Interactions of the active components of Punica granatum (pomegranate) with the essential renal and hepatic human Solute Carrier transporters

Abstract

Context: Solute carrier transporters (SLCs) are membrane proteins responsible for cellular influx of various substances including many pharmaceutical agents; therefore, they largely impact on drug disposition and elimination in body. Punica granatum Linnaeus (Lythraceae), pomegranate, is a fruit with antidiabetic potential. Oleanolic acid (OA), ursolic acid (UA), and gallic acid (GA) are the major bioactive components of pomegranate. Co-administration of these compounds with other drugs could result in altered drug pharmacokinetics, possibly due to competing for transporter proteins.

Objective: We investigated the interactions of these three compounds with the essential hepatic and renal SLC transporters.

Materials and methods: Uptake of radiolabeled transporter model substrates was assessed in HEK293 cells over-expressing SLC transporters including the organic anion transporters (OATs), organic anion transporting polypeptides (OATPs) and organic cation transporters (OCTs), in the presence or absence of 10.0 µM UA, OA, or GA. Their IC₅₀ values on specific SLC transporters were also evaluated using varying concentrations of the particular compound (ranging from 0.10 nM to 80.0 µM).

Results: Our results demonstrated UA could significantly inhibit OAT3 and OATP2B1 uptake (IC₅₀: 18.9 ± 8.20 µM and 11.0 ± 5.00 µM, respectively) and GA has a pronounced inhibitory effect on OATP1B3 uptake (IC₅₀: 1.60 ± 0.60 μM).

Discussion and conclusion: Our study reports the interactions of OA, UA, and GA with the essential SLC transporters. This information may contribute to elucidating the drug–drug/herb interactions involved with these three compounds and form the basis of therapeutic optimization when drugs are co-administered.

• Drug–drug/herb interaction
• gallic acid
• oleanolic acid
• ursolic acid

Introduction

The solute carrier transporters (SLCs) are a superfamily of membrane proteins responsible for the translocation of various substances crossing biological membranes (Degorter et al., 2012). Previous studies showed that individual members of the SLC superfamily have specific tissue localization suggesting their distinct roles in drug absorption, distribution, and elimination. Two primary subfamilies of SLCs, SLCO and SLC22A, represent the essential membrane transporters expressed in the kidney and liver. The SLCO family refers to the organic anion transporting polypeptides (OATPs), whereas the SLC22A family includes the organic cation transporters (OCTs) and the organic anion transporters (OATs) (Roth et al., 2012). It is known that the substrate spectra of these transporters are remarkably broad ranging from endogenous substances to various xenobiotics such as many pharmaceutical drugs (Roth et al., 2012). Among all the SLCO and SLC22A transporters, several isomembers have been well studied and proved to play critical roles in drug performance. OATP1A2, OAT4, OCTN1, and OCTN2 were expressed at the apical membrane of renal tubule epithelium in charge of the reabsorption from or the secretion of xenobiotics into urine (Roth et al., 2012). OAT1, OAT2, OAT3, OCT2, and OCT3 localized at the basolateral membrane of proximal tubule cells were responsible for the uptake of xenobiotics from blood (Roth et al., 2012). OATP1B1, OATP1B3, OATP2B1, OAT2, OCT1, and OCT3 present at the basolateral membrane of hepatocytes assist the absorption of drugs into liver (Roth et al., 2012), while OATP1A2 was also found in cholangiocytes, where it was involved with the reabsorption of xenobiotics from the bile duct (Roth et al., 2012). Overall, these SLC transporters are the key determinants of drug disposition and elimination in the body, and, therefore, largely impact the therapeutic efficacy and toxicity of agents.

