflavone was reported. In their investigation, the flavonoids were demonstrated to be potent inhibitors of human BCRP but weak inhibitors of mouse BCRP [23]; one possible explanation for this discrepancy may be species differences between the human and rodent transporters. Although a further study regarding food–drug interaction is required in humans, this may also be true for quercetin, in which case the clinical impact of the modulation of BCRP activity in humans may be much greater than that estimated from pharmacokinetic studies performed in rats.
