- Email: [email protected]
Transepithelial transport and accumulation of flavone in human intestinal CACO-2 cells
Life Sciences, Vol. 63, No. 26, pp. 23252331, 1998 Copyright0 1998 Elsevier Science Inc. Printed in the USA. All riehts mewed @X4-3205/91X-$19.00 t .30 ELSEVIER
PII SOO24-3205(98)00521-9
TRANSEPITHELIAL TRANSPORT AND ACCUMULATION IN HUMAN INTESTINAL CACO-2 CELLS
OF FLAVONE
Shiu-Ming Kuo’ Nutrition Program, Department of Physical Therapy, Exercise and Nutrition State University of New York, Buffalo, NY 14214, USA (Received
Sciences
in final form October 13, 1998)
Summary Flavonoids are found in many food items of plant origin. Intake of flavonoids has been linked to the prevention of human diseases including cancer. However, little is known about the intestinal absorption of flavonoids in the cellular level. This study was designed to study the absorption of dietary flavonoids using cultured human intestinal epithelial ceil monolayer as a model system and 14C-flavone as a model compound. 14C-flavone at 10 uM was found to move across the cell monolayer rapidly both from the luminal to basolateral direction and from the basolateral to luminal direction. The rate of transport from the luminal to basolateral direction was 5 times of the rate for phenylalanine, an aromatic amino acid. Flavone also accumulated substantially in the cells. Replacing sodium in the transport buffer with potassium did not affect the transport but reducing the incubation temperature significantly decreased the initial rate of transport. The presence of protein in the transport buffer reduced the initial rate of transport to half. Other flavonoids and hydrophobic chemicals at 100 pM had no effects on the transport. Together with the evidence from microscopic observation (Cancer Letts. 110: 41-48, 1996) this study supports that rapid difisional transport may be the main route for flavonoid absorption. The ability of intestinal cells to accumulate flavone is consistent with the role of flavonoids in colon cancer prevention. &y Words: intestine, tlavonoid, flavone Dietary flavonoids are widely present in food items of plant origin (1). They are known to be anticarcinogenic (2-4) and have been shown to have properties of antioxidants (S-9). In epidemiological studies, flavonoid intake has also been linked to a reduced incidence of some forms of cancers including colon cancer (10, 11). To be able to determine the level of intake that will maximize the health-promoting effects of this class of compounds, some knowledge in the bioavailability of flavonoids is required. Previous studies performed with human subjects indicated
Tel: 716-829-3680;
Fax: 716-829-3700;
E-mail: [email protected]
2324
Intestinal Transport of Flavonoid
Vol. 63, No. 26, 1998
that flavonoids, in general, were absorbed to appreciable amounts (12-16). The absorption process was rapid (12, 16) and flavonoids remained quite stable in the luminal fluid (15). However, the mode of transport of this class of compound across the intestinal epithelial cells was not characterized. The goal of this study was to use cultured human intestinal cells to investigate the absorption event. Human intestinal Caco-2 cells were known to differentiate under proper culture condition and exhibit the phenotype of mature enterocytes (17, 18). Upon growing on permeable membranes, Caco-2 cells have been established in previous studies as a model of human intestine to study the transepithelial absorption and secretion of chemicals (19-22). In this study, they were used to characterize the mode of transepithelial transport of flavone, a member of the flavonoid family. Flavone was chosen in this study because of its known chemical stability under the incubation condition (23); and the availability of radiolabeled compound. In addition to the transepithelial transport, cellular accumulation of flavone at the end of each transport study was measured. Since flavonoid intake has been linked to the prevention of colon cancer, the hypothesis of the study was that intestinal cells accumulate flavone in significant amount. Several factors that may affect the transport and accumulation of flavone were also investigated.
Material and Methods_ Cell Culture Caco-2 cells were purchased from American Type Culture Collection and passages 24-35 were used for the experiments. These cells were cultivated in supplemented high-glucose (4.5 g/L) Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum as described before (23). Culture cups with permeable membrane inserts (Costar Transwell clear, 24 mm diameter, 0.4 uM pore size) placed in 6-well plates were used for transport studies. Cells were seeded at 2 x 1OS/cm2 on the membrane inserts with 1.5 ml medium in the luminal side and 2.5 ml medium in the bottom well facing the basolateral side. Transepithelial resistance of cells grown in the Transwell was measured using chopstick electrodes (World Precision Instruments). The increase in resistance was used as an indicator of epithelial cell differentiation. Intestinal cell lines, for example, HT-29 or IEC-6 that are not capable of spontaneous differentiation do not develop transepithelial resistance (results not shown). Only monolayers with resistance higher than 250 n/cup were used for experiments (mostly at 14 days post seeding). Transnort studv Transport studies were performed as described before (22). Briefly, growth medium in the luminal and basolateral sides of the Transwell were replaced with 1.5 ml and 2.5 ml, respectively, of transport buffer (Hanks Balanced Salt Solution supplemented with 25 mM HEPES and 100 unit penicillin, 100 pg streptomycin per ml). After 30 minutes of preincubation with transport buffer, freshly prepared [14C]-flavone solution was added to the luminal side or basolateral side to start the transport reaction. In general, the concentration of flavone was 10 pM with the total radioactivity in each well at 0.05 pCi. All transport experiments were performed on a platform rocker (60 rpm) inside a humidified 37°C chamber. Duplicate samples of 50 pl transport buffer were taken from the opposite side of the well at each time point to determine the rate of transepithelial transport. The radioactivity in the sample was measured by liquid scintillation
Intestinal Transport of FIavonoid
Vol. 63, No. 26, 1998
2325
technique and normalized by total protein content of cells on the membrane insert. Determination of cellular accumulation Cellular accumulation of radioactivity was determined as described before (22). At the end of the experiment, membrane inserts with cells were removed from Transwell cups, rinsed with icecold phosphate buffer saline stop solution and subjected to overnight solubilization in 3 ml 0.1% SDS at 4°C. Radioactivities in the lysate were then determined by liquid scintillation technique. The total recovery of radioactivity in the luminal chamber, basolateral chamber and membrane was usually >95%. Total protein content on each Transwell membrane was determined by a modified Lowry assay using bovine serum albumin as the standard (24). Chemicals All chemicals used were of analytical grade with greater than 95% purity. Reagents used for cell culture were all from GIBCO-BRL (Grand Island, NY) with the exception of fetal bovine serum which was from Hyclone (Logan, UT). [t4C]-flavone (25) was a gift from Proctor & Gamble Company (Cincinnati, OH) with greater than 99% purity. It was stored as ethanol solution in -7O’C freezer prior to usage. Statistical analvsis The data at each time point were analyzed by ANOVA and post-hoc Student-Newman-Keuls multiple comparison using StatView 4.01. P-value less than 0.05 was considered statistically significant.