Punica granatum Linnaeus (Lythraceae), pomegranate, is an ancient and mystical fruit, which has been widely consumed by various cultural systems for centuries (American Diabetes Association, 2012; Viuda-Martos et al., 2010). In addition to its historical uses, pomegranate also has anti-inflammatory and antioxidant effects (Faria & Calhau, 2011). In ancient Chinese medicine, pomegranate and its extracts were employed to treat acidosis, hemorrhage, diarrhea, helminthiasis, and microbial infections (Lei et al., 2007). More recently, numerous scientific studies have shown that pomegranate also exerts antidiabetic potential (Esmaillzadeh, 2004; Huang et al., 2005a; Li et al., 2005, 2008; Parmar & Kar, 2007). Studies indicated that the major bioactive components of pomegranate include oleanolic acid (OA), ursolic acid (UA), and gallic acid (GA) (Figure 1) (Jang et al., 2009; Katz et al., 2007; Mondal et al., 2012; Teodoro et al., 2008). OA and UA are triterpenoid compounds with similar chemical structures and widely distributed in the kingdom of plants (Liu, 1995). OA has been shown to have glucose-lowering properties in animals via stimulating insulin secretion (Teodoro et al., 2008; Zhang et al., 2003), reducing insulin resistance (Sato et al., 2007), enhancing acetylcholine release (Hsu et al., 2006) and activating PPAR-α (Huang et al., 2005a). Insulin level could be elevated by UA through attenuating the influx of glucose (Jayprakasam et al., 2005), hyperglycemia, or hepatic glucose production (Jang et al., 2010). GA, a type of phenol, is found abundantly in tea, grapes, different berries, fruits, as well as wine (Manach et al., 2005). In addition to its anti-inflammatory (Kroes et al., 1992) and antioxidant effects (Kim et al., 2002), GA also exhibits antidiabetic activity in various animal studies. Therefore, it has been considered to be beneficial for diabetes mellitus treatment as well (Makihara et al., 2012; Mondal et al., 2012; Punithavathi et al., 2011).

Figure 1. Chemical structures of ursolic acid, oleanolic acid, and gallic acid.

However, co-administration of these compounds with other drugs have been reported to be problematic possibly due to drug–drug/herb interactions, which lead to altered pharmacokinetics of drugs (Zhou et al., 2004). When co-administered, rosuvastatin (a known substrate of OATPs (Ho et al., 2006; Kitamura et al., 2008)) together with UA or GA, an increased systemic exposure and reduced clearance of rosuvastatin have been observed in rats (Basu et al., 2012; Wen & Xiong, 2011). In the previous study of Xu et al. (2013a,b), OA significantly altered the pharmacokinetic behavior of swertiamarin in rats. Increasing evidence indicated that the drug–drug/herb interactions involved with UA, OA, and GA are clinically important, which might lead to an altered therapeutic outcome in patients.

Drug–drug/herb interactions may result from competing for membrane transporters such as OATPs and OATs (Wang & Sweet, 2012; Xu et al., 2013a, b; Zhao et al., 2012). Thus, exploiting how these SLCs interact with natural compounds is important to prevent adverse events associated with drug–drug/herb interactions. Currently, limited information is available regarding the interactions of UA, OA, and GA with renal and hepatic SLC transporters. Early studies showed that both OA and UA had predominant inhibitory effects on the transport of estradiol-17β-glucuronide (E17β) mediated via OATP1B1 (Roth et al., 2011; Wu et al., 2012). Gui et al. (2010) found that UA could also potently inhibit OATP1B3-mediated fluorescein-methotrexate (FMTX) uptake. Until now, GA has only been reported to inhibit the uptake of OAT1 and OAT3 (Wang & Sweet, 2012). Apart from these, there is no further information indicating the interactions between these compounds and the other essential renal and hepatic SLCs. The present study then explored the influence of UA, OA, and GA on the uptake of specific substrates mediated through a number of essential renal and hepatic SLCs not previously investigated. Elucidating the interaction between these bioactive compounds and the essential SLC transporters will greatly contribute to understanding drug–drug/herb interactions. It will also form the basis of optimizing therapeutic outcomes when drugs are co-administered with these three compounds in treating diseases, including diabetes.

Materials and methods

Materials

[³H]-4-Aminohippuric acid (PAH, 60.0 Ci/mmol), [³H]-l-ergothioneine (1.70 Ci/mmol), and [¹⁴C]-l-carnitine (56.0 mCi/mmol) were purchased from BioScientific Pty. Ltd., (Gymea, NSW, Australia). [³H]-Estrone-3-sulfate (E3S; 57.3 Ci/mmol), [³H]-cholecystokinin octapeptide (CCK-8, 97.5 Ci/mmol), and [³H]-methyl-4-phenylpyridinium acetate (MPP⁺, 82.1 Ci/mmol) were obtained from PerkinElmer (Melbourne, VIC, Australia). Culture media were purchased from Thermo Scientific (Lidcombe, NSW, Australia). OA, UA, and GA were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). All other chemicals were purchased from Sigma-Aldrich (Castle Hill, NSW, Australia).

The plasmids containing the full-length human OAT1 (reference sequence: NM 004790.4), OAT2 (reference sequence: NM 006672.2), OAT3 (reference sequence: NM 004254.2), OCT1 (reference sequence: NM 003057.2), OCT2 (reference sequence: NM 003058.2), and OCT3 (reference sequence: NM 021977.2) cDNAs were purchased or cloned by us as described before (Xu et al., 2013a,b; Zhou et al., 2010).

