- Email: [email protected]
Influence of an estrone-desulfating intestinal flora on the enterohepatic circulation of estrone-sulfate in rats
J. sreroid Biochem.
Vol. 26, No. 2, pp. 235-239,
0022-4731/87
1987
Copyright 0
Printed in Great Britain. All rights reserved
$3.00 + 0.00
1987 Pergamon Journals Ltd
INFLUENCE OF AN ESTRONE-DESULFATING INTESTINAL FLORA ON THE ENTEROHEPATIC CIRCULATION OF ESTRONE-SULFATE IN RATS J. VAN ELDERE*, G. PARMENTIER,J. ROBBEN and H. EY~~EN The Rega Institute, University of Leuven, Minderbroedersstraat 10, B-3000 Leuven, (Received
22 May
Belgium
1986)
Summary-The fecal and urinary excretion of orally administered [4-“‘Clestrone-3-sulfate was studied in germfree (GF) rats, conventional (CV) rats and gnotobiotic rats selectively associated with estronedesulfating and/or cecal-volume reducing microorganisms. The time required to excrete 50% of the total label recovered (t I/2) was 22 h in CV rats vs 32 h in GF rats. Gnotobiotic rats selectively associated with a cecal volume-reducing flora (CRF rats) excreted the label even faster (t l/2 = 13 h) than CV rats. Association of GF rats as well as CRF rats with estrone-desulfating microorganisms (termed S, + S, + R rats and CRF + S, + S, + & rats, respectively) led to a slower excretion of labeled products (t I/2 = 38 h in S, + S, + & rats and t l/2 = 27 h in CFR + S, + S, + & rats). Intestinal microbial desulfation also increased the relative part of the urinary excretion from 4% in GF rats to 8% in S, + S, + & rats and from 3% in CRF rats to 9% in CFR + S, + S, + I& rats. We conclude that intestinal microbial desulfation enhances the enterohepatic circulation of orally administered estrone-3-sulfate.
INTRODUCTION The use of estrogens in clinical situations has led to a considerable interest in the regulation of the enterohepatic circulation and excretion of these substances. Conjugation with sulfate or glucuronic acid in the liver enhances the water-solubility and the fecal excretion of estrogens. During their passage through the intestines, these estrogen conjugates can however be hydrolysed. Deglucuronidation can be of mammalian or microbial origin; desulfation is probably solely of microbial origin [ 1,2]. Experiments with germfree (GF) rats have shown that estrone-3-sulfate can be reabsorbed intact from the small intestine but at a much lower rate than estrone [3]. In the cecum, reabsorption was only possible after desulfation of estrone-sulfate [3]. Conventional (CV) rats pretreated with antibiotics reabsorbed less estrone-sulfate, probably because of the elimination of estrone desulfating bacteria [4]. Administration of antibiotics to pregnant women has been shown to decrease plasma and urinary estriol levels [S, 61. Similarly, there have been reports of pregnancy in women on oral contraceptives who were also taking antibiotics [7, 81. The use of antimicrobial drugs in rat studies on the enterohepatic circulation of estrogen-conjugates entails however certain limitations. The drug itself, apart from its antimicrobial activity, may in some way influence the absorption and metabolism of the estrogens [9] and it is very difficult to define what part
*J. van Eldere is a Research National Fund for Scientific
Assistant Research
of the Belgian (NFWO). 235
of the normal flora has been eliminated. The use of GF animals is also subject to noncontrolled variability. GF rodents, such as mice and rats, because of the absence of the intestinal flora, present with several physiological anomalies that form potential sources of error [lo]. Lack of microbial degradation leads to the accumulation of mutinous material in the lower bowel, causing retention and even secretion of water in the lumen of the bowel. In addition, the reduced intestinal muscle tone and motility lead to dilatation of the cecum and a slower intestinal transit. The comparison of GF and CV rats is, moreover, hindered by the fact that the intestinal flora performs also other reactions besides deconjugation, such as oxidation-reduction reactions, that may influence the excretion of the estrogens. By isolating pure cultures of estrone-desulfating bacteria from rat cecal material, and associating GF rats with these bacteria, we were able to define quantitatively the influence of estrone-desulfation in comparable situations. Because a close relationship has been demonstrated between the volume of the cecum and the rate of the intestinal transit in mice [l l] two other groups of gnotobiotic rats were associated with a defined flora, termed CRF flora, consisting of four strains of unidentified anaerobic intestinal bacteria that were able to reduce the cecal volume and the intestinal transit time of GF rats by 60%. One of these two groups of rats was, in addition, associated with the estrone-desulfating strains. Estrone-3-sulfate was administered orally to prevent partial desulfation in the liver prior to the first passage through the intestines.
236
Animals
J.
VAN ELDERE et al.
