Estriol metabolism in the baboon: analysis of urinary and biliary metabolites

The metabolism of i.v. estriol was investigated in two intact baboons and four with biliary fistulas. Urine and bile samples were collected periodically and the radioactivity extracted by Amberlite-XAD resin. Metabolites were separated and purified by a combination of DEAE-Sephadex chromatography, celite partition, specific enzyme hydrolysis of the conjugates and identification of the aglycones. The excretion and metabolism of estriol in the animals closely resembled those of the human. Intact animals excreted an average of 50% of the radioactivity in the urine during 12 hours and two animals with biliary drainage excreted an average of 40% in the urine and 49% in the bile. When the steroid was injected into the portal vein an average of 11.5% and 84% were excreted in the urine and bile, respectively.In the urine of intact animals, approximately 65.8% of the radioactivity was in the form of E₃-16G; 14.2% as E₃-3G; 13.4% as E₃-3S and 5.1% as E₃-3S-16G. Over 73% of biliary radioactivity from the peripheral injections was made up of E₃-3S-16G and 3.6% as E₃-16G and 8.3% as 3-sulfate. In the urine,however, 57% of the label was made up of E₃-16G. No radioactive E₃-3G was detected in the bile of any of the animals. Following simultaneous injection of ³H-E₃ peripherally and ¹⁴C-E₃ intraportally, the 3-glucosiduronate excreted in the urine was derived exclusively from the ³H-label. Based on the results obtained, the baboon has been shown to metabolize estriol in the same fashion as the human, with E³-3S-16G as the predominant biliary metabolite and E³-16G as the major urinary metabolite. As in the human, evidence was also found for an enterohepatic circulation of e³ in the baboon, 16-glucuronidation in the kidney, and extrahepatic (enteric?) formation of E³-3G. $\text{In vitro}$ incubation of the baboon liver yielded 94% of the total conjugate as E³-16G without any trace of E³-3G.     

Enterohepatic circulation of N-acetyl-leukotriene E4

N-Acetyl-leukotriene E₄, the end product of leukotriene C₄ metabolism in the mercapturic acid pathway, was rapidly eliminated from the blood circulation into the bile of rats. Part of the N-acethyl-leukotriene E₄ secreted from bile into the intestine undewent enterohepatic circulation. Leukotriene absorption occurred from the small intestine and from the colon. Biliary and urinary excretion within 5.5 h amounted to 15 and 2%, respectively, of the intraduodenally administered N-acetyl- H leukotriene E₄ in animals anesthetized with ketamine. HPLC analyses indicated that 35% of the biliary radioactivity corresponded to unchanged N-acetyl- H leukotriene E₄, while 65% in bile and 100% in urine were polar metabolites. Enterohepatic circulation extends the biological half-life of N-acetyl-leukotriene E₄.     

The metabolism of estriol in rabbits

Rabbits have been shown to excrete 6, 7-³H-estriol, its conjugates and metabolites preponderantly in the bile during the initial 4 hours following the I.V. injection of the labeled steroid. The amount of radioactivity excreted in the urine was $\text{1}\text{3}$ of that in the bile. Since in intact rabbits most of the injected radioactivity of ³H-estriol is excreted in the urine over a period of days (and very little in the feces), it appears that estriol and its conjugates and metabolites are involved in an efficient enterohepatic. circulation. In the bile, the preponderant metabolite of ³H-estriol was the 3-glucosiduronate. Even though the latter constituted a substantial part of the urinary metabolites, other conjugates and metabolites of estriol were present in considerable amounts. It is possible that the latter have resulted from gastro-intestinal and/or renal metabolism. Incubation of rabbit liver with estriol led to 75% conjugation with glucuronic acid in the 3-position.     

The metabolic fate of medroxyprogesterone acetate in the baboon.

