Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Food ingredients may interact with the absorption of drugs by a variety of mechanisms and may substantially affect their pharmacokinetics. Besides the effects that food exerts on gastric pH, gastric emptying etc., components of food can alter drug absorption, e.g., through alterations in drug solubility, by affecting presystemic intestinal metabolism, or by interaction with drug transporters (Evans, 2000; Aronson, 2001). Ingredients of grapefruit juice, especially, have achieved much attention when it was found that grapefruit juice enhanced the oral and systemic availability of a variety of drugs when compared with other juices or water (Ameer and Weintraub, 1997; Bailey et al., 1998). The ingredients responsible for those effects were suggested to be flavonoids, such as naringin and furanocoumarines, like 6',7'-dihydroxybergamottin (Eagling et al., 1999). Effects on systemic availability by grapefruit juice components were explained by the acute inhibition and down-regulation of phase-I drug-metabolizing enzymes, mainly CYP3A4 (Eagling et al., 1999). Although the in vitro findings supported naringin or 6',7'-dihydroxybergamottin as being active ingredients, neither of these compounds contributed significantly to the observed grapefruit juice-drug interactions in human volunteers (Bailey et al., 1993, 1998). It therefore was suggested that other plant components, mainly other flavonoids, could be the active principle. In addition to affecting the presystemic metabolism of drugs, flavonoids may also be able to alter the activity of epithelial drug transporters and thereby modulate the absorption of drugs from the intestine (Kane and Lipsky, 2000). This has so far been shown for a variety of flavonoids in the cases of the P-glycoprotein (Castro and Altenberg, 1997; Conseil et al., 1998; Mitsunaga et al., 2000; Sadzuka et al., 2000) and the multidrug resistance-associated proteins (Hooijberg et al., 1997; Walgren et al., 2000), which are drug efflux pumps transporting a huge variety of xenobiotics (Seelig, 1998; Seelig et al., 2000).

Especially for drugs displaying generally low oral availabilities, the enhancement of absorption by naturally occurring food ingredients, such as flavonoids, might be beneficial. We have therefore focused on dianionic -lactam antibiotics such as cefixime, which is mainly used for treatment of respiratory tract infections (Verghese et al., 1990). It shows a comparably low oral availability in vivo (Bergan, 1984; Faulkner et al., 1988), and its intestinal absorption is like that of all orally active -lactam antibiotics, determined by the interaction with the intestinal peptide transporter PEPT1 (Bretschneider et al., 1999). We have previously shown that the lower oral availability of cefixime compared with that of zwitterionic -aminocephalosporins is due to a low transport rate by PEPT1 under physiological pH conditions (Wenzel et al., 1996). Therefore, increasing the PEPT1-mediated transport for cefixime by flavonoids could provide a principle for the enhancement of its oral availability. Moreover, as interactions of flavonoids with the intestinal peptide transporter PEPT1 have never been studied, nothing is known on potential interactions of such components with drugs that utilize PEPT1 for uptake into the body.

We therefore investigated 33 selected flavonoids of the flavone, flavonol, flavanone, and isoflavone subgroups for their ability to modulate uptake of [¹⁴C]cefixime by PEPT1 in Caco-2 cells. Moreover, by measurements of intracellular pH (pH_in) we assessed whether effects on transport were directly caused by interaction of the test compounds with the PEPT1-protein or secondarily caused by alterations in pH_in homeostasis and the proton-motive driving force required for cefixime uptake.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials

[¹⁴C]Cefixime (44.7 mCi/mmol) and unlabeled cefixime were provided by Klinge-Pharma (Munich, Germany) and Fujisawa (Osaka, Japan). Flavonoids were purchased from Sigma (Deisenhofen, Germany) or Phytochem (Ischenhausen, Germany). Luteolin, tangeretin, and diosmetin were from Roth (Karlsruhe, Germany), and lavendustin A, genistein, and daidzein were from Calbiochem (Bad Soden, Germany). Tyrphostin 25 and amiloride were obtained from Sigma. Cell culture plates and flasks were obtained from Nunc (Wiesbaden, Germany). All other materials needed for cell culture were from Invitrogen (Eggenstein, Germany). Rat tail collagen R was purchased from Serva (Heidelberg, Germany).