and diet 1 ml water and 5 ml of I M NadH. Thin-layer chromatography (TLC) on Silicagel 60 (E. Merck, Female inbred Fisher rats, approx. 3 months old at Darmstadt, W. Germany) in ethylacetate-butanolthe start of the experiments were used. GF, gnoacetic acid-water (8:6:3:3; v/v) showed that no free tobiotic and CV animals were kept in Trexler’s estrone remained. The pyridine was evaporated and flexible-film plastic isolators (Standard Safety Equipthe residue dissolved in 80 ml of water. This solution ment, Palatine, Ill., U.S.A.) [12] and were fed a was desalted on Baker octadecyl extraction columns. sterilized commercial diet (N.V. Trouw, Gent, Belgium) and water ad l~b~turn. Each group contained 5 The [4-‘4C]estrone-sulfate was eluted with methanol and more than 99% of the [4-‘4C]estrone was recovrats all kept individually in cages. Cages had double ered as sulfate-ester. mesh-wire bottoms and absorbing paper on the bottom to prevent coprophagy and facilitate collection Quantifcation qf label excreted in .feces and urine of feces and urine. At the start of the experiment, one pCi of labeled Feces were collected in 50 ml Falcon tubes (Becton [4-t4C]estrone-3-sulfate was administered intraDickinson, Oxnard, CA, U.S.A.) and homogenized gastrically through a catheter. For the first 55 h, feces in pre-cooled water. Steroids were extracted from and urine were collected at intervals of 7 and 17 h 0.5 g of freeze-dried feces with 30 ml of aqueous 80% and, later on, at intervals of 3 days until the cumuethanol for 4 h at 65’C. Three samples of 0.5 ml of lative amount of label excreted reached a plateau. the extract were counted twice in a Packard Tri-carb The cumulative amounts of label excreted were transLiquid Scintillation Counter 2660 (Packard, Ill., formed into excretion curves via -log (1 - p,/p,,,,,) U.S.A.) To correct for quenching, quench-correlation with pu,= total amount of label excreted at time t and curves were stored in the memory of the Liquid &ax = total amount of label recovered, as described Scintillation System and all results were consequently by Lindstedt and Norman[ 131.Fifty percent excretion expressed as disintegrations per minute. of the total amount of label recovered (t l/2) corTo determine the urinary excretion of labeled related with -log{1 -p,/p,.,) = 0.301 on best fit products. absorbing papers were placed on the botcurves. At the end of the experiment, the animals tom of the rat-cages, collected, dried, and then extracwere sacrificed and total body weight and cecum ted with 50 ml of 80% ethanol. From the extract, weight were determined. three aliquots of 1 ml were counted. GF animals were associated with estronedesulfating bacteria (S, + S, + Rs rats), or cecal Fractionation qf the fecal labeled compounds volume-reducing strains (CRF rats), or cecal volumeThe amounts of unconjugated label, glucuronic reducing strains plus estrone-desulfating strains acid conjugated label and sulfoconjugated label were (CRF + S, + Sz + R, rats), by oral inoculation of determined on the aqueous 80% ethanol extract of liquid cultures. This was repeated until cultures of feces excreted between 7 and 24 h after the start of the fecal material showed the presence of the viable experiment. Five ml of extract were evaporated to microorganisms. The strains S,, Sz and R, were dryness and redissolved in 10ml of 0.2 M acetate isolated from the intestinal flora of the rat; strains S, and Sz possessed bile salt-sulfatase activity 114, 151; buffer, pH 5. Three successive extractions with 8 ml diethylether yielded the unconjugated labeled fracstrain R, possessed iodothyronine-sulfatase tion. Glucuronide-esters in the residue were enactivity (161. In addition, all were found to have zymatically deglucuronidated with 3302 Fishmann estrone-sulfatase activity. The cecal volume-reducing Units of ~-glucuronidase (Calbiochem no. 34743). flora consisted of four unidentified strictly anaerobic The liberated label was then extracted 3 times with microorganisms and was isolated from the mouse. 8 ml diethylether. The residue was further used for The CRF strains did not transform estrone or the determination of sulfate-esters. Enzymatic deestrone-sulfate but reduced the cecal size of gnosulfation of these was performed by addition of 198 tobiotic mice and rats by 60-&S%. I.U. of sulfatase-enzyme (Sigma S9626), 1 ml EDTA 1.86% and 1 ml mercapto-ethanol 0.78%. The dePreparation of [4-‘4C]estrone-3-sulfate sulfated estrone was extracted three times with 8 ml [4-“‘Clestrone was obtained from Amersham Interdiethylether. All fractions and the residual nonnational (Amersham, England), chlorosulfonic acid extractable labeled product in the acetate buffer were was purchased from Fluka AG (Buchs, Switzerland) counted in duplicate as described above. and Baker extraction columns for single use (OctaTo determine the relative amounts of estrone, decyl, 6ml) were from J. T. Baker Chemical Co. estradiol and estriol or other estrogens in the feces, (Phillipsburg, N.Y., U.S.A.). One mg of labeled the non-conjugated labeled fractions of CV rat feces, [4-“‘C]estrone (100 pCi/mg) was diluted with 50 mg the sulfate-conjugated label of GF, CRF and of non-labeled estrone, dissolved in 2 ml of dry CRF + S, + S? + R, rat feces and the non-conjugated pyridine and placed into an ice-bath. To this was label of CRF + S, f S, + R, rat feces, were subjected added, dropwise, 0.1 ml of chlorosuifonic acid. After to TLC analysis in isooctane~thylacetate-acetic acid 24 h the reaction was terminated by the addition of (5:5:1; by vol).
Influence of intestinal microbial desulfation
237
RESULTS
TLC analysis of labeled steroids excreted via the feces
TLC analysis of the feces collected between 7 and 24 h after the oral administration of 1 PCi of [4-i4C]estrone-sulfate, showed that more than 95% of the radioactivity corresponded to reference estrogens. The radioactivity of the non-conjugated labeled estrogens in CV rats was distributed between a first peak that corresponded to estrone (4&45%) and a second peak that corresponded to estradiol (55-60%). In all other fractions examined (sulfateester fraction in GF, CRF and CRF + St + S2 + R, rats and unconjugated fraction in CRF + S, + Sz + Ry rats) nearly 100% of the labeled product corresponded to estrone. The other fractions were not analysed because they contained insu~cient label for identification.
S,+S,+R9 GF CRF+S,+S,+RS
CRF
Time
Table
administration
As shown in Table 2, the composition of the fecal estrogen-fraction was fairly similar in GF and CRF rats. In GF rats free estrogen represented 2.8% of the labeled estrogen versus 5.1% in CRF rats; estrogenglucuronide was 10.6% in GF and 7.4% in CRF rats, while estrogen-sulfate was 73.8 and 79.5% in GF and CRF rats respectively. This suggests that, although CRF rats excreted the administered estrone-sulfate rapidly, and even faster than CV rats, their intestinal metabolism of estrone-sulfate did not differ from that of GF rats. In CV rats, however, the fractionation of the fecal label yielded markedly different results; free estrogen represented 81.5%, estrogen-glucuronide 1.7% and estrogen-sulfate 2.7% of the fecal labeled estrogens. Injluence of microbial desulfation on the excretion of labeled estrone-sulfate
Comparison of the fecal plus urinary excretion of label in GF and S, -t Sz + R,-associated rats showed that, after 8 days, the total excretion expressed as a percentage of the total amount administered was only slightly but still signifi~ntly (P c 0.01) higher in GF
I. Fecal and urinary excretion of label after oral administration of [4-‘4C]estrone-sulfate to enotobiotic and conventional rats* Total r112t
rats S, t S, + R, rats CRF rats CRF + S, + S2 + R, rats CV rats GF
32h”(iZh) 38hd(lt:3h) 13h”(s_7h) 27h’(f3h) 22h”(k3h)
recoveryt
82%b(k2%) 77%b(+2%) 83%’ (14%) 71°,Gc(+ 6%) 71%b.Cff4%)
Urinary excretionlj
4%d ( + I .O%) 8%d (10.8%) 3%‘(&0.8%) 9%C( f I .5%) 24%d.C(i 3.0%)
a, b, c, d, e: All values in the same column that were statistically different (P < 0.01) have the same alphabetical superscript. *Three-month old rats; 5 animals per group. tThe r1/2 value was calculated from the “best-fit” curve after transformation of the cumulative excretion-curve according to -lo& 1 L r,/&*,) [see ref. 13). $Percentage of the amount administe~d; & one standard deviation. @Percentage of total amount recovered in feces + urine: & one standard deviation. sa
%‘?.--E
1hr 1
Fig. 1. Semilogarithmic plot of the cumulative excretion curves after administration of labeled estrone-sulfate. Transformation according to -log(l - p,/p,,,) [see ref. 131. Fifty percent excretion ([l/2) is reached at -log/l -a/ p,,,) = 0.301. pI: total amount of label excreted at time t. p_: total amount of label recovered.