Results
+ B q
- 0
15
30
L
to L (flavone) to B (phenylalanine
45
60
75
TIME (min) Fig. 1
Transport of 1OuM l 4C-flavone across Caco-2 cell monolayers in the luminal to basolateral direction (L to B), and basolateral to luminal direction (B to L) and a comparison to the rate of luminal to basolateral (L to B) transport of 10 uM phenylalanine. Data shown were meanskS.D. of three wells. *, b indicate significant differences at PcO.05 within the same time point.
2326
Intestinal Transport of Flavonoid
Vol. 63, No. 26, 1998
Caco-2 cells cultured in Transwell cups have been shown to differentiate spontaneously (17, The cell surface facing the top medium developed brush border which resembled the luminal membrane of the intestinal epithelium. The cell surface attaching to the permeable membrane and facing the bottom medium developed into the basolateral membrane. The differentiated cell monolayers can thus be used for the study ofunidirectional transport of substrates. The transport of flavone from luminal to basolateral side was measured by including t4C-flavone in the top transport buffer and measuring the appearance of isotope in the bottom transport buffer. The transport of flavone from basolateral to luminal side was measured by including t4C-flavone in the bottom buffer and measuring the appearance of isotope in the top buffer. As shown in Fig. 1, the rates of flavone transport in these two directions were almost the same. Transepithelial transport of aromatic amino acid, phenylalanine, has been demonstrated previously in Caco-2 cell monolayer (19). It was found to be carrier-mediated and IO-times faster in the luminal to basolateral direction compared to the basolateral to the luminal direction (19). The luminal to basolateral transport of phenylalanine was measured here as well. Compared to phenylalanine, the rate of transport for flavone was 5 times greater (Fig. 1). 18).
To further characterize the mode of transport for flavone, the effects of temperature and inward sodium radient were also examined. In the luminal to basolateral transport of flavone, 5 substituting Na m transport buffer with K+ did not affect the rate of transport (Fig. 2A), but lowering the incubation temperature decreased the rate of transport (Fig. 2A). Similar temperature effect was also observed for the transport of flavone from the basolateral to luminal direction (Fig. 2B). Substituting Na+ in transport buffer with K+ did not affect the rate of transport in the basolateral to luminal direction, either (Fig. 2B)
TIME (min) Fig. 2
Effect of sodium gradient and incubation temperature on the transepithelial transport of 14C-flavone. L-B: 10uM 14C-flavone transport from the luminal side of the cells to the basolateral side; B-L: 10uM 14C-flavone transport horn the basolateral side of the cells to the luminal side; Na: sodium gradient in the direction of transport; K: no sodium gradient, Na replaced by K. Each point represents mean*S.D. of three Transwells. *,b, c indicate significant differences at P
Vol. 63, No. 26, 1998
Intestinal Transport of Flavonoid
2327
There were extensive cell-associated radioactivities at the end of one hour transport study. In the luminal to basolateral transport experiment conducted for Fig. 2A, there was 2.14 f 0.15 mnol flavone/mg protein at 60 minute incubation. Incubation temperatures and the presence or absence of inward sodium gradient had no effects on the amount of flavone accumulated in the cells. In the basolateral to luminal transport study conducted for Fig. 2B, there was 2.67 f 0.16 run01 flavone/mg protein at the 60 minute incubation in the two 37°C groups (no diierence between the two); and 1.90 f 0.08 mnol flavone/mg protein in the two lower temperature groups (no diierence between the two). In comparison, the amount of cell-associated phenylalanine was 381*41 pmol/mg protein at the end of one hour transport study of phenylalanine.
TIME (min)
Fig. 3
Effect of luminal addition of other agents on the transepithelial transport of 10 pM 14C-flavone from the luminal side of the cells to the basolateral side. BSAl: 1 mg/ml bovine serum albumin; BSAS: 5 mg/ml bovine serum albumin; PHI? 5 mM phenylalnine; MET: 5 mM methionine; ASA: 5 mM ascorbic acid. Each point represents meanG.D. of three transwells. *,b, C indicate significant differences at PcO.05 within the same time point. There were no differences among the treatment groups for all time points in Fig. 3B and 3C.
