- Email: [email protected]
Metabolism of stevioside in pigs and intestinal absorption characteristics of stevioside, rebaudioside A and steviol
Food and Chemical Toxicology 41 (2003) 1599–1607 www.elsevier.com/locate/foodchemtox
Metabolism of stevioside in pigs and intestinal absorption characteristics of stevioside, rebaudioside A and steviol Jan M.C. Geunsa,*, Patrick Augustijnsb, Raf Molsb, Johan G. Buysec, Bert Driessend a
Laboratory of Plant Physiology, Catholic University of Leuven, Kasteelpark Arenberg 31, B 3001 Leuven, Belgium b Laboratory of Pharmacotechnology and Biopharmacy, O&N, Gasthuisberg, 3000 Leuven, Belgium c Laboratory of Physiology and Immunology of domestic animals, Kasteelpark Arenberg 30, B-3001 Leuven, Belgium d Zootechnical Centre, Bijzondere Weg 12, B-3360 Lovenjoel, Belgium Accepted 16 June 2003
Abstract Stevioside orally administered to pigs was completely converted into steviol by the bacteria of the colon. However, no stevioside or steviol could be detected in the blood of the animals, even not after converting steviol into the (7-methoxycoumarin-4-yl)methyl ester of steviol, a very sensitive fluorescent derivative with a detection limit of about 50 pg. The intestinal transport characteristics of stevioside, rebaudioside A and steviol were also studied in the Caco-2 system. Only a minor fraction of stevioside and rebaudioside A was transported through the Caco-2 cell layer giving a Papp value of 0.16106 and 0.11106 cm/s, respectively. The Papp value for the absorptive transport of steviol was about 38.6106 cm/s while the Papp value for the secretory transport of steviol was only about 5.32106 cm/s suggesting carrier-mediated transport. The discrepancy between the relatively high absorptive transport of steviol and the lack of steviol in the blood may be explained by the fact that in the Caco-2 study, steviol is applied as a solution facilitating the uptake, whereas in the colon steviol probably is adsorbed to the compounds present in the colon of which the contents is being concentrated by withdrawal of water. # 2003 Elsevier Ltd. All rights reserved. Keywords: Stevia rebaudiana Bertoni; Stevioside; Rebaudioside A; Steviol; Metabolism; Pig; Transport Caco-2; Toxicology
1. Introduction Stevioside, the main sweet component in the leaves of Stevia rebaudiana (Bertoni) Bertoni (Compositae), tastes about 300 times sweeter than sucrose (0.4% solution). Structures of the sweet components of Stevia occurring mainly in the leaves are given in Fig. 1. Their content varies between 4 and 20% of the dry weight of the leaves depending on the cultivar and growing conditions. Stevioside 3 is the main sweet component. Other Abbreviations: AcCN, acetonitrile; ADI, allowable daily intake; BW, body weight; DMSO, Dimethylsulfoxide; EtOH, ethanol; MeOH, methanol; NOEL, no observable effect level; Papp, apparent permeability; TEER, transepithelial electrical resistance; TM, transport medium; Vit E TGPS, vitamin E tocopheryl polyethylene glycol 1000 succinate. * Corresponding author. Tel.: +32-16-321510; fax: +32-16321509. E-mail address: [email protected] (J.M.C. Geuns). 0278-6915/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0278-6915(03)00191-1
compounds present but in lower concentration are: steviolbioside 2, rebaudioside A 4, B 5, C 6, D 7, E 8, F 9 and dulcoside A 10 (Kennelly, 2002, Starrat et al., 2002). The presence of steviolbioside and rebaudioside B in extracts might be due to artifacts of the extraction procedure (refs. in Kennelly, 2002). Both the Stevia plant, its extracts, and stevioside have been used for several years as a sweetener in South America, Asia, Japan, China, and in different countries of the EU. In Brazil, Korea and Japan Stevia leaves, stevioside and highly refined extracts are officially used as a low-calorie sweetener. In the USA, powdered Stevia leaves and refined extracts from the leaves have been used as a dietary supplement since 1995. In 2000, the European Commission refused to accept Stevia or stevioside as a novel food because of a lack of critical scientific reports on Stevia and the discrepancies between cited studies with respect to possible toxicological effects of stevioside and especially its aglycone steviol 1 (Fig. 1)
1600
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
Fig. 1. Structures of stevioside and related compounds. In rebaudioside D and E R1 is composed of 2 b-Glc-b-Glc(2!1). In rebaudioside A, B, C, D, E and F in group R2 an additional sugar moiety is added on carbon 3 of the first b-Glc. In rebaudioside F one b-Glc is substituted for by-b-Xyl.
(Kinghorn, 2002; Geuns, unpublished). The advantages of stevioside as a dietary supplement for human subjects are manifold: it is stable, it is non-calorific, it helps maintain good dental health by reducing the intake of sugar and opens the possibility for use by diabetic and phenylketonuria patients and obese persons. Many papers describe the safety of stevioside used as a sweetener (see Geuns, 2002, for a review). In humans, an acceptable daily intake (ADI) of 7.9 mg stevioside/kg BW was calculated (Xili et al., 1992). However, this ADI should be considered as a minimum value as the authors did not test concentrations of stevioside higher than 793 mg/kg BW (safety factor 100). Considering many reports from the literature, an ADI of more than 20 mg stevioside/kg BW is likely (Geuns, 2002). However, mutagenic effects of steviol, the aglycone of stevioside, and/or its metabolites were reported in the forward mutation test using Salmonella typhimurium TM677 (Pezzuto et al., 1985; Campadre et al., 1988; Matsui et al., 1996a; Temcharoen et al., 1998; Terai et al., 2002). After metabolic activation it was shown that so far unknown steviol metabolites caused mutations in S. typhimurium TM677, i.e. transitions, transversions, duplications and deletions at the guanine phosphoribosyltransferase (gpt) gene (Matsui et al., 1996b). However, stevioside and steviol were inactive in various TA strains of S. typhimurium, Escherichia coli WP2 uvrA/pKM101 and the rec-assay using Bacillus subtilis even when activation S9 mix was present (Matsui et al.,
1996a; Klongpanichpak et al., 1997). The direct mutagenic activity of 15-oxo-steviol was refuted by Procinska et al. (1991), but confirmed by Terai et al. (2002). The activity of steviol in S. typhimurium TM677 was only about 1/3000 that of 3,4-benzopyrene and that of steviol methyl ester 8,13 lactone was 1/24,500 that of furylfuramide (Terai et al., 2002). Although a weak activity of steviol and some of its derivatives was found in the very sensitive S. typhymurium TM677 strain, the authors concluded that the daily use of stevioside as a sweetener is safe. Moreover, the presence in the blood of the chemically synthesised steviol derivatives after feeding stevioside has not been demonstrated so far. Very high doses of steviol (90% purity) intubated to hamsters (4 g/ kg bw), rats and mice (8 g/kg bw) did not induce micronucleus in bone marrow erythrocytes of both male and female animals. However, these doses showed some cytotoxic effect to the female, but not to the male of all treated animal species (Temcharoen et al., 2000). It is not excluded that the toxicity is due to the 10% impurities present. After metabolic activation of steviol, some gene mutation and chromosomal aberration was found in Chinese hamster lung fibroblasts (Matsui et al., 1996a). It has to be said that of all animals tested hamsters had the most sensitive response. Moreover, in hamster, several metabolites of stevioside were found that are not formed in rats or humans. Therefore, the relevance of experiments with hamsters should be questioned. It has been shown that oral stevioside is not taken up by the human body (Yamamoto et al., 1985; Bracht et al., 1985) and none of the digestive enzymes from the gastro-intestinal tract of different animals and man are able to degrade stevioside into steviol, the aglycone of stevioside (Wingard et al., 1980, Hutapea et al., 1997; Koyama et al., 2001, 2003a). Nevertheless, in feeding experiments with rats and hamsters stevioside was metabolised to steviol by the bacterial flora of the caecum. Steviol was found in the blood of the animals with the maximum concentration occurring after 8 h (Nakayama et al., 1986; Koyama et al., 2003a). In the cited studies, it was not indicated that coprophagy, occurring in rodents, was prevented, so it is not clear whether the steviol occurring in the blood was taken up directly from the colon or indirectly from the ingested faeces (after passing through the intestines again). Although bacteria isolated from the human colon are able to transform stevioside into steviol in vitro (Hutapea et al., 1997; Koyama et al., 2001, 2003a), it has never been proven that this is also the case in vivo nor that the steviol possibly formed is taken up directly from the colon. Moreover, studies with roosters (Pomaret and Lavieille, 1931) and chickens (laying hens and broylers, Geuns et al., 2003) indicated that stevioside was rapidly eliminated from the body, largely untransformed. Only the bacteria from the caecum or
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
colon were able to degrade stevioside into steviol (caecum of mice, rats, hamsters and chickens; colon of pigs and man). The bacteria from the human colon also formed steviol 16,17a-epoxid in vitro, that was again metabolised to steviol. However, in vivo this epoxid formation most probably will not occur due to the anaerobic conditions of the human colon. It was correctly concluded that steviol is the only metabolite in faeces and is not further metabolised (Hutapea et al., 1997; Koyama et al., 2001, 2003a; Gardana et al., submitted for publication; Geuns et al., 2003). Anyway, steviol epoxid has been tested in mutagenicity studies and showed to be inactive (Pezzuto et al., 1985). In hamsters the LD50 of steviol (90% purity) was 5.2 and 6.1 g/kg bw for respectively, male and female animals. In rats and mice the LD50 was above 15 g/kg bw demonstrating that of the tested animals hamsters are more sensitive to steviol (Toskulkao et al., 1997). As mutagenic effects of steviol and/or its metabolites were published, one of the most urgent problems to solve is the possible degradation of stevioside into steviol and other metabolites in vivo. In this study, pigs were used because their metabolism resembles that of humans. A second important issue is whether steviol, if produced, is taken up by the intestine and to what extent. Therefore, in addition we studied the transport of steviol, stevioside and rebaudioside A through Caco2 monolayers as more information is needed with respect to the possible uptake of steviol by the colon. The Caco-2 cell layer is a valuable transport model for the small intestinal epithelium (Hidalgo et al., 1989).
1601
moment both groups received normal pig food. After the animals were adapted to the new pig-sties, 1 group of pigs received stevioside at a dose of 1.67 g stevioside kg1 feed. Two days after the start of the stevioside feeding, faeces were daily collected directly from the rectum and immediately frozen in solid carbon dioxide. Blood was taken and frozen in solid carbon dioxide. Control animals were given the same treatment. After freeze-drying of the samples, stevioside and steviol were analysed in the faeces and blood samples. 2.3. Measurement of stevioside Weighed aliquots of about 100 mg of powdered faeces or blood samples, pulverised in a Retsch mixer mill MM200, were extracted with methanol (41 ml). Then 6 ml H2O was added to the pooled extract fractions and the solution was acidified with HCl to pH 6.5 to enable the binding of steviol and steviolbioside onto C18-cartridges. This solution was applied to a conditioned 500 mg C18-cartridge (Alltech) which was rinsed with 5 ml 40% MeOH. Stevioside and steviolbioside were eluted with 3 ml of a 70% MeOH solution. The extracts were evaporated at 50 C under a flow of nitrogen and the residues were redissolved in 0.2 ml of ethanol containing 3% diethylether. Extracts of faeces samples could directly be injected onto the HPLC column for stevioside quantification (ODS-silica column, 25 cm length, 4.6 mm ID using 35% AcCN in water as solvent; detection of stevioside proceeds by UV at 210 nm). Extracts of blood samples were purified by TLC with ethyl acetate:EtOH:water (130:27:20) as solvent. The stevioside bands (Rf 0.20) were eluted by MeOH.
