Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

An ever-increasing body of evidence suggests that dietary flavonoids are beneficial to human health and can result in significant reductions both in the risk of developing certain cancers and in the risk of mortality from coronary heart disease and stroke through antiproliferative, antioxidant, and other mechanisms (for review, see Walgren et al., 2000). However, a major hurdle in linking in vitro and in vivo findings has been generating evidence for oral bioavailability of the flavonoids. Estimates from human studies have indicated that the dietary flavonoids display poor and variable bioavailability (Gugler et al., 1975; Hollman et al., 1995, 1997; Manach et al., 1998; Aziz et al., 1999). In part, the observed variability is a result of the complexity of the in vivo system, which involves limited absorption as well as extensive metabolism and degradation with significant losses attributable to the intestinal microflora. An additional factor is the relative absence of molecularly specific methodologies for analysis of the various forms of the flavonoids in this complex system.

We have used the human Caco-2 cell model of intestinal absorption (Artursson and Karlsson, 1991; Yee, 1997), together with molecularly specific analysis, in an attempt to better understand the extent of and the mechanisms governing flavonoid absorption. In particular, we have focused on the most prevalent flavonoid in the Western diet, quercetin (Hertog et al., 1993). With this model, we have previously shown that quercetin aglycone is capable of crossing the intestinal epithelium, but that its major dietary form, quercetin 4'--glucoside, is not absorbed (Walgren et al., 1998). This latter finding was surprising because it has been previously suggested that the quercetin glucosides are absorbed within the intestine by the active glucose transporter SGLT1 (Hollman et al., 1995). Subsequently, we demonstrated that quercetin 4'--glucoside is effluxed from Caco-2 cell monolayers by the apically expressed multidrug resistance-associated protein MRP2 (Walgren et al., 2000).

Despite a lack of direct evidence for absorptive transport of flavonoid glucosides, work by Gee et al. (1998) supports an interaction of quercetin glucosides with the intestinal glucose transporters. In our previous studies, the activity of MRP2 could have masked evidence for SGLT1-dependent absorption. In this study, we have directly tested the hypothesis that quercetin 4'--glucoside is a substrate for the intestinal active glucose transporter by examining transport of this glucoside across the apical membrane of Caco-2 cells, which express human SGLT1 (Blais et al., 1987). In addition, we have examined transport of this glucoside in Chinese hamster ovary (CHO) cells stably transfected with rabbit SGLT1 (G6D3 cells; Lin et al., 1998) and compared it with transport in parental CHO cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

MK-571 was a generous gift from A. W. Ford-Hutchinson, Merck-Frosst Center for Therapeutic Research, Pointe Claire-Dorval, Quebec, Canada. Dulbecco's PBS with 0.1 g l¹ calcium chloride was purchased from Life Technologies (Grand Island, NY). Quercetin 3-O-sulfate potassium salt was purchased from Extrasynthèse (Genay, France). Quercetin 4'--glucoside was isolated from the red onion as described previously (Walgren et al., 1998). Except where noted, all other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO).

Cell Cultures. Caco-2 cells obtained from American Type Culture Collection (Rockville, MD) were cultured in Eagle's minimum essential medium (MEM) (Cellgro; Mediatech, Herndon, VA) supplemented with 1% MEM nonessential amino acids (Mediatech), 10% fetal bovine serum (Summit Biotechnology, Fort Collins, CO), 100 U ml¹ penicillin, and 0.1 mg ml¹ streptomycin (Sigma Chemical Co.) and were grown in a humidified atmosphere of 5% CO₂ at 37°C. Cells were subcultured at 80% confluence.