Fecal and urinary excretion of label in GF, CV and CRF rats
In rats selectively associated with the CRF flora the cecal volume was substantially decreased as compared to GF rats; the cecum-weight as a percentage of total body-weight decreased from 13.5% in GF rats to 5.7% in CRF rats. Association of these rats with the desulfating strains S, + S, + b did not significantly change the cecal weight (11.5% of total body weight in S, + S, + Rq rats and 5.8% in CRF + S, + S, + R, rats). In CV rats the cecal weight represented 2.3% of total body weight. In CRF rats, the total amount of label recovered in feces plus urine after 8 days was 83% of the amount administered, nearly equal to the 82% found in GF rats, but significantly (P < 0.01) different from the 71% found in CV rats. The urinary excretion, expressed as percentage of the total fecal plus urinary excretion of label, amounted to 3% in CRF and 4% in GF rats versus 24% in CV rats. The excretion of labeled steroids in feces and urine of CRF rats was significantly faster than in GF rats and in CV rats, as deduced both from the cumulative excretion curves (Fig. 1) and from the t l/2 values (Table I). Of the total label recovered, 50% (= t l/2) had been excreted by GF rats after 32 h and by CV rats after 22 h; in CRF rats the t l/2 was only 13 h.
after
238
J. Table
2.
Percentages
of
the
VAN ELDERE et al.
different types of labeled fecal estrogen conjugates after [4-‘“Clestrone-sulfate to gnotobiotic and conventional rats* Free estrogen
GF rats S, + S, + R, rats CRF rats CRF + S, + S, + R, rats cv rats
2.8% 24.7% 5.1% 39.4% 81.5%
(*o.s%) (+3%) (kO.5%) (f 5%) (&3%)
Estrogen glucuronide
Estrogen sulfate
10.6%(?2%) 32.8% (+ I I %) 7.4% ( * I %) 27.8% (k 7%) 1.7% (kO.3%)
73.8%(+5%) 24.6% (+9%) 79.5%(kl%) 19.5% (f 10%) 2.7% (+0.5%)
oral
administration
of
Residue 12.8% 18.0% 7.9% 13.3% 14.2%
(k4%) (+2%) (kO.5%) (?0.5%) (+2%)
*Feces collected between 7 h and 24 h after administration of label; values are expressed as percentage of the total amount of label recovered. In the same column, the differences between GF and S, + S, + & rats, between CRF and CRF + S, + S,_+ R., rats and between CV rats and all the other groups were statistically significant at a level of P < 0.01, except for the residues.
rats (82%) than in S, + Sr + R, rats (77%). Of the total amount excreted, GF rats excreted only 4% via the urine as compared to 8% in S, + S2 + R, rats. The excretion, as can be deduced from the cumulative excretion curves (Fig. 1) and the 11/2 (38 h in S, + S, + R, rats and 32 h in GF rats) (Table I), was clearly slower in the S, + S, + R, rats. Analysis of the fecal label (Table 2) excreted between 7 and 24 h showed that free estrogen had increased from 2.8% in GF rats to 25% in S, + S, + R, rats. Estrogen-sulfate represented 73.8% of radioactivity in GF rats but only 24.6% in S, + S, + R, rats; estrogen-glucuronide accounted for 10.6% in GF rats and for 33% in S, + S, + R, rats. To determine the influence of microbial desulfation in rats with a reduced cecal volume, the sum of fecal and urinary excretion was compared in CRF and CRF + S, + SZ + R, rats. The recovery of administered label was 83% in CRF rats but only 77% in CRF + S, + S, + R, rats. Urinary excretion represented 3% of the total label excreted in CRF rats and 9% in CRF + S, + SZ + R, rats. Label excretion was significantly slower in CRF + S, + S, + R, rats than in CRF rats as illustrated by the cumulative excretion curves (Fig. 1). The t l/2 values were 27 h in CRF + S, + S, + R, rats versus 13 h in CRF rats. Analysis of the fecal estrogen-fractions showed that free estrogen increased from 5% in CRF rats to 39% in CRF + S, + S, + R, rats; estrogen-sulfate decreased from 79.5% in CRF rats to 19.5% in CRF + S, + S, + R, rats; estrogen-glucuronide represented 27.8 and 7.4% in CRF + S, + S, + R, and CRF rats respectively. Statistical analysis showed that the differences between GF and S, + S, + R, rats and between CRF and CRF + S, + S2 + R, rats were highly significant (P < 0.01).
DISCUSSION
Sim et a[.[31 and Back et aZ.[4] have already established the important role of the cecal microflora in the enterohepatic circulation of estrone-sulfate by studying the gastrointestinal absorption of estrone-sulfate in both GF and antibiotic-treated CV rats. In the present investigation, we studied the fecal and urinary excretion of estrone-sulfate not only in GF and CV
rats but also in gnotobiotic rats selectively associated with estrone-desulfating and/or cecal volume reducing bacteria. Thus, we avoided the potential errors associated with comparing results in GF rats or antibiotic-treated CV rats with normal CV rats. In doing so, we were able to define quantitatively the impact of microbial desulfation on the excretion of estrone-sulfate. From the results of this study, it is clear that microbial desulfation of estrone-sulfate leads to a substantial delay in the excretion of orally administered estrone-sulfate in gnotobiotic rats with an enlarged cecum and abnormal intestinal physiology as well as in rats with a partially normalized cecal volume. Microbial desulfation caused a greater delay in the excretion of labeled products in rats with a cecal partially normalized volume (CRF + S, + S, + R, rats) than in rats with a fully enlarged cecum (S, + S, + R, rats). Partial normalization by the CRF flora of the gastro-intestinal physiology, of which the reduced cecal volume is only one aspect, is most likely the cause of this difference. Rats with a partially reduced cecal volume but no estrone-desulfating flora had a substantially higher excretion rate for estrone-sulfate than both GF and CV rats. This rapid excretion was not due to intestinal metabolism of estrone-sulfate since analysis of the fecal label showed that CRF rats and GF rats had a nearly identical pattern of estrone-sulfate metabolites. The total label excretion as well as its urinary component were similar in GF and CRF rats. Moreover, the CRF microorganisms did not metabolize estrone or estrone-sulfate in vitro. We therefore assume that the rapid excretion of labeled products in CRF rats was solely due to the more rapid intestinal transit rate and/or the reduced cecal volume. A relationship between cecal volume and intestinal transit rate has been proposed by some authors [l I]. Although the cecal volume was only partially reduced in the CRF rats, their rate of excretion of label was more rapid than that of CV rats, presumably because the normal intestinal flora of CV rats also desulfated estrone-sulfate and thus promoted the enterohepatic circulation of the compound. Association of germfree rats with both the cecal volume reducing flora and the estrone-sulfate desulfating flora slowed the excretion of orally administered estrone-sulfate down to a rate somewhat below that found in CV rats.