2. Material and methods 2.4. Measurement of steviol 2.1. Chemicals The experiments were performed using stevioside that was purified by repeated crystallisation from MeOH to a purity level of more than 96%. Steviolbioside (around 3%) and rebaudioside A (around 0.5%) were the main impurities. Steviol was made and repeatedly crystallised from MeOH as described (Ogawa et al., 1980). Solvents of HPLC grade were from Acros (H2O, acetonitrile, CHCl3), BDH (MeOH, EtOH, N,N-dimethylformamide) and Biosolve (acetone). N,N-diisopropylethylamine was from Acros and 4-(bromomethyl)-7-methoxycoumarin from Fluka. All chemicals used for culturing the Caco-2 cells were purchased from Invitrogen (Merelbeke, Belgium). Vitamin E-TPGS was kindly provided by Eastman Chemical Company (Kingsport, TN). 2.2. Experiments with animals Two groups of each six female pigs weighing about 23 kg each were numbered and housed in pig-sties. At that
Weighed samples of about 50 mg dried and powdered faeces or 100 mg dry blood were extracted with acetone (41 ml). The extracts were concentrated and purified by TLC (CHCl3:MeOH, 90:10). The bands corresponding to steviol (Rf=0.35) were eluted with CHCl3:MeOH (50:50) and following evaporation of the solvent the residues were dried under a stream of nitrogen (traces of water interfere with the subsequent derivatisation reaction). To enable sensitive fluorometric analysis of steviol in biological samples, steviol was derivatised by esterification of the free carboxyl group with the alkylating reagent 4-(bromomethyl)-7-methoxycoumarin (IUPAC name: 4-(bromomethyl)-7-methoxy-2H-chromen-2-one; Fig. 2). This reaction was carried out in the aprotic solvent N,N-dimethylformamide (DMF). To ensure a quantitative derivatisation, a five-fold excess of reagent was used, e.g. 100 mg steviol was dissolved in 200 ml DMF containing 500 mg of reagent and 1 ml N,N-diisopropylethylamine. For smaller samples ( < 5 mg steviol) a larger excess of reagent (25 mg in 200 ml DMF) was
1602
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
Fig. 2. Derivatisation of steviol with 4-(bromomethyl)-7-methoxycoumarin and N,N-diisopropylethylamine in an aprotic solvent (N,N-dimethylformamide).
used. After heating at 75 C for 20 min, the reaction mixture can be directly injected onto the HPLC column (ODS column, 25 cm4.6 mm ID, AcCN:water, 80:20; detection: fluorescence detector, exc. at 321 nm, em. 391 nm). 2.5. Transport experiments 2.5.1. Stock solutions Culture medium for Caco-2 cells consisted of high glucose (4.5 g/l) Dulbecco’s minimum essential medium containing glutaMAXTM, 100 U/ml of penicillin, 100 mg/ml of streptomycin, 1% non-essential amino acids and 10% fetal bovine serum. Transport medium (TM) consisted of 500 ml Hanks’ balanced salt solution containing 25 mm glucose and 10 mM Hepes, adjusted to pH 7.4 with NaOH 1 n. A solution of Na fluorescein was prepared in TM (1 mg/ml). This solution was added as such to the Caco-2 cell monolayers. The sodium fluorescein was detected by UV at 486 nm. 2.5.2. Maintenance of cells Cells were grown in 75 cm2 Nunc flasks in culture medium in an incubator at 37 C with controlled atmosphere containing 5% CO2 and 90% relative humidity. The transfer of cells from an ongoing cell culture flask to a new culture vessel (subculture or passage) as well as the feeding of the cells was standardised and performed following well-defined procedures. Subculturing was performed every 3–4 days (at 80–90% confluence). To this end, cells were rinsed twice with 2 ml of a prewarmed Trypsin–EDTA solution (37 C); after a third addition of 1.5 ml of the Trypsin–EDTA solution to the culture flask, the flask was incubated at 37 C. After about 5 min, 10 ml of culture medium was added to suspend the cells. Cell clusters were broken up by repeated pipetting. From this cell suspension 1.5–2 ml were transferred to a new 75 cm2 flask containing 25 ml fresh culture medium.
2.5.3. Growth of cells on membrane inserts For transport experiments, Caco-2 cells were cultured on Costar inserts of 10 mm diameter with a pore size of 0.2 mm. Cells were plated at a density of 100,000 cells/ cm2. Passage number 93 was used for the experiment. The experiment was performed using monolayers at 24 days post seeding. The confluence and integrity of the cell monolayer were controlled by measuring the transepithelial electrical resistance (TEER) and fluxes of a hydrophilic marker (sodium fluorescein). Epithelial resistance was measured with the EVOMTM epithelial voltohmmeter (WPI Inc., UK) and a EndohmTM Tissue resistance measurement chamber (WPI Inc., UK). Only monolayers with TEER values above 150 cm2 were used for transport experiments. Volumes added during experiments amounted to 0.5 ml for the apical and 1.5 ml for the basolateral compartment. 2.5.4. Standard procedure for transport studies The experiment was performed as described by Van Gelder et al. (1999). For determination of the transepithelial flux of the different molecules, Caco-2 monolayers were first rinsed twice with TM. Inserts were then preincubated with TM for 30 min, after which TEER values were measured to control cell monolayer integrity. The medium was replaced by TM containing the test compound at the donor side. The test compounds were added in the following concentrations: stevioside and rebaudioside A (1 mm), steviol (30, 100, 300 and 1000 mm) and benzoic acid (1, 5 and 10 mM). All solutions were made in 2% DMSO in Tm that also served as control. VitE TGPS (0.1%) was included in the receiver compartment to create sink conditions. Each experiment was performed in triplicate. Transport was initiated by adding a solution of the test compound to the apical or basolateral compartment. After incubation (2 h), samples were collected and TEER values were measured to control if any toxic effect on cell monolayer
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
integrity had occurred. Samples were analysed by HPLC or by UV spectrophotometry. 2.5.5. Data presentation of transport experiments The results of transport experiments are expressed as% of the amount added to the donor compartment or as apparent permeability coefficient, calculated according to: Papp ¼
DQ 1 1 DT A C0
with DQ DT : transport rate; A: surface area of the insert; and C0: initial target concentration in the donor compartment. Post-experiment validation. After each experiment, an incubation step with sodium fluorescein in TM (0.1% w/v) was performed for 60 min, followed by TEER measurement. The amount of sodium fluorescein appearing in the acceptor compartment was measured by UV spectrophotometry at 486 nm. These tests were carried out as an additional control of monolayer integrity and to assess possible effects on tight junctions by the components. Typical sodium fluorescein flux values across Caco-2 monolayers after the transport experiment with test compound are below 1%.
3. Results 3.1. Detection limits and recovery experiments The detection limit of stevioside was 50 ng per injection. The recovery of stevioside from spiked faeces samples (1, 6.5 or 20 mg g1 dry wt.) was 72.1 1.3% (n=20). Recovery could not be improved by using absolute EtOH or 97:3 EtOH:diethylether as extraction solvent. The recovery of stevioside from spiked blood samples (6.67 mg g1 dry wt.) after TLC clean-up was 68.4 2.6% (n=9). In order to obtain such a recovery it is very important to rinse the sample vials containing the stevioside extracts three times to allow for a quantitative transfer of stevioside to the TLC plates. The detection limit of (7-methoxycoumarin-4-yl)methyl ester of steviol was about 50 pg per injection. Using this procedure, steviol can be measured with a sensitivity that is about 1000 better than that obtained for stevioside. Recovery of steviol from spiked faeces (100, 170 and 1000 mg g1 dry wt.) was 34.5 2.1% for the lower and 55.2 11.8% for the higher concentrations. Use of other solvents for the extraction did not improve recovery (MeOH, diethyl ether). Recovery of steviol from spiked blood (340 mg g1 dry wt.) was 39.7 3.0%. The addition of 1% or higher concentrations of acetic acid to acetone did not significantly increase the recovery. In fact, care must be taken, as acid conditions can
1603
cause conversion of steviol into isosteviol. Therefore, in our experiments steviol was always extracted with acetone. As in vivo steviol might bind to different blood fractions, we preferred to extract the total blood fraction. Therefore, in all experiments, blood samples were freeze-dried before extraction. In spite of the poor recovery of steviol at the lower concentration levels, very minute amounts of steviol can still be measured due to the very sensitive detection after derivatisation, e.g. at the lowest concentration of spiked faeces (100 mg steviol g1 dry wt.) the steviol content in 50 mg faeces is 5 mg. At 35% recovery 1.7 mg will be detected, i.e. 35,000 the detection limit (50 pg). 3.2. Analysis of pig faeces During the period of the stevioside feeding (14 days) the control pigs consumed 1.09 kg day1, the stevioside group 0.96 kg day1 (daily intake of about 1.6 g stevioside 70 mg/kg bw). The somewhat reduced feed intake is probably due to the high stevioside content which makes the feed about 20 times sweeter than necessary for normal sweetening purposes and therefore might be repelling. No stevioside could be detected in the faeces samples, suggesting that bacteria of the colon metabolised all the stevioside present. For each kg of feed about 0.5 kg of dry faeces was produced. The maximum amount of stevioside to be expected in the faeces was about 3.2 mg g1 dry wt. It follows that the maximum concentration of steviol to be expected is around 1.28 mg g1 dry wt. Analysis of steviol in the faeces demonstrated hugh amounts of steviol, in concentrations around 853 48 mg g1 dry wt. This is about 65% of the maximum amount and implies a recovery of around 65% as found with spiked samples. No other constituents were detected. It follows that of the daily uptake of 1.6 g stevioside at the most 224 mg (35% of the steviol formed) is available for uptake via the colon (9.7 mg/kg bw). The results demonstrate that stevioside was completely converted into steviol. These results are opposed to the low conversion of stevioside (about 20%) by bacteria of the chicken caecum (Pomaret and Lavieille, 1931; Geuns et al., 2003), but are in agreement with the incubations of pig (Geuns et al., unpublished) and human faeces (Hutapea et al., 1997; Koyama et al., 2003a; Gardana et al., submitted for publication) under anaerobic conditions in which a complete conversion of stevioside to steviol took place. 3.3. Analysis of blood The complete conversion of stevioside into steviol makes the blood analyses a crucial aspect of the work. No stevioside or steviol could be found in the blood samples analysed. As the steviol analyses in the blood are
1604
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
very important, analyses of larger amounts of blood were done to be absolutely sure that no steviol is present. Up to 5.5 ml samples of blood were exhaustively extracted, but no indication for the occurrence of steviol was found. 3.4. Transport studies The control experiment with fluorescein shows that after the transport experiments with stevioside, rebaudioside A and steviol were performed, the monolayer integrity was not compromised, the flux values of sodium fluorescein being below 0.5% (Fig. 3). Previous experiments had shown that DMSO (up to 2%) did not affect the monolayer integrity (results not shown). The transport of stevioside and rebaudioside A through the Caco-2 monolayers was very low (Papp 0.16106 and 0.11106 cm/s, respectively, Table 1). No effect on TEER-values or on sodium fluorescein flux could be observed (Fig. 3). Transport of stevioside and rebaudioside A from the apical to basolateral side is lower than the transport rate of Na-fluorescein, a marker for paracellular transport. This is consistent with the fact that the transport is inversely related to the molecular weight (Hidalgo et al., 1989). Transport of steviol in the absorptive direction was concentratrion independent and was much higher than
Fig. 3. Effect of different test conditions on transport of sodium fluorescein (open bars) and on TEER-values (average of three determinationsS.D.). 1: 1 mm stevioside, 2: 1 mm rebaudioside A, 3: 30 mm steviol, 4: 100 mm steviol, 5: 300 mm steviol, 6: 1 mm steviol, 7: 30 mm steviol, 8: 100 mm steviol; conditions 1–6: absorptive transport, conditions 7–8: secretory transport. Table 1 Apparent permeability coefficients in Caco-2 cell layers of the tested compounds in the absorptive and secretory direction (n=3) Condition
Pappabs 106 (cm/s)S.D.