Methods

Cell Culture. Caco-2 cells (ATCC, HTB 37, passage 31) were cultured and passaged in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal calf serum, 2 mM glutamine, 1% minimum essential medium nonessential amino acids (Invitrogen), and 70 µg/ml gentamicin (Invitrogen) in a humidified incubator at 37°C under an atmosphere of 5% CO₂. Cells between passages 40 and 65 were seeded at a density of 1 · 10⁵ cells/well onto Transwell polycarbonate membranes (24-wells, Nunc) and were used 14 days after reaching confluency. Fresh medium was given every 2nd day, and on the day before uptake measurements were performed. Transepithelial electrical resistance was measured with an epithelial voltmeter (EVOM, World Precision Instruments, Berlin, Germany) and resistances of 300 /cm² indicated the presence of an intact monolayer (Artursson, 1990; Riley et al., 1991).

Transport Studies. Flux studies in Caco-2 cells were performed in a modified Krebs' buffer containing 137 mM NaCl, 5.4 mM KCl, 2.8 mM CaCl₂, 1.0 mM MgSO₄, 0.3 mM NaH₂PO₄, 0.3 mM KH₂PO₄, 10 mM glucose, and 10 mM HEPES-Tris for incubation buffer pH 7.4 or 10 mM 2-(N-morpholino) ethane-sulfonic-acid-Tris for incubation buffer pH 5.0, respectively. For uptake, cell monolayers were washed free of serum-containing medium and incubated from the apical side of the monolayer with radiolabeled [¹⁴C]cefixime in the presence or absence (control) of flavonoids for 30 min at 37°C. The basolateral side of the monolayers was incubated with incubation buffer pH 7.4. After the incubation period, the cells were washed three times with ice-cold incubation buffer and subsequently digested with a tissue solubilizer. Cellular accumulation of [¹⁴C]cefixime was measured after the addition of scintillation cocktail by liquid scintillation spectroscopy. Binding of tracer to the cells was determined as the residual radioactivity associated with the cells in the presence of excess unlabeled substrate. The basic characteristics of cefixime transport in the Caco-2 cell model have been reported previously (Wenzel et al., 1996).

pH_in Measurements. For pH_in measurements, the Caco-2 monolayers were loaded with BCECF by preincubating the cells with 5 µM the lipophilic acetoxymethyl-ester at 37°C for 30 min. Subsequently, the monolayers were washed with buffer pH 7.4, and the buffers with or without substrates were changed by superfusion at the time points indicated in the graphs. Intracellular H⁺ activity was determined by intensity of emission at 538 nm after excitation of the fluorophore at 444 nm (isosbestic point) and 490 nm (pH sensitive wavelength), respectively, using a microtiter plate reader (Fluoroskan Ascent, Labsystems, Merlin Diagnostika, Bornheim-Hersel, Germany). The 444/490 fluorescence ratio was converted to pH_in from a calibration curve generated by estimation of the fluorescence ratio in buffers of different pH (Boyarsky et al., 1996).

Calculations and Statistics. All calculations (linear as well as nonlinear regression analysis) were performed by using Prism 2.01 (GraphPad Software, Los Angeles, CA). For each variable, 4 to 8 independent experiments were carried out. Data are given as the mean ± S.E.M. Significance of differences between control and treated cells were determined by an unpaired Student's t test.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

In a first set of experiments, 33 flavonoids of the flavone, flavonol, flavanone, and isoflavone subclasses (Fig. 1) were tested for their effects on [¹⁴C]cefixime uptake into Caco-2 cells at apical pH 5.0, where transport of the dianionic -lactam is optimal (Wenzel et al., 1996). At a concentration of 100 µM, the flavones acacetin, chrysin, and diosmin, the flavonol quercetin, the flavanone naringin, and the isoflavone genistein increased uptake of 1 mM cefixime by 25 to 37% (Table 1). All other flavonoids tested were either without significant effects or even reduced [¹⁴C]cefixime absorption (Table 1).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Four subclasses of flavonoids. The basic structure of flavonoids consists of an O-heterocyclic ring (C) fused to an aromatic ring (A) and by a carbon-carbon bond to a second aromatic ring (B). The classification is based on variations on the heterocyclic C-ring.

View this table: [in this window] [in a new window]   TABLE 1 Effects of flavonoids on the uptake of [14C]cefixime into Caco-2 cells Uptake of 1 mM cefixime was measured at apical buffer pH 5.0 for 30 min in the absence (control) or presence of 100 µM the respective flavonoid. Control uptakes were 11.8 ± 3.2 nmol · cm2 · 30 min1. Values are expressed as the mean ± S.E.M. (n = 4).

