Inhibitory Effect of Flavonoids on the Efflux of N-Acetyl 5-Aminosalicylic Acid Intracellularly Formed in Caco-2 Cells

N-acetyl 5-aminosalicylic acid (5-AcASA) that was intracellularly formed from 5-aminosalicylic acid (5-ASA) at 200 μM was discharged 5.3, 7.1, and 8.1-fold higher into the apical site than into the basolateral site during 1, 2, and 4-hour incubations, respectively, in Caco-2 cells grown in Transwells. The addition of flavonols (100 μM) such as fisetin and quercetin with 5-ASA remarkably decreased the apically directed efflux of 5-AcASA. When 5-ASA (200 μM) was added to Caco-2 cells grown in tissue culture dishes, the formation of 5-AcASA decreased, and, in addition, the formed 5-AcASA was found to be accumulated within the cells in the presence of such flavonols. Thus, the decrease in 5-AcASA efflux by such flavonols was attributed not only to the inhibition of N-acetyl-conjugation of 5-ASA but to the predominant cellular accumulation of 5-AcASA. Various flavonoids also had both of the effects with potencies that depend on their specific structures. The essential structure of flavonoids was an absence of a hydroxyl substitution at the C5 position on the A-ring of flavone structure for the inhibitory effect on the N-acetyl-conjugation of 5-ASA, and a presence of hydroxyl substitutions at the C3′ or C4′ position on the B-ring of flavone structure for the promoting effect on the cellular accumulation of 5-AcASA. Both the decrease in 5-AcASA apical efflux and the increase in 5-AcASA cellular accumulation were also caused by MK571 and indomethacin, inhibitors of MRPs, but not by quinidine, cyclosporin A, P-glycoprotein inhibitors, and mitoxantrone, a BCRP substrate. These results suggest that certain flavonoids suppress the apical efflux of 5-AcASA possibly by inhibiting MRPs pumps located on apical membranes in Caco-2 cells.


Introduction
Sulfasalazine used in the therapy of inflammatory bowel diseases, such as ulcerative colitis and Crohn's disease [1,2]. Ingested sulfasalazine passes to the colon without being absorbed in intestine and is split into 5-aminosalicylic acid (5-ASA) and sulfapyridine by colonic bacteria [1,2]. Most of 5-ASA is metabolized by N-acetyl-conjugation in the form of N-acetyl 5-aminosalicylic acid (5-AcASA) in the colonic epithelia, while sulfapyridine is quickly absorbed from the colon and metabolized in the liver [3][4][5]. It has been proposed that 5-ASA, the active moiety of sulfasalazine, exerts an antiinflammatory activity by inhibiting prostaglandin synthesis in colonic mucosa [6,7]. Some reports have shown that 5-AcASA has a potency as an inhibitor of prostaglandin synthesis comparable to that of 5-ASA [7], and therapeu-tically active when administered by enema to patients with ulcerative colitis [8]. However, 5-AcASA formed in colonic epithelia is immediately secreted into mucosal lumen and excreted in feces [9][10][11]. Thus, 5-AcASA is considered the portion that has already exerted therapeutical action within the bowel tissue [1][2][3][9][10][11]. Zhou et al. reported that 5-AcASA was exclusively transported from the basolateral to the apical direction using human colon-derived Caco-2 cells [11]. However, the mechanism underlying the cellular transport of 5-AcASA has not extensively elucidated. It is well known that flavonoids (Figure 1), plant-derived compounds, alter the function of efflux transporters such as P-glycoprotein, that is, present in epithelium cells [12][13][14]. Recently, several researchers reported the inhibitory interaction of flavonoids with multidrug resistance-associated proteins (MRPs) that are responsible for active secretion of  pharmacologically relevant drugs [15][16][17][18][19][20]. In this study, the effect of flavonoids and transporter inhibitors on the cellular efflux of 5-AcASA that was intracellularly formed from 5-ASA was examined in Caco-2 cells. Certain flavonoids and MRPs inhibitors displayed strong potency in decreasing the preferential apical efflux of 5-AcASA and in increasing the cellular accumulation of 5-AcASA in Caco-2 cells. Com. and Wako Pure Chemical Co. 5-AcASA was synthesized by the reaction of 5-ASA with acetic anhydride, as described by other researchers [21].