Pappsecr106 (cm/s)S.D.
1 mm stevioside 1 mm rebaudioside A 30 mm steviol 100 mm steviol 300 mm steviol 1 mm steviol
0.16 0.01 0.11 0.03 38.7 4.7 44.5 1.4 39.5 1.7 31.9 3.4
– – 6.57 0.06 7.93 0.15 – –
the transport of stevioside and rebaudioside A. Steviol appears to easily cross the Caco-2 monolayers, exhibiting transport behavior of a compound with transcellular passive diffusion characteristics. Steviol at 300 mm and 1 mm slightly compromised the integrity of the Caco-2 monolayers as seen by decreased TEER-values (Fig. 3) and increased sodium fluorescein flux as compared with control conditions. At 100 mm integrity appeared to be maintained. However, the steviol transport itself was not influenced. Steviol transport in the secretory direction (basolateral to apical) was lower than the absorptive transport. At 100 mm the absorptive apparent permeability coefficient (Papp) amounted to 44.5106 while the secretory apparent permeability coefficient amounted to 7.93106 cm/s (Table 1). In additional experiments, it was shown that benzoic acid inhibited the absorptive transport of steviol: Papp for steviol (100 mm) decreased from 30.56 to 20.47106 cm/s upon coincubation with 10 mm benzoic acid.
4. Discussion Steviol was the only compound that we could detect in pig faeces. This is in agreement with published results where steviol was found as the only possible metabolite produced by the intestinal microflora from various animal species and humans under strictly anaerobic conditions (Hutapea et al., 1997; Koyama et al., 2003a; Gardana et al., submitted for publication; Geuns et al., 2003). The steviol formed in the faeces was not found in the blood. These results suggest that steviol, which is only sparingly soluble, was not taken up from the colon of pigs. Transport studies with Caco-2 cells on the contrary showed that the absorptive transepithelial flux of steviol from a solution was relatively high. This discrepancy will be discussed. The lack of steviol in the blood samples can probably not be attributed to metabolism during or after uptake as was the case with soy isoflavones that after uptake were metabolised to compounds that were hydrolysable with a combined b-glucuronidase and sulfatase enzyme preparation (Setchell et al., 2002); indeed free steviol was detected in the plasma of rats up to at least 8 h after feeding stevioside or steviol (Compadre et al., 1988; Koyama et al., 2003b). Moreover, the in vitro conversion of steviol by liver S-9 fraction from Aroclor 1254-pretreated rats was rather low (about 0.3% after 2 h; Compadre et al., 1988). The intrinsic clearance of steviol by human microsomes was about 4 times lower than that of rat microsomes (Koyama et al., 2003b). Taking into account the very low detection limits of steviol when analysed as the (7-methoxycoumarin-4yl)methyl ester (50 pg) the amount of steviol possibly remaining undetected in the blood samples can be estimated to be very low (below 1 mM, i.e. below 318
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
ng/ml). This hypothetical maximum steviol concentration in the blood would probably not be toxic, as in hamsters fed steviol up to 250 mg/kg bw no toxic effects were found (Wasuntarawat et al., 1998). In this case steviol would have been easily taken up by the intestines. When steviol was fed to rats (45 mg/kg bw) a fast uptake was found and the highest plasma concentration of 18.3 mg/ml was observed after 15 min (the first data point, Table 2; Koyama et al., 2003b). The plasma concentration declined to about 2–3 mg/ml at 8 h. Although we are aware that species differences might occur, we have extrapolated the data obtained in rats to hamsters and pigs. Assuming a similar uptake and metabolism in hamsters the reported NOEL of 250 mg steviol/kg bw would correspond to a plasma concentration of 102 mg/ ml 15 min after intubating steviol and about 13.6 mg/ml after 8 h. In our experiment this would then be 3.94 and 0.5 mg/ml, respectively. These concentrations are above the detection limit of (7-methoxycoumarin-4-yl)methyl ester of steviol. Therefore, we suggest that in vivo the uptake of the carboxy acid steviol from the colon is neglectible and that it rather remains adsorbed to the compounds present in the colon (pH 7–7.5) of which the contents is being concentrated by withdrawal of water. The results of the transport study with Caco-2 cells agree with those obtained with everted gastrointestinal sacs (Koyama et al., 2003b). In these experiments only a very small amount of the Stevia mixture (rebaudioside A, 28.2%, stevioside, 17%, rebaudioside C, 25.2%, dulcoside A, 10.2% and unknowns, 18.8%) was transported from the mucosal to the serosal side (less than 0.5% equivalent of salicylic acid, the positive control). The lack of stevioside uptake was also reported earlier (Yamamoto et al., 1985; Bracht et al., 1985). However, the steviol transport through everted gastrointestinal sacs amounted to about 70% of that of salicylic acid (Koyama et al., 2003b). A complete in vivo absorption from the lower rat bowel was reported after oral or intracecal administration of 17-[14C]-steviol (Wingard et al., 1980). However, in this study, it was not mentioned Table 2 Steviol concentration (mg/ml plasma) measured in the blood of rats (Koyama et al., 2003b) or estimated to be present in the blood of hamsters (Wasuntarawat et al., 1998) and pigs after the administration of steviol to the animalsa Amount administered mg/kg bw 45 mg (rats) 250 mg (hamsters) 9.7 mg (pigs) a
mg/ml 15 min
2h
4h
8h
16 hb
18.3 102 3.94
2.5 13.6 0.5
3.5 19 0.7
2.5 13.6 0.5
1 5.4 0.2
The amount available for uptake in pigs is given as the difference between the total amount of steviol expected in the faeces after feeding stevioside and that really found in the faeces (see Section 3). b Amount estimated from the decay of the previous data points.
1605
that coprophagy was prevented. After oral administration of steviol 96.36% was excreted in the feces, 1.48% in the urine. In bile duct ligated animals, 96% was excreted in the urine, 3.31% in the feces after oral administration, and 94.3 and 6.03% in the urine respectively in the faeces after intracaecal administration. In bile duct cannulated animals dosed intracaecally 105.5% of the radioactivity was found in the bile. These results demonstrated that nearly all the steviol administered was excreted. After administration of 3H-stevioside (125 mg/kg bw) to Wistar rats the radioactivity in the blood increased to reach a maximum of 4.83 mg stevioside equivalents/ml. At 120 h the percentages of radioactivity excreted into the feces and expired air were 68.4 and 23.9%, respectively, while radioactivity excreted into the urine was only 2.3%. Radioactivity excreted into the bile at 72 h was 40.9% of the original dose. The authors concluded that an entero-hepatic circulation occurs in the body (Nakayama et al., 1986). However, it is not always possible to discriminate between label originating from the sugar moiety or from the steviol skeleton of the tritium-labelled stevioside. The chemical identification of the radioactive compounds present was not straightforward. The hydrophilic nature of stevioside and rebaudioside A may explain the poor transport through the Caco-2 monolayers and everted gastrointestinal sacs, as well as the poor absorption in an in vivo situation. The absorption of steviol might be explained partly by passive diffusion and partly by carrier-mediated transport via the monocarboxylic acid transporter (Takanaga et al., 1994). Indeed, we found a 33% inhibition of steviol transport by benzoic acid. This last possibility might also explain the much lower steviol transport in the secretory direction. We speculate that the low in vivo absorption may be attributed to low diffusion out of the caecal content. A weak mutagenic effect of steviol (only 90% purity) in one sensitive S. typhimurium TM 677 strain (see Section 1) does not mean that stevioside used as a sweetener should be carcinogenic per se, even if the stevioside might be transformed to steviol by bacteria in the colon. The safety of oral stevioside in relation to carcinogenic activity is evidenced by the following reports. Male and female F344 rats were daily fed with a ration containing 0.1, 0.3 or 1% of stevioside and rebaudioside for a period of 22 (males) or 24 (female rats) months (Yamada et al., 1985). The animals were then killed, and the researchers conducted biochemical, anatomical, pathological and carcinogenic tests on 41 organs following autopsy. In addition they performed ongoing hematologic and urine tests on the same animals. Each of the animals was matched to a control animal that experienced exactly the same treatment, except for the stevioside. No significant dose-related changes were found in the growth, general appearance, hematological and
1606
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
blood biochemical findings, organ weights, and macroscopic or microscopic observations. It was also concluded that any neoplasms that occurred were not attributable to the administration of stevioside. Even at the highest dose of 1% no significant effects were found. This high dose is equivalent to 125 times the average daily dose of sweeteners that a normal human would require for sweetening purposes. The effects of a chronic oral feeding of stevioside during 24 months were studied in rats (Xili et al., 1992). The concentrations used were: 0, 0.2, 0.6 and 1.2%. Growth, food utilisation and consumption, general appearance and mortality were similar in treated and control groups. The mean lifespan of rats given stevioside was not significantly different from that of the controls. No treatment-related changes were observed in haematological, urinary or clinical biochemical values at any stage of the study. The incidence and severity of non-neoplastic and neoplastic changes were unrelated to the level of stevioside in the diet. In a chronic toxicity study with F344 rats during 2 years it was concluded that there were no significant increases in the incidence of neoplastic lesions in any organ or tissue in the stevioside treated groups (2.5 and 5%, i.e. daily dose of 385 and 775 mg per rat, i.e. 1 and 2 g kg1 bw, which is a very high dose; Toyoda et al., 1997). In male animals, the number of testicular tumours had the tendency to decrease. Moreover, the incidence of adenomas of the mammary gland in the stevioside-treated female rats was significantly lower than that in the controls. The severity of chronic nephropathy in males was also clearly reduced by both stevioside concentrations. There was no increase in the incidence of urinary bladder carcinogenesis in rats by stevioside (Hagiwara et al., 1984). Very significant inhibitory effects of stevioside were reported on tumor promotion by 12-O-tetradecanoylphorbol-13-acetate in carcinogenesis in mouse skin (Yasukawa et al., 2002). In 1999, the Joint FAO/ WHO Expert Committee on Food Additives (JECFA) clearly stated: ‘‘Stevioside has a very low acute oral toxicity. Oral administration of stevioside at a dietary concentration of 2.5% to rats for two years, equal to 970 and 1100 mg kg1 BW per day in males and females, respectively, had no significant effect. Reduced body-weight gain and survival rate were observed at a dietary concentration of 5% stevioside. There was no indication of carcinogenic potential in a long-term study. . .’’ (WHO, 1999). Moreover, there have never appeared any clinical adverse reports on Stevia rebaudiana or stevioside in any of the countries in which these products were used over a long period of time (e.g. Paraguay: at least over 100 years, Japan: more than 25 years, South-Korea: 16 years, Brazil: 13 years, China: 12 years or the USA: since 1995 admitted as a dietary supplement). The above literature data, mainly with rodents, the lack of reports on cancerigenous activity in populations, and our data showing
additionally that no traces of stevioside or its controversial metabolite steviol were found in the blood of pigs are supportive for the safe use of stevioside as a sweetener.