2328
Intestinal
Transport
of Flavonoid
Vol. 63, No. 26, 1998
Most flavonoids including flavone are hydrophobic. The effect of other hydrophobic compounds on the transepithelial transport of 10 pM i4C-flavone was studied. Due to the limited solubility of the hydrophobic compounds used, they were only tested at 100 uM in most of the cases. When basolateral to luminal transport of 10 uM *4C-flavone were followed for 90 minutes, additional unlabeled 100 uM flavone in the luminal or basolateral side had no effect on the transport. Additional 100 uM quercetin, genistein (both members of the flavonoid family), diethylstilbestrol (synthetic estrogen), nifedipine (calcium channel blocker), or vinblastine (antitumor alkaloid) also had no effect on the transport. In addition, acebutolol @-blocker) at 1 mM had no effect on the transport (results not shown). These hydrophobic compounds in the buffer did not affect the amount of cell-associated flavone at the end of the transport study (results not shown). Flavonoids were found to bind to extracellular proteins (26). The effect of protein in the transport of flavone was studied. As shown in Fig. 3A, when bovine serum albumin (BSA) was included in the luminal side, it decreased the luminal to basolateral transport of 14C-flavone in a dose-dependent fashion. The cell-associated i4C-flavone at the end of one hour incubation was also decreased in this study when BSA was included in the incubation (1.52*0.23, 1.30-+0.04, and 0.81*0.04 nmol/mg protein for control, 1 mg/ml BSA, and 5 mg/ml BSA group, individually). The effect of neutral amino acids on the luminal to basolateral transport and accumulation of r4Cflavone was then studied. As shown in Fig. 3B, 5 mM phenylalanine or methionine had no effect on the transport of *4C-flavone. The cell-associated radioactivity at the end of one hour incubation also remained the same when cells were incubated in buffers containing additional phenylalanine or methionine (1.52+0.23, 1.61&0.07 and 1.54kO.07 nmol/mg protein for control, phenylalanine and methionine group, individually). We have found that flavone decreased the accumulation of ascorbic acid in Caco-2 cells (22). Here, the effect of 5 mM ascorbic acid on the transport of 10 uM i4C-flavone was studied. As shown in Fig. 3C, ascorbic acid had no effect on the transport of flavone. The cell-associated radioactivity also remained the same in the ascorbic acid group at the end of one hour incubation (1.52*0.05 and 1.53*0.07 nmol/mg protein for control and ascorbic acid group, respectively). In order to understand whether the cell-associated radioactivity represents high-affinity binding of flavone to intracellular components, efflux experiments were performed. Caco-2 cells were first incubated with 10 uM 14C-flavone in the luminal side for 30 minutes. The total cell-associated flavone was found to be 1.02*0.07 nmol/mg protein (n=3). The 14C-flavone-containing buffer was then removed and replaced with 2 ml fresh non-flavone containing buffer for efflux measurement. At the end of the 30 minute et&x, 89*4% of the original cell-associated radioactivity was recovered in the buffer. Another experiment was conducted by incubating cells with 100 uM 14Cflavone for 1 hour. The total ceil-associated flavone was 12.3*1.8 nmol/mg protein at the end of one-hour incubation. The efflux of i4C-flavone was then followed. 68*4% of 14C-flavone was found in the buffer at the end of 3Ominute eIIlux and 74*4% of 14C-flavone was found in the buffer at the end of 60-minute efflux. Results from both studies showed a rapid efflux of the accumulated flavone. Discussions This is the first study exploring the transepithehal transport of flavonoids at the cellular level. Several lines of evidences support that the transepithelial transport of flavone across intestinal
Vol. 63, No. 26, 1998
Intestinal Transport of Flavonoid
2329
cells is a passive diffusion event: the transport is much more rapid compared to that of phenylalanine (Fig. 1) or ascorbic acid (22); is bidirectional at about the same rate (Fig. 1, 2); can not be inhibited by any of the small hydrophobic compounds we tested including flavone itself (Fig. 3 and see results); is not affected by the removal of inward sodium gradient (Fig. 2). Lower incubation temperature did lead to a decrease in the rate of transport, however, substantial transport was still observed even when the incubation temperature was decreased to 1” C Although the focus of this study is intestinal cells, it would not be surprising that passive diffusion of flavone and other flavonoids also occurs in other cell types. Similar transepithelial transport study should be applied in the future to other flavonoids with hydroxyl groups on the ring structure, for example, flavonols and flavanols; and to flavonoids with glycoside conjugates. Using a different approach with fluorescence microscope and cells grown on glass coverslips, quercetin (a flavonol) and rutin (quercetin with glycoside conjugate) were found to rapidly move into Caco-2 cells (23). These observation are consistent with the current radioisotope study in supporting a diffusional membrane transport for flavonoids in general. Flavone accumulated extensively in the cells (see results). The fluorescence microscopy study described above also supports an accumulation of quercetin and rutin in the cells (23). The protein-binding property of flavone as reported by another group (26) and noted in this study (Fig. 3A) may explain why largeamounts of cell-associated radioactivity was observed. Although ceil volume was not measured in this study, based on the reported cells volume of 3.49-6 ul per mg protein for Caco-2 cells (18, 27, 28) the net concentration of cell-associated flavone was clearly above that of the medium. Although the subcellular distribution of 14C-flavone, i.e whether the radioactivity is on the cell surface or inside the cells, can not be determined here, in our microscopy study quercetin was found to distribute throughout the cell with no particular enrichment on the cell surface. Neither the accumulation of 14C-flavone in this study or quercetin in the microscopic study could be affected by additional flavonoids at concentrations within their solubility limit (100 PM). These observations support the presence of high-capacity binding site(s) for flavone in the cells. Although the rapid efflux of flavone from cells indicates that majority of the binding is of lowaffinity, significant amount of flavone remained in the cells. For example, the cells incubated with 10 uM flavone for 30 min lost 89% of the_flavone during the subsequent efflux but the cellassociated flavone still had a concentration of 19pM (assuming 6 ul/mg protein). The presence of high-afhnity binding sites for flavone is consistent with the known biological activities of this class of compounds (4). The binding of flavone to BSA in the incubation has led to an impedance in the transepithelial transport and cellular accumulation of flavone in a dose-dependent fashion (Fig. 3A). This binding is not due to a specific interaction with the hydrophobic side-chain of amino acids, phenylalanine or methionine, as these amino acids themselves have no effects on the transepithelial transport of *4C-flavone (Fig. 3B). We have observed that flavonoids, including flavone, inhibited the accumulation of ascorbic acid in intestinal cells (22). With differences in structure and hydrophobicity between flavonoids and ascorbic acid, it is not expected that they compete for the same transporter. In this study, ascorbic acid at 5 mM failed to affect the transport or accumulation of 1OuM t4C-flavone (Fig. 3). It thus supports our original hypothesis that the inhibitory effect of flavonoids is not mediated
2330
Intestinal Transport of Flavonoid
through a competition
Vol. 63, No. 26, 1998
at the transport site (29).