The metabolic fate of medroxyprogesterone acetate (6alpha-methyl-17a lpha-acetoxyprogesterone; MAP)was studied in intact baboons and in those with bile fistulas. The steroid moiety was labeled with tritiated hydrogen at positions 1 and 2 and the 17alpha-acetate with carbon-14, thus affording the opportunity to ascertain the loss of the 17-acetoxy group and the fate of both labels. Following the iv administration of labeled MAP only a small percentage (less than 15%) of the administered dose was recovered in the urine in 7 hours in intact baboons, as well as in the urine of baboons with biliary fistulas. Higher amounts of radioa ctivity were excreted in the bile (approximately 25%), amounting to almost double the percentage excreted into the urine. The similarity in the urinary excretion of radioactivity in intact and biliary fistula animals indicates that, even though a substantial biliary excretion of the labeled MAP occurred, the amount involved in an enterohepatic circulation is probably small. Glucosiduronates were the predominant conjugates, both in the urine and bile. The loss of the 17alpha-acetate group appeared to be rather extensive, ranging 30-70% among different co njugated and unconjugated metabolities of MAP. The degree of in vivo hydrolysis of the axial 17alpha-acetate of MAP, though extensive appeared to be of a significantly lesser magnitude than that exhibited toward the equatorial 3beta and 17beta acetate groups of labeled ethynodiol diacetate injected into baboons. The deacetylation of the 17alpha-acetate in MAP was similar to that observed in humans given the drug. Oxygenation of MAP at positions 1 and/or 2 appeared to be rather minimal (less than 5%).     

Cholestasis induced by lithocholate and its glucuronide: their biliary excretion and metabolism

We studied the effects of the infusion of lithocholate and lithocholate-3-sulfate and 3-glucuronide in rats (0.29 mumol/min per 100 g body weight for 40 min) on bile flow, together with their biliary excretion and metabolism. Lithocholate-glucuronide had a higher cholestatic effect than lithocholate, whereas lithocholate-sulfate had almost no effect on bile flow. Lithocholate was mainly converted to taurine or glucuronide conjugates in the bile, serum and liver and hydroxylation of the tauro-conjugate proceeded. Lithocholate-sulfate was almost completely excreted in the bile, mainly as tauro-conjugate. Lithocholate-glucuronide was excreted in bile almost without conjugation, while some taurine conjugation occurred in the serum and liver. These results suggest that the poor biotransformation of lithocholate-glucuronide is related to its higher cholestatic potency than lithocholate.     

Metabolic fate of ethynodiol diacetate in the baboon.

The enterohepatic circulation and metabolism of ethynodiol diacetate (3beta,17beta-diacetoxy-17alpha-ethynyl-estr-4-ene) in baboons were studied following the intravenous injection of this contraceptive steroid labeled with 14C (4-position) and with 3H (in either the 3- or 17-acetoxy moieties). Bile and urine from four baboons with biliary fistulas and urine from four intact baboons were collected for 7 hours. On the average, 40% and 44% of the injected dose were excreted in the bile and urine, respectively. Only 48% was recovered in the urine of intact baboons. Analysis of these excretion rates indicates an insignificant enterohepatic circulation of this compound. The steroid was excreted mostly (over 80%) as a glucosiduronate in urine and bile. Very little excretion of the 3-acetoxy compound was detected in the urine or bile at any time interval. 17-Monoacetoxy compounds, however, were detected both in urine and bile, suggesting a difference in the rate of in vivo hydrolysis of the 17beta- vs. the 3beta-acetate.     

Bile acid-oligodeoxynucleotide conjugates: synthesis and liver excretion in rats

The synthesis of bile acid oligodeoxynucleotide conjugates via the 3-OH group of the bile acids is described. When used in vivo in rats, covalent conjugation of an oligodeoxynucleotide via a linker to cholic acid resulted in an increased biliary excretion of bile acid-oligodeoxynucleotide conjugates compared to unconjugated oligodeoxynucleotides.     

Metabolism of sodium oestrone [35S]sulphate in the guinea pig

Intraperitoneal administration of sodium oestrone [(35)S]sulphate to male and female free-ranging guinea pigs is followed by excretion of most of the radioactivity mainly as inorganic [(35)S]sulphate in the urine within 72h. The remainder of the radioactivity in the urine was found in oestrone [(35)S]sulphate, two unidentified metabolites (A and B) and traces of oestradiol-17beta 3-[(35)S]sulphate. When injected intraperitoneally into animals with bile-duct and bladder cannulae, most of the dose was excreted in the bile. Unchanged oestrone [(35)S]sulphate was the main biliary component excreted in males and females, but the latter also excreted appreciable amounts of oestradiol-17beta 3-[(35)S]sulphate and metabolites A and B. The urine from these animals also contained these metabolites, inorganic [(35)S]sulphate and also oestrone [(35)S]sulphate, but in small amounts. Metabolite A was present only in samples from males. Whole body radioautography pinpointed the liver and kidney as the possible sites of metabolism of the ester. The ester underwent little desulphation in the isolated perfused female guinea-pig liver and in animals in which kidney function had been eliminated, and was excreted unchanged in the bile. These results and the observed low oestrogen sulphatase and arylsulphatase C activities found in guinea-pig liver and kidney support the view that the two enzymes are identical.     