Acknowledgements The authors acknowledge Professor D. Kinghorn for the gift or pure stevioside and Specchiasol for financial support. Professor R. Geers is acknowledged for use of the Zootechnical Centre and Jan M.C. Geuns and Patrick Augustijns acknowledge the ‘‘Onderzoeksraad KULeuven’’ for grants OT/00/15 and OT/01/14, respectively and the FWO for grant number G.0111.01. We thank Hilde Verlinden and Tom Struyf for their skilful technical assistance.
References Bracht, K.A.M., Kemmelmeier, F.S., Ishii, E.L., Alvarez, M., Bracht, A., 1985. Effect of Stevia rebaudiana natural products on cellular and sub-cellular metabolism. Arq. Biol. Tecnol. 28, 431–455. Compadre, C.M., Hussain, R.A., Nanayakkara, N.P.D., Pezzuto, J.M., Kinghorn, A.D., 1988. Mass spectral analysis of some derivatives and in vitro metabolites of steviol, the aglycone of the natural sweeteners, stevioside, rebaudioside A and rubusoside. Biomed. Environ. Mass Spectrom. 15, 211–222. Gardana, C., Simonetti, P., Canzi, E., Zanchi, R., Pietta, P.G. 2003. Metabolism of stevioside and rebaudioside A from Stevia rebaudiana extracts by human microflora. J. Agric. Food Chem. (submitted). Geuns, J.M.C., 2002. Safety evaluation of Stevia and stevioside. In: Atta-ur-Rahman (Ed.), Studies in Natural Products Chemistry, vol. 27: Bioactive Natural Products (Part H). Elsevier, Amsterdam, pp. 299–319. Geuns, J.M.C., Malheiros, R.D., Moraes, V.M.B., Decuypere, E.M.P., Compernolle, F., Buyse, J.G., 2003. Metabolism of stevioside by chickens. J. Agric. Food Chem. 51, 1095–1101. Hagiwara, A., Fukushima, S., Kitaori, M., 1984. Effects of the three sweeteners on rats urinary bladder carcinogenesis initiated by N-butyl-N-(4-hydroxybutyl)-nitrosamine. Gann 75, 763–768. Hidalgo, I.J., Raub, T.J., Borchardt, R.T., 1989. Characterization of the human colon carcinoma cell line (Caco2) as a model system for intestinal epithelial permeability. Gastroenterology 96, 736–749. Hutapea, A.M., Toskulkao, C., Buddhasukh, D., Wilairat, P., Glinsukon, T., 1997. Digestion of Stevioside, a natural sweetener, by various digestive enzymes. J. Clin. Biochem. Nutr. 23, 177–186. Kennelly, E.J., 2002. Sweet and non-sweet constituents of Stevia rebaudiana (Bertoni) Bertoni. In: Kinghorn, A.D. (Ed.), Stevia, the genus Stevia. Medicinal and Aromatic Plants—Industrial Profiles, Vol. 19. Taylor and Francis, London, New York, pp. 68–85. Kinghorn, A.D., 2002. Overview. In: Kinghorn, A.D. (Ed.), Stevia, the genus Stevia. Medicinal and Aromatic Plants—Industrial Profiles, Vol. 19. Taylor and Francis, London, New York, pp. 1–17. Klongpanichpak, S., Temcharoen, P., Toskulkao, C., Apibal, S., Glinsukon, T., 1997. Lack of mutagenicity of stevioside and steviol in Salmonella typhimurium TA98 and TA100. J. Med. Assoc. Thailand 80, 121–128. Koyama, E., Kitazawa, K., Ohori, Y., Sakai, N., Izawa, O., Kakegawa, K., Fujino, A., Ui, M., 2001. Intestinal degradation, absorption, and hepatic metabolism of the glycosidic sweeteners, Stevia
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607 mixture. In: Proceedings Second IUPA-International Symposium on Sweeteners, Hiroshima, p. 12. Koyama, E., Kitazawa, K., Ohori, Y., Izawa, O., Kakegawa, K., Fujino, A., Ui, M., 2003a. In vitro metabolism fo the glycosidic sweeteners, Stevia mixture and enzymatically modified Stevia in human intestinal microflora. Food Chem. Toxicol. 41, 359–374. Koyama, E., Sakai, N., Ohori, Y., Kitazawa, K., Izawa, O., Kakegawa, K., Fujino, A., Ui, M., 2003b. Absorption and metabolism of the glycosidic sweeteners, Stevia related compounds in human and rat. Food Chem. Toxicol. 41, 875–883. Matsui, M., Matsui, K., Kawasaki, Y., Oda, Y., Noguchi, T., Kitagawa, Y., Sawada, M., Hayashi, M., Nohmi, T., Yoshihira, K., Ishidate, M., Sofuni, T., 1996a. Evaluation of the genotoxicity of stevioside and steviol using six in vitro and one in vivo mutagenicity assays. Mutagenesis 11, 573–579. Matsui, M., Sofuni, T., Nohmi, T., 1996b. Regionally-targeted mutagenesis by metabolically-activated steviol: DNA sequence analysis of steviol-induced mutants of guanine phosphoribosyltransferase (gpt) gene of Salmonella typhimurium TM677. Mutagenesis 11, 565–572. Nakayama, K., Kasahara, D., Yamamoto, F., 1986. Absorption, distribution, metabolism and excretion of Stevioside in rats. J. Food Hyg. Soc. Jpn 27, 1–8. Ogawa, T., Nozaki, M., Matsui, M., 1980. Total synthesis of stevioside. Tetrahedron 36, 2641–2648. Pezzuto, J.M., Compadre, C.M., Swanson, S.M., Nanayakkara, N.P.D., Kinghorn, A.D., 1985. Metabolically activated steviol, the aglycone of stevioside, is mutagenic. Proc. Nat. Acad. Sci. USA 82, 2478–2482. Pomaret, M., Lavieille, R., 1931. Le principe a` saveur sucre´e du Kaa`he`-e´ (Stevia rebaudiana Bertoni). Quelques proprie´te´s physiologiques du Ste´vioside. Bull. Soc. Chim. Biol. 13, 1248–1252. Procinska, E., Bridges, B.A., Hanson, J.R., 1991. Interpretation of results with 8-azaguanine resistance system in Salmonella typhimurium: no evidence for direct acting mutagenesis by 15-oxosteviol, a possible metabolite of steviol. Mutagenesis 6, 165–167. Setchell, K.D.R., Brown, N.M., Zimmer-Nechemias, L., Brasher, W.T., Wolfe, B.E., Kirschner, A.S., Heubi, J.E., 2002. Evidence for lack of absorption of soy isoflavone glycosides in humans, supporting the crucial role of intestinal metabolism for bioavailability. Am. J. Clin. Nutr. 76, 447–453. Starrat, A.N., Kirby, C.W., Pocs, R., Brandle, J.E., 2002. Rebaudioside F, a diterpene glycoside from Stevia rebaudiana. Phytochemistry 59, 367–370. Takanaga, H., Tamai, I., Tsuji, A., 1994. pH-Dependent and carriermediated transport of salicylic acid across caco-2 cells. J. Pharm. Pharmacol. 46, 567–570.
1607
Temcharoen, P., Pimbua, J., Glinsukon, T., Rojanapo, W., Apibal, S., 1998. Mutagenic activity of steviol to Salmonella typhimurium TM 677: comparison of the activity of S9 liver fractions from five laboratory animal species. Bull. Health Sci. Tech. 1, 38–45. Temcharoen, P., Suwannatrai, M., Klongpanichpak, S., Apibal, S., Glinsukon, T., Toskulkao, C., 2000. Evaluation of the effect of steviol on chromosomal damage using micronucleus test in three laboratory animal species. J. Med. Assoc. Thailand 83, 101–108. Terai, T., Ren, H., Mori, G., Yamaguchi, Y., Hayashi, T., 2002. Mutagenicity of steviol and its oxidative derivatives in Salmonella typhimurium TM677. Chem. Pharm. Bull. 50, 1007–1010. Toskulkao, C., Chaturat, L., Temcharoen, P., Glinsukon, T., 1997. Acute toxicity of stevioside, a natural sweetener, and its metabolite, steviol, in several animal species. Drug Chem. Toxicol. 20, 31–44. Toyoda, K., Matsui, H., Shoda, T., Uneyama, C., Takada, K., Takahashi, M., 1997. Assessment of the carcinogenicity of Stevioside in F344 rats. Food Chem. Toxicol. 35, 597–603. Van Gelder, J., Witvrouw, M., Pannecouque, C., Henson, G., Bridger, G., Naesens, L., De Clercq, E., Annaert, P., Shafiee, M., Van den Mooter, G., Kinget, R., Augustijns, P., 1999. Evaluation of the potential of ion pair formation to improve the oral absorption of twe potent antiviral compounds, AMD3100 and PMPA. Int. J. Pharm. 186, 127–136. Wasuntarawat, C., Temcharoen, P., Toskulkao, C., Mungkornkarn, P., Suttajit, M., Glinsukon, T., 1998. Developmental toxicity of steviol, a metabolite of stevioside, in the hamster. Drug Chem. Toxicol. 21, 207–222. WHO Food Additives Series 42, 1999. Safety evaluation of certain food additives ’’Stevioside’’. Genf 119–143. Wingard, R.E., Brown, J.P., Enderlin, F.E., Dale, J.A., Hale, R.L., Seitz, C.T., 1980. Intestinal degradation and absorption of the glycosidic sweeteners stevioside and rebaudioside A. Experientia 36, 519–520. Xili, L., Chengjiany, B., Eryi, X., Reiming, S., Yuengming, W., Haodong, S., Zhiyian, H., 1992. Chronic oral toxicity and carcinogenicity study of stevioside in rats. Food Chem. Toxicol. 30, 957–965. Yamada, A., Ohgaki, S., Noda, T., Shimizu, M., 1985. Chronic toxicity of dietary Stevia extracts. J. Food Hyg. Soc. Jpn. 26, 169–183. Yamamoto, N.S., Bracht, K.A.M., Ishii, E.L., Kemmelmeier, F.S., Alvarez, M., Bracht, A., 1985. Effect of steviol and its structural analogues on glucose production and oxygen uptake in rat renal tubules. Experientia 41, 55–57. Yasukawa, K., Kitanaka, S., Seo, S., 2002. Inhibitory effect of stevioside on tumor promotion by 12-O-tetradecanoylphorbol-13-acetate in two-stage carcinogenesis in mouse skin. Biol. Pharm. Bull. 25, 1488–1490.