1

1

1

Fluorescence Studies. Caco-2 cells were seeded on 18-mm² glass coverslips at a density of 1.0 × 10⁵ cells cm² and grown to confluence. Culture medium was replaced three times a week for 10 to 14 days and 24 h before experiments. Before uptake studies, coverslips were washed twice for 30 min with warm transport buffer, PBS (Life Technologies) containing 0.1 g l¹ calcium, pH 7.4 (PBS). Cells were incubated for 15 min at 37°C in PBS containing 50 µM quercetin, quercetin 3-sulfate, or quercetin 4'-glucoside. Transport was halted rapidly by aspiration of buffer and three sequential washings with ice-cold PBS. Cells were fixed for 5 min with 3.5% paraformaldehyde and washed three times. Coverslips were mounted on slides with FluorSave (Calbiochem, La Jolla, CA) and allowed to dry. Cells were examined and images were acquired with Image-Pro Plus (Media Cybernetics, Silver Spring, MD) with a DAGE-MTI CCD100 camera (Michigan City, IN) on a Zeiss Axioplan microscope (Carl Zeiss, Inc., Thornwood, NY).

Cellular Uptake Studies. For all cellular uptake studies, confluent monolayers of Caco-2 cells, G6D3 cells, and CHO cells were grown on 6-well plastic plates (Corning Costar Corp., Cambridge, MA). Culture medium was replaced three times a week and cells were used 10 to 12 days post seeding. Fresh culture medium was replaced 24 h before transport experiments. One hour before transport experiments, culture medium was aspirated, monolayers were quickly rinsed twice with warm PBS, and monolayers were preincubated twice for 30 min each time in transport buffer. Where applicable, transport inhibitors were included in the final 30-min preincubation. In transport experiments designed to examine sodium dependence, PBS (140 mM Na⁺) was replaced with sodium-free PBS brought to equal osmolarity with choline chloride (140 mM). Preincubation buffer was aspirated and replaced with 1 ml of transport buffer containing substrates ± inhibitors. Stock solutions of quercetin and quercetin 4'--glucoside in ethanol were diluted with transport buffer before transport experiments. The resulting maximum final concentration of ethanol, 0.5%, did not affect the transport of mannitol, a marker of paracellular transport. Nystatin, an ionophore for monovalent cations (Vemuri et al., 1989), was dissolved in dimethyl sulfoxide (final concentration <0.1%) and controls with identical concentrations of dimethyl sulfoxide were examined. All other compounds were dissolved in transport medium. After 4 min, uptake was halted by rapidly aspirating uptake buffer and cells were rinsed three times with ice-cold transport buffer.

1

-Glucosidase Assay. Cell-free extracts were prepared from 11-day-old confluent Caco-2 monolayers grown in 100-mm Petri dishes. Cells were washed twice with PBS and scraped into PBS containing 1 mM phenylmethylsulfonyl fluoride and 10 mM 2-mercaptoethanol. The cell suspension was homogenized by fine needle aspiration and centrifuged at 19,000g for 10 min at 4°C. The resulting supernatant was assayed for protein content with the method of Smith et al. (1985) and stored on ice before use. For enzymatic assay, 200 µg of extract protein in phosphate buffer was preincubated for 15 min at 37°C. A final reaction volume of 0.5 ml was obtained by the addition of prewarmed quercetin 4'-glycoside in PBS at time zero. The reaction was halted by the addition of 0.5 ml of ice-cold methanol, followed by centrifugation at 3000g for 10 min. Samples were immediately analyzed by reversed phase HPLC. Apparent kinetic constants were obtained by fitting data to the Henri-Michealis-Menton equation (Segal, 1975) with the solver function in Microsoft Excel 2000.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Fluorescence Studies. In studies designed to determine whether cellular uptake of quercetin 4'--glucoside occurs, we took advantage of the intrinsic fluorescence of the quercetin ring system (Kuo, 1996). In these experiments, we used indirect fluorescent microscopy to examine confluent Caco-2 cell monolayers, grown for 14 days on glass coverslips. In control cells, incubated with buffer alone, no fluorescent signal was observed within the green spectrum (Fig. 1A). Cells incubated for 15 min with 50 µM quercetin aglycone, a molecule that crosses the Caco-2 cell monolayer (Walgren et al., 1998), demonstrated cytoplasmic and nuclear staining (Fig. 1B). The negatively charged and highly polar quercetin metabolite quercetin 3-sulfate demonstrated a complete absence of staining (Fig. 1C). In contrast, with our studies that demonstrated a lack of transcellular absorption of quercetin 4'--glucoside after apical loading (Walgren et al., 1998), Caco-2 cells that were incubated with 50 µM quercetin 4'--glucoside for 15 min demonstrated intracellular staining. This provided direct evidence for absorption of this glucoside across the apical membrane of Caco-2 cells (Fig. 1D).