The biochemical basis for the conjugation of bile acids with either glycine or taurine

All animals, except for the placental mammals, conjugate their bile acids exclusively with taurine. However, in certain of the placental mammals, glycine conjugates are also found. The basis for the appearance of glycine conjugation among the placental mammals was investigated. The reaction of choloyl-CoA with glycine and taurine, as catalysed by the soluble fraction from guinea-pig liver, had a high affinity for taurine and a poor affinity for glycine. The predominant synthesis of glycine conjugates in the guinea pig can be related to the fact that guinea-pig liver contains an unusually low concentration of taurine and a high concentration of glycine. Rabbits make exclusively glycine conjugates and their livers also contain low concentrations of taurine. However, the biochemical basis for their glycine conjugation is more straightforward than in the guinea pig in that the soluble fraction from rabbit liver has a high affinity for glycine and a poor affinity for taurine. Alternative-substrate-inhibition studies with glycine and taurine in soluble fractions from guinea-pig and rabbit liver revealed that glycine and taurine were mutually inhibitory. This suggests that there is only one enzyme for glycine and taurine conjugation in these tissues. The soluble fractions from bovine liver and human liver also made both glycine and taurine conjugates and evidence is presented that suggests that there is only one enzyme in these tissues too. Even the rat, which excretes mostly taurine conjugates, could make both glycine and taurine conjugates in vitro. However, in contrast with all of the placental mammals studied, the supernatant fraction from liver of the chicken, and other non-mammals, could not make glycine conjugates even in the presence of very high concentrations of glycine.     

The metabolic fate of chlormadinone acetate in the baboon.

The metabolic fate of chlormadinone acetate (17alpha-acetoxy-6-chloro-4, 6-pregnadiene-3, 20-dione; CAP) was studied in intact and biliary fistula baboons. The steroid was labeled with 3H at position 1 and with 14C at the carboxyl moiety of the 17alpha-acetate, thus affording the opportunity to ascertain the loss of the 17alpha-acetoxy group and the fate of both labels. The averages of the radioactivity excreted, given as percentages of the amounts injected, and the standard deviations were as follows: In the urine of intact animals after 6 hours, 5.7 +/- 0.2% and 5.5 +/- 0.7% of the 3H and 14C were recovered, respectively. After 6 days, there was 17.5% of the 3H and 16.2% of the 14C in the urine plus 15.3% of the 3H and 16.4% of the 14C in the feces. In baboons with biliary fistulas, the total radioactivity excreted was 7.8 +/- 0.7% of the 3H and 11.6% of the 14C in the urine, and 30.9 +/- 4.4% of the 3H and 30.7% of the 14C in the bile after 6 hours. Glucosiduronates were the predominant conjugates in the urine and bile. The similarity in the urinary excretion of radioactivity in the first 6 hours in intact and biliary fistula animals, the relatively low excretion of radioactivity in the bile and after 6 days in the urine, and the low fecal excretion suggest that the metabolites of CAP are not involved in an extensive enterohepatic circulation in the baboon. Deacetylation of the 17alpha-acetate in CAP was detected in the early collection periods of the urine and bile and constituted a very small percentage of the injected compound. No significant oxygenation of CAP at position 1 was detected. The metabolism of CAP is discussed and compared to our previously reported data on the metabolism of progesterone, ethynodiol diacetate and medroxyprogesterone acetate and the data on other progestogens reported in the literature. It appears that the excretion of CAP is significantly slower in the baboon than that of the other progestogens. The amounts of glucosiduronates of CAP and/or its metabolites formed in vivo are less than those formed with the other progestogens. Also, the extent of deacetylation of the 17alpha-acetate of CAP is much less than that of the 3beta-acetate of ethynodiol diacetate.     

Comparison in different species of biliary bilirubin-IX alpha conjugates with the activities of hepatic and renal bilirubin-IX alpha-uridine diphosphate glycosyltransferases.