Expression of the essential renal and hepatic SLC transporters in HEK-293 cells

HEK293 cells were seeded onto human fibronectin coated 48-well plates at an initial density of 1.60 × 10⁵ cells/well and maintained at 37 °C and a humidified 5% CO₂ atmosphere in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. After 24 h, the cells were transfected with plasmid DNA using Lipofectamine 2000 Reagent (Invitrogen, Mount Waverley, VIC, Australia) following the manufacturer's instructions. Twenty-four hours after transfection, substrate transport activities were measured.

Transport functional study

Cellular uptake of the radiolabeled typical substrate of each SLC transporter was initiated at room temperature in phosphate-buffered saline (PBS; pH7.4, containing 154 mM NaCl, 3.00 mM Na₂HPO₄, 1.10 mM K₂HPO₄, 1.00 mM MgCl₂, and 1.00 mM CaCl₂) in the presence or absence of 10.0 μM concentration of test compounds. Substrate conditions used in the study were compatible with previous publications: 300 nM [³H]-E3S for OAT3, OAT4, OATP1A2, OATP1B1, and OATP2B1 (Noé et al., 2007; Ueo et al., 2005; Ugele et al., 2008; Zhou et al., 2010, 2011); 5.00 µM [³H]-PAH for OAT1 (Race et al., 1999); 2.00 nM [³H]-CCK-8 for OATP1B3 (Hirano et al., 2004); 100 nM [³H]-MPP⁺ for OCT1, OCT2, and OCT3 (Amphoux et al., 2010); 5.00 µM [³H]-l-ergothioneine for OCTN1 (Gründemann et al., 2005); 5.00 µM [¹⁴C]-l-carnitine for OCTN2 (Tamai et al., 1998; Toh et al., 2011). After 8 min, cells were quickly rinsed two times with ice-cold PBS, lysed in 0.20 M NaOH, neutralized with 0.20 M HCl, and aliquoted for liquid scintillation counting. The uptake was corrected for substrate accumulation in corresponding empty vector control cells and expressed as pmol/mg/protein⁻¹/min⁻¹. Exposure to up to 80.0 µM of OA, UA, and GA had no influence on the substrate signal observed in HEK empty vector control cells relative to unexposed cells (data not shown), indicating that the accumulation of prototypical substrates in corresponding empty vector control cells in the presence of test compounds represented appropriate values for background correction. Data are presented as mean ± SD (n = 3).

For each potent inhibitor, concentration-dependent inhibition assays were conducted to derive the IC₅₀ value (the concentration of test compound required to inhibit 50% of transporter function), using varying concentrations of the particular compound (ranged from 0.10 nM to 80.0 µM). Results were confirmed by repeating for three times with triplicate wells for each data group in every experiment.

Statistics

Statistical difference was detected by Student's t-test. The criterion of significance was taken to be p < 0.01.

Results

Functional characterization of SLCs-expressing HEK293 cells

In the present study, HEK293 cells transiently expressing the essential hepatic and renal SLC transporters were successfully established, particularly for OAT1, OAT2, OAT3, OAT4, OATP1A2, OATP1B1, OATP1B3, OATP2B1, OCT1, OCT2, OCT3, OCTN1, and OCTN2. As illustrated in Figure 2, the HEK293-SLC cells were characterized by a significant accumulation of [³H]- or [¹⁴C]-labeled prototypical substrates compared with that of vector transfected HEK293 cells. Furthermore, time dependence of SLC transporter-mediated uptake was evaluated in the cells mentioned above, which indicated that the uptake velocity of all prototypical substrates maintained linear up to 10 min (data not shown).

Figure 2. Functional characterization of the HEK293 cells over-expressing various hepatic and renal SLC transporters. Eight minutes cellular uptake of 5.00 μM [³H]-PAH (for hOAT1); or 0.30 μM [³H]-E3S (for hOAT3, hOAT4, OATP1A2, OATP1B1, and OATP2B1); or 0.50 μM [³H]-E3S (for hOAT2); or 2.00 nM [³H]-CCK-8 (for OATP1B3); or 0.10 μM [³H]-MPP⁺ (for hOCT1, hOCT2, and hOCT3); or 5.00 μM [³H]-l-ergothioneine (for hOCTN1); or 5.00 μM [¹⁴C]-l-carnitine (for hOCTN2) were measured in HEK293 cells transfected with particular transporter plasmids (black bars) and vector (white bars). All data were presented as mean ± SD of three independent experiments (triplicate repeats in each experiment).