Dose-response relationships for those flavonoids that increased the uptake of 1 mM cefixime revealed EC₅₀ values ranging between 12.6 and 155.7 µM with maximal stimulation of transport rates exceeding that in control cells by 50 to 60% (Fig. 2). Similar stimulatory effects on cefixime uptake by these flavonoids were found at apical pH 6.0, indicating that the stimulatory mechanism is also operational under less acidic pH conditions (data not shown). Since quercetin showed a pronounced effect and occurs in comparably large quantities in fruits and vegetables (Hertog et al., 1992) we selected this compound for further analysis of the underlying mechanisms of transport activation. Kinetics of [¹⁴C]cefixime transport into Caco-2 cells determined in the absence or the presence of 100 µM quercetin revealed that V_max was significantly increased by the flavonoid without any effects on the K_m (Fig. 3). Transformation of the data according to Eadie-Hofstee revealed a maximal transport rate increased by 53% in the presence of quercetin (Fig. 3, inset). Such an increase could either be due to direct modifications of the PEPT1 protein allowing a higher turnover rate or by alterations of its driving force. Since the transmembrane pH gradient across the apical plasma membrane is required for the H⁺-coupled uptake of short-chain peptides and -lactam antibiotics mediated by PEPT1 (Thwaites et al., 1993a,b; Wenzel et al., 1996), quercetin and other flavonoids might stimulate cefixime absorption indirectly by alterations of the proton-motive driving force.

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Screening for the effects of selected flavonoids on [14C]cefixime uptake into Caco-2 cells. Cells were incubated for 30 min with 1 mM cefixime at apical buffer pH 5.0 either in the absence (control) or presence of 0.1 to 750 µM flavonoid (A, acacetin; B, chrysin; C, diosmin; D, quercetin; E, naringin; F, genistein). EC50 values for stimulation of cefixime transport derived from the dose-response curves were 109.0 ± 2.6 µM for acacetin, 114.5 ± 3.2 µM for chrysin, 155.7 ± 5.6 µM for diosmin, 17.5 ± 1.6 µM for quercetin, 66.6 ± 1.3 µM for naringin, and 12.6 ± 0.8 µM for genistein, respectively. Uptake of cefixime into control cells was 7.6 ± 2.1 nmol · cm2 · 30 min1. Results are expressed as the mean ± S.E.M. of four independent experiments.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Uptake of [14C]cefixime into Caco-2 cells as a function of substrate concentration. Transport of 0.05 to 5 mM cefixime was measured for 30 min at pH 5.0 either in the absence () or in the presence () of 100 µM quercetin. Transformation of the transport rates according to Eadie-Hofstee analysis (inset) revealed Km and Vmax values of 1.5 ± 0.2 mM and 30.0 ± 1.7 nmol · cm2 · 30 min1 in control cells and 1.4 ± 0.2 mM and 45.9 ± 3.2 nmol · cm2 · 30 min1 in quercetin-treated cells, respectively.

We therefore used BCECF to determine whether exposure of cells to the test compounds altered the intracellular pH in Caco-2 cells, which is a critical determinant of the driving force of PEPT1. The measurement of pH_in in Caco-2 cells superfused either with apical buffer pH 5.0 alone or in addition with 5 mM cefixime revealed that transport of cefixime caused as expected a further pH_in decline as compared with perfusion with buffer alone (Fig. 4). However, when quercetin was coadministered, the pH_in decline by cefixime was less pronounced (Fig. 4), although under this conditions H⁺/cefixime cotransport is significantly higher (Table 1; Fig. 3). This strongly suggested that transport systems responsible for the maintenance of the transmembrane pH gradient, such as apical Na⁺/H⁺-exchange (NHE), could be activated by quercetin. To show that quercetin indeed enhances cefixime absorption by activation of apical Na⁺/H⁺-exchange, we used 100 µM the NHE blocker amiloride. In the presence of amiloride, quercetin was no longer able to stimulate cefixime transport (Fig. 4, inset).