Metabolism of stevioside in pigs and intestinal absorption characteristics of stevioside, rebaudioside A and steviol Jan M.C. Geunsa,*, Patrick Augustijnsb, Raf Molsb, Johan G. Buysec, Bert Driessend a
Laboratory of Plant Physiology, Catholic University of Leuven, Kasteelpark Arenberg 31, B 3001 Leuven, Belgium b Laboratory of Pharmacotechnology and Biopharmacy, O&N, Gasthuisberg, 3000 Leuven, Belgium c Laboratory of Physiology and Immunology of domestic animals, Kasteelpark Arenberg 30, B-3001 Leuven, Belgium d Zootechnical Centre, Bijzondere Weg 12, B-3360 Lovenjoel, Belgium Accepted 16 June 2003
Abstract Stevioside orally administered to pigs was completely converted into steviol by the bacteria of the colon. However, no stevioside or steviol could be detected in the blood of the animals, even not after converting steviol into the (7-methoxycoumarin-4-yl)methyl ester of steviol, a very sensitive fluorescent derivative with a detection limit of about 50 pg. The intestinal transport characteristics of stevioside, rebaudioside A and steviol were also studied in the Caco-2 system. Only a minor fraction of stevioside and rebaudioside A was transported through the Caco-2 cell layer giving a Papp value of 0.16106 and 0.11106 cm/s, respectively. The Papp value for the absorptive transport of steviol was about 38.6106 cm/s while the Papp value for the secretory transport of steviol was only about 5.32106 cm/s suggesting carrier-mediated transport. The discrepancy between the relatively high absorptive transport of steviol and the lack of steviol in the blood may be explained by the fact that in the Caco-2 study, steviol is applied as a solution facilitating the uptake, whereas in the colon steviol probably is adsorbed to the compounds present in the colon of which the contents is being concentrated by withdrawal of water. # 2003 Elsevier Ltd. All rights reserved. Keywords: Stevia rebaudiana Bertoni; Stevioside; Rebaudioside A; Steviol; Metabolism; Pig; Transport Caco-2; Toxicology
1. Introduction Stevioside, the main sweet component in the leaves of Stevia rebaudiana (Bertoni) Bertoni (Compositae), tastes about 300 times sweeter than sucrose (0.4% solution). Structures of the sweet components of Stevia occurring mainly in the leaves are given in Fig. 1. Their content varies between 4 and 20% of the dry weight of the leaves depending on the cultivar and growing conditions. Stevioside 3 is the main sweet component. Other Abbreviations: AcCN, acetonitrile; ADI, allowable daily intake; BW, body weight; DMSO, Dimethylsulfoxide; EtOH, ethanol; MeOH, methanol; NOEL, no observable effect level; Papp, apparent permeability; TEER, transepithelial electrical resistance; TM, transport medium; Vit E TGPS, vitamin E tocopheryl polyethylene glycol 1000 succinate. * Corresponding author. Tel.: +32-16-321510; fax: +32-16321509. E-mail address: [email protected] (J.M.C. Geuns). 0278-6915/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0278-6915(03)00191-1
compounds present but in lower concentration are: steviolbioside 2, rebaudioside A 4, B 5, C 6, D 7, E 8, F 9 and dulcoside A 10 (Kennelly, 2002, Starrat et al., 2002). The presence of steviolbioside and rebaudioside B in extracts might be due to artifacts of the extraction procedure (refs. in Kennelly, 2002). Both the Stevia plant, its extracts, and stevioside have been used for several years as a sweetener in South America, Asia, Japan, China, and in different countries of the EU. In Brazil, Korea and Japan Stevia leaves, stevioside and highly refined extracts are officially used as a low-calorie sweetener. In the USA, powdered Stevia leaves and refined extracts from the leaves have been used as a dietary supplement since 1995. In 2000, the European Commission refused to accept Stevia or stevioside as a novel food because of a lack of critical scientific reports on Stevia and the discrepancies between cited studies with respect to possible toxicological effects of stevioside and especially its aglycone steviol 1 (Fig. 1)
1600
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
Fig. 1. Structures of stevioside and related compounds. In rebaudioside D and E R1 is composed of 2 b-Glc-b-Glc(2!1). In rebaudioside A, B, C, D, E and F in group R2 an additional sugar moiety is added on carbon 3 of the first b-Glc. In rebaudioside F one b-Glc is substituted for by-b-Xyl.
(Kinghorn, 2002; Geuns, unpublished). The advantages of stevioside as a dietary supplement for human subjects are manifold: it is stable, it is non-calorific, it helps maintain good dental health by reducing the intake of sugar and opens the possibility for use by diabetic and phenylketonuria patients and obese persons. Many papers describe the safety of stevioside used as a sweetener (see Geuns, 2002, for a review). In humans, an acceptable daily intake (ADI) of 7.9 mg stevioside/kg BW was calculated (Xili et al., 1992). However, this ADI should be considered as a minimum value as the authors did not test concentrations of stevioside higher than 793 mg/kg BW (safety factor 100). Considering many reports from the literature, an ADI of more than 20 mg stevioside/kg BW is likely (Geuns, 2002). However, mutagenic effects of steviol, the aglycone of stevioside, and/or its metabolites were reported in the forward mutation test using Salmonella typhimurium TM677 (Pezzuto et al., 1985; Campadre et al., 1988; Matsui et al., 1996a; Temcharoen et al., 1998; Terai et al., 2002). After metabolic activation it was shown that so far unknown steviol metabolites caused mutations in S. typhimurium TM677, i.e. transitions, transversions, duplications and deletions at the guanine phosphoribosyltransferase (gpt) gene (Matsui et al., 1996b). However, stevioside and steviol were inactive in various TA strains of S. typhimurium, Escherichia coli WP2 uvrA/pKM101 and the rec-assay using Bacillus subtilis even when activation S9 mix was present (Matsui et al.,
1996a; Klongpanichpak et al., 1997). The direct mutagenic activity of 15-oxo-steviol was refuted by Procinska et al. (1991), but confirmed by Terai et al. (2002). The activity of steviol in S. typhimurium TM677 was only about 1/3000 that of 3,4-benzopyrene and that of steviol methyl ester 8,13 lactone was 1/24,500 that of furylfuramide (Terai et al., 2002). Although a weak activity of steviol and some of its derivatives was found in the very sensitive S. typhymurium TM677 strain, the authors concluded that the daily use of stevioside as a sweetener is safe. Moreover, the presence in the blood of the chemically synthesised steviol derivatives after feeding stevioside has not been demonstrated so far. Very high doses of steviol (90% purity) intubated to hamsters (4 g/ kg bw), rats and mice (8 g/kg bw) did not induce micronucleus in bone marrow erythrocytes of both male and female animals. However, these doses showed some cytotoxic effect to the female, but not to the male of all treated animal species (Temcharoen et al., 2000). It is not excluded that the toxicity is due to the 10% impurities present. After metabolic activation of steviol, some gene mutation and chromosomal aberration was found in Chinese hamster lung fibroblasts (Matsui et al., 1996a). It has to be said that of all animals tested hamsters had the most sensitive response. Moreover, in hamster, several metabolites of stevioside were found that are not formed in rats or humans. Therefore, the relevance of experiments with hamsters should be questioned. It has been shown that oral stevioside is not taken up by the human body (Yamamoto et al., 1985; Bracht et al., 1985) and none of the digestive enzymes from the gastro-intestinal tract of different animals and man are able to degrade stevioside into steviol, the aglycone of stevioside (Wingard et al., 1980, Hutapea et al., 1997; Koyama et al., 2001, 2003a). Nevertheless, in feeding experiments with rats and hamsters stevioside was metabolised to steviol by the bacterial flora of the caecum. Steviol was found in the blood of the animals with the maximum concentration occurring after 8 h (Nakayama et al., 1986; Koyama et al., 2003a). In the cited studies, it was not indicated that coprophagy, occurring in rodents, was prevented, so it is not clear whether the steviol occurring in the blood was taken up directly from the colon or indirectly from the ingested faeces (after passing through the intestines again). Although bacteria isolated from the human colon are able to transform stevioside into steviol in vitro (Hutapea et al., 1997; Koyama et al., 2001, 2003a), it has never been proven that this is also the case in vivo nor that the steviol possibly formed is taken up directly from the colon. Moreover, studies with roosters (Pomaret and Lavieille, 1931) and chickens (laying hens and broylers, Geuns et al., 2003) indicated that stevioside was rapidly eliminated from the body, largely untransformed. Only the bacteria from the caecum or
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
colon were able to degrade stevioside into steviol (caecum of mice, rats, hamsters and chickens; colon of pigs and man). The bacteria from the human colon also formed steviol 16,17a-epoxid in vitro, that was again metabolised to steviol. However, in vivo this epoxid formation most probably will not occur due to the anaerobic conditions of the human colon. It was correctly concluded that steviol is the only metabolite in faeces and is not further metabolised (Hutapea et al., 1997; Koyama et al., 2001, 2003a; Gardana et al., submitted for publication; Geuns et al., 2003). Anyway, steviol epoxid has been tested in mutagenicity studies and showed to be inactive (Pezzuto et al., 1985). In hamsters the LD50 of steviol (90% purity) was 5.2 and 6.1 g/kg bw for respectively, male and female animals. In rats and mice the LD50 was above 15 g/kg bw demonstrating that of the tested animals hamsters are more sensitive to steviol (Toskulkao et al., 1997). As mutagenic effects of steviol and/or its metabolites were published, one of the most urgent problems to solve is the possible degradation of stevioside into steviol and other metabolites in vivo. In this study, pigs were used because their metabolism resembles that of humans. A second important issue is whether steviol, if produced, is taken up by the intestine and to what extent. Therefore, in addition we studied the transport of steviol, stevioside and rebaudioside A through Caco2 monolayers as more information is needed with respect to the possible uptake of steviol by the colon. The Caco-2 cell layer is a valuable transport model for the small intestinal epithelium (Hidalgo et al., 1989).
1601
moment both groups received normal pig food. After the animals were adapted to the new pig-sties, 1 group of pigs received stevioside at a dose of 1.67 g stevioside kg1 feed. Two days after the start of the stevioside feeding, faeces were daily collected directly from the rectum and immediately frozen in solid carbon dioxide. Blood was taken and frozen in solid carbon dioxide. Control animals were given the same treatment. After freeze-drying of the samples, stevioside and steviol were analysed in the faeces and blood samples. 2.3. Measurement of stevioside Weighed aliquots of about 100 mg of powdered faeces or blood samples, pulverised in a Retsch mixer mill MM200, were extracted with methanol (41 ml). Then 6 ml H2O was added to the pooled extract fractions and the solution was acidified with HCl to pH 6.5 to enable the binding of steviol and steviolbioside onto C18-cartridges. This solution was applied to a conditioned 500 mg C18-cartridge (Alltech) which was rinsed with 5 ml 40% MeOH. Stevioside and steviolbioside were eluted with 3 ml of a 70% MeOH solution. The extracts were evaporated at 50 C under a flow of nitrogen and the residues were redissolved in 0.2 ml of ethanol containing 3% diethylether. Extracts of faeces samples could directly be injected onto the HPLC column for stevioside quantification (ODS-silica column, 25 cm length, 4.6 mm ID using 35% AcCN in water as solvent; detection of stevioside proceeds by UV at 210 nm). Extracts of blood samples were purified by TLC with ethyl acetate:EtOH:water (130:27:20) as solvent. The stevioside bands (Rf 0.20) were eluted by MeOH.