View larger version (107K): [in this window] [in a new window]   Fig. 1.   Cellular uptake of quercetin and quercetin conjugates in Caco-2 cell monolayers. Confluent monolayers, grown for 14 days on glass coverslips, were incubated for 15 min at 37°C with warm transport buffer (A), 50 µM quercetin (B), 50 µM quercetin 3-sulfate (C), or 50 µM quercetin 4'--glucoside (D). Transport was halted by aspiration of buffer and cells were washed three times with ice-cold PBS. Coverslips were mounted on slides and examined by indirect fluorescence (original magnification, 100× oil). Scale bar, 10 µm.

Uptake Studies in Caco-2 Monolayers. To examine the mechanism of transport, we used a method in which methanolic extracts from the monolayers were examined by reversed phase HPLC analysis. This molecularly specific technique demonstrated that the intracellular fluorescence observed after apical exposure to quercetin 4'--glucoside was due to uptake of the glucoside, and not the result of metabolites such as the aglycone quercetin or its glucuronides (Walgren et al., 2000). As shown in Fig. 2, the observed rate of uptake after exposure to 100 µM quercetin 4'--glucoside for 4 min was 2.47 ± 0.12 pmol min¹ cm² for control (n = 3). This value is well above the minimum detectable amount for this glucoside of 0.001 pmol min¹ cm² but is less than 3% of the rate of uptake for 100 µM quercetin aglycone, 100.53 ± 2.80 pmol min¹ cm² (n = 3). In a previous study, we have shown that the MRP2 inhibitor MK-571 blocked the apical efflux of quercetin 4'--glucoside (Walgren et al., 2000). Consistent with this observation, the rate of cellular accumulation of quercetin 4'--glucoside increased 3.5-fold to 8.79 ± 1.04 pmol min¹ cm² (n  3, P < .05) in the presence of 50 µM MK-571 (Fig. 2). Subsequent experiments with Caco-2 cells were done in the presence of 50 µM MK-571 to enhance the observed rate of uptake by minimizing MRP2-mediated efflux.

View larger version (28K): [in this window] [in a new window]   Fig. 2.   Cellular uptake of quercetin 4'--glucoside in Caco-2 monolayers. Cells, grown on 6-well plates, were incubated for 4 min with 100 µM quercetin 4'--glucoside alone (control) or in the presence of 50 µM MK-571. Cellular uptake of quercetin 4'--glucoside was measured by a molecularly specific HPLC method. *P < .05, n  3.

1

2

1

2

1

2

1

2

1

2

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Cellular uptake of quercetin 4'--glucoside in Caco-2 monolayers. A, cells treated with 50 µM MK-571 were incubated for 4 min with 100 µM quercetin 4'--glucoside in the presence of a normal sodium buffer (control), a sodium-free buffer, or in the presence of 50 µM nystatin in a normal sodium buffer. *P < .01, n  5. B, cells treated with 50 µM MK-571 were incubated for 4 min with 100 µM quercetin 4'--glucoside (control) or with the addition of either 0.5 mM phloridzin (PLZ) or 30 mM glucose. **P < .001, n  9; P < .01, n = 3.