The bilrubin-IXalpha conjugates in bile and the activities of bilirubin-IX alpha--UDP-glycosyltransferases in liver and kidney were determined for ten species of mammals and for the chicken. 1. In the mammalian species, bilirubin-IX alpha glucuronide was the predominant bile pigment. Excretion of neutral glycosides was unimportant, except in the cat, the mouse, the rabbit and the dog, where glucose and xylose represented 12--41% of total conjugating groups bound to bilirubin-IX alpha. In chicken bile, glucoside and glucuronide conjugates were of equal importance. They probably represent only a small fraction of the total bile pigment. 2. The transferase activities in liver showed pronounced species variation. This was also apparent with regard to activation by digitonin, pH optimum and relative activities of transferases acting on either UDP-glucuronic acid or neutral UDP-sugars. 3. Man, the dog, the cat and the rat excrete bilirubin-IX alpha largely as diconjugated derivatives. In general, diconjugated bilirubin-IX alpha could also be synthesized in vitro with liver homogenate, bilirubin-IX alpha and UDP-sugar. In contrast, for the other species examined, bilirubin pigments consisted predominantly of monoconjugated bilirubin-IX alpha. Synthesis in vitro with UDP-glucuronic acid, UDP-glucose or UDP-xylose as the sugar donor led exclusively to the formation of monoconjugated bilirubin-IX alpha. 4. The transferase activities in the kidney were restricted to the cortex and were important only for the rat and the dog. No activity at all could be detected for several species, including man. 5. Comparison of the transferase activities in liver with reported values of the maximal rate of excretion in bile suggests a close linkage between conjugation and biliary secretion of bilirubin-IX alpha.     

Bile formation and hepatic plasma membrane composition in guinea-pigs and rats

We compared bile formation, and biliary and liver plasma membrane composition in guinea-pigs and rats in an attempt to explain the observation that the bile flow rate and the bile acid independent fraction of bile flow (BAIF) in guinea-pigs is about five to seven times higher than in rats. Analysis of electrolytes in bile showed that bicarbonate was significantly [acid] higher in guinea-pigs while Cl⁻, phosphate and Ca²⁺ were markedly lower than in rats. High bile independent secretion in guinea-pigs was associated with a significantly lower concentration of total bile acid, phospholipid and cholesterol than in rats. Bile acid distribution studies showed that glycine conjugated chenodeoxycholate and ketolithocholate were the main bile acids in guinea-pigs, while taurine conjugated cholate and muricholate were the predominant bile acids in rats. Total fatty acid analysis of bile indicated that in rats the major fatty acids were palmitic acid (C16:0) and linoleic acid (C18:2, n-6). In guinea-pigs, the contribution of these fatty acids was lower than in rats and compensated with a significantly higher percentage of oleic acid (C18:1, n-9). Concentrations of anionic polypeptide fraction (APF), an acidic calcium binding apoprotein closely associated with biliary phospholipid and cholesterol secretion was also significantly lower in guinea-pigs. Canalicular plasma membrane analysis showed that as compared with rats, specific activities of Na⁺,K⁺ ATPase, and cholesterol and phospholipid content were markedly lower in guinea-pigs. Total fatty acid analysis of the membrane revealed that palmitic acid (C16:0), stearic acid (C18:0) and linoleic acid (C18:2, n-6) were the predominant fatty acids in guinea-pigs, while palmitic acid (C16:0), stearic acid (C18:0) and arachidonic acid (C20:4, n-6) were the most important in rats. Thus, high bile flow rate and BAIF in the guinea-pig may be attributed to the low bile acid concentration (below the critical micellar concentration), secretion of hypercholeretic bile acids (e.g. ketolithocholate) and high bicarbonate output.     