Influence of intestinal microbial desulfation Association of GF as well as CRF rats with the estrone-desulfating flora increased the fraction of label recovered in the urine by 100 and 200%, respectively. CV rats, where desulfation was most intensive as reflected by the fractionation of the fecal label, excreted 24% of the total label via the urine. The differences in excretion rates between the different groups of rats corresponded to equal changes in the fecal estrone-sulfate metabolites. Association of GF or CRF rats with the estrone desulfating strains increased the percentage of labeled fecal free estrone. It should be noted that fecal estrone-glucuronide also increased when GF or CRF rats were associated with desulfating floras. A possible explanation can be as follows; in the intestinal mucosa and in the liver, estrone is preferentially conjugated to glucuronide [4]. Labeled estrone appearing in the bile after enterohepatic circulation is thus mainly present as estrone-glucuronide. In rats associated with desulfating floras, the labeled estronesulfate is quickly desulfated and thus more rapidly reabsorbed into the intestinal mucosa and enterohepatic circulation than in rats without the desulfating flora. After 1 day we can therefore expect to find more labeled estrone-glucuronide in the bile and consequently in the feces in the S, + S, + & and the CRF + S, + S, + R, rats than in the GF and CFR rats. In the present study we found only very little estrone-glucuronide in the feces of CV rats, supporting the notion that, although the intestinal wall possesses b-glucuronidase activity, the main source of intestinal ,Q-glucuronidase is of bacterial origin [17]. CV rats also were found to excrete the highest amount of free estrogen (81.5%) and the lowest part of estrogen-sulfate (2.7%). This indicates that desulfation in and S, + S, + R, CRF + S, + S, + R, rats was still incomplete and suggests that in CV rats many steroid desulfating microorganisms might be cooperating. Our results are in accordance with those found for bile acid sulfates in gnotobiotic rats selectively associated with bile acid desulfating strains [18]. They confirm that the intestinal metabolism, and in particular intestinal desulfation, plays an important role in the pharmacokinetics of estrone in rat, and probably man.
L. Verlooy
for skillful
technical
J. Mertens, G. De Pauw and assistance.
REFERENCES I. Eriksson
H. and Gustafsson J.-A.: Steroids in germfree and conventional rats. Sulpho- and glucuronohydrolase activities of caecal contents from conventional rats. Eur. J. Biochem. 13 (1970) 198-202.
239
S. M., Sim S. M., Back D. J. and Eyssen 2. Huijghebaert H. J.: Distribution of estrone sulfatase activity in the intestine of germfree and conventional rats. J. steroid Biochem. 20 (1984) 1175-l 179. S., Back D. J. and Eyssen H. 3. Sim S. M., Huijghebaert J.: Gastrointestinal absorption of estrone sulfate in germfree and conventional rats. J. steroid Biochem. 18 (1983) 499-503. C. R., May S. A. and Rowe P. 4. Back D. J., Chapman H.: Absorption of oestrone sulphate from the gastrointestinal tract of the rat. J. steroid Biochem. 14 (1981) 347-356. 5. Adlercreutz H., Martin F., Lehtinen T., Tikkanen M. J. and Pulkkinen M. 0.: Effect of ampicillin administration on plasma conjugated and unconjugated estrogen and progesterone levels in pregnancy. Am. J. Obstet. Gynec. 128 (1977) 266271. H.: 6. Tikkanen M. J., Pulkkinen M. 0. and Adlercreutz Effect of ampicillin treatment on the urinary excretion of estriol conjugates in pregnancy. J. steroid Biochem. 4 (1973) 43940. with oral contraceptives. I. Dossetor J.: Drug interactions Br. med. J. ii (1975) 467468. with oral 8. Orme M. L’E. and Back D. J.: Therapy contraceptive steroids and antibiotics. J.Antimicrob. Chemother. 20 (1979) 126125. and associ9. Illing H. P. A.: Techniques for microfloral ated metabolic studies in relation to the absorption and enterohepatic circulation of drugs. Xenobiotica 11 (1981) 815-830. H. A. and Bruckner G.: Anomalous lower 10. Gordon bowel function and related phenomena in germ-free animals. In The Germ-Free Animal in Biomedical Research (Edited by M. E. Coates and B. E. Gustafsson). Laboratory Animals Ltd, London (1984) pp. 193-213. J. and Ito T.: Effects 11. Iwai H., Ishihara Y., Yamanaka of bacterial flora on cecal size and transit rate of intestinal contents in mice. Jap. J. exp. Med. 43 (1973) 297-305. 12. Trexler P. C.: The use of plastics in the design of isolator systems. Ann. N. Y. Acad. Sri. 78 (1959) 29-36. 13. Lindstedt S. and Norman A.: The turnover of bile acids in the rat. Bile acids and steroids 39. Acfaphysiol. stand. 38 (1956) 121-128. S. M., Mertens J. A. and Eyssen H. J.: 14. Huijghebaert Isolation of a bile salt sulfatase-producing Clostridium strain from rat intestinal microflora. Appl. enuir. Microbiol. 43 (1982) 1855192. G. and Eyssen H.: Isolation of 15. Robben J., Parmentier a rat intestinal CIostridium strain producing Sa- and Sfi-bile salt 3a-sulfatase activity. Appl. enuir. Microbial. 51 (1986) 32-38. 16. Otten M. H., de Herder W. W., Hazenberg M. P., van de Boom M. and Hennemann G.: Iodothyronine sulfatase activity of two anaerobic bacterial strains from rat intestinal microflora. FEMS Microbial. Len. 18 (1983) 75-77. P. H., Marsh H. H., Wei17. Williams J. R., Grantham sburger J. H. and Weisburger E. K.: Participation of liver fractions and of intestinal bacteria in the metabolism of N-hydroxy-N-2-fluorenylacetamide in the rat. Biochem. Pharmac. 19 (1970) 173-188. G., Huijghebaert 18. Eyssen H., van Eldere J., Parmentier S. and Mertens J.: Influence of microbial bile salt desulfation upon the fecal excretion of bile salts in gnotobiotic rats. J. steroid Biochem. 22 (1985) 547-554.