2. Material and methods 2.4. Measurement of steviol 2.1. Chemicals The experiments were performed using stevioside that was purified by repeated crystallisation from MeOH to a purity level of more than 96%. Steviolbioside (around 3%) and rebaudioside A (around 0.5%) were the main impurities. Steviol was made and repeatedly crystallised from MeOH as described (Ogawa et al., 1980). Solvents of HPLC grade were from Acros (H2O, acetonitrile, CHCl3), BDH (MeOH, EtOH, N,N-dimethylformamide) and Biosolve (acetone). N,N-diisopropylethylamine was from Acros and 4-(bromomethyl)-7-methoxycoumarin from Fluka. All chemicals used for culturing the Caco-2 cells were purchased from Invitrogen (Merelbeke, Belgium). Vitamin E-TPGS was kindly provided by Eastman Chemical Company (Kingsport, TN). 2.2. Experiments with animals Two groups of each six female pigs weighing about 23 kg each were numbered and housed in pig-sties. At that
Weighed samples of about 50 mg dried and powdered faeces or 100 mg dry blood were extracted with acetone (41 ml). The extracts were concentrated and purified by TLC (CHCl3:MeOH, 90:10). The bands corresponding to steviol (Rf=0.35) were eluted with CHCl3:MeOH (50:50) and following evaporation of the solvent the residues were dried under a stream of nitrogen (traces of water interfere with the subsequent derivatisation reaction). To enable sensitive fluorometric analysis of steviol in biological samples, steviol was derivatised by esterification of the free carboxyl group with the alkylating reagent 4-(bromomethyl)-7-methoxycoumarin (IUPAC name: 4-(bromomethyl)-7-methoxy-2H-chromen-2-one; Fig. 2). This reaction was carried out in the aprotic solvent N,N-dimethylformamide (DMF). To ensure a quantitative derivatisation, a five-fold excess of reagent was used, e.g. 100 mg steviol was dissolved in 200 ml DMF containing 500 mg of reagent and 1 ml N,N-diisopropylethylamine. For smaller samples ( < 5 mg steviol) a larger excess of reagent (25 mg in 200 ml DMF) was
1602
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
Fig. 2. Derivatisation of steviol with 4-(bromomethyl)-7-methoxycoumarin and N,N-diisopropylethylamine in an aprotic solvent (N,N-dimethylformamide).
used. After heating at 75 C for 20 min, the reaction mixture can be directly injected onto the HPLC column (ODS column, 25 cm4.6 mm ID, AcCN:water, 80:20; detection: fluorescence detector, exc. at 321 nm, em. 391 nm). 2.5. Transport experiments 2.5.1. Stock solutions Culture medium for Caco-2 cells consisted of high glucose (4.5 g/l) Dulbecco’s minimum essential medium containing glutaMAXTM, 100 U/ml of penicillin, 100 mg/ml of streptomycin, 1% non-essential amino acids and 10% fetal bovine serum. Transport medium (TM) consisted of 500 ml Hanks’ balanced salt solution containing 25 mm glucose and 10 mM Hepes, adjusted to pH 7.4 with NaOH 1 n. A solution of Na fluorescein was prepared in TM (1 mg/ml). This solution was added as such to the Caco-2 cell monolayers. The sodium fluorescein was detected by UV at 486 nm. 2.5.2. Maintenance of cells Cells were grown in 75 cm2 Nunc flasks in culture medium in an incubator at 37 C with controlled atmosphere containing 5% CO2 and 90% relative humidity. The transfer of cells from an ongoing cell culture flask to a new culture vessel (subculture or passage) as well as the feeding of the cells was standardised and performed following well-defined procedures. Subculturing was performed every 3–4 days (at 80–90% confluence). To this end, cells were rinsed twice with 2 ml of a prewarmed Trypsin–EDTA solution (37 C); after a third addition of 1.5 ml of the Trypsin–EDTA solution to the culture flask, the flask was incubated at 37 C. After about 5 min, 10 ml of culture medium was added to suspend the cells. Cell clusters were broken up by repeated pipetting. From this cell suspension 1.5–2 ml were transferred to a new 75 cm2 flask containing 25 ml fresh culture medium.
2.5.3. Growth of cells on membrane inserts For transport experiments, Caco-2 cells were cultured on Costar inserts of 10 mm diameter with a pore size of 0.2 mm. Cells were plated at a density of 100,000 cells/ cm2. Passage number 93 was used for the experiment. The experiment was performed using monolayers at 24 days post seeding. The confluence and integrity of the cell monolayer were controlled by measuring the transepithelial electrical resistance (TEER) and fluxes of a hydrophilic marker (sodium fluorescein). Epithelial resistance was measured with the EVOMTM epithelial voltohmmeter (WPI Inc., UK) and a EndohmTM Tissue resistance measurement chamber (WPI Inc., UK). Only monolayers with TEER values above 150 cm2 were used for transport experiments. Volumes added during experiments amounted to 0.5 ml for the apical and 1.5 ml for the basolateral compartment. 2.5.4. Standard procedure for transport studies The experiment was performed as described by Van Gelder et al. (1999). For determination of the transepithelial flux of the different molecules, Caco-2 monolayers were first rinsed twice with TM. Inserts were then preincubated with TM for 30 min, after which TEER values were measured to control cell monolayer integrity. The medium was replaced by TM containing the test compound at the donor side. The test compounds were added in the following concentrations: stevioside and rebaudioside A (1 mm), steviol (30, 100, 300 and 1000 mm) and benzoic acid (1, 5 and 10 mM). All solutions were made in 2% DMSO in Tm that also served as control. VitE TGPS (0.1%) was included in the receiver compartment to create sink conditions. Each experiment was performed in triplicate. Transport was initiated by adding a solution of the test compound to the apical or basolateral compartment. After incubation (2 h), samples were collected and TEER values were measured to control if any toxic effect on cell monolayer
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
integrity had occurred. Samples were analysed by HPLC or by UV spectrophotometry. 2.5.5. Data presentation of transport experiments The results of transport experiments are expressed as% of the amount added to the donor compartment or as apparent permeability coefficient, calculated according to: Papp ¼
DQ 1 1 DT A C0
with DQ DT : transport rate; A: surface area of the insert; and C0: initial target concentration in the donor compartment. Post-experiment validation. After each experiment, an incubation step with sodium fluorescein in TM (0.1% w/v) was performed for 60 min, followed by TEER measurement. The amount of sodium fluorescein appearing in the acceptor compartment was measured by UV spectrophotometry at 486 nm. These tests were carried out as an additional control of monolayer integrity and to assess possible effects on tight junctions by the components. Typical sodium fluorescein flux values across Caco-2 monolayers after the transport experiment with test compound are below 1%.
3. Results 3.1. Detection limits and recovery experiments The detection limit of stevioside was 50 ng per injection. The recovery of stevioside from spiked faeces samples (1, 6.5 or 20 mg g1 dry wt.) was 72.1 1.3% (n=20). Recovery could not be improved by using absolute EtOH or 97:3 EtOH:diethylether as extraction solvent. The recovery of stevioside from spiked blood samples (6.67 mg g1 dry wt.) after TLC clean-up was 68.4 2.6% (n=9). In order to obtain such a recovery it is very important to rinse the sample vials containing the stevioside extracts three times to allow for a quantitative transfer of stevioside to the TLC plates. The detection limit of (7-methoxycoumarin-4-yl)methyl ester of steviol was about 50 pg per injection. Using this procedure, steviol can be measured with a sensitivity that is about 1000 better than that obtained for stevioside. Recovery of steviol from spiked faeces (100, 170 and 1000 mg g1 dry wt.) was 34.5 2.1% for the lower and 55.2 11.8% for the higher concentrations. Use of other solvents for the extraction did not improve recovery (MeOH, diethyl ether). Recovery of steviol from spiked blood (340 mg g1 dry wt.) was 39.7 3.0%. The addition of 1% or higher concentrations of acetic acid to acetone did not significantly increase the recovery. In fact, care must be taken, as acid conditions can
1603
cause conversion of steviol into isosteviol. Therefore, in our experiments steviol was always extracted with acetone. As in vivo steviol might bind to different blood fractions, we preferred to extract the total blood fraction. Therefore, in all experiments, blood samples were freeze-dried before extraction. In spite of the poor recovery of steviol at the lower concentration levels, very minute amounts of steviol can still be measured due to the very sensitive detection after derivatisation, e.g. at the lowest concentration of spiked faeces (100 mg steviol g1 dry wt.) the steviol content in 50 mg faeces is 5 mg. At 35% recovery 1.7 mg will be detected, i.e. 35,000 the detection limit (50 pg). 3.2. Analysis of pig faeces During the period of the stevioside feeding (14 days) the control pigs consumed 1.09 kg day1, the stevioside group 0.96 kg day1 (daily intake of about 1.6 g stevioside 70 mg/kg bw). The somewhat reduced feed intake is probably due to the high stevioside content which makes the feed about 20 times sweeter than necessary for normal sweetening purposes and therefore might be repelling. No stevioside could be detected in the faeces samples, suggesting that bacteria of the colon metabolised all the stevioside present. For each kg of feed about 0.5 kg of dry faeces was produced. The maximum amount of stevioside to be expected in the faeces was about 3.2 mg g1 dry wt. It follows that the maximum concentration of steviol to be expected is around 1.28 mg g1 dry wt. Analysis of steviol in the faeces demonstrated hugh amounts of steviol, in concentrations around 853 48 mg g1 dry wt. This is about 65% of the maximum amount and implies a recovery of around 65% as found with spiked samples. No other constituents were detected. It follows that of the daily uptake of 1.6 g stevioside at the most 224 mg (35% of the steviol formed) is available for uptake via the colon (9.7 mg/kg bw). The results demonstrate that stevioside was completely converted into steviol. These results are opposed to the low conversion of stevioside (about 20%) by bacteria of the chicken caecum (Pomaret and Lavieille, 1931; Geuns et al., 2003), but are in agreement with the incubations of pig (Geuns et al., unpublished) and human faeces (Hutapea et al., 1997; Koyama et al., 2003a; Gardana et al., submitted for publication) under anaerobic conditions in which a complete conversion of stevioside to steviol took place. 3.3. Analysis of blood The complete conversion of stevioside into steviol makes the blood analyses a crucial aspect of the work. No stevioside or steviol could be found in the blood samples analysed. As the steviol analyses in the blood are
1604
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
very important, analyses of larger amounts of blood were done to be absolutely sure that no steviol is present. Up to 5.5 ml samples of blood were exhaustively extracted, but no indication for the occurrence of steviol was found. 3.4. Transport studies The control experiment with fluorescein shows that after the transport experiments with stevioside, rebaudioside A and steviol were performed, the monolayer integrity was not compromised, the flux values of sodium fluorescein being below 0.5% (Fig. 3). Previous experiments had shown that DMSO (up to 2%) did not affect the monolayer integrity (results not shown). The transport of stevioside and rebaudioside A through the Caco-2 monolayers was very low (Papp 0.16106 and 0.11106 cm/s, respectively, Table 1). No effect on TEER-values or on sodium fluorescein flux could be observed (Fig. 3). Transport of stevioside and rebaudioside A from the apical to basolateral side is lower than the transport rate of Na-fluorescein, a marker for paracellular transport. This is consistent with the fact that the transport is inversely related to the molecular weight (Hidalgo et al., 1989). Transport of steviol in the absorptive direction was concentratrion independent and was much higher than
Fig. 3. Effect of different test conditions on transport of sodium fluorescein (open bars) and on TEER-values (average of three determinationsS.D.). 1: 1 mm stevioside, 2: 1 mm rebaudioside A, 3: 30 mm steviol, 4: 100 mm steviol, 5: 300 mm steviol, 6: 1 mm steviol, 7: 30 mm steviol, 8: 100 mm steviol; conditions 1–6: absorptive transport, conditions 7–8: secretory transport. Table 1 Apparent permeability coefficients in Caco-2 cell layers of the tested compounds in the absorptive and secretory direction (n=3) Condition
Pappabs 106 (cm/s)S.D.