Uptake Studies in SGLT1-Transfected G6D3 Cells. G6D3 cells, a CHO cell line that has been stably transfected with SGLT1 (Lin et al., 1998), were used to examine the transport of quercetin 4'--glucoside. In our initial experiments with these cells, we examined the uptake of quercetin 4'--glucoside alone or in the presence of 30 mM glucose. A low rate of cellular uptake was observed in the presence of glucose, 0.30 ± 0.04 pmol min¹ cm², or about one-quarter that observed in the control treatment, 1.14 ± 0.20 pmol min¹ cm² (n  6, P < .001; Fig. 4). While we were doing these experiments, Barnouin et al. (1998) reported finding expression of the hamster homolog of MRP1 (Mrp1) in parental CHO-K1 cells. Because similar substrate specificities have been found for MRP1 and MRP2 (König et al., 1999), we also examined the influence of 50 µM MK-571 on the uptake of quercetin 4'--glucoside in these SGLT1-transfected CHO cells. As seen in Fig. 4, the rate of uptake was almost 6.5 times greater with the MRP inhibitor MK-571, 7.35 ± 0.55 pmol min¹ cm² (n  15, P < .0001). Next, we examined uptake of quercetin 4'--glucoside in G6D3 cells for inhibition by phloridzin and glucose. In G6D3 cells, the presence of either 0.5 mM phloridzin or 30 mM glucose reduced the uptake of quercetin 4'--glucoside by approximately one-half, 3.75 ± 0.25 pmol min¹ cm² for phloridzin and 3.83 ± 0.27 for glucose versus 7.35 ± 0.55 pmol min¹ cm² for control (n  14, P < .0001; Fig. 4), analogous to the findings in Caco-2 cells (Fig. 3).

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Cellular uptake of quercetin 4'--glucoside in G6D3 cells. G6D3 cells were incubated for 4 min with 100 µM quercetin 4'--glucoside alone (control) or in the presence of 30 mM glucose, 50 µM MK-571, 50 µM MK-571 and 0.5 mM phloridzin (PLZ), or 50 µM MK-571 and 30 mM glucose. *P< .001, significantly lower than control, n  6; **P < .0001, significantly greater than control, n  15; P < .0001, significantly lower than in the presence of MK-571 alone, n  14.

1

2

1

2

1

2

1

2

View larger version (39K): [in this window] [in a new window]   Fig. 5.   Sodium-dependent uptake of [14C]glucose and quercetin 4'--glucoside in CHO-K1 and G6D3 cells. Both G6D3 cells and CHO cells were incubated with 100 µM [14C]glucose (A) or 100 µM quercetin 4'--glucoside (B) for 4 min in either a sodium (dark bars) or sodium free (light bars) buffer. *P < .001, n = 6, significantly less than uptake in G6D3 cells in the presence of sodium; P < .0001, n  3, significantly less than uptake uptake in G6D3 cells in the presence of sodium.

-Glucosidase Assay. The ability of Caco-2 cells to hydrolyze quercetin 4'-glucoside to quercetin aglycone was examined in cell-free extracts. The reaction was linear over 1 h (Fig. 6A) and recovery was >99%. No activity was observed in heat-inactivated cell extracts (data not shown). The -glucosidase activity at various concentrations was measured based on the formation of the aglycone and used to determine apparent kinetic constants (Fig. 6B). The cell extract demonstrated an apparent K_m for the hydrolysis of quercetin 4'-glucoside of 78.4 µM and a V_max of 0.27 mU mg of protein¹ (enzymatic unit of activity where one unit is defined as 1 µmol of product formed per minute at 37°C).

View larger version (11K): [in this window] [in a new window]   Fig. 6.   Hydrolysis of quercetin 4'--glucoside to quercetin aglycone was examined in cell-free extracts from Caco-2 cells. A, prewarmed cell-free extracts were incubated with 30 µM quercetin 4'--glucoside for various times at 37°C. B, prewarmed cell free extracts were incubated with various concentration of quercetin 4'--glucoside for 15 min at 37°C. Points represent mean ± S.E. of triplicate determinations.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

To examine the relative absorption of dietary flavonoids and to gain insight into the mechanisms governing their absorption we have used human Caco-2 cell monolayers, a widely accepted model of human intestinal absorption (Artursson and Karlsson, 1991; Yee, 1997). In our previous studies, we have demonstrated that although the minor dietary flavonoid quercetin aglycone is well absorbed, the major dietary flavonoid quercetin 4'--glucoside is not (Walgren et al., 1998). This finding was in sharp contrast to indirect evidence that this glucoside is absorbed via the active intestinal glucose transporter (Hollman et al., 1995; Gee et al., 1998). Further studies with the Caco-2 cell model have demonstrated that quercetin 4'--glucoside is, in fact, effluxed by the apically expressed MRP2 (Walgren et al., 2000). Although these studies did not provide evidence for absorption of quercetin 4'--glucoside, MRP2-dependent efflux may have prevented detection of absorption, including possible transport by the sodium-dependent glucose transporter SGLT1.