Formation and biliary excretion of glutathione conjugates of bile acids in the rat as shown by liquid chromatography/electrospray ionization-linear ion trap mass spectrometry

Acyl-adenylates and acyl-CoA thioesters of bile acids (BAs) are reactive acyl-linked metabolites that have been shown to undergo transacylation-type reactions with the thiol group of glutathione (GSH), leading to the formation of thioester-linked GSH conjugates. In the current study, we examined the transformation of cholyl-adenylate (CA-AMP) and cholyl-coenzyme A thioester (CA-CoA) into a cholyl-S-acyl GSH (CA-GSH) conjugate by rat hepatic glutathione S-transferase (GST). The reaction product was analyzed by liquid chromatography (LC)/electrospray ionization (ESI)-linear ion trap mass spectrometry (MS). The GST-catalyzed formation of CA-GSH occurred with both CA-AMP and CA-CoA. Ursodeoxycholic acid, lithocholic acid, and 2,2,4,4-²H₄-labeled lithocholic acid were administered orally to biliary fistula rats, and their corresponding GSH conjugates were identified in bile by LC/ESI-MS². These in vitro and in vivo studies confirm a new mode of BA conjugation in which BAs are transformed into their GSH conjugates via their acyl-linked intermediary metabolites by the catalytic action of GST in the liver, and the GSH conjugates are then excreted into the bile.     

Direct quantitative estimation of several iodothyronines in rat bile by radioimmunoassay and basal data on their biliary excretion

A method was developed for the hydrolysis of conjugated iodothyronines in bile with the aid of β-glucuronidase / arylsulfatase and for subsequent direct estimation of total and free iodothyronines with the aid of specific radioimmunoassay. The amount of conjugated fraction could then be calculated from the difference. Thus, basal biliary excretion of several iodothyronines was measured in 31 normal, fed rats in which the bile duct was drained with polyethylene tubing under pentobarbiturate anesthesia and the bile was collected for 2 h. The free fraction of thyroxine, 3,5,3′-triiodothyronine and 3,3′-diiodothyronine was approx. 30% of total content, while that of 3,3′,5′-triiodothyronine and 3,5-diiodothyronine was approx. 20% and that of 3′,5′-diiodothyronine was less than 10%. This suggests some considerable diffrences in the conjugation of individual iodothyronines was about 3–8 times higher than in plasma. This shows close interrelations between the iodothyronine deiodinating pathway in liver cells in vivo and the spectrum of iodothyronine in bile. The average ratio of T₃/rT₃ as found in bile was about 4.     

Prostacyclin and prostagladin synthesis in isolated brain capillaries

The synthesis of prostacyclin and prostaglandins was examined in isolated blood-free brain capillaries of guinea-pigs and rats using 1-¹⁴C-arachidonic acid as a precursor. The main prostaglandins synthesized by guinea-pig microvessels were prostaglandin D₂ and prostaglandin E₂. Substantially less prostaglandin F_2α or the prostacyclin stable metabolite, 6-oxo-prostaglandin F_1α was synthesized. Rat capillary prostaglandin distribution differed substantially from that of the guinea-pigs although the principle prostaglandin was also PGD₂. Total prostaglandin conversion was greater in guinea-pig capillaries than in the rat.Norepinephrine stimulated the prostaglandin forming capacity of blood free cerebral microvasculature of guinea-pigs. Prostacyclin and prostaglandins could be involved in the activity dependent regulation of regional cerebral blood flow and permeability.     

In vitro liver clearance of tritiated estradiol-17β in the female rat after retrochiasmatic transection and ovariectomy

It was reported in a previous study that serum estradiol-17β (E₂) was elevated in rats after retrochiasmatic transection (FC). Serum E₂ was also higher in estradiol cypionate treated, ovariectomized (ovx) rats that had been subjected to FC than in those that had not. This suggested that increased secretion of E₂ was not the only factor responsible for elevated serum E₂ after FC. To ascertain the contribution of decreased metabolism of E₂ to this response, liver tissue slices were incubated with 3H-estradiol-17β, and the rates of ³h uptake and conversion to water-soluble conjugates were measured.The rate of uptake of ³H by the tissue was not indicative of the rate of conjugate formation. Livers of rats with high serum E₂ exhibited lower rates of ³h uptake than those of rats with low serum E₂. Even so, the formation of ³H conjugates was greater in liver of rats with high serum E₂. As hypothesized, livers of rats with FC formed conjugates at a significantly lower rate than those of similarly treated rats without FC. Thus, FC of an intact rat leads to an increase in serum E₂ by increasing the secretion of E₂, and apparently also by decreasing the rate of E₂ metabolism by the liver.     

beta-Adrenergic receptor induction in HeLa cells: synergistic effect of 5-azacytidine and butyrate