View larger version (30K): [in this window] [in a new window]   Fig. 4.   pHin measurements in Caco-2 cells using the pH sensitive fluorophore BCECF. Cells were superfused from the apical side with buffer pH 7.4 for 5 min. Subsequently, cells were perfused for 30 min with buffer pH 5.0 () containing either buffer alone () or in addition to () 5 mM cefixime or a mixture of 5 mM cefixime and 100 µM quercetin (). All substrates were washed out after 35 min followed by perfusion with buffer pH 7.4 (). Inset, relative uptake rates of [14C]cefixime into Caco-2 cells after 30 min of incubation with either 100 µM amiloride () or 100 µM amiloride plus 100 µM quercetin () at an apical pH of 5.0 and expressed as a percentage of control uptake (). Uptake rates in cells incubated either with amiloride or with amiloride plus quercetin did not display significant differences.

It has been shown recently that NHE-3 is the Na⁺/H⁺-exchanger isoform in the apical membrane of differentiated Caco-2 cells that is preferentially activated for pH_in control following PEPT1-mediated H⁺ influx into cells (Thwaites et al., 1999). The NHE-3 protein is subject to regulation by protein kinases such as protein kinase C, protein kinase A, phosphatidylinositol 3-kinase, and protein tyrosine kinases (PTK) of the receptor and nonreceptor subtypes (Good, 1995; Kandasamy et al., 1995; Yamaji et al., 1995; Khurana et al., 1996; Good et al., 2000). Since genistein proved to be a specific inhibitor of the EGF-receptor (EGFR) PTK (Akiyama et al., 1987), and since quercetin is also used as an inhibitor of EGFR PTK (Traxler et al., 1999), one possible mechanism for stimulation of cefixime transport by those flavonoids might be the activation of NHE-3 via inhibition of EGFR PTK that in turn would allow a more efficient pH_in control. To prove this hypothesis, we used the specific EGFR PTK inhibitor tyrphostin 25 and determined its effects on H⁺/cefixime cotransport. As shown in Fig. 5, tyrphostin 25 reduced the acidification induced by cefixime but at the same time increased cefixime transport rate by 43% (Fig. 5, inset). This finding suggests a common mode of action for tyrphostin 25 and also for the flavonoids in maintaining the proton-motive driving force for PEPT1 through an activation of NHE-3 activity via inhibition of EGFR PTK.

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Effects of tyrphostin 25 on cefixime-mediated intracellular acidification and on cefixime transport. After perfusing the cells with buffer pH 7.4 for 5 min, superfusion was performed for 30 min with buffer pH 5.0 () containing either 5 mM cefixime () or a mixture of 5 mM cefixime and 10 µM tyrphostin 25 (). All substrates were washed out after 35 min of incubation with buffer pH 7.4 (). Inset, uptake of [14C]cefixime into Caco-2 cells as measured for 30 min at pH 5.0 either in the absence (control, ) or in the presence () of 10 µM tyrphostin 25. **P < 0.01 versus the control.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Our present findings provide the first evidence for an interaction of food-derived flavonoids with the intestinal peptide transporter PEPT1, which mediates the absorption of a variety of drugs including -lactams (Wenzel et al., 1995), angiotensin-converting enzyme inhibitors (Boll et al., 1994), bestatins (Saito and Inui, 1993), as well as special prodrugs (Sawada et al., 1999). Selected flavonoids may therefore have the potential to significantly alter the oral availability and pharmacokinetics of these PEPT1 substrates when coadministered in humans. Similar interactions have been shown to exist between several flavonoids and the intestinal drug efflux pumps P-glycoprotein (Castro and Altenberg, 1997; Conseil et al., 1998; Mitsunaga et al., 2000; Sadzuka et al., 2000) and the multidrug resistance-associated proteins (Hooijberg et al., 1997; Walgren et al., 2000) leading to alterations in the extent of absorption of those drugs that are substrates of these drug efflux pumps (Seelig, 1998; Seelig et al., 2000).

We show here that a variety of flavonoids out of the flavone, flavonol, flavanone, and isoflavone subgroups can increase or decrease PEPT1-mediated cefixime transport in Caco-2 cells. Of the 33 flavonoids tested, the flavones acacetin, chrysin, and diosmin, the flavonol quercetin, the flavanone naringin, and the isoflavone genistein significantly increased the uptake of cefixime into Caco-2 human intestinal epithelial cells. In contrast, the flavonols 3-OH-flavone, kaempferol, luteolin, and morin, the flavanone didymin, and the isoflavone daidzein significantly reduced cefixime absorption. All other flavonoids had no significant effects on cefixime uptake. With emphasis on genistein and quercetin, which are present in larger quantities in a variety of fruits and vegetables, we focused on the mechanisms by which they transmit the effects on PEPT1. In addition, increasing the oral availability of -lactams such as cefixime that possess an intrinsically low oral availability by means of selected flavonoids could be beneficial for more effective antimicrobial treatment.