This generaldiffisive be readily available upon by the appearance in the observation here supports
property of flavonoids observed implies that the dietary flavonoids will ingestion. Indeed, rapid intestinal absorption of flavonoids as determined bile or plasma has been observed in rats (30, 31) and human (16). The that intestine transport is not a rate-limiting step for flavonoids.
Acknowledement The work was supported partly by a grant from Buffalo Foundation, Buffalo, New York. The generosity of Proctor & Gamble Company (Cincinnati, OH) in providing l 4C-flavone is greatly appreciated. The author thanks Chi-Ping Lin for performing the efflux experiment and Penny Leavitt for editing the manuscript. References 1. W, S PIERPOINT,
Plant Flavonoids in Biology and Medicine: Biochemical, Pharmacological, andStructure-ActivityRelationships, V. Cody, E. Middleton and J. B. Harbome (Eds), 125
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
140, Alan R. Liss, New York (1986). A.K. VERMA, J.A. JOHNSON, M.N. GOULD and M.A. TANNER Cancer Res. 48 57545758 (1988). E.E. DESCHNER, J. RUPERTO, G. WONG and H.L. NEWMARK, Carcinogenesis 12 1193-1196 (1991). S.-M. KUO, Crit. Rev. Oncogenesis 8 47-69 (1997). M.J. SANZ, M.L. FERRANDIZ, M. CEJUDO, M.C. TERENCIO, B. GIL, B. BUSTOS, A, UBEDA, R. GUNASEGARAN and M.J. ALCARAZ, Xenobiotica 24 689-699 (1994). H. WEI, R. BOWEN, Q.Y. CAI, S. BARNES and Y. WANG, Proc. Sot. Exp. Biol. Med. 208 124-130 (1995). A. SAIJA, M. SCALESE, M. LANZA, D. MARZULLO, F. BONINA and F. CASTELLI, Free Radic. Biol. Med. 19 481-486 (1995). I.R. RECORD, I.E. DREOSTI and J.K.MCINERNEY, J. Nutr. Biochem. 6 481-485 (1995). N. COTELLE, J. BERNIER, J. CATTEAU, J. POMMERY, J. WALLET and E. M. GAYDOU, Free Radic. Biol. Med. 20 35-43 (1996). D. INGRAM, K. SANDERS, M. KOLYBABA and D. LOPEZ, Lancet 350 990-994 (1997). P. KNEKT, R. JARVINEN, R. SEPPANEN, M. HELIOVAARA, L. TEPPO, E. PUKKALA and A. AROMAA, Am. J. Epidemiol. 146 223-230 (1997). D. COVA, L. DE ANGELIS, F. GIAVARINI, G. PALLADINI and R. PEREGQ, Inter. J. Clin. Pharmacol. Thera. Toxicol. 30 29-33 (1992). X. XU, H.-J. WANG, P.A. MURPHY, L. COOK and S. HENDRICH, J. Nutr. 124 825-832 (1994). M.S. MORTON, G. WILCOX, M. L. WAHLQVIST and K GRIFFITHS, J. Endocrinol. 142 251-259 (1994). P.C.H. HOLLMAN, J.H.M. DE VRIES, S.D. VAN LEEUWENT, M.J.B. MENGELERS and M.B. KATAN, Am. J. Clin. Nutr. 62 1276-1282 (1995). R. A. KING and D.B. BURSILL, Am. J. Clin. Nutr. 67 867-872 (1998) M. PINTO, S. ROBINE-LEON, M.D. APPAY, M. KEDINGER, N. TRIADOU, E.
Vol. 63, No. 26, 1998
18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.
