If [carbonyl-¹⁴C]biotin at relatively physiological doses is administered intraperitoneally to rats, ~60% of the dose is recovered in urine within 24 h (Lee et al. 1972, Wang et al. 1996). In contrast, only a trace amount of intraperitoneally administered [¹⁴C]biotin is recovered in feces (Lee et al. 1972, Wang et al. 1996). However, the small rate of fecal excretion after intraperitoneal administration does not necessarily exclude the possibility that biotin is excreted via bile into the intestine, metabolized by enteral microorganisms and efficiently reabsorbed (enterohepatic circulation). For example, the enterohepatic circulation of both folic acid and vitamin B-12 contributes importantly to the homeostasis of these vitamins (El Kholty et al. 1991, Green et al. 1981, Steinberg et al. 1982). Approximately 50% of enterally administered folate circulates through the enterohepatic cycle within 6 h (Steinberg et al. 1982). For vitamin B-12, biliary excretion exceeds the daily requirement for the vitamin, and enteral reabsorption after biliary excretion is important for the maintenance of normal cobalamin status (El Kholty et al. 1991, Green et al. 1981).

We performed the following two experiments to study the biliary excretion of biotin and biotin metabolites: 1 ) in rats, the biliary excretion profiles of radiolabeled biotin and metabolites were compared with the urinary excretion and 2 ) in pigs, the endogenous biliary concentrations of biotin and metabolites were compared with the serum concentrations.

MATERIALS AND METHODS

Crystalline [carbonyl-¹⁴C]biotin (specific radioactivity, 2.00 GBq/mmol) was purchased from Amersham (Arlington Heights, IL). The [¹⁴C]biotin for intravenous injection was dissolved in 154 mmol/L sodium chloride at 0.28 mmol biotin/L. The radiochemical purity of [¹⁴C]biotin was 95.3% as determined by HPLC. A solution of sodium taurocholate (Sigma Chemical, St. Louis, MO) for bile salt replacement was prepared in 154 mmol/L sodium chloride at 100 mmol sodium taurocholate/L.

Table 1. Biliary and urinary excretion of biotin and biotin metabolites in rats1 [View Table]

The animal protocols were approved by the Institutional Animal Care and Use Committees at the University of Texas Health Science Center (rats) and the University of Arkansas for Medical Sciences (pigs).

To monitor for the presence of sulfate conjugates or -glucuronic acid esters of [¹⁴C]biotin and metabolites, bile collections from two rats were pooled and analyzed by HPLC/radiometric flow detection after incubation with either microbial -glucuronidase or with sulfatase (Sigma Chemical). For -glucuronidase (from Escherichia coli), bile samples were incubated at pH 6.8 (4 mmol/L phosphate buffer) and 37°C for 30 min. For sulfatase (from Aerobacter aerogenes), the samples were incubated at pH 7.1 (5 mmol/L Tris buffer) and 37°C for 30 min. As a negative control, bile was incubated with -glucuronidase and sulfatase that had been heated for 10 min at 100°C. As a positive control, the activity of the two enzymes was confirmed qualitatively by their ability to cleave p-nitrophenyl sulfate and phenolphthalein glucuronide.

To monitor for the presence of biotinyl peptides, bile collections from three rats were pooled and incubated with hydrochloric acid (final concentration at 1 mol/L) for 1 h at 100°C. After incubation, samples were analyzed by HPLC/radiometric flow detection. As a control, an aliquot of the pooled bile was analyzed without prior acid treatment. The stability of biotin during acid treatment was confirmed by adding [¹⁴C]biotin to nonradioactive bile and measuring radiopurity with and without acid treatment (purity > 96%).

Bile and serum samples from pigs were deproteinated with the use of a Centriprep concentrator with a molecular mass cutoff of 30 kDa (Amicon, Beverly, MA). The samples were centrifuged at 1500 × g and 15°C in two intervals. After 60 min of centrifugation, the ultrafiltrate was removed, and the retentate was centrifuged for another 30 min; both ultrafiltrates were combined. Unlabeled biotin and biotin metabolites in serum and bile from pigs were analyzed by HPLC/avidin-binding assay as previously described (Mock et al. 1993, Zempleni et al. 1996). Serum and bile from pigs was analyzed for total cholesterol (cholesterol esterase/cholesterol oxidase method, no. 352-20), total bilirubin (caffeine-benzoate-acetate/diazotized sulfanilic acid method, no. 605-C), and glucose (glucose oxidase-method, no. 510-DA) by commercially available test kits (Sigma Chemical).