The results from this study demonstrate that quercetin 4'--glucoside is indeed transported by SGLT1. With the intrinsic fluorescence of quercetin and its derivatives, we have been able to examine the uptake of these compounds across the apical membrane of Caco-2 cells grown on glass coverslips. Consistent with our transcellular absorption studies, clear evidence was obtained for the uptake of quercetin aglycone. Despite the lack of transcellular absorption observed in our previous studies (Walgren et al., 1998), cells incubated with quercetin 4'--glucoside also demonstrated an intracellular staining, supporting absorption of the glucoside across the apical membrane (Fig. 1D). Transport of the intact and unaltered glucoside was confirmed by HPLC analysis of extracts from Caco-2 monolayers incubated with quercetin 4'--glucoside. As predicted based on the subcellular localization of MRP2 to the apical membrane of Caco-2 cells (Walgren et al., 2000), the cellular uptake of the glucoside was enhanced in the presence of the MRP2 inhibitor MK-571 (Fig. 2). We have demonstrated that the uptake of quercetin 4'--glucoside across the apical membrane of Caco-2 cells is sodium-dependent and is inhibited by both glucose, a substrate of SGLT1, and phloridzin, a competitive inhibitor of SGLT1 (Fig. 3).

As additional evidence for transport of quercetin 4'--glucoside by SGLT1 we have examined transport of the glucoside in G6D3 cells, a CHO cell line that has been stably transfected with rabbit SGLT1 (Lin et al., 1998). Uptake of quercetin 4'--glucoside in G6D3 cells was inhibited by both glucose and phloridzin. The unexpected presence of the MRP1 efflux pump also limited cellular absorption in this cell line. However, the use of MK-571 to inhibit this pump enabled us to selectively examine SGLT1-dependent transport (Fig. 4). Furthermore, the observed transport rate for quercetin 4'--glucoside uptake was greater in transfected G6D3 cells than in parental CHO-K1 cells, and the transport demonstrated sodium dependence in the G6D3 cells but not in the CHO-K1 cells (Fig. 5).

Based on our accumulated results, the intestinal absorption of dietary flavonoids is becoming more clearly defined (Fig. 7). The major dietary flavonoid, quercetin 4'--glucoside, is absorbed across the brush-border membrane by SGLT1. MRP2, also localized to the apical membrane, is capable of effluxing quercetin 4'--glucoside and effectively opposing absorption and intracellular accumulation. However, upon gaining access to the cytosol, quercetin 4'--glucoside may undergo metabolism such as hydrolysis, yielding quercetin aglycone. Hydrolysis by a broad specificity -glucosidase is supported by our results obtained with extracts from Caco-2 cells (Fig. 6) and by the recent identification of a similar enzymatic activity within the human intestine (Day et al., 1998). Quercetin aglycone may then undergo further metabolism (Boulton et al., 1998; Crespy et al., 1999; Walle et al., 1999b) or it may cross the basolateral membrane, resulting in absorption (Walgren et al., 1998).

View larger version (19K): [in this window] [in a new window]   Fig. 7.   Model of quercetin 4'--glucoside absorption in the enterocyte. Quercetin 4'--glucoside (Q4'G), the predominant dietary form of quercetin, is actively absorbed across the apical membrane by SGLT1 and is actively effluxed across the apical membrane by MRP2. Although present in the cytosol, quercetin 4'--glucoside is unable to exit the cell at the basolateral membrane. However, it is able to enter the cell across the basolateral membrane, suggesting the presence of an unidentified transporter. Once within the cell, the glucoside may undergo metabolism such as hydrolysis to produce quercetin aglycone (Q).