Vol. 26, No. 2, pp. 235-239,
0022-4731/87
1987
Copyright 0
Printed in Great Britain. All rights reserved
$3.00 + 0.00
1987 Pergamon Journals Ltd
INFLUENCE OF AN ESTRONE-DESULFATING INTESTINAL FLORA ON THE ENTEROHEPATIC CIRCULATION OF ESTRONE-SULFATE IN RATS J. VAN ELDERE*, G. PARMENTIER,J. ROBBEN and H. EY~~EN The Rega Institute, University of Leuven, Minderbroedersstraat 10, B-3000 Leuven, (Received
22 May
Belgium
1986)
Summary-The fecal and urinary excretion of orally administered [4-“‘Clestrone-3-sulfate was studied in germfree (GF) rats, conventional (CV) rats and gnotobiotic rats selectively associated with estronedesulfating and/or cecal-volume reducing microorganisms. The time required to excrete 50% of the total label recovered (t I/2) was 22 h in CV rats vs 32 h in GF rats. Gnotobiotic rats selectively associated with a cecal volume-reducing flora (CRF rats) excreted the label even faster (t l/2 = 13 h) than CV rats. Association of GF rats as well as CRF rats with estrone-desulfating microorganisms (termed S, + S, + R rats and CRF + S, + S, + & rats, respectively) led to a slower excretion of labeled products (t I/2 = 38 h in S, + S, + & rats and t l/2 = 27 h in CFR + S, + S, + & rats). Intestinal microbial desulfation also increased the relative part of the urinary excretion from 4% in GF rats to 8% in S, + S, + & rats and from 3% in CRF rats to 9% in CFR + S, + S, + I& rats. We conclude that intestinal microbial desulfation enhances the enterohepatic circulation of orally administered estrone-3-sulfate.
INTRODUCTION The use of estrogens in clinical situations has led to a considerable interest in the regulation of the enterohepatic circulation and excretion of these substances. Conjugation with sulfate or glucuronic acid in the liver enhances the water-solubility and the fecal excretion of estrogens. During their passage through the intestines, these estrogen conjugates can however be hydrolysed. Deglucuronidation can be of mammalian or microbial origin; desulfation is probably solely of microbial origin [ 1,2]. Experiments with germfree (GF) rats have shown that estrone-3-sulfate can be reabsorbed intact from the small intestine but at a much lower rate than estrone [3]. In the cecum, reabsorption was only possible after desulfation of estrone-sulfate [3]. Conventional (CV) rats pretreated with antibiotics reabsorbed less estrone-sulfate, probably because of the elimination of estrone desulfating bacteria [4]. Administration of antibiotics to pregnant women has been shown to decrease plasma and urinary estriol levels [S, 61. Similarly, there have been reports of pregnancy in women on oral contraceptives who were also taking antibiotics [7, 81. The use of antimicrobial drugs in rat studies on the enterohepatic circulation of estrogen-conjugates entails however certain limitations. The drug itself, apart from its antimicrobial activity, may in some way influence the absorption and metabolism of the estrogens [9] and it is very difficult to define what part
*J. van Eldere is a Research National Fund for Scientific
Assistant Research
of the Belgian (NFWO). 235
of the normal flora has been eliminated. The use of GF animals is also subject to noncontrolled variability. GF rodents, such as mice and rats, because of the absence of the intestinal flora, present with several physiological anomalies that form potential sources of error [lo]. Lack of microbial degradation leads to the accumulation of mutinous material in the lower bowel, causing retention and even secretion of water in the lumen of the bowel. In addition, the reduced intestinal muscle tone and motility lead to dilatation of the cecum and a slower intestinal transit. The comparison of GF and CV rats is, moreover, hindered by the fact that the intestinal flora performs also other reactions besides deconjugation, such as oxidation-reduction reactions, that may influence the excretion of the estrogens. By isolating pure cultures of estrone-desulfating bacteria from rat cecal material, and associating GF rats with these bacteria, we were able to define quantitatively the influence of estrone-desulfation in comparable situations. Because a close relationship has been demonstrated between the volume of the cecum and the rate of the intestinal transit in mice [l l] two other groups of gnotobiotic rats were associated with a defined flora, termed CRF flora, consisting of four strains of unidentified anaerobic intestinal bacteria that were able to reduce the cecal volume and the intestinal transit time of GF rats by 60%. One of these two groups of rats was, in addition, associated with the estrone-desulfating strains. Estrone-3-sulfate was administered orally to prevent partial desulfation in the liver prior to the first passage through the intestines.
236
Animals
J.
VAN ELDERE et al.
and diet 1 ml water and 5 ml of I M NadH. Thin-layer chromatography (TLC) on Silicagel 60 (E. Merck, Female inbred Fisher rats, approx. 3 months old at Darmstadt, W. Germany) in ethylacetate-butanolthe start of the experiments were used. GF, gnoacetic acid-water (8:6:3:3; v/v) showed that no free tobiotic and CV animals were kept in Trexler’s estrone remained. The pyridine was evaporated and flexible-film plastic isolators (Standard Safety Equipthe residue dissolved in 80 ml of water. This solution ment, Palatine, Ill., U.S.A.) [12] and were fed a was desalted on Baker octadecyl extraction columns. sterilized commercial diet (N.V. Trouw, Gent, Belgium) and water ad l~b~turn. Each group contained 5 The [4-‘4C]estrone-sulfate was eluted with methanol and more than 99% of the [4-‘4C]estrone was recovrats all kept individually in cages. Cages had double ered as sulfate-ester. mesh-wire bottoms and absorbing paper on the bottom to prevent coprophagy and facilitate collection Quantifcation qf label excreted in .feces and urine of feces and urine. At the start of the experiment, one pCi of labeled Feces were collected in 50 ml Falcon tubes (Becton [4-t4C]estrone-3-sulfate was administered intraDickinson, Oxnard, CA, U.S.A.) and homogenized gastrically through a catheter. For the first 55 h, feces in pre-cooled water. Steroids were extracted from and urine were collected at intervals of 7 and 17 h 0.5 g of freeze-dried feces with 30 ml of aqueous 80% and, later on, at intervals of 3 days until the cumuethanol for 4 h at 65’C. Three samples of 0.5 ml of lative amount of label excreted reached a plateau. the extract were counted twice in a Packard Tri-carb The cumulative amounts of label excreted were transLiquid Scintillation Counter 2660 (Packard, Ill., formed into excretion curves via -log (1 - p,/p,,,,,) U.S.A.) To correct for quenching, quench-correlation with pu,= total amount of label excreted at time t and curves were stored in the memory of the