RESULTS

Two major unknown peaks (referred to as unknowns A and B at retention times 10 and 19 min, respectively) in rat bile and urine were detected. The incubation of a pooled bile sample with either -glucuronidase or sulfatase did not shift or alter either peak (data not shown). Acid treatment did not shift or alter either of the two peaks. We conclude that the two unknown peaks do not contain major amounts of sulfate conjugates, -glucuronic acid esters, or peptides of biotin or biotin metabolites. The retention times of the two unidentified peaks on HPLC did not match the retention times of tetranorbiotin-d,l-sulfoxide, tetranorbiotin, biotin sulfone, bisnorbiotin-d,l-sulfoxide, tetranorbiotin methyl ketone, -hydroxybiotin-l-sulfoxide, ,-dehydrobisnorbiotin, biocytin-d,l-sulfoxide, biocytin, ,-dehydrobiotin or biotin methyl ester (Zempleni et al. 1996). The retention times of the two unidentified peaks on HPLC did not match the retention times of either uracil or urea. The retention time of [¹⁴C]uracil on HPLC was 5.80 min. The retention time of urea on HPLC was 5 min (Zempleni et al. 1996). The identities of the two compounds eluting at 10 and 19 min remain unknown. On a percentage basis, unknowns A and B were much more abundant in bile than in urine, accounting for a total of 33.5% of the biliary radioactivity, compared with 6.1% of urinary radioactivity (Fig. 1).

When compared with the time courses of urinary excretion, the biliary excretion of biotin and its metabolites decreased rapidly after injection. Figure 2 depicts excretion rates of biotin and metabolites expressed as a percentage of the rate during the first collection interval (0-3.0 h) plotted against the midpoints of the times of the collection intervals (Gibaldi and Perrier 1982). In bile, the excretion rates of biotin and biotin metabolites decreased consistently from the first collection interval through the third interval (Fig. 2). No radioactivity was detectable over background in the last interval (6.01-24.0 h). Similarly, the urinary excretion rates of biotin, bisnorbiotin methyl ketone and biotin-d,l-sulfoxide decreased from the first to the third collection interval. The time course of the urinary excretion of some metabolites was quite different from that in bile; the urinary excretion rates of bisnorbiotin and unknowns A and B increased or remained at approximately their peak levels through the second collection interval (Fig. 2).

Table 2. Biotin and biotin metabolites in blood serum and bile from the gall bladder in pigs1 [View Table]

The metabolite profile of biotin was different in pig bile compared with serum. In bile, biotin accounted for 72 ± 11% of the total of biotin plus metabolites, whereas it accounted for 55 ± 4% in serum (P < 0.05). Bisnorbiotin accounted for 22 ± 7% in bile and for 36 ± 4% in serum (P < 0.05). The percentage contribution of biotin-d,l-sulfoxide was similar in bile and serum (6 ± 5% in bile vs. 10 ± 3% in serum; P = 0.22).

The concentration ratios of bile to serum for biotin and metabolites were compared with those of three endogenous reference compounds: glucose, total cholesterol and total bilirubin. Previous studies have shown that the biliary excretion ratios of these reference compounds span a range from small or nondetectable biliary concentrations (glucose), to similar concentrations in bile and serum (cholesterol), to greater concentrations in bile than in serum (bilirubin) (Klaassen and Watkins 1984, Schein et al. 1968 and 1969, Vondruska and Greco 1973). As expected, the bile to serum concentration ratios were glucose < cholesterol < bilirubin (Table 3). The bile to serum concentration ratios for biotin and metabolites (Table 2) were greater than that of cholesterol but substantially less than that of bilirubin.

Table 3. Cholesterol, glucose and bilirubin in blood serum and in bile from the gall bladder of pigs1 [View Table]

DISCUSSION

The biliary excretion of biotin is small compared with its urinary excretion. In contrast to folic acid and vitamin B-12, the enterohepatic circulation of biotin does not play a quantitatively important role in the homeostasis of the vitamin on the basis of the data of this study. Consequently, metabolism of biotin by gut microorganisms should not contribute importantly to the metabolite profile of biotin. Thus biotin resembles vitamin B-6; only 1% of the dietary B-6 and 2.1% of intravenous B-6 will be excreted in bile (Heard and Annison 1986, Lui et al. 1983).

The profile of biotin metabolites differed considerably between rat bile and urine. The two major biotin metabolites (bisnorbiotin and biotin-d,l-sulfoxide) in human serum and urine and in rat urine were detectable in rat bile in only small quantities (Lee et al. 1972, Mock et al. 1993 and 1995, Wang et al. 1996). The main metabolites in bile were bisnorbiotin methyl ketone and the unknowns A and B. Bisnorbiotin methyl ketone accounts for ~8% of the total biotin plus metabolites in human urine (Zempleni et al. 1997). Its relatively small excretion into rat urine might be a species-specific effect.