Pappsecr106 (cm/s)S.D.
1 mm stevioside 1 mm rebaudioside A 30 mm steviol 100 mm steviol 300 mm steviol 1 mm steviol
0.16 0.01 0.11 0.03 38.7 4.7 44.5 1.4 39.5 1.7 31.9 3.4
– – 6.57 0.06 7.93 0.15 – –
the transport of stevioside and rebaudioside A. Steviol appears to easily cross the Caco-2 monolayers, exhibiting transport behavior of a compound with transcellular passive diffusion characteristics. Steviol at 300 mm and 1 mm slightly compromised the integrity of the Caco-2 monolayers as seen by decreased TEER-values (Fig. 3) and increased sodium fluorescein flux as compared with control conditions. At 100 mm integrity appeared to be maintained. However, the steviol transport itself was not influenced. Steviol transport in the secretory direction (basolateral to apical) was lower than the absorptive transport. At 100 mm the absorptive apparent permeability coefficient (Papp) amounted to 44.5106 while the secretory apparent permeability coefficient amounted to 7.93106 cm/s (Table 1). In additional experiments, it was shown that benzoic acid inhibited the absorptive transport of steviol: Papp for steviol (100 mm) decreased from 30.56 to 20.47106 cm/s upon coincubation with 10 mm benzoic acid.
4. Discussion Steviol was the only compound that we could detect in pig faeces. This is in agreement with published results where steviol was found as the only possible metabolite produced by the intestinal microflora from various animal species and humans under strictly anaerobic conditions (Hutapea et al., 1997; Koyama et al., 2003a; Gardana et al., submitted for publication; Geuns et al., 2003). The steviol formed in the faeces was not found in the blood. These results suggest that steviol, which is only sparingly soluble, was not taken up from the colon of pigs. Transport studies with Caco-2 cells on the contrary showed that the absorptive transepithelial flux of steviol from a solution was relatively high. This discrepancy will be discussed. The lack of steviol in the blood samples can probably not be attributed to metabolism during or after uptake as was the case with soy isoflavones that after uptake were metabolised to compounds that were hydrolysable with a combined b-glucuronidase and sulfatase enzyme preparation (Setchell et al., 2002); indeed free steviol was detected in the plasma of rats up to at least 8 h after feeding stevioside or steviol (Compadre et al., 1988; Koyama et al., 2003b). Moreover, the in vitro conversion of steviol by liver S-9 fraction from Aroclor 1254-pretreated rats was rather low (about 0.3% after 2 h; Compadre et al., 1988). The intrinsic clearance of steviol by human microsomes was about 4 times lower than that of rat microsomes (Koyama et al., 2003b). Taking into account the very low detection limits of steviol when analysed as the (7-methoxycoumarin-4yl)methyl ester (50 pg) the amount of steviol possibly remaining undetected in the blood samples can be estimated to be very low (below 1 mM, i.e. below 318
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
ng/ml). This hypothetical maximum steviol concentration in the blood would probably not be toxic, as in hamsters fed steviol up to 250 mg/kg bw no toxic effects were found (Wasuntarawat et al., 1998). In this case steviol would have been easily taken up by the intestines. When steviol was fed to rats (45 mg/kg bw) a fast uptake was found and the highest plasma concentration of 18.3 mg/ml was observed after 15 min (the first data point, Table 2; Koyama et al., 2003b). The plasma concentration declined to about 2–3 mg/ml at 8 h. Although we are aware that species differences might occur, we have extrapolated the data obtained in rats to hamsters and pigs. Assuming a similar uptake and metabolism in hamsters the reported NOEL of 250 mg steviol/kg bw would correspond to a plasma concentration of 102 mg/ ml 15 min after intubating steviol and about 13.6 mg/ml after 8 h. In our experiment this would then be 3.94 and 0.5 mg/ml, respectively. These concentrations are above the detection limit of (7-methoxycoumarin-4-yl)methyl ester of steviol. Therefore, we suggest that in vivo the uptake of the carboxy acid steviol from the colon is neglectible and that it rather remains adsorbed to the compounds present in the colon (pH 7–7.5) of which the contents is being concentrated by withdrawal of water. The results of the transport study with Caco-2 cells agree with those obtained with everted gastrointestinal sacs (Koyama et al., 2003b). In these experiments only a very small amount of the Stevia mixture (rebaudioside A, 28.2%, stevioside, 17%, rebaudioside C, 25.2%, dulcoside A, 10.2% and unknowns, 18.8%) was transported from the mucosal to the serosal side (less than 0.5% equivalent of salicylic acid, the positive control). The lack of stevioside uptake was also reported earlier (Yamamoto et al., 1985; Bracht et al., 1985). However, the steviol transport through everted gastrointestinal sacs amounted to about 70% of that of salicylic acid (Koyama et al., 2003b). A complete in vivo absorption from the lower rat bowel was reported after oral or intracecal administration of 17-[14C]-steviol (Wingard et al., 1980). However, in this study, it was not mentioned Table 2 Steviol concentration (mg/ml plasma) measured in the blood of rats (Koyama et al., 2003b) or estimated to be present in the blood of hamsters (Wasuntarawat et al., 1998) and pigs after the administration of steviol to the animalsa Amount administered mg/kg bw 45 mg (rats) 250 mg (hamsters) 9.7 mg (pigs) a
mg/ml 15 min
2h
4h
8h
16 hb
18.3 102 3.94
2.5 13.6 0.5
3.5 19 0.7
2.5 13.6 0.5
1 5.4 0.2
The amount available for uptake in pigs is given as the difference between the total amount of steviol expected in the faeces after feeding stevioside and that really found in the faeces (see Section 3). b Amount estimated from the decay of the previous data points.
1605
that coprophagy was prevented. After oral administration of steviol 96.36% was excreted in the feces, 1.48% in the urine. In bile duct ligated animals, 96% was excreted in the urine, 3.31% in the feces after oral administration, and 94.3 and 6.03% in the urine respectively in the faeces after intracaecal administration. In bile duct cannulated animals dosed intracaecally 105.5% of the radioactivity was found in the bile. These results demonstrated that nearly all the steviol administered was excreted. After administration of 3H-stevioside (125 mg/kg bw) to Wistar rats the radioactivity in the blood increased to reach a maximum of 4.83 mg stevioside equivalents/ml. At 120 h the percentages of radioactivity excreted into the feces and expired air were 68.4 and 23.9%, respectively, while radioactivity excreted into the urine was only 2.3%. Radioactivity excreted into the bile at 72 h was 40.9% of the original dose. The authors concluded that an entero-hepatic circulation occurs in the body (Nakayama et al., 1986). However, it is not always possible to discriminate between label originating from the sugar moiety or from the steviol skeleton of the tritium-labelled stevioside. The chemical identification of the radioactive compounds present was not straightforward. The hydrophilic nature of stevioside and rebaudioside A may explain the poor transport through the Caco-2 monolayers and everted gastrointestinal sacs, as well as the poor absorption in an in vivo situation. The absorption of steviol might be explained partly by passive diffusion and partly by carrier-mediated transport via the monocarboxylic acid transporter (Takanaga et al., 1994). Indeed, we found a 33% inhibition of steviol transport by benzoic acid. This last possibility might also explain the much lower steviol transport in the secretory direction. We speculate that the low in vivo absorption may be attributed to low diffusion out of the caecal content. A weak mutagenic effect of steviol (only 90% purity) in one sensitive S. typhimurium TM 677 strain (see Section 1) does not mean that stevioside used as a sweetener should be carcinogenic per se, even if the stevioside might be transformed to steviol by bacteria in the colon. The safety of oral stevioside in relation to carcinogenic activity is evidenced by the following reports. Male and female F344 rats were daily fed with a ration containing 0.1, 0.3 or 1% of stevioside and rebaudioside for a period of 22 (males) or 24 (female rats) months (Yamada et al., 1985). The animals were then killed, and the researchers conducted biochemical, anatomical, pathological and carcinogenic tests on 41 organs following autopsy. In addition they performed ongoing hematologic and urine tests on the same animals. Each of the animals was matched to a control animal that experienced exactly the same treatment, except for the stevioside. No significant dose-related changes were found in the growth, general appearance, hematological and
1606
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607
blood biochemical findings, organ weights, and macroscopic or microscopic observations. It was also concluded that any neoplasms that occurred were not attributable to the administration of stevioside. Even at the highest dose of 1% no significant effects were found. This high dose is equivalent to 125 times the average daily dose of sweeteners that a normal human would require for sweetening purposes. The effects of a chronic oral feeding of stevioside during 24 months were studied in rats (Xili et al., 1992). The concentrations used were: 0, 0.2, 0.6 and 1.2%. Growth, food utilisation and consumption, general appearance and mortality were similar in treated and control groups. The mean lifespan of rats given stevioside was not significantly different from that of the controls. No treatment-related changes were observed in haematological, urinary or clinical biochemical values at any stage of the study. The incidence and severity of non-neoplastic and neoplastic changes were unrelated to the level of stevioside in the diet. In a chronic toxicity study with F344 rats during 2 years it was concluded that there were no significant increases in the incidence of neoplastic lesions in any organ or tissue in the stevioside treated groups (2.5 and 5%, i.e. daily dose of 385 and 775 mg per rat, i.e. 1 and 2 g kg1 bw, which is a very high dose; Toyoda et al., 1997). In male animals, the number of testicular tumours had the tendency to decrease. Moreover, the incidence of adenomas of the mammary gland in the stevioside-treated female rats was significantly lower than that in the controls. The severity of chronic nephropathy in males was also clearly reduced by both stevioside concentrations. There was no increase in the incidence of urinary bladder carcinogenesis in rats by stevioside (Hagiwara et al., 1984). Very significant inhibitory effects of stevioside were reported on tumor promotion by 12-O-tetradecanoylphorbol-13-acetate in carcinogenesis in mouse skin (Yasukawa et al., 2002). In 1999, the Joint FAO/ WHO Expert Committee on Food Additives (JECFA) clearly stated: ‘‘Stevioside has a very low acute oral toxicity. Oral administration of stevioside at a dietary concentration of 2.5% to rats for two years, equal to 970 and 1100 mg kg1 BW per day in males and females, respectively, had no significant effect. Reduced body-weight gain and survival rate were observed at a dietary concentration of 5% stevioside. There was no indication of carcinogenic potential in a long-term study. . .’’ (WHO, 1999). Moreover, there have never appeared any clinical adverse reports on Stevia rebaudiana or stevioside in any of the countries in which these products were used over a long period of time (e.g. Paraguay: at least over 100 years, Japan: more than 25 years, South-Korea: 16 years, Brazil: 13 years, China: 12 years or the USA: since 1995 admitted as a dietary supplement). The above literature data, mainly with rodents, the lack of reports on cancerigenous activity in populations, and our data showing
additionally that no traces of stevioside or its controversial metabolite steviol were found in the blood of pigs are supportive for the safe use of stevioside as a sweetener.