Intestinal Transport of FIavonoid
2331
DUSSAULX, B. LACROIX, P. SIMON-ASSMANN, K. HAFFEN, J. FOGH and A. ZWEIBAUM, Bio. Cell 47 323-330 (1983). I.J. HIDALGQ, T.J. RAUB and R.T. BORCHARDT, Gastroenterology 96 736-749 (1989). I.J. HIDALGO and R.T. BORCHARDT, Biochim. Biophys. Acta 1028 25-30 (1990). P. ARTURSSON, J. Pharma. Sci.79 476-482 (1990). J. KARLSSON, S.-M. KUO, J. ZIEMNIAK and P. ARTURSSON, Br. J. Pharmacol. 110 1009-1016 (1993). S.-M. KUO, H.F. MOREHOUSE and C.-P. LIN, Cancer Lett. 116 131-137 (1997). S.-M. KUO, Cancer Lett. 110 41-48 (1996). G.L. PETERSON, Methods Enzymol. 9195-l 19 (1983). H. LIU and K. R. WEHMEYER, J. Chromatogr. 577 61-67 (1992) S.F. ARNOLD, B.M. COLLINS, M.K. ROBINSON, L.J. GUILLETTE, JR. and J.A. MCLACHLAN, Steroids 61642-646 (1996) A. BLAIS, P. BISSONNETTE and A. BERTELOOT, J. Membrane Biol. 99 113-125 (1987). D.B. BURNHAM and J.D. FONDACARO, Am. J. Physiol. 256 G808-G816 (1989). C.-P. LlN and S.-M. KUO, FASEB J. 10 A770 (1996) R.A. KING, J.L. BROADBEND and R.J. HEAD, J. Nutr. 126 1761182 (19%) J. SFAKIANOS, L. COWARD, M. KIRK and S. BARNES, J. Nutr. 127 1260-1268 (1997)
PII SOO24-3205(98)00521-9
TRANSEPITHELIAL TRANSPORT AND ACCUMULATION IN HUMAN INTESTINAL CACO-2 CELLS
OF FLAVONE
Shiu-Ming Kuo’ Nutrition Program, Department of Physical Therapy, Exercise and Nutrition State University of New York, Buffalo, NY 14214, USA (Received
Sciences
in final form October 13, 1998)
Summary Flavonoids are found in many food items of plant origin. Intake of flavonoids has been linked to the prevention of human diseases including cancer. However, little is known about the intestinal absorption of flavonoids in the cellular level. This study was designed to study the absorption of dietary flavonoids using cultured human intestinal epithelial ceil monolayer as a model system and 14C-flavone as a model compound. 14C-flavone at 10 uM was found to move across the cell monolayer rapidly both from the luminal to basolateral direction and from the basolateral to luminal direction. The rate of transport from the luminal to basolateral direction was 5 times of the rate for phenylalanine, an aromatic amino acid. Flavone also accumulated substantially in the cells. Replacing sodium in the transport buffer with potassium did not affect the transport but reducing the incubation temperature significantly decreased the initial rate of transport. The presence of protein in the transport buffer reduced the initial rate of transport to half. Other flavonoids and hydrophobic chemicals at 100 pM had no effects on the transport. Together with the evidence from microscopic observation (Cancer Letts. 110: 41-48, 1996) this study supports that rapid difisional transport may be the main route for flavonoid absorption. The ability of intestinal cells to accumulate flavone is consistent with the role of flavonoids in colon cancer prevention. &y Words: intestine, tlavonoid, flavone Dietary flavonoids are widely present in food items of plant origin (1). They are known to be anticarcinogenic (2-4) and have been shown to have properties of antioxidants (S-9). In epidemiological studies, flavonoid intake has also been linked to a reduced incidence of some forms of cancers including colon cancer (10, 11). To be able to determine the level of intake that will maximize the health-promoting effects of this class of compounds, some knowledge in the bioavailability of flavonoids is required. Previous studies performed with human subjects indicated
Tel: 716-829-3680;
Fax: 716-829-3700;
E-mail: [email protected]
2324
Intestinal Transport of Flavonoid
Vol. 63, No. 26, 1998
that flavonoids, in general, were absorbed to appreciable amounts (12-16). The absorption process was rapid (12, 16) and flavonoids remained quite stable in the luminal fluid (15). However, the mode of transport of this class of compound across the intestinal epithelial cells was not characterized. The goal of this study was to use cultured human intestinal cells to investigate the absorption event. Human intestinal Caco-2 cells were known to differentiate under proper culture condition and exhibit the phenotype of mature enterocytes (17, 18). Upon growing on permeable membranes, Caco-2 cells have been established in previous studies as a model of human intestine to study the transepithelial absorption and secretion of chemicals (19-22). In this study, they were used to characterize the mode of transepithelial transport of flavone, a member of the flavonoid family. Flavone was chosen in this study because of its known chemical stability under the incubation condition (23); and the availability of radiolabeled compound. In addition to the transepithelial transport, cellular accumulation of flavone at the end of each transport study was measured. Since flavonoid intake has been linked to the prevention of colon cancer, the hypothesis of the study was that intestinal cells accumulate flavone in significant amount. Several factors that may affect the transport and accumulation of flavone were also investigated.
Material and Methods_ Cell Culture Caco-2 cells were purchased from American Type Culture Collection and passages 24-35 were used for the experiments. These cells were cultivated in supplemented high-glucose (4.5 g/L) Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum as described before (23). Culture cups with permeable membrane inserts (Costar Transwell clear, 24 mm diameter, 0.4 uM pore size) placed in 6-well plates were used for transport studies. Cells were seeded at 2 x 1OS/cm2 on the membrane inserts with 1.5 ml medium in the luminal side and 2.5 ml medium in the bottom well facing the basolateral side. Transepithelial resistance of cells grown in the Transwell was measured using chopstick electrodes (World Precision Instruments). The increase in resistance was used as an indicator of epithelial cell differentiation. Intestinal cell lines, for example, HT-29 or IEC-6 that are not capable of spontaneous differentiation do not develop transepithelial resistance (results not shown). Only monolayers with resistance higher than 250 n/cup were used for experiments (mostly at 14 days post seeding). Transnort studv Transport studies were performed as described before (22). Briefly, growth medium in the luminal and basolateral sides of the Transwell were replaced with 1.5 ml and 2.5 ml, respectively, of transport buffer (Hanks Balanced Salt Solution supplemented with 25 mM HEPES and 100 unit penicillin, 100 pg streptomycin per ml). After 30 minutes of preincubation with transport buffer, freshly prepared [14C]-flavone solution was added to the luminal side or basolateral side to start the transport reaction. In general, the concentration of flavone was 10 pM with the total radioactivity in each well at 0.05 pCi. All transport experiments were performed on a platform rocker (60 rpm) inside a humidified 37°C chamber. Duplicate samples of 50 pl transport buffer were taken from the opposite side of the well at each time point to determine the rate of transepithelial transport. The radioactivity in the sample was measured by liquid scintillation
Intestinal Transport of FIavonoid
Vol. 63, No. 26, 1998
2325
technique and normalized by total protein content of cells on the membrane insert. Determination of cellular accumulation Cellular accumulation of radioactivity was determined as described before (22). At the end of the experiment, membrane inserts with cells were removed from Transwell cups, rinsed with icecold phosphate buffer saline stop solution and subjected to overnight solubilization in 3 ml 0.1% SDS at 4°C. Radioactivities in the lysate were then determined by liquid scintillation technique. The total recovery of radioactivity in the luminal chamber, basolateral chamber and membrane was usually >95%. Total protein content on each Transwell membrane was determined by a modified Lowry assay using bovine serum albumin as the standard (24). Chemicals All chemicals used were of analytical grade with greater than 95% purity. Reagents used for cell culture were all from GIBCO-BRL (Grand Island, NY) with the exception of fetal bovine serum which was from Hyclone (Logan, UT). [t4C]-flavone (25) was a gift from Proctor & Gamble Company (Cincinnati, OH) with greater than 99% purity. It was stored as ethanol solution in -7O’C freezer prior to usage. Statistical analvsis The data at each time point were analyzed by ANOVA and post-hoc Student-Newman-Keuls multiple comparison using StatView 4.01. P-value less than 0.05 was considered statistically significant.