Although this study has revealed the interplay in the flux of quercetin 4'--glucoside across the apical membrane of Caco-2 cells between the inwardly directed SGLT1 and the outwardly directed MRP2, it is still not clear why transcellular absorption does not occur. In fact, even in the presence of the MRP2 inhibitor MK-571, transepithelial absorption of the glucoside has been detected only after apical loading with very high concentrations of the glucoside. Even then, only a very minute level of transport was observed, suggesting the presence of a saturable, intracellularly directed transporter at the basolateral membrane. In contrast, after basolateral loading, the rate-limiting step in the transcellular flux of quercetin 4'--glucoside appears to be transport by MRP2 on the apical side (Walgren et al., 2000).

Interestingly, a similar multitransporter system has recently been described within the renal proximal tubule. In proximal tubule cells, organic anions enter the cell across the basolateral membrane via the organic anion transporter OAT, or via another as yet unidentified transporter. Once within the cell, they are subsequently effluxed across the apical membrane by MRP2 or potentially by a second unidentified transporter (Masereeuw et al., 1999). We have been able to rule out the possibility that quercetin 4'--glucoside crosses the basolateral membrane of the Caco-2 cell via a sodium-dependent process because the transcellular efflux of this glucoside is not reduced in the absence of sodium (R. A. Walgren and T. Walle, unpublished data).

Absorption of the flavonoid glucosides, across the apical membrane, appears to be governed by a balance between uptake by SGLT1 and efflux by MRP2. Although this balance may favor efflux by MRP2 for quercetin 4'--glucoside and genistein-7-glucoside (Walle et al., 1999a), it may be shifted to favor absorption by SGLT1 for other glucosides. For example, a recent study demonstrated clear absorption of the flavonoid cyanidine 3--glucoside in the intact rat (Tsuda et al., 1999). It may be that this flavonoid, which carries a positive charge on the aglycone moiety, is not a substrate for the MRP2 efflux pump and, hence, is efficiently absorbed via SGLT1. Aside from flavonoids, a normal diet contains a number of other glucosides, including -glucoside derivatives of niacin in wheat bran and pyridoxine 5'--glucose, the major dietary form of vitamin B₆ (Gregory, 1998). A number of therapeutic agents also are used orally as glycosides, including the aminoglycoside antibiotic neomycin, which is unabsorbed, and the orally absorbed cardiac glycoside digoxin. It is interesting to note that although attempts have been made to improve the intestinal absorption of peptides and nucletide analogs by glycosylation with glucose (Nomoto et al., 1998; Mizuma et al., 1999), the addition of glucose may not only produce a substrate for SGLT1 but also a substrate for MRP2. This possibility, which is of great relevance to new drug development, will require further studies.

A lack of intestinal absorption of the flavonoid glucosides does not imply a lack of biological effects from these compounds in humans. It has long been argued that enzymes from the colonic microflora can hydrolyze flavonoid glucosides (Griffiths and Barrow 1972; Bokkenheuser et al., 1987). In addition, a recent study in ileostomy patients demonstrated that the dietary quercetin glucosides are completely hydrolyzed before the colon,² potentially suggesting a nonbacterial source of -glucosidase in the human intestine. As shown by Day et al. (1998) and confirmed in this study (Fig. 6), enterocytes have the ability to hydrolyze absorbed quercetin 4'--glucoside within the cytosol. The fate of the resulting quercetin aglycone may include transport out of the epithelial cell (Walgren et al., 1998) or metabolism by a variety of enzymes (Boulton et al., 1998; Crespy et al., 1999; Walle et al., 1999b), with presently unknown consequences.

In summary, this study demonstrates for the first time that quercetin 4'--glucoside, the most abundant dietary flavonoid, is transported by SGLT1 across the apical membrane of enterocytes. However, its transcellular absorption is limited by MRP2-mediated efflux across the apical membrane as well as by an unknown transporter on the basolateral membrane.