Interaction of oleanolic acid with the essential hepatic and renal SLCs

The inhibitory effect of OA on the uptake of the essential hepatic and renal SLCs was assessed (Figure 3A). Considering the plasma concentrations of UA, OA and GA reported in the literature were lower than 10.0 μM (Shahrzad et al., 2001; Song et al., 2006; Zhang et al., 2013; Zhu et al., 2013), we selected 10.0 μM as the inhibitor concentration in our analysis. The presence of 10.0 μM OA slightly inhibited the uptake of OAT1, OAT2, and OAT3 (∼35.0% reduction in transport function). It showed minimal inhibition on the influx of the other SLC transporters evaluated in our study. Interestingly, OA even showed slightly inductive effect on hOAT4-mediated [³H]-E3S uptake and hOCTN1-mediated [³H]-l-ergothioneine uptake (28.8% and 14.3%, respectively).

Figure 3. Inhibitory effects of OA, UA, and GA on cellular uptake mediated by the essential hepatic and renal SLCs. Cellular uptake was measured in HEK293 cells expressing particular transporter in the presence (black bars) or absence of OA (A), UA (B) and GA (C) at 10.0 μM (white bars). The results were expressed as the percentage of control. Each column were shown as the mean ± SD, three independent experiments were conducted (triplicate repeats in each experiment) (*p < 0.05; **p < 0.01).

Interaction of ursolic acid with the essential hepatic and renal SLCs

To determine whether UA could influence the uptake of these SLCs, inhibition studies were performed accordingly (Figure 3B). The significant inhibition of OAT3- and OATP2B1-mediated [³H]-E3S uptake (60.9% and 56.0% inhibition, respectively) and mild to no inhibition on the uptake of the other transporters were observed in the presence of 10.0 μM UA. Our dose–response studies obtained the IC₅₀ values of UA on [³H]-E3S transport mediated by OAT3 and OATP2B1, which were 18.9 ± 8.20 μM and 11.0 ± 5.00 μM, respectively (Figure 4).

Figure 4. Inhibitory potency of UA and GA on the uptake of specific SLC transporters. Cellular uptake of radiolabeled substrates was measured in the absence or presence of UA or GA (concentration ranged from 0.10 nM to 80.0 µM) in HEK293 cells over-expressing OAT3, OATP2B1, or OATP1B3. The data were expressed as the percentage of control. All data were presented as mean ± SD; three independent experiments were conducted (triplicate repeats in each experiment).

Interaction of gallic acid with the essential hepatic and renal SLCs

The influence of GA on the essential hepatic and renal SLCs transport function was also examined. As shown in Figure 3(C), GA remarkably inhibited OATP1B3-mediated [³H]-CCK-8 transport by 54%, and mildly reduced the transport function of OAT2, OATP2B1, and OCTN2 by 23.3, 21.0, and 20.0%, respectively. Additionally, with the presence of GA, the uptake of OCTN1 was stimulated by 31.3%. Our dose–response study revealed that the inhibition of GA on the transport of [³H]-CCK-8 mediated by OATP1B3 was with an IC₅₀ value of 1.60 ± 0.60 μM (Figure 4).

Discussion

Pomegranate, an edible fruit (Jain et al., 2012), is widely used as a traditional medicine to treat a variety of ailments, such as ulcers, diarrhea, and acidosis (Jain et al., 2012; Lei et al., 2007). More recently, an increasing number of scientific studies have demonstrated that pomegranate-derived products exhibit great potential to treat diabetes mellitus, which is characterized by hyperglycemia with long-term complications including retinopathy and nephropathy (American Diabetes Association, 2012; Huang et al., 2005b; Katz et al., 2007; Li et al., 2005). Although the molecular mechanisms underlying this effect have yet to be adequately defined, it is known that the bioactive compounds of pomegranate, OA, UA, and GA have glucose lowering (Ban et al., 2008; Jang et al., 2010; Teodoro et al., 2008), neuroprotective (Ban et al., 2008; Hong et al., 2012) and nephroprotective properties (Nabavi et al., 2013; Pai et al., 2012). Thus, they have been considered as antidiabetic agents (Katz et al., 2007). However, when co-administrated with other drugs, there is great concern that drug–drug/herb interactions involved with three compounds may occur, possibly via competing for transporter proteins.