Results
+ B q
- 0
15
30
L
to L (flavone) to B (phenylalanine
45
60
75
TIME (min) Fig. 1
Transport of 1OuM l 4C-flavone across Caco-2 cell monolayers in the luminal to basolateral direction (L to B), and basolateral to luminal direction (B to L) and a comparison to the rate of luminal to basolateral (L to B) transport of 10 uM phenylalanine. Data shown were meanskS.D. of three wells. *, b indicate significant differences at PcO.05 within the same time point.
2326
Intestinal Transport of Flavonoid
Vol. 63, No. 26, 1998
Caco-2 cells cultured in Transwell cups have been shown to differentiate spontaneously (17, The cell surface facing the top medium developed brush border which resembled the luminal membrane of the intestinal epithelium. The cell surface attaching to the permeable membrane and facing the bottom medium developed into the basolateral membrane. The differentiated cell monolayers can thus be used for the study ofunidirectional transport of substrates. The transport of flavone from luminal to basolateral side was measured by including t4C-flavone in the top transport buffer and measuring the appearance of isotope in the bottom transport buffer. The transport of flavone from basolateral to luminal side was measured by including t4C-flavone in the bottom buffer and measuring the appearance of isotope in the top buffer. As shown in Fig. 1, the rates of flavone transport in these two directions were almost the same. Transepithelial transport of aromatic amino acid, phenylalanine, has been demonstrated previously in Caco-2 cell monolayer (19). It was found to be carrier-mediated and IO-times faster in the luminal to basolateral direction compared to the basolateral to the luminal direction (19). The luminal to basolateral transport of phenylalanine was measured here as well. Compared to phenylalanine, the rate of transport for flavone was 5 times greater (Fig. 1). 18).
To further characterize the mode of transport for flavone, the effects of temperature and inward sodium radient were also examined. In the luminal to basolateral transport of flavone, 5 substituting Na m transport buffer with K+ did not affect the rate of transport (Fig. 2A), but lowering the incubation temperature decreased the rate of transport (Fig. 2A). Similar temperature effect was also observed for the transport of flavone from the basolateral to luminal direction (Fig. 2B). Substituting Na+ in transport buffer with K+ did not affect the rate of transport in the basolateral to luminal direction, either (Fig. 2B)
TIME (min) Fig. 2
Effect of sodium gradient and incubation temperature on the transepithelial transport of 14C-flavone. L-B: 10uM 14C-flavone transport from the luminal side of the cells to the basolateral side; B-L: 10uM 14C-flavone transport horn the basolateral side of the cells to the luminal side; Na: sodium gradient in the direction of transport; K: no sodium gradient, Na replaced by K. Each point represents mean*S.D. of three Transwells. *,b, c indicate significant differences at P
Vol. 63, No. 26, 1998
Intestinal Transport of Flavonoid
2327
There were extensive cell-associated radioactivities at the end of one hour transport study. In the luminal to basolateral transport experiment conducted for Fig. 2A, there was 2.14 f 0.15 mnol flavone/mg protein at 60 minute incubation. Incubation temperatures and the presence or absence of inward sodium gradient had no effects on the amount of flavone accumulated in the cells. In the basolateral to luminal transport study conducted for Fig. 2B, there was 2.67 f 0.16 run01 flavone/mg protein at the 60 minute incubation in the two 37°C groups (no diierence between the two); and 1.90 f 0.08 mnol flavone/mg protein in the two lower temperature groups (no diierence between the two). In comparison, the amount of cell-associated phenylalanine was 381*41 pmol/mg protein at the end of one hour transport study of phenylalanine.
TIME (min)
Fig. 3
Effect of luminal addition of other agents on the transepithelial transport of 10 pM 14C-flavone from the luminal side of the cells to the basolateral side. BSAl: 1 mg/ml bovine serum albumin; BSAS: 5 mg/ml bovine serum albumin; PHI? 5 mM phenylalnine; MET: 5 mM methionine; ASA: 5 mM ascorbic acid. Each point represents meanG.D. of three transwells. *,b, C indicate significant differences at PcO.05 within the same time point. There were no differences among the treatment groups for all time points in Fig. 3B and 3C.