Liquid &ax = total amount of label recovered, as described Scintillation System and all results were consequently by Lindstedt and Norman[ 131.Fifty percent excretion expressed as disintegrations per minute. of the total amount of label recovered (t l/2) corTo determine the urinary excretion of labeled related with -log{1 -p,/p,.,) = 0.301 on best fit products. absorbing papers were placed on the botcurves. At the end of the experiment, the animals tom of the rat-cages, collected, dried, and then extracwere sacrificed and total body weight and cecum ted with 50 ml of 80% ethanol. From the extract, weight were determined. three aliquots of 1 ml were counted. GF animals were associated with estronedesulfating bacteria (S, + S, + Rs rats), or cecal Fractionation qf the fecal labeled compounds volume-reducing strains (CRF rats), or cecal volumeThe amounts of unconjugated label, glucuronic reducing strains plus estrone-desulfating strains acid conjugated label and sulfoconjugated label were (CRF + S, + Sz + R, rats), by oral inoculation of determined on the aqueous 80% ethanol extract of liquid cultures. This was repeated until cultures of feces excreted between 7 and 24 h after the start of the fecal material showed the presence of the viable experiment. Five ml of extract were evaporated to microorganisms. The strains S,, Sz and R, were dryness and redissolved in 10ml of 0.2 M acetate isolated from the intestinal flora of the rat; strains S, and Sz possessed bile salt-sulfatase activity 114, 151; buffer, pH 5. Three successive extractions with 8 ml diethylether yielded the unconjugated labeled fracstrain R, possessed iodothyronine-sulfatase tion. Glucuronide-esters in the residue were enactivity (161. In addition, all were found to have zymatically deglucuronidated with 3302 Fishmann estrone-sulfatase activity. The cecal volume-reducing Units of ~-glucuronidase (Calbiochem no. 34743). flora consisted of four unidentified strictly anaerobic The liberated label was then extracted 3 times with microorganisms and was isolated from the mouse. 8 ml diethylether. The residue was further used for The CRF strains did not transform estrone or the determination of sulfate-esters. Enzymatic deestrone-sulfate but reduced the cecal size of gnosulfation of these was performed by addition of 198 tobiotic mice and rats by 60-&S%. I.U. of sulfatase-enzyme (Sigma S9626), 1 ml EDTA 1.86% and 1 ml mercapto-ethanol 0.78%. The dePreparation of [4-‘4C]estrone-3-sulfate sulfated estrone was extracted three times with 8 ml [4-“‘Clestrone was obtained from Amersham Interdiethylether. All fractions and the residual nonnational (Amersham, England), chlorosulfonic acid extractable labeled product in the acetate buffer were was purchased from Fluka AG (Buchs, Switzerland) counted in duplicate as described above. and Baker extraction columns for single use (OctaTo determine the relative amounts of estrone, decyl, 6ml) were from J. T. Baker Chemical Co. estradiol and estriol or other estrogens in the feces, (Phillipsburg, N.Y., U.S.A.). One mg of labeled the non-conjugated labeled fractions of CV rat feces, [4-“‘C]estrone (100 pCi/mg) was diluted with 50 mg the sulfate-conjugated label of GF, CRF and of non-labeled estrone, dissolved in 2 ml of dry CRF + S, + S? + R, rat feces and the non-conjugated pyridine and placed into an ice-bath. To this was label of CRF + S, f S, + R, rat feces, were subjected added, dropwise, 0.1 ml of chlorosuifonic acid. After to TLC analysis in isooctane~thylacetate-acetic acid 24 h the reaction was terminated by the addition of (5:5:1; by vol).
Influence of intestinal microbial desulfation
237
RESULTS
TLC analysis of labeled steroids excreted via the feces
TLC analysis of the feces collected between 7 and 24 h after the oral administration of 1 PCi of [4-i4C]estrone-sulfate, showed that more than 95% of the radioactivity corresponded to reference estrogens. The radioactivity of the non-conjugated labeled estrogens in CV rats was distributed between a first peak that corresponded to estrone (4&45%) and a second peak that corresponded to estradiol (55-60%). In all other fractions examined (sulfateester fraction in GF, CRF and CRF + St + S2 + R, rats and unconjugated fraction in CRF + S, + Sz + Ry rats) nearly 100% of the labeled product corresponded to estrone. The other fractions were not analysed because they contained insu~cient label for identification.
S,+S,+R9 GF CRF+S,+S,+RS
CRF
Time
Table
administration
As shown in Table 2, the composition of the fecal estrogen-fraction was fairly similar in GF and CRF rats. In GF rats free estrogen represented 2.8% of the labeled estrogen versus 5.1% in CRF rats; estrogenglucuronide was 10.6% in GF and 7.4% in CRF rats, while estrogen-sulfate was 73.8 and 79.5% in GF and CRF rats respectively. This suggests that, although CRF rats excreted the administered estrone-sulfate rapidly, and even faster than CV rats, their intestinal metabolism of estrone-sulfate did not differ from that of GF rats. In CV rats, however, the fractionation of the fecal label yielded markedly different results; free estrogen represented 81.5%, estrogen-glucuronide 1.7% and estrogen-sulfate 2.7% of the fecal labeled estrogens. Injluence of microbial desulfation on the excretion of labeled estrone-sulfate
Comparison of the fecal plus urinary excretion of label in GF and S, -t Sz + R,-associated rats showed that, after 8 days, the total excretion expressed as a percentage of the total amount administered was only slightly but still signifi~ntly (P c 0.01) higher in GF
I. Fecal and urinary excretion of label after oral administration of [4-‘4C]estrone-sulfate to enotobiotic and conventional rats* Total r112t
rats S, t S, + R, rats CRF rats CRF + S, + S2 + R, rats CV rats GF
32h”(iZh) 38hd(lt:3h) 13h”(s_7h) 27h’(f3h) 22h”(k3h)
recoveryt
82%b(k2%) 77%b(+2%) 83%’ (14%) 71°,Gc(+ 6%) 71%b.Cff4%)
Urinary excretionlj
4%d ( + I .O%) 8%d (10.8%) 3%‘(&0.8%) 9%C( f I .5%) 24%d.C(i 3.0%)
a, b, c, d, e: All values in the same column that were statistically different (P < 0.01) have the same alphabetical superscript. *Three-month old rats; 5 animals per group. tThe r1/2 value was calculated from the “best-fit” curve after transformation of the cumulative excretion-curve according to -lo& 1 L r,/&*,) [see ref. 13). $Percentage of the amount administe~d; & one standard deviation. @Percentage of total amount recovered in feces + urine: & one standard deviation. sa
%‘?.--E
1hr 1
Fig. 1. Semilogarithmic plot of the cumulative excretion curves after administration of labeled estrone-sulfate. Transformation according to -log(l - p,/p,,,) [see ref. 131. Fifty percent excretion ([l/2) is reached at -log/l -a/ p,,,) = 0.301. pI: total amount of label excreted at time t. p_: total amount of label recovered.