Efflux of 5-AcASA from Caco-2 Cells.
Caco-2 cells were purchased from the Riken (no. RCB0988) and used as previously described [22]. The cell line was cultured in Dulbecco's modified Eagle's medium containing 12% fetal calf serum and penicillin-streptomycin-amphotericin B.
The suspended cells were seeded on 6-well-polycarbonate Transwell inserts (0.4 μm mean pore size, 4.7 cm 2 growth area) at a density of 5 × 10 4 cells/dish, and then placed in an incubator in an atmosphere of 5% CO 2 -95% air at 37 • C. The Caco-2 cells in the Transwell were grown for 3 weeks in Dulbecco's modified Eagle's medium containing fetal calf serum. The monolayers with transepithelial electric resistance of more than 250 Ω cm 2 were used for transport studies. 5-ASA in a stock solution at 50 mM was added to the apical chamber at a final concentration of 200 μM after 10 minutes of the addition of flavonols. After incubation for 2 and 4 hours at 37 • C, 50 μL of the medium from both of the chambers was mixed with 50 μL of 0.5 M perchloric acid.

Cellular Accumulation of 5-AcASA.
Caco-2 cell line at passage of 40 was used for the experiments. The suspended cells in Dulbecco's modified Eagle's medium containing 12% fetal calf serum and penicillin-streptomycin-amphotericin B were seeded on 35 mm plastic culture dishes at a density of 5 × 10 4 cells/dish. After seeding, the cells were cultured in a 37 • C incubator under 5% CO 2 -95% air at 37 • C for two weeks until the cells were fully differentiated into confluent enterocyte-like monolayers. Flavonoids, 5-ASA and other chemicals were dissolved in dimethyl sulfoxide and added to the medium at definite concentrations, with the final concentration of dimethyl sulfoxide about 1%. After incubation for 2 hours, the cell monolayers were washed twice with Hanks balanced solution and harvested. The adequate volume of the medium and cell suspensions was treated with the same volume of 0.5 M perchloric acid.

HPLC Analysis.
Chromatographic separation and quantitative determination were carried out according to the HPLC analytical methods described previously [23]. A 0.1 mL aliquot of perchloric acid-treated sample was neutralized with 25 μL of 1 M NaOH solution and 25 μL of 0.5 M Tris-HCl buffer (pH 7.4), and the total volume was adjusted to 0.5 mL with HPLC elution solvent. A 50 μL aliquot of sample was injected onto a Develosil C-30-UG-3 (4.6 I.D. × 150 mm) column adjusted to 40 • C, and 5-AcASA was separated by solution with a mixture of acetonitrile (4%) and 20 mM phosphate buffer (pH 5.0 solution) using a CCPD HPLC system equipped with an FS-8020 fluorescence detector (Tosoh Co., Japan). The flow rate of the mobile phase was 1.0 mL/min, and elution of 5-ASA and 5-AcASA was monitored at a fluorescence excitation wavelength of 310 nm and an emission wavelength of 480 nm. 5-ASA and 5-AcASA were eluted at 2.7 and 11.5 minutes, respectively. The quantitative determination of 5-AcASA was based upon the integration of fluorescence peak areas.