Our data showed that OA minimally impacted the uptake of the hepatic and renal SLC transporters assessed in our study (Figure 3A). Interestingly, its isomer, UA, notably inhibited OAT3- and OATP2B1-mediated [³H]-E3S transport with IC₅₀ values of 18.9 ± 8.20 μM and 11.0 ± 5.00 μM, respectively (Figures 3B and 4), which suggested that the presence of UA, but not OA, could greatly influence the cellular uptake of molecules into hepatocytes via OATP2B1 and significantly decrease the influx of substances such as toxins into renal tubular cells through OAT3. Additionally, our data indicated that the isomerized structures of OA and UA differentiated the recognition of OAT3 and OATP2B1, which might contribute to predicting the structural properties of selective inhibitors of these two transporter proteins in the future.

Previous studies demonstrated both OA and UA displayed significant inhibitory effect on the influx of [³H]-E17β via OATP1B1 at 100 μM concentration (Roth et al., 2011). Additionally, UA was found to be an effective inhibitor of OATP1B3-mediated FMTX transport (Gui et al., 2010). However, no pronounced inhibition has been observed on these transporters in the current study, which may be because we assessed the transporter uptake of different model substrates from these previous studies and we focused on a lower concentration scale (10.0 µM used in our inhibition study) as to illustrate the clinical relevance of our study.

It was reported that after intravenous infusion of UA nano-liposomes at the dose of 98 mg/m², the average maximum plasma concentration could reach 7.50 ± 1.60 μM (Zhu et al., 2013), indicating that in vivo inhibition of OAT3 and OATP2B1 may not be realized. However, in vitro studies revealed that antidiabetic effect of UA favors higher concentrations (Jung et al., 2007; Rao et al., 2012), therefore, precautions will be needed when treating diabetes with elevated dose of UA in patients in the future.

Our study was the first to report the inhibitory effect of GA on OATP1B3 uptake. Based on in vitro data, De Bruyn et al. (2013) successfully developed an in silico proteochemometric model for OATP1B3 inhibitors. This model suggested that the presence of anionic atoms or a number of hydrogen bond donors and the absence of cationic atoms are key descriptors of OATP1B3 inhibitors, which prediction was in agreement with the previous literature and the current study (Chang et al., 2005; De Bruyn et al., 2013; Karlgren, 2012). As shown in Figure 1, GA contains several hydrogen bond donors with no cationic atoms; hence it perfectly fits into the proposed model.

The previous literature reported that the mean maximum plasma concentrations of GA were 1.83 ± 0.16 and 2.09 ± 0.22 µM, respectively, when healthy humans consume 0.3 mM GA in the form of acidum gallicum tablets or a black tea brew (Shahrzad et al., 2001). The oral administration of processed radix polygoni multiflori (P-RPM) could achieve a C_max of 4.86 ± 0.90 µM GA in humans (Zhang et al., 2013). Therefore, with a high inhibitory potency on OATP1B3 (IC₅₀ value of 1.60 ± 0.60 μM), GA is very likely to compete with other agents for cellular uptake in hepatocytes. And precautions when co-administering GA with drugs substrates of OATP1B3 are mandatory.

It was also noticed previously that GA was a potent inhibitor of OAT1 and OAT3 (IC₅₀ values: 1.24 and 9.02 μM, respectively) (Wang & Sweet, 2012). However, this was not observed in the current study, which might be due to the different in vitro cell models used.

As shown in Figure 3, transport activities of OAT4 and OCTN1 were stimulated by OA; whereas similar observations were made on UA and GA to the OCTN1-mediated uptake. Such sporadic phenomenon has been previously reported, which was likely due to altering allosteric sites within transporters with binding to their chemical inhibitors and consequently, modulating the uptake kinetics of transporter substrates (Kindla et al., 2011; Pedersen et al., 2008). Future studies will be needed to further investigate these discoveries.

Conclusions

In summary, our study comprehensively demonstrated the interactions of the bioactive compounds of pomegranate, UA, OA, and GA, with the essential SLC transporters expressed in the liver and kidney. Our findings indicated that co-administration of these agents with therapeutic drugs may lead to potential drug–drug/herb interactions via competing for particular SLC transporters, which will result in altered pharmacokinetics of drugs, and in turn, greatly impact therapeutic outcomes. Our study provided molecular explanations to the drug–drug/herb interactions involved with UA, OA, and GA, which formed the basis of therapeutic optimization when drugs were co-administered with these compounds or herbal preparations containing these natural compounds are used in treating diseases such as diabetes mellitus.

Declaration of interest

The authors declare no conflicts of interest. This work is supported by grants from the Natural Science Foundation of Jiangsu Province (BK2011168 and BK2012105).