Fecal and urinary excretion of label in GF, CV and CRF rats
In rats selectively associated with the CRF flora the cecal volume was substantially decreased as compared to GF rats; the cecum-weight as a percentage of total body-weight decreased from 13.5% in GF rats to 5.7% in CRF rats. Association of these rats with the desulfating strains S, + S, + b did not significantly change the cecal weight (11.5% of total body weight in S, + S, + Rq rats and 5.8% in CRF + S, + S, + R, rats). In CV rats the cecal weight represented 2.3% of total body weight. In CRF rats, the total amount of label recovered in feces plus urine after 8 days was 83% of the amount administered, nearly equal to the 82% found in GF rats, but significantly (P < 0.01) different from the 71% found in CV rats. The urinary excretion, expressed as percentage of the total fecal plus urinary excretion of label, amounted to 3% in CRF and 4% in GF rats versus 24% in CV rats. The excretion of labeled steroids in feces and urine of CRF rats was significantly faster than in GF rats and in CV rats, as deduced both from the cumulative excretion curves (Fig. 1) and from the t l/2 values (Table I). Of the total label recovered, 50% (= t l/2) had been excreted by GF rats after 32 h and by CV rats after 22 h; in CRF rats the t l/2 was only 13 h.
after
238
J. Table
2.
Percentages
of
the
VAN ELDERE et al.
different types of labeled fecal estrogen conjugates after [4-‘“Clestrone-sulfate to gnotobiotic and conventional rats* Free estrogen
GF rats S, + S, + R, rats CRF rats CRF + S, + S, + R, rats cv rats
2.8% 24.7% 5.1% 39.4% 81.5%
(*o.s%) (+3%) (kO.5%) (f 5%) (&3%)
Estrogen glucuronide
Estrogen sulfate
10.6%(?2%) 32.8% (+ I I %) 7.4% ( * I %) 27.8% (k 7%) 1.7% (kO.3%)
73.8%(+5%) 24.6% (+9%) 79.5%(kl%) 19.5% (f 10%) 2.7% (+0.5%)
oral
administration
of
Residue 12.8% 18.0% 7.9% 13.3% 14.2%
(k4%) (+2%) (kO.5%) (?0.5%) (+2%)
*Feces collected between 7 h and 24 h after administration of label; values are expressed as percentage of the total amount of label recovered. In the same column, the differences between GF and S, + S, + & rats, between CRF and CRF + S, + S,_+ R., rats and between CV rats and all the other groups were statistically significant at a level of P < 0.01, except for the residues.
rats (82%) than in S, + Sr + R, rats (77%). Of the total amount excreted, GF rats excreted only 4% via the urine as compared to 8% in S, + S2 + R, rats. The excretion, as can be deduced from the cumulative excretion curves (Fig. 1) and the 11/2 (38 h in S, + S, + R, rats and 32 h in GF rats) (Table I), was clearly slower in the S, + S, + R, rats. Analysis of the fecal label (Table 2) excreted between 7 and 24 h showed that free estrogen had increased from 2.8% in GF rats to 25% in S, + S, + R, rats. Estrogen-sulfate represented 73.8% of radioactivity in GF rats but only 24.6% in S, + S, + R, rats; estrogen-glucuronide accounted for 10.6% in GF rats and for 33% in S, + S, + R, rats. To determine the influence of microbial desulfation in rats with a reduced cecal volume, the sum of fecal and urinary excretion was compared in CRF and CRF + S, + SZ + R, rats. The recovery of administered label was 83% in CRF rats but only 77% in CRF + S, + S, + R, rats. Urinary excretion represented 3% of the total label excreted in CRF rats and 9% in CRF + S, + SZ + R, rats. Label excretion was significantly slower in CRF + S, + S, + R, rats than in CRF rats as illustrated by the cumulative excretion curves (Fig. 1). The t l/2 values were 27 h in CRF + S, + S, + R, rats versus 13 h in CRF rats. Analysis of the fecal estrogen-fractions showed that free estrogen increased from 5% in CRF rats to 39% in CRF + S, + S, + R, rats; estrogen-sulfate decreased from 79.5% in CRF rats to 19.5% in CRF + S, + S, + R, rats; estrogen-glucuronide represented 27.8 and 7.4% in CRF + S, + S, + R, and CRF rats respectively. Statistical analysis showed that the differences between GF and S, + S, + R, rats and between CRF and CRF + S, + S2 + R, rats were highly significant (P < 0.01).
DISCUSSION
Sim et a[.[31 and Back et aZ.[4] have already established the important role of the cecal microflora in the enterohepatic circulation of estrone-sulfate by studying the gastrointestinal absorption of estrone-sulfate in both GF and antibiotic-treated CV rats. In the present investigation, we studied the fecal and urinary excretion of estrone-sulfate not only in GF and CV
rats but also in gnotobiotic rats selectively associated with estrone-desulfating and/or cecal volume reducing bacteria. Thus, we avoided the potential errors associated with comparing results in GF rats or antibiotic-treated CV rats with normal CV rats. In doing so, we were able to define quantitatively the impact of microbial desulfation on the excretion of estrone-sulfate. From the results of this study, it is clear that microbial desulfation of estrone-sulfate leads to a substantial delay in the excretion of orally administered estrone-sulfate in gnotobiotic rats with an enlarged cecum and abnormal intestinal physiology as well as in rats with a partially normalized cecal volume. Microbial desulfation caused a greater delay in the excretion of labeled products in rats with a cecal partially normalized volume (CRF + S, + S, + R, rats) than in rats with a fully enlarged cecum (S, + S, + R, rats). Partial normalization by the CRF flora of the gastro-intestinal physiology, of which the reduced cecal volume is only one aspect, is most likely the cause of this difference. Rats with a partially reduced cecal volume but no estrone-desulfating flora had a substantially higher excretion rate for estrone-sulfate than both GF and CV rats. This rapid excretion was not due to intestinal metabolism of estrone-sulfate since analysis of the fecal label showed that CRF rats and GF rats had a nearly identical pattern of estrone-sulfate metabolites. The total label excretion as well as its urinary component were similar in GF and CRF rats. Moreover, the CRF microorganisms did not metabolize estrone or estrone-sulfate in vitro. We therefore assume that the rapid excretion of labeled products in CRF rats was solely due to the more rapid intestinal transit rate and/or the reduced cecal volume. A relationship between cecal volume and intestinal transit rate has been proposed by some authors [l I]. Although the cecal volume was only partially reduced in the CRF rats, their rate of excretion of label was more rapid than that of CV rats, presumably because the normal intestinal flora of CV rats also desulfated estrone-sulfate and thus promoted the enterohepatic circulation of the compound. Association of germfree rats with both the cecal volume reducing flora and the estrone-sulfate desulfating flora slowed the excretion of orally administered estrone-sulfate down to a rate somewhat below that found in CV rats.