2328
Intestinal
Transport
of Flavonoid
Vol. 63, No. 26, 1998
Most flavonoids including flavone are hydrophobic. The effect of other hydrophobic compounds on the transepithelial transport of 10 pM i4C-flavone was studied. Due to the limited solubility of the hydrophobic compounds used, they were only tested at 100 uM in most of the cases. When basolateral to luminal transport of 10 uM *4C-flavone were followed for 90 minutes, additional unlabeled 100 uM flavone in the luminal or basolateral side had no effect on the transport. Additional 100 uM quercetin, genistein (both members of the flavonoid family), diethylstilbestrol (synthetic estrogen), nifedipine (calcium channel blocker), or vinblastine (antitumor alkaloid) also had no effect on the transport. In addition, acebutolol @-blocker) at 1 mM had no effect on the transport (results not shown). These hydrophobic compounds in the buffer did not affect the amount of cell-associated flavone at the end of the transport study (results not shown). Flavonoids were found to bind to extracellular proteins (26). The effect of protein in the transport of flavone was studied. As shown in Fig. 3A, when bovine serum albumin (BSA) was included in the luminal side, it decreased the luminal to basolateral transport of 14C-flavone in a dose-dependent fashion. The cell-associated i4C-flavone at the end of one hour incubation was also decreased in this study when BSA was included in the incubation (1.52*0.23, 1.30-+0.04, and 0.81*0.04 nmol/mg protein for control, 1 mg/ml BSA, and 5 mg/ml BSA group, individually). The effect of neutral amino acids on the luminal to basolateral transport and accumulation of r4Cflavone was then studied. As shown in Fig. 3B, 5 mM phenylalanine or methionine had no effect on the transport of *4C-flavone. The cell-associated radioactivity at the end of one hour incubation also remained the same when cells were incubated in buffers containing additional phenylalanine or methionine (1.52+0.23, 1.61&0.07 and 1.54kO.07 nmol/mg protein for control, phenylalanine and methionine group, individually). We have found that flavone decreased the accumulation of ascorbic acid in Caco-2 cells (22). Here, the effect of 5 mM ascorbic acid on the transport of 10 uM i4C-flavone was studied. As shown in Fig. 3C, ascorbic acid had no effect on the transport of flavone. The cell-associated radioactivity also remained the same in the ascorbic acid group at the end of one hour incubation (1.52*0.05 and 1.53*0.07 nmol/mg protein for control and ascorbic acid group, respectively). In order to understand whether the cell-associated radioactivity represents high-affinity binding of flavone to intracellular components, efflux experiments were performed. Caco-2 cells were first incubated with 10 uM 14C-flavone in the luminal side for 30 minutes. The total cell-associated flavone was found to be 1.02*0.07 nmol/mg protein (n=3). The 14C-flavone-containing buffer was then removed and replaced with 2 ml fresh non-flavone containing buffer for efflux measurement. At the end of the 30 minute et&x, 89*4% of the original cell-associated radioactivity was recovered in the buffer. Another experiment was conducted by incubating cells with 100 uM 14Cflavone for 1 hour. The total ceil-associated flavone was 12.3*1.8 nmol/mg protein at the end of one-hour incubation. The efflux of i4C-flavone was then followed. 68*4% of 14C-flavone was found in the buffer at the end of 3Ominute eIIlux and 74*4% of 14C-flavone was found in the buffer at the end of 60-minute efflux. Results from both studies showed a rapid efflux of the accumulated flavone. Discussions This is the first study exploring the transepithehal transport of flavonoids at the cellular level. Several lines of evidences support that the transepithelial transport of flavone across intestinal
Vol. 63, No. 26, 1998
Intestinal Transport of Flavonoid
2329
cells is a passive diffusion event: the transport is much more rapid compared to that of phenylalanine (Fig. 1) or ascorbic acid (22); is bidirectional at about the same rate (Fig. 1, 2); can not be inhibited by any of the small hydrophobic compounds we tested including flavone itself (Fig. 3 and see results); is not affected by the removal of inward sodium gradient (Fig. 2). Lower incubation temperature did lead to a decrease in the rate of transport, however, substantial transport was still observed even when the incubation temperature was decreased to 1” C Although the focus of this study is intestinal cells, it would not be surprising that passive diffusion of flavone and other flavonoids also occurs in other cell types. Similar transepithelial transport study should be applied in the future to other flavonoids with hydroxyl groups on the ring structure, for example, flavonols and flavanols; and to flavonoids with glycoside conjugates. Using a different approach with fluorescence microscope and cells grown on glass coverslips, quercetin (a flavonol) and rutin (quercetin with glycoside conjugate) were found to rapidly move into Caco-2 cells (23). These observation are consistent with the current radioisotope study in supporting a diffusional membrane transport for flavonoids in general. Flavone accumulated extensively in the cells (see results). The fluorescence microscopy study described above also supports an accumulation of quercetin and rutin in the cells (23). The protein-binding property of flavone as reported by another group (26) and noted in this study (Fig. 3A) may explain why largeamounts of cell-associated radioactivity was observed. Although ceil volume was not measured in this study, based on the reported cells volume of 3.49-6 ul per mg protein for Caco-2 cells (18, 27, 28) the net concentration of cell-associated flavone was clearly above that of the medium. Although the subcellular distribution of 14C-flavone, i.e whether the radioactivity is on the cell surface or inside the cells, can not be determined here, in our microscopy study quercetin was found to distribute throughout the cell with no particular enrichment on the cell surface. Neither the accumulation of 14C-flavone in this study or quercetin in the microscopic study could be affected by additional flavonoids at concentrations within their solubility limit (100 PM). These observations support the presence of high-capacity binding site(s) for flavone in the cells. Although the rapid efflux of flavone from cells indicates that majority of the binding is of lowaffinity, significant amount of flavone remained in the cells. For example, the cells incubated with 10 uM flavone for 30 min lost 89% of the_flavone during the subsequent efflux but the cellassociated flavone still had a concentration of 19pM (assuming 6 ul/mg protein). The presence of high-afhnity binding sites for flavone is consistent with the known biological activities of this class of compounds (4). The binding of flavone to BSA in the incubation has led to an impedance in the transepithelial transport and cellular accumulation of flavone in a dose-dependent fashion (Fig. 3A). This binding is not due to a specific interaction with the hydrophobic side-chain of amino acids, phenylalanine or methionine, as these amino acids themselves have no effects on the transepithelial transport of *4C-flavone (Fig. 3B). We have observed that flavonoids, including flavone, inhibited the accumulation of ascorbic acid in intestinal cells (22). With differences in structure and hydrophobicity between flavonoids and ascorbic acid, it is not expected that they compete for the same transporter. In this study, ascorbic acid at 5 mM failed to affect the transport or accumulation of 1OuM t4C-flavone (Fig. 3). It thus supports our original hypothesis that the inhibitory effect of flavonoids is not mediated
2330
Intestinal Transport of Flavonoid
through a competition
Vol. 63, No. 26, 1998
at the transport site (29).