Frequently, compounds are secreted into bile as conjugates with sulfate or -glucuronic acid. On the basis of the incubation experiments with sulfatase and glucuronidase, the unknowns A and B in the present study are not sulfate or glucuronate conjugates. Unknowns A and B were not detectable in our pig samples, which were assayed for biotin metabolites based on avidin affinity. Potentially, the unknowns A and B could bind too weakly to avidin to allow detection in the avidin-binding assay (Zempleni et al. 1996). In a previous study on the metabolism of intravenously administered [carbonyl-¹⁴C]biotin in pigs, compounds eluting at 10 min (trace amounts) and 19 min were detected using HPLC and radiometric flow detection (Mock, D. M. and Wang, K. S., unpublished observations). Unknowns A and B in the present study also eluted at 10 and 19 min, respectively. On the basis of HPLC retention times or acid treatment, the unknowns A and B in the present study are not urea, uracil or peptide esters of biotin and biotin metabolites.

The excretion rate of biotin and biotin metabolites in bile decreased quickly during the three collection intervals. In contrast, urinary excretion was roughly constant during the first two collection intervals. The mechanism of the more sustained urinary excretion is unknown. Three time points for a disappearance curve do not allow reliable pharmacokinetic modeling of excretion.

Our pig experiments examined concentrations of biotin and biotin metabolites in bile and serum at a probable steady state. Pigs were used for these studies because of local pig availability, and sufficient quantities of blood and bile from pigs could be collected for assay of endogenous biotin and metabolites that are not radioactively labeled. Pigs had been deprived of feed for 4 h before they were killed. The concentrations of biotin and metabolites were 11 times greater in bile than in serum. Nevertheless, the urinary route of excretion appears to be the major one. The normal bile flow in pigs is ~23 mL/(kg body weight·d) (Mathisen and Omland 1989). On the basis of this bile flow and the concentrations of biotin and metabolites in bile as determined in the present study, the average biliary excretion is 0.86 nmol biotin plus metabolites/(kg body weight·d). This equals 2.2% of the estimated dietary biotin intake of our pigs and agrees well with the biliary excretion that was measured in our rat studies.

In the present study, the concentration ratios of biotin and biotin metabolites (bile vs. serum) were at least an order of magnitude smaller than the ratio measured for the cholephil compound bilirubin. Apparently, the excretion process for biotin and metabolites is less efficient than that of bilirubin, which is excreted predominantly via the biliary route. The molecular mass of a substance is a factor that determines distribution among routes of excretion. Substances at molecular masses < 300-400 Da (e.g., biotin at 244 Da) are preferentially excreted into urine, whereas larger molecules (molecular mass 850 Da) are preferentially excreted into bile (Klaassen and Watkins 1984). However, the concentration ratio of 11 for biotin and metabolites (bile vs. serum) is probably not attributable simply to ductular reabsorption of water. Studies with the marker substance erythritol have shown that the concentrating effect due to the ductular reabsorption or secretion of water is usually small (Klaassen and Watkins 1984). Moreover, if the accumulation of biotin and metabolites in bile was caused by reabsorption of water, all biotin compounds analyzed should be accumulated at the same ratio, yielding a metabolite profile in bile similar to that seen in serum. This was not the case. We propose the existence of an active carrier-mediated transporter that excretes biotin at the canaliculus of hepatocytes, from the bile duct cell into the lumen, or both. On the basis of concentration ratios, we would further hypothesize that the metabolites bisnorbiotin and biotin-d,l-sulfoxide can compete for this transporter, although less efficiently than biotin. Several transport systems have been identified that are involved in the excretion of compounds by the liver, including transporters for organic anions (Klaassen and Watkins 1984). The isoelectric point of biotin is pH 3.5, i.e., biotin is present as an anion at pH 7.4. One might speculate that the presence of dead hepatocytes in bile might be an alternative source of elevated biotin concentrations (Gueant et al. 1984). However, intracellular biotin is bound to carboxylases, including a biotin reservoir with a mitochondrial form of acetyl-CoA carboxylase (Mock 1996, Shriver et al. 1993). Protein-bound intracellular biotin is removed during membrane filtration of the bile in the course of sample preparation. We saw no evidence of radioactivity removed by ultrafiltration and conclude that biotin from dead hepatocytes (indeed biotin bound to proteins >30 kDa) could not account for large proportions of total biotin in bile.

ACKNOWLEDGMENTS