Results
The incubation of Caco-2 cells with 5-ASA formed only one peak of 5-ASA metabolite, which was identified as 5-AcASA by the same retention time as the synthesized standard in HPLC. The N-acetyl-conjugative reaction of 5-ASA in Caco-2 cells was saturated above 1 mM of 5-ASA. The effect of flavonols and inhibitors of transporters on 5-AcASA efflux was examined using Caco-2 cell monolayers grown in Transwells which contained 1.5 and 2.6 mL Dulbecco's modified Eagle's medium in the apical and basolateral chambers, respectively. 5-ASA was loaded at 200 μM in the apical chamber and 5-AcASA discharged from both of the apical and basolateral sites was measured. After 1, 2, and 4-hour incubation, amounts of 5-AcASA were 1.01, 2.05, and 5.04 nmoles in the apical chamber and 0.19, 0.29, and 0.62 nmoles in the basolateral chamber, respectively ( Table 1). The apical efflux of 5-AcASA was 5.32, 7.07, and 8.13-fold higher than the basolateral efflux at 1, 2, and 4hour incubation, respectively. When fisetin and quercetin (100 μM) were added with 5-ASA to Caco-2 cells, the apical efflux of 5-AcASA decreased remarkably ( Table 1). The basolateral efflux of 5-AcASA rather increased in the presence of such flavonols. The ratios for the apical to the basolateral efflux of 5-AcASA actually decreased to 0.52 and 0.68 at 1 hour, 0.77 and 0.85 at 2 hours, and 0.92 and 1.05 at 4hour incubation, in the presence of fisetin and quercetin, respectively. Morin had a weaker effect than fisetin and quercetin. MK571, a MRPs inhibitor, showed a similar effect to quercetin; however, quinidine, a P-glycoprotein inhibitor, had no effects. grown in tissue culture dishes. 5-AcASA was formed at the rate of 4 nmol/h/1 × 10 6 cells during a 4 h-incubation period in the control cells. Flavonoids are potent inhibitors of N-acetyltransferase [23]. Fisetin remarkably decreased the formation of 5-AcASA from 5-ASA in Caco-2 cells. Furthermore, a large amount of 5-AcASA was found in the cells treated by quercetin. The amount of 5-AcASA inside the control cells was 12 percent of the total 5-AcASA at a 1 h-incubation and decreased to 6.3 and 3.2 percents at 2 and 4-hour incubation, respectively. The cellular accumulation rate increased by several-fold than that of the control cells by quercetin and fisetin, and increased slightly by morin during a 4 h-incubation period.  Table 2). Flavonoids that lack a C2-3 double bond or a carboxyl group at the C4 position on the C-ring, such as catechins and taxifolin, had no effects. The structural feature required for the potent effect on the cellular 5-AcASA accumulation was a presence of hydroxyl group on the B-ring of flavone structure. The effect of inhibitors or substrate of transporters on the cellular 5-AcASA accumulation was compared with flavonols at a 2 h-incubation with 200 μM of 5-ASA in Caco-2 cells (Figure 4). MK-571 and indomethacin, MRPs inhibitors [24][25][26], increased in concentration-dependent manner the cellular 5-AcASA accumulation, while they did not affect the formation of 5-AcASA. MK-571 was more effective than indomethacin and showed equivalent efficacy to quercetin and fisetin. On the other hand, qunidine, a Journal of Biomedicine and Biotechnology P-glycoprotein inhibitor, and cyclosporine A, an inhibitor of both P-glycoprotein and MRPs [27,28], did not affect the cellular 5-AcASA accumulation. Mitoxantrone, a breast cancer resistance protein (BCRP) substrate [29], had no effects either at the concentration of 20 μM (data not shown).