This generaldiffisive be readily available upon by the appearance in the observation here supports
property of flavonoids observed implies that the dietary flavonoids will ingestion. Indeed, rapid intestinal absorption of flavonoids as determined bile or plasma has been observed in rats (30, 31) and human (16). The that intestine transport is not a rate-limiting step for flavonoids.
Acknowledement The work was supported partly by a grant from Buffalo Foundation, Buffalo, New York. The generosity of Proctor & Gamble Company (Cincinnati, OH) in providing l 4C-flavone is greatly appreciated. The author thanks Chi-Ping Lin for performing the efflux experiment and Penny Leavitt for editing the manuscript. References 1. W, S PIERPOINT,
Plant Flavonoids in Biology and Medicine: Biochemical, Pharmacological, andStructure-ActivityRelationships, V. Cody, E. Middleton and J. B. Harbome (Eds), 125
2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
140, Alan R. Liss, New York (1986). A.K. VERMA, J.A. JOHNSON, M.N. GOULD and M.A. TANNER Cancer Res. 48 57545758 (1988). E.E. DESCHNER, J. RUPERTO, G. WONG and H.L. NEWMARK, Carcinogenesis 12 1193-1196 (1991). S.-M. KUO, Crit. Rev. Oncogenesis 8 47-69 (1997). M.J. SANZ, M.L. FERRANDIZ, M. CEJUDO, M.C. TERENCIO, B. GIL, B. BUSTOS, A, UBEDA, R. GUNASEGARAN and M.J. ALCARAZ, Xenobiotica 24 689-699 (1994). H. WEI, R. BOWEN, Q.Y. CAI, S. BARNES and Y. WANG, Proc. Sot. Exp. Biol. Med. 208 124-130 (1995). A. SAIJA, M. SCALESE, M. LANZA, D. MARZULLO, F. BONINA and F. CASTELLI, Free Radic. Biol. Med. 19 481-486 (1995). I.R. RECORD, I.E. DREOSTI and J.K.MCINERNEY, J. Nutr. Biochem. 6 481-485 (1995). N. COTELLE, J. BERNIER, J. CATTEAU, J. POMMERY, J. WALLET and E. M. GAYDOU, Free Radic. Biol. Med. 20 35-43 (1996). D. INGRAM, K. SANDERS, M. KOLYBABA and D. LOPEZ, Lancet 350 990-994 (1997). P. KNEKT, R. JARVINEN, R. SEPPANEN, M. HELIOVAARA, L. TEPPO, E. PUKKALA and A. AROMAA, Am. J. Epidemiol. 146 223-230 (1997). D. COVA, L. DE ANGELIS, F. GIAVARINI, G. PALLADINI and R. PEREGQ, Inter. J. Clin. Pharmacol. Thera. Toxicol. 30 29-33 (1992). X. XU, H.-J. WANG, P.A. MURPHY, L. COOK and S. HENDRICH, J. Nutr. 124 825-832 (1994). M.S. MORTON, G. WILCOX, M. L. WAHLQVIST and K GRIFFITHS, J. Endocrinol. 142 251-259 (1994). P.C.H. HOLLMAN, J.H.M. DE VRIES, S.D. VAN LEEUWENT, M.J.B. MENGELERS and M.B. KATAN, Am. J. Clin. Nutr. 62 1276-1282 (1995). R. A. KING and D.B. BURSILL, Am. J. Clin. Nutr. 67 867-872 (1998) M. PINTO, S. ROBINE-LEON, M.D. APPAY, M. KEDINGER, N. TRIADOU, E.
Vol. 63, No. 26, 1998
18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.
Intestinal Transport of FIavonoid
2331
DUSSAULX, B. LACROIX, P. SIMON-ASSMANN, K. HAFFEN, J. FOGH and A. ZWEIBAUM, Bio. Cell 47 323-330 (1983). I.J. HIDALGQ, T.J. RAUB and R.T. BORCHARDT, Gastroenterology 96 736-749 (1989). I.J. HIDALGO and R.T. BORCHARDT, Biochim. Biophys. Acta 1028 25-30 (1990). P. ARTURSSON, J. Pharma. Sci.79 476-482 (1990). J. KARLSSON, S.-M. KUO, J. ZIEMNIAK and P. ARTURSSON, Br. J. Pharmacol. 110 1009-1016 (1993). S.-M. KUO, H.F. MOREHOUSE and C.-P. LIN, Cancer Lett. 116 131-137 (1997). S.-M. KUO, Cancer Lett. 110 41-48 (1996). G.L. PETERSON, Methods Enzymol. 9195-l 19 (1983). H. LIU and K. R. WEHMEYER, J. Chromatogr. 577 61-67 (1992) S.F. ARNOLD, B.M. COLLINS, M.K. ROBINSON, L.J. GUILLETTE, JR. and J.A. MCLACHLAN, Steroids 61642-646 (1996) A. BLAIS, P. BISSONNETTE and A. BERTELOOT, J. Membrane Biol. 99 113-125 (1987). D.B. BURNHAM and J.D. FONDACARO, Am. J. Physiol. 256 G808-G816 (1989). C.-P. LlN and S.-M. KUO, FASEB J. 10 A770 (1996) R.A. KING, J.L. BROADBEND and R.J. HEAD, J. Nutr. 126 1761182 (19%) J. SFAKIANOS, L. COWARD, M. KIRK and S. BARNES, J. Nutr. 127 1260-1268 (1997)