Acknowledgements The authors acknowledge Professor D. Kinghorn for the gift or pure stevioside and Specchiasol for financial support. Professor R. Geers is acknowledged for use of the Zootechnical Centre and Jan M.C. Geuns and Patrick Augustijns acknowledge the ‘‘Onderzoeksraad KULeuven’’ for grants OT/00/15 and OT/01/14, respectively and the FWO for grant number G.0111.01. We thank Hilde Verlinden and Tom Struyf for their skilful technical assistance.
References Bracht, K.A.M., Kemmelmeier, F.S., Ishii, E.L., Alvarez, M., Bracht, A., 1985. Effect of Stevia rebaudiana natural products on cellular and sub-cellular metabolism. Arq. Biol. Tecnol. 28, 431–455. Compadre, C.M., Hussain, R.A., Nanayakkara, N.P.D., Pezzuto, J.M., Kinghorn, A.D., 1988. Mass spectral analysis of some derivatives and in vitro metabolites of steviol, the aglycone of the natural sweeteners, stevioside, rebaudioside A and rubusoside. Biomed. Environ. Mass Spectrom. 15, 211–222. Gardana, C., Simonetti, P., Canzi, E., Zanchi, R., Pietta, P.G. 2003. Metabolism of stevioside and rebaudioside A from Stevia rebaudiana extracts by human microflora. J. Agric. Food Chem. (submitted). Geuns, J.M.C., 2002. Safety evaluation of Stevia and stevioside. In: Atta-ur-Rahman (Ed.), Studies in Natural Products Chemistry, vol. 27: Bioactive Natural Products (Part H). Elsevier, Amsterdam, pp. 299–319. Geuns, J.M.C., Malheiros, R.D., Moraes, V.M.B., Decuypere, E.M.P., Compernolle, F., Buyse, J.G., 2003. Metabolism of stevioside by chickens. J. Agric. Food Chem. 51, 1095–1101. Hagiwara, A., Fukushima, S., Kitaori, M., 1984. Effects of the three sweeteners on rats urinary bladder carcinogenesis initiated by N-butyl-N-(4-hydroxybutyl)-nitrosamine. Gann 75, 763–768. Hidalgo, I.J., Raub, T.J., Borchardt, R.T., 1989. Characterization of the human colon carcinoma cell line (Caco2) as a model system for intestinal epithelial permeability. Gastroenterology 96, 736–749. Hutapea, A.M., Toskulkao, C., Buddhasukh, D., Wilairat, P., Glinsukon, T., 1997. Digestion of Stevioside, a natural sweetener, by various digestive enzymes. J. Clin. Biochem. Nutr. 23, 177–186. Kennelly, E.J., 2002. Sweet and non-sweet constituents of Stevia rebaudiana (Bertoni) Bertoni. In: Kinghorn, A.D. (Ed.), Stevia, the genus Stevia. Medicinal and Aromatic Plants—Industrial Profiles, Vol. 19. Taylor and Francis, London, New York, pp. 68–85. Kinghorn, A.D., 2002. Overview. In: Kinghorn, A.D. (Ed.), Stevia, the genus Stevia. Medicinal and Aromatic Plants—Industrial Profiles, Vol. 19. Taylor and Francis, London, New York, pp. 1–17. Klongpanichpak, S., Temcharoen, P., Toskulkao, C., Apibal, S., Glinsukon, T., 1997. Lack of mutagenicity of stevioside and steviol in Salmonella typhimurium TA98 and TA100. J. Med. Assoc. Thailand 80, 121–128. Koyama, E., Kitazawa, K., Ohori, Y., Sakai, N., Izawa, O., Kakegawa, K., Fujino, A., Ui, M., 2001. Intestinal degradation, absorption, and hepatic metabolism of the glycosidic sweeteners, Stevia
J.M.C. Geuns et al. / Food and Chemical Toxicology 41 (2003) 1599–1607 mixture. In: Proceedings Second IUPA-International Symposium on Sweeteners, Hiroshima, p. 12. Koyama, E., Kitazawa, K., Ohori, Y., Izawa, O., Kakegawa, K., Fujino, A., Ui, M., 2003a. In vitro metabolism fo the glycosidic sweeteners, Stevia mixture and enzymatically modified Stevia in human intestinal microflora. Food Chem. Toxicol. 41, 359–374. Koyama, E., Sakai, N., Ohori, Y., Kitazawa, K., Izawa, O., Kakegawa, K., Fujino, A., Ui, M., 2003b. Absorption and metabolism of the glycosidic sweeteners, Stevia related compounds in human and rat. Food Chem. Toxicol. 41, 875–883. Matsui, M., Matsui, K., Kawasaki, Y., Oda, Y., Noguchi, T., Kitagawa, Y., Sawada, M., Hayashi, M., Nohmi, T., Yoshihira, K., Ishidate, M., Sofuni, T., 1996a. Evaluation of the genotoxicity of stevioside and steviol using six in vitro and one in vivo mutagenicity assays. Mutagenesis 11, 573–579. Matsui, M., Sofuni, T., Nohmi, T., 1996b. Regionally-targeted mutagenesis by metabolically-activated steviol: DNA sequence analysis of steviol-induced mutants of guanine phosphoribosyltransferase (gpt) gene of Salmonella typhimurium TM677. Mutagenesis 11, 565–572. Nakayama, K., Kasahara, D., Yamamoto, F., 1986. Absorption, distribution, metabolism and excretion of Stevioside in rats. J. Food Hyg. Soc. Jpn 27, 1–8. Ogawa, T., Nozaki, M., Matsui, M., 1980. Total synthesis of stevioside. Tetrahedron 36, 2641–2648. Pezzuto, J.M., Compadre, C.M., Swanson, S.M., Nanayakkara, N.P.D., Kinghorn, A.D., 1985. Metabolically activated steviol, the aglycone of stevioside, is mutagenic. Proc. Nat. Acad. Sci. USA 82, 2478–2482. Pomaret, M., Lavieille, R., 1931. Le principe a` saveur sucre´e du Kaa`he`-e´ (Stevia rebaudiana Bertoni). Quelques proprie´te´s physiologiques du Ste´vioside. Bull. Soc. Chim. Biol. 13, 1248–1252. Procinska, E., Bridges, B.A., Hanson, J.R., 1991. Interpretation of results with 8-azaguanine resistance system in Salmonella typhimurium: no evidence for direct acting mutagenesis by 15-oxosteviol, a possible metabolite of steviol. Mutagenesis 6, 165–167. Setchell, K.D.R., Brown, N.M., Zimmer-Nechemias, L., Brasher, W.T., Wolfe, B.E., Kirschner, A.S., Heubi, J.E., 2002. Evidence for lack of absorption of soy isoflavone glycosides in humans, supporting the crucial role of intestinal metabolism for bioavailability. Am. J. Clin. Nutr. 76, 447–453. Starrat, A.N., Kirby, C.W., Pocs, R., Brandle, J.E., 2002. Rebaudioside F, a diterpene glycoside from Stevia rebaudiana. Phytochemistry 59, 367–370. Takanaga, H., Tamai, I., Tsuji, A., 1994. pH-Dependent and carriermediated transport of salicylic acid across caco-2 cells. J. Pharm. Pharmacol. 46, 567–570.
1607
Temcharoen, P., Pimbua, J., Glinsukon, T., Rojanapo, W., Apibal, S., 1998. Mutagenic activity of steviol to Salmonella typhimurium TM 677: comparison of the activity of S9 liver fractions from five laboratory animal species. Bull. Health Sci. Tech. 1, 38–45. Temcharoen, P., Suwannatrai, M., Klongpanichpak, S., Apibal, S., Glinsukon, T., Toskulkao, C., 2000. Evaluation of the effect of steviol on chromosomal damage using micronucleus test in three laboratory animal species. J. Med. Assoc. Thailand 83, 101–108. Terai, T., Ren, H., Mori, G., Yamaguchi, Y., Hayashi, T., 2002. Mutagenicity of steviol and its oxidative derivatives in Salmonella typhimurium TM677. Chem. Pharm. Bull. 50, 1007–1010. Toskulkao, C., Chaturat, L., Temcharoen, P., Glinsukon, T., 1997. Acute toxicity of stevioside, a natural sweetener, and its metabolite, steviol, in several animal species. Drug Chem. Toxicol. 20, 31–44. Toyoda, K., Matsui, H., Shoda, T., Uneyama, C., Takada, K., Takahashi, M., 1997. Assessment of the carcinogenicity of Stevioside in F344 rats. Food Chem. Toxicol. 35, 597–603. Van Gelder, J., Witvrouw, M., Pannecouque, C., Henson, G., Bridger, G., Naesens, L., De Clercq, E., Annaert, P., Shafiee, M., Van den Mooter, G., Kinget, R., Augustijns, P., 1999. Evaluation of the potential of ion pair formation to improve the oral absorption of twe potent antiviral compounds, AMD3100 and PMPA. Int. J. Pharm. 186, 127–136. Wasuntarawat, C., Temcharoen, P., Toskulkao, C., Mungkornkarn, P., Suttajit, M., Glinsukon, T., 1998. Developmental toxicity of steviol, a metabolite of stevioside, in the hamster. Drug Chem. Toxicol. 21, 207–222. WHO Food Additives Series 42, 1999. Safety evaluation of certain food additives ’’Stevioside’’. Genf 119–143. Wingard, R.E., Brown, J.P., Enderlin, F.E., Dale, J.A., Hale, R.L., Seitz, C.T., 1980. Intestinal degradation and absorption of the glycosidic sweeteners stevioside and rebaudioside A. Experientia 36, 519–520. Xili, L., Chengjiany, B., Eryi, X., Reiming, S., Yuengming, W., Haodong, S., Zhiyian, H., 1992. Chronic oral toxicity and carcinogenicity study of stevioside in rats. Food Chem. Toxicol. 30, 957–965. Yamada, A., Ohgaki, S., Noda, T., Shimizu, M., 1985. Chronic toxicity of dietary Stevia extracts. J. Food Hyg. Soc. Jpn. 26, 169–183. Yamamoto, N.S., Bracht, K.A.M., Ishii, E.L., Kemmelmeier, F.S., Alvarez, M., Bracht, A., 1985. Effect of steviol and its structural analogues on glucose production and oxygen uptake in rat renal tubules. Experientia 41, 55–57. Yasukawa, K., Kitanaka, S., Seo, S., 2002. Inhibitory effect of stevioside on tumor promotion by 12-O-tetradecanoylphorbol-13-acetate in two-stage carcinogenesis in mouse skin. Biol. Pharm. Bull. 25, 1488–1490.