³H-Labeled leukotriene C₃ was efficiently taken up by the isolated, perfused rat liver and excreted into the bile. The isolated, perfused kidney eliminated leukotriene C₃ from the perfusate slower and excreted only a fraction of the radioactivity into the urine. Isolated hepatic, intestinal and renal cells also took up leukotriene C₃, the renal cells being the most effective in accumulating the label. Anthglutin, an inhibitor of γ-glutamyl transferase, decreased the uptake by kidney cells but had no effect on the uptake by the other cell types. In liver cells, the uptake rate was sensitive to temperature and to cellular ATP content. Chromatographic analyses indicated that renal cells metabolized leukotriene C₃ more rapidly than hepatic and intestinal cells. Leukotriene D₃ and E₃ were formed during the incubations with kidney cells, whereas intestinal cells produced mainly more polar metabolites.     

Biliary and urinary excretion of peptide leukotrienes in the domestic pig

The metabolism of leukotriene (LT)C₄ and its major routes of elimination have been studied in four anesthetized domestic pigs administered intravenous [³H]-LTC₄ (0.5 μCi/kg). The kinetic profile of LTC₄ in the blood was followed for 60 min after administration while the biliary and urinary excretion of LTC₄ and its metabolites were determined over a 120 min interval. The total recovery of radioactivity in bile and urine was 45% ± 1 (n = 3) and 18% (n = 2) respectively. Examination of the radioactive metabolites in bile showed LTD₄ (44% of biliary content) and LTE₄ (21% of biliary content) as the major identified lipoxygenase products at t (27 min). The only identified cysteinyl leukotriene observed in the urine was LTE₄ (13% of urinary content). In both bile and urine substantial amount of radioactivity were detected at the solvent front of the reverse phase chromatographic system indicating the presence of additional unidentified metabolites. We suggest that measurement of metabolites using these sampling methods may be useful for the detection and measurement of peptide leukotriene production .     

Excretion of cholate glucuronide

[3-3H]Cholic acid glucuronide [7 alpha,12 alpha-dihydroxy-3 alpha-O-(beta-D-glucopyranosyluronate)-5 beta- cholan-24-oate] was synthesized and administered to rats prepared with either an external biliary fistula or a ligated bile duct. When bile fistula animals were given either microgram or milligram amounts of the glucuronide, biliary secretion of label was rapid and efficient: greater than 90% of the administered label was secreted within 60 min and total recovery of label in bile was 98.6 +/- 1.2%. Studies in which [14C]taurocholate was included in the dose indicated that this bile acid was secreted into bile significantly more rapidly than was the glucuronide. In animals with ligated bile ducts, urinary excretion was the major route of elimination: after 20 hr, 83.4 +/- 9.3% of the administered dose had been excreted in urine. Urinary excretion of cholate glucuronide was significantly more rapid than that of taurocholate. Gas-liquid chromatographic analysis of the methyl ester acetate derivatives of labeled compounds isolated from bile and urine by chromatography established that the bulk (greater than 70%) of the administered material was secreted in bile or excreted in urine as the intact cholate glucuronide. From these results, we conclude that the glucuronidation of cholic acid produces a derivative which is rapidly and effectively cleared from the circulation and excreted.     

Biliary excretion of cyclohexylphenyl 4-[35S]sulphate in the guinea pig.

The metabolic fate and mode of excretion of cyclohexylphenyl 4-[35S]sulphate were studied in the guinea pig. Up to 54.8% of the dose appeared in the bile, the majority as unchanged ester. Substantial amounts of hydroxylated cyclohexylphenyl 4-[35S]sulphate were also excreted in the bile together with minor amounts of the corresponding glucuronic acid conjugate. When isolated guinea-pig livers were perfused with cyclohexylphenyl 4-[35S]sulphate the biliary components were the same as those in the intact animal, although the relative concentration of the hydroxylated derivative was significantly greater. When the hydroxylated derivative was re-injected into guinea pigs it was excreted almost entirely unchanged in the bile. However, in the rat, it was excreted in the bile as a glucuronic acid conjugate. These findings are discussed in relation to studies carried out in the rat [Hearse, Powell, Olavesen & Dodgson (1969) Biochem. Pharmacol. 18, 181--195] and to differences in enzyme activities in rat and guinea-pig liver. The results are also discussed in terms of the molecular-weight threshold for the excretion of anions in guinea-pig bile.