Discussion
5-AcASA that was formed from 5-ASA in the interior of cells was discharged preferentially to the apical direction compared to the basolateral direction in Caco-2 cells grown in Transwells. Quercetin and fisetin remarkably decreased the apical efflux of 5-AcASA, while morin did with a less potency. The amount of 5-AcASA in Caco-2 cells and the medium was measured during a 4 h-incubation with 5-ASA in the presence of such flavonols. Flavonoids are effective inhibitors of N-acetyl-conjugation of 5-ASA in rat liver cytosol preparation [23]. Fisetin, in particular, exhibited strong inhibitory activity on 5-AcASA formation in Caco-2 cells. Thus, the inhibition of 5-AcASA formation is likely to contribute largely to the decrease in the 5-AcASA efflux in the case of fisetin. However, quercetin showed a much weaker inhibitory effect on the 5-AcASA formation than fisetin. Surprisingly, the formed 5-AcASA was found to be accumulated inside the cells treated by flavonols. For quercetin, the cellular accumulation of 5-AcASA coincides with the decrease in 5-AcASA apical efflux. An increase in the basolateral efflux of 5-AcASA during an incubation of Transwells is probably due to the extensive cellular accumulation of 5-AcASA particularly in quercetin-treated cells.
A large group of flavonoids were examined for their inhibitory effects on the 5-AcASA formation as well as their promoting effects on the cellular 5-AcASA accumulation. A key chemical determinant necessary for exerting the strong inhibitory effect on the N-acetyl-conjugation of 5-ASA was a lack of hydroxyl substitution at the C5 position on the A-ring of flavone structure such as fisetin and 7,3 ,4 -OH favone. On the other hand, the structural requirement for the promoting effect on cellular 5-AcASA accumulation was a presence of hydroxyl substitution at the C3 or C4 position on the Bring of flavone structure. Therefore, the inhibition of 5-AcASA formation and the promotion of cellular 5-AcASA accumulation by flavonoids seem to be caused by different mechanisms.
The results mentioned above suggest that 5-AcASA is pumped out by an active efflux transporter located on the apical membrane and certain flavonoids appear to play an important replacing role in the apical-directed transport of 5-AcASA in Caco-2 cells. Flavonoids are well-known modulators of the cellular transport of various substances mediated by P-glycoprotein which is localized on apical membranes in polarized cells [12][13][14]. Recently, several researchers reported the interaction of flavonoids with MRPs transporters. Walgren et al. reported that the efflux of quercetin 4 -betaglucoside across Caco-2 cell monolayers was mediated by MRP2 [24]. Van Zanden et al. studied on the inhibitory effect of quercetin on MRPs pump-mediated efflux of calcein and vincristine, well-known MRPs substrates, in the MRP1 and MRP2 transfected MDCK cells [18][19][20]. They mentioned that MRP2 displayed higher selectivity for flavonoid-type inhibition than MRP1. Phase II metabolites of various drugs conjugated to glutathione, glucuronate, or sulfate are generally considered to be transported by MRPs-like transporters [30][31][32]. MRPs were characterized as the canalicular multispecific organic anion transporters that function in terminal secretion into bile canaliculus of endo-and xenobiotics such as acetaminophen metabolites, bilirubin glucuronides, 2,4dinitrophoenyl-S-glutathione, 17β-glucuronosyl estradiol, and 4-methylumbelliferyl glucuronide that are conjugated in hepatocytes [33][34][35]. The transcellular transport of acetylconjugated 5-ASA from the basolateral site to the apical site in Caco-2 cell was first reported by Zhou et al. [11]. However, the transporter-mediated efflux of 5-AcASA has not been investigated thoroughly. To address the interest in involvement of transporters that are responsible for the 5-AcASA apical efflux in Caco-2 cells, several inhibitors of transporters were examined for their suppressing effect on the 5-AcASA apical efflux and promoting effect on the cellular 5-AcASA accumulation. MK571 and indomethacin, inhibitors of MRPs had similar effects to flavonoids. Quinidine, a P-glycoprotein inhibitor, and Cyclosporine A, an inhibitor of P-glycoprotein and MRPs [27,28], showed no effects. Absence of inhibitory activity of Cyclosporine A may be explained by substrate specificity of 5-AcASA for MRPs. Mitoxantrone, a substrate of BCRP [29], had no effects either. These results suggest that 5-AcASA is possibly pumped out by an MRPs-like transporter and certain flavonoids inhibit their efflux-pump activity in Caco-2 cells.
Flavonoids are part of the human diet and possess many health benefits with low toxicity [36,37]. However, flavonoids are poorly absorbable compounds from the  digestive tract in vertebrates [38,39]. When quercetin was given p.o. to the rats (630 mg/kg), approximately 20% of the total dose was absorbed from the digestive tract, more than 30% was decomposed in the intestinal microflora, and approximately 30% was excreted unchanged in the feces during 72 hours [38]. After a single oral dose of quercetin in humans (4 g), approximately 53% of the dose was recovered unchanged in the feces. Thus it was concluded that 1% of the original 4 g dose of quercetin was absorbed [39]. In this study, flavonoids were added at the concentration range from 20 to 100 μM only into the apical compartment of Caco-2 cells in Transwells that faces to intestinal lumen in vivo. A high luminal level around 100 μM of flavonoids is expected to be achieved with a single oral administration of a few hundred mg of flavonoids in humans.
5-ASA, an active moiety of sulfasalazine, is immediately secreted into the luminal side from intestinal epithelia following extensive N-acetyl-conjugation, and is finally excreted into feces [3][4][5]. Zhou et al. [11] reported that at luminal levels below 200 μg/mL (concentrations that are typically achieved by controlled release dosage forms), intestinal secretion of 5-AcASA accounts for more than 50% of the total 5-ASA elimination. Thus, 5-AcASA has been considered to be therapeutically nonactive portion [1][2][3][9][10][11]. However, 5-AcASA has still antiinflammatory potential if the drug retains within the intestinal tissues [8]. The efficacy of 5-ASA therapy correlates with tissue delivery of 5-ASA, that is, determined by N-acetylation and cellular discharge. The present study showed that certain flavonoids have the inhibitory effect on N-acetyl-conjugation of 5-ASA and the suppressive effect on the 5-AcASA apical efflux in Caco-2 cells. Viewed in this light, both of these effects of flavonoids seem to be desirable in the treatment of inflammatory bowel diseases, since coadministration of flavonoids with 5-ASA is expected to increase the tissue levels of 5-ASA and 5-AcASA in intestine.