Influence of intestinal microbial desulfation Association of GF as well as CRF rats with the estrone-desulfating flora increased the fraction of label recovered in the urine by 100 and 200%, respectively. CV rats, where desulfation was most intensive as reflected by the fractionation of the fecal label, excreted 24% of the total label via the urine. The differences in excretion rates between the different groups of rats corresponded to equal changes in the fecal estrone-sulfate metabolites. Association of GF or CRF rats with the estrone desulfating strains increased the percentage of labeled fecal free estrone. It should be noted that fecal estrone-glucuronide also increased when GF or CRF rats were associated with desulfating floras. A possible explanation can be as follows; in the intestinal mucosa and in the liver, estrone is preferentially conjugated to glucuronide [4]. Labeled estrone appearing in the bile after enterohepatic circulation is thus mainly present as estrone-glucuronide. In rats associated with desulfating floras, the labeled estronesulfate is quickly desulfated and thus more rapidly reabsorbed into the intestinal mucosa and enterohepatic circulation than in rats without the desulfating flora. After 1 day we can therefore expect to find more labeled estrone-glucuronide in the bile and consequently in the feces in the S, + S, + & and the CRF + S, + S, + R, rats than in the GF and CFR rats. In the present study we found only very little estrone-glucuronide in the feces of CV rats, supporting the notion that, although the intestinal wall possesses b-glucuronidase activity, the main source of intestinal ,Q-glucuronidase is of bacterial origin [17]. CV rats also were found to excrete the highest amount of free estrogen (81.5%) and the lowest part of estrogen-sulfate (2.7%). This indicates that desulfation in and S, + S, + R, CRF + S, + S, + R, rats was still incomplete and suggests that in CV rats many steroid desulfating microorganisms might be cooperating. Our results are in accordance with those found for bile acid sulfates in gnotobiotic rats selectively associated with bile acid desulfating strains [18]. They confirm that the intestinal metabolism, and in particular intestinal desulfation, plays an important role in the pharmacokinetics of estrone in rat, and probably man.
L. Verlooy
for skillful
technical
J. Mertens, G. De Pauw and assistance.
REFERENCES I. Eriksson
H. and Gustafsson J.-A.: Steroids in germfree and conventional rats. Sulpho- and glucuronohydrolase activities of caecal contents from conventional rats. Eur. J. Biochem. 13 (1970) 198-202.
239
S. M., Sim S. M., Back D. J. and Eyssen 2. Huijghebaert H. J.: Distribution of estrone sulfatase activity in the intestine of germfree and conventional rats. J. steroid Biochem. 20 (1984) 1175-l 179. S., Back D. J. and Eyssen H. 3. Sim S. M., Huijghebaert J.: Gastrointestinal absorption of estrone sulfate in germfree and conventional rats. J. steroid Biochem. 18 (1983) 499-503. C. R., May S. A. and Rowe P. 4. Back D. J., Chapman H.: Absorption of oestrone sulphate from the gastrointestinal tract of the rat. J. steroid Biochem. 14 (1981) 347-356. 5. Adlercreutz H., Martin F., Lehtinen T., Tikkanen M. J. and Pulkkinen M. 0.: Effect of ampicillin administration on plasma conjugated and unconjugated estrogen and progesterone levels in pregnancy. Am. J. Obstet. Gynec. 128 (1977) 266271. H.: 6. Tikkanen M. J., Pulkkinen M. 0. and Adlercreutz Effect of ampicillin treatment on the urinary excretion of estriol conjugates in pregnancy. J. steroid Biochem. 4 (1973) 43940. with oral contraceptives. I. Dossetor J.: Drug interactions Br. med. J. ii (1975) 467468. with oral 8. Orme M. L’E. and Back D. J.: Therapy contraceptive steroids and antibiotics. J.Antimicrob. Chemother. 20 (1979) 126125. and associ9. Illing H. P. A.: Techniques for microfloral ated metabolic studies in relation to the absorption and enterohepatic circulation of drugs. Xenobiotica 11 (1981) 815-830. H. A. and Bruckner G.: Anomalous lower 10. Gordon bowel function and related phenomena in germ-free animals. In The Germ-Free Animal in Biomedical Research (Edited by M. E. Coates and B. E. Gustafsson). Laboratory Animals Ltd, London (1984) pp. 193-213. J. and Ito T.: Effects 11. Iwai H., Ishihara Y., Yamanaka of bacterial flora on cecal size and transit rate of intestinal contents in mice. Jap. J. exp. Med. 43 (1973) 297-305. 12. Trexler P. C.: The use of plastics in the design of isolator systems. Ann. N. Y. Acad. Sri. 78 (1959) 29-36. 13. Lindstedt S. and Norman A.: The turnover of bile acids in the rat. Bile acids and steroids 39. Acfaphysiol. stand. 38 (1956) 121-128. S. M., Mertens J. A. and Eyssen H. J.: 14. Huijghebaert Isolation of a bile salt sulfatase-producing Clostridium strain from rat intestinal microflora. Appl. enuir. Microbiol. 43 (1982) 1855192. G. and Eyssen H.: Isolation of 15. Robben J., Parmentier a rat intestinal CIostridium strain producing Sa- and Sfi-bile salt 3a-sulfatase activity. Appl. enuir. Microbial. 51 (1986) 32-38. 16. Otten M. H., de Herder W. W., Hazenberg M. P., van de Boom M. and Hennemann G.: Iodothyronine sulfatase activity of two anaerobic bacterial strains from rat intestinal microflora. FEMS Microbial. Len. 18 (1983) 75-77. P. H., Marsh H. H., Wei17. Williams J. R., Grantham sburger J. H. and Weisburger E. K.: Participation of liver fractions and of intestinal bacteria in the metabolism of N-hydroxy-N-2-fluorenylacetamide in the rat. Biochem. Pharmac. 19 (1970) 173-188. G., Huijghebaert 18. Eyssen H., van Eldere J., Parmentier S. and Mertens J.: Influence of microbial bile salt desulfation upon the fecal excretion of bile salts in gnotobiotic rats. J. steroid Biochem. 22 (1985) 547-554.