Quercetin was dose dependently able to increase cefixime transport by increasing the V_max of PEPT1 but was at the same time able to reduce the intracellular acidification caused by cefixime influx and thereby maintained a higher electrochemical proton gradient across the apical plasma membrane, allowing higher transport rates. Since quercetin and especially genistein are frequently used as inhibitors of EGFR PTK (Akiyama et al., 1987; Traxler et al., 1999), we supposed that such an activity might be responsible for the stimulation of cefixime transport caused by the flavonoids. Blocking EGFR PTK with the specific inhibitor tyrphostin 25 simulated the effects of the flavonoids most likely by stabilizing the transmembrane pH gradient required for PEPT1 function and thereby allowing increased H⁺/cefixime influx. Maintenance of the proton-motive driving force for cefixime uptake is most likely achieved by activation of NHE-3 because this exchanger subtype was identified functionally in the apical membrane of differentiated Caco-2 cells to be the prime pH recovery system activated by PEPT1-mediated acid loading of cells (Thwaites et al., 1999). Although it was demonstrated that inhibitors of EGFR PTK, such as genistein or herbimycin A, stimulated NHE-3 in the apical membrane of cells from the rat medullary thick ascending limb (Good, 1995; Good et al., 1999), EGF was found to stimulate Na⁺/H⁺-exchange in the apical membrane of rabbit ileum as well as in Caco-2 cells stably transfected with NHE-3 (Khurana et al., 1996; Janecki et al., 1999). Irrespective of this apparent contradiction, our studies show that the specific inhibition of EGFR PTK in Caco-2 cells enhanced H⁺ extrusion out of the cells and allowed a higher PEPT1 transport rate. It is noteworthy in this context that the physiological regulation and function of epithelium-specific NHEs are dependent on tissue-specific factors and/or conditional requirements as well (McSwine et al., 1998).

For inhibition of EGFR PTK, some structural requirements must be fulfilled by the respective flavonoids. According to a recently proposed pharmacophore model, hydrogen bonding between functional groups of inhibitors and two residues of the hinge region of EGFR PTK (dual hydrogen bond donor-acceptor system; Traxler et al., 1999) is an important requirement. When applied to the flavonoids, the keto function at position 4 of ring C (see Fig. 1) could generally act as a hydrogen bond acceptor, whereas the 5-OH group in genistein and the 3-OH group in quercetin could serve as donor groups. Moreover, the 7-OH group was found to be important for the EGFR PTK inhibitory activity of isoflavones (Traxler et al., 1999). These structural features were also found as a common feature in the five flavonoids that stimulated cefixime absorption (Table 2). All of these transport-enhancing structures bear OH groups in the 5- and 7-positions of ring A. In contrast, flavonoids missing one or both of the 5- and 7-OH substituents generally were not able to stimulate cefixime absorption (Table 2). However, such a restriction to the effects of OH groups at positions 5, 7, and 3 (position 3 in the case of flavonols only) would mean an oversimplification since there were flavonoids bearing OH groups at these positions but showed no or only minor transport effects (Table 2). That other structural features and most likely a specific hydroxylation pattern of the flavonoids must play an important role becomes obvious by inspection of the flavonol subgroup where quercetin was the only compound found to stimulate cefixime uptake (Table 2).

In conclusion, five flavonoids of 33 tested were able to significantly and dose dependently increase cefixime uptake into Caco-2 cells. Comparing the chemical structures of the different flavonoids reveals that their effects on cefixime transport match the requirements for their ability to inhibit EGFR PTK. The effects of tyrphostin 25, a specific EGFR PTK inhibitor, on maintenance of the proton-motive driving force for cefixime uptake resemble those observed for the flavonoid quercetin and suggest a common mode of action. Inhibition and stimulation of PEPT1-mediated drug transport by flavonoids should be considered as another example for drug-food interactions of therapeutical importance.