Main

Under physiologic conditions, bilirubin is taken up from plasma by the hepatocytes, conjugated mainly with glucuronic acid, and excreted in bile. Failure of conjugation results in accumulation of UCB. Kernicterus represents the major complication of severe unconjugated hyperbilirubinemia of new-born infants and in the CN syndrome(1).

The study of bilirubin metabolism was facilitated by the availability of the Gunn rat, a mutant strain of Wistar rats(2). Gunn rats are characterized by a deficiency of UDP-glucuronosyltransferase (EC 2.4.1.17). These rats maintain relatively high but constant levels of plasma unconjugated bilirubin throughout life. Because UCB is continuously produced, alternate pathways of pigment disposition must be present to compensate for the deficient glucuronide formation and excretion.

After i.v. injection of radiolabeled bilirubin in Gunn rats, radioactivity was detected in urine, bile, and feces(3–5). In Gunn rats with an external bile drain, approximately 35% of the calculated isotope turnover was recovered in bile; 3-13% was identified as bilirubin. The major fraction consisted of diazo-negative pigments, which have not been up to now identified. In addition, 3% of the injected radioactivity was detected in the urine as polar derivatives, and 17% was found in the feces (25-50% of this was recovered as unconjugated bilirubin). It was suggested that Gunn rats transferred a minor part of disposed UCB directly from plasma into the intestinal lumen(3). However, in rats with bile fistulas the daily bilirubin turnover dropped to about 50% of the turnover of intact rats(3) due to an unknown phenomenon.

Small amounts of bilirubin can be oxidized to yet unidentified compounds; indeed, a variety of enzymatic oxidation systems have been described in the liver and intestine(6–10), including a cytochrome P450 isoenzyme(9, 10). However, proof that oxidation represents a significant pathway of bilirubin disposal “in vivo” is still lacking(11).

Other investigators have shown that Gunn rat bile contains more urobilin than Wistar rat bile in spite of biliary drainage(12). Its origin is presumably intestinal, because urobilin excretion was decreased by the administration of neomycin. Our recent work identified Ugen as the major bilirubin-derived product in Gunn rat bile(13). Because Ugen is formed in the intestine from bilirubin and UCB represents only a minor fraction in Gunn rat bile, a majority of UCB should be passed to the intestine by other than the biliary route to supply a sufficient amount of UCB needed for formation of Ugen. We have, therefore, reinvestigated the role of the intestine as an extrabiliary route for the disposal of bilirubin in man and rats that cannot form bilirubin glucuronides.

METHODS

Chemicals. 14C-Labeled bilirubin was prepared from14 C-δ-aminolevulinic acid according to Ostrow et al.(14). Labeled bilirubin was rechromatographed using TLC before injection(15). Bilirubin IXα and urobilin-I were purchased from Sigma Chemical Co., St. Louis, MO. Bilirubin diconjugate was isolated from the bile of Wistar rats injected with bilirubin according to Spivak and Carey(16) and repurified according to Blanckaert et al.(17).

Patients. Six patients with CN disease (five type II and one type I) at the age of 5, 15, 16, 19, and 20 y (two patients), respectively, were investigated. All with the exception of two were receiving phototherapy. In addition, six patients with GS at the age of 17 (two), 18 (three patients), and 19 y, respectively, and six normal volunteers at the age of 19,20 (two), and 21 (three) y were studied. The stools were collected over three consecutive 24-h periods, stored in the dark, and frozen at -20 °C immediately after collection. In addition, stool samples from a female patient with biliary atresia were collected on the 20th and 35th day after birth. Informed consent was obtained from each patient.

Animals. Male jaundiced and control Wistar RA/JJ rats weighing 250-350 g were used. The properties of the jaundiced strains are described in Table 1. The RA/jj and RHA/jj are Gunn rat strains with unconjugated hyperbilirubinemia due to UDP-glucuronosyltransferase deficiency(18, 19). The TR-/JJ strain is a mutant Wistar rat that cannot secrete conjugated bilirubin and other organic anions in bile and that exhibits conjugated hyperbilirubinemia(20). The TR-/jj strain has both the UDP-glucuronosyltransferase and the transport defect and has unconjugated hyperbilirubinemia(19).

Table 1 Rat strains with congenital jaundice
Full size table

The rats were provided with an external biliary drain under pentobarbital anesthesia. Bile was collected over 1-h periods on ice and in the dark. The oxidation of pigments was prevented by collection of the bile under a layer of paraffin oil and an argon atmosphere. After exsanguination, the intestine was separated into three parts (proximal, middle, and distal), and the content was removed. For stool collection the rats were placed into metabolic cages; they had free access to drinking water and food during the experiments. Stool was collected every 24 h.

To locate the site of transfer of UCB across the intestinal wall, the different parts of the small intestine (duodenum, jejunum, and ileum) and large intestine (cecum and colon) were ligated and transected, and the rats were provided with external biliary drainage. These rats were injected with14 C-bilirubin 2 h after the operation. The labeled pigment was first dissolved in 0.1 mL of NaOH (O.1 mol/L) before addition to 2 mL of Wistar rat serum. Before use the solution of 14C-bilirubin (10-30 nmol with a specific radioactivity of 13,000 dpm/nmol) was sterilized through a 0.22-μm Millex-GV filter (Millipore Corp., Bedford, MA). Immediately after the experiments the rats were exsanguinated.

As far as possible, all manipulations were done under dim light or in the dark, and the samples were processed immediately after collection. All study protocols comply with ethical guidelines.

Determination of bilirubin and its conjugates. Alkaline methanolysis and HPLC. Samples of bile, urine, serum, or of intestinal content, homogenized in physiologic saline, were submitted to alkaline methanolysis and extracted into chloroform as described by Muraca and Blanckaert(21). After evaporation, the residue was redissolved in chloroform and finally analyzed by reverse phase HPLC(21). Blanks were processed according to the same procedure, but alkaline methanolysis was omitted.

Azopigment analysis. Samples were treated with diazotized p- iodoaniline(22) to determine total bilirubin and with diazotized ethyl anthranilate(23) for determination of conjugated bilirubins. Subsequently, the azopigments were extracted twice into n-butyl acetate or in pentane-2 one, respectively. After washing with glycine/HCl buffer, pH 2.4, the extracts were concentrated by evaporation and applied to TLC plates together with reference standards(24). After development, the pigments were scraped off the plates and eluted with methanol, and their amount was determined spectrophotometrically(22, 25). If still contaminated with other substances or not clearly separated from each other, the pigments were reseparated in the same solvent system.

Determination of Ugen. Ugen was determined spectrophotometrically or fluorometrically as a Zn2+ complex of urobilin formed in DMSO by oxidation with iodine(26). The method used yielded more than 80% recovery.

TLC analysis of extract of intestinal lumen. A homogenate of intestinal content was buffered to pH 2.4 with a glycine/HCl buffer (0.4 mol/L HCl adjusted to pH 2.4 with solid glycine), and pigments were extracted into chloroform. The organic extract was evaporated under vacuum and redissolved in a small amount of chloroform. It was then applied to thin-layer plates(DC-Kieselgel F 254, 5715/0025; Merck, Darmstadt, Germany), and the pigments together with standards of UCB and urobilin were separated by development with CHCl3:methanol:H2O (1:1:1; by volume)(27). Five zones, arbitrarily chosen according to the Rf value, were scraped off.

Radioassays. The crude stool samples were lyophilized and subsequently combusted on TRI-CARB SAMPLE OXIDIZER B306 (Packard Instrument Co. Inc., Downers Grove, IL). In parallel, these samples were submitted to alkaline methanolysis(17) and subsequently analyzed by HPLC. The separated fractions were directly mixed with a scintillation mixture(Instagel, Packard Instrument Co.). Finally, the samples were counted in TRI-CARB 4539 Liquid Scintillation System (Packard Instrument Co.).

Statistical analysis. One-way analysis of variance was used to test the differences between the examined group with respect to parameter changes during the experiments. This test was used because of relatively small numbers of samples examined. Differences were considered as significant if P values were <0.05. All results are expressed as means±SD.

RESULTS

The biliary excretion of pigments was determined in three strains of rats with unconjugated hyperbilirubinemia and in control Wistar rats(Table 2). The jaundiced strains excreted Ugen in amounts comparable to controls but the excretion of bilirubins was reduced to only 2-13% of that excreted by Wistar rats. When provided with external bile drains jaundiced rats (RA/jj) continued to excreted Ugen in bile at a constant rate during the entire 3-d period: 2.9 ± 0.6 nmol of Ugen/h/100 g of body weight (n = 8). In contrast, in Wistar rats the excretory rate of Ugen progressively decreased, to become undetectable after 30 h. After 3 d of external biliary drainage, the rats were killed, and the intestinal content was analyzed. In Wistar rats, both Ugen and bilirubin were undetectable. In jaundiced RA/jj rats, both pigments were still present in the intestinal lumen despite the absence of bile influx (Table 3).

Table 2 Bile pigments in serum and bile of Gunn and normal rats
Full size table
Table 3 Content of Ugen and UCB in intestinal lumen of Wistar (RA/JJ) and Gunn (RA/jj) rats after 3-d external biliary drainage
Full size table

The intestine of Gunn RA/jj rats contained more UCB than that of Wistar rats (66.8 ± 13.1 versus 21.3 ± 6.4 nmol/100 g of body weight, p = 2.4 × 10-4, Table 4). The relative amount of UCB, expressed as a percentage of total tetrapyrroles present, was also significantly higher: 65% versus 7% in the small intestine and 22% versus 9% in the large intestine of Gunn and Wistar rats, respectively. Conjugated bilirubin was present in the small but not in the large intestine of Wistar rats. TLC analysis of p- iodoaniline and ethylantranilate azopigments confirmed the presence of UCB in the small and in the large intestine of Wistar and RA/jj rats. It was shown to have the IXα structure(24).

Table 4 Content of Ugen, UCB, bilirubin monoconjugate(BMC), and diconjugated bilirubin (BDC) in intestinal lumen of Wistar (RA/JJ) and Gunn (RA/jj) rats
Full size table

The content of total bile pigments in the intestinal lumen was equal in Gunn and Wistar strains and averaged 250 ± 60 nmol/100 g of body weight, despite the fact that in Gunn rat strains, the biliary excretion of total pigment amounted only to 10-30% of that of Wistar rats (Fig. 1). In the transport-deficient strain(TR-/JJ) both the biliary excretion and gut content of pigments was decreased to approximately half the values obtained in Wistar rats.

Figure 1
figure 1

Comparison of biliary excretion (A) and intestinal content (B) of bile pigments (bilirubins and Ugen) in rats. Panel A: The biliary excretion in Gunn rats (RA/jj, RHA/jj), transport mutant (TR-/JJ) double mutant (TR-/jj), and Wistar(RA/JJ) strains was determined by collecting bile over a 1-h period from bile duct canulated rats. The bile was processed as described under“Methods.” Panel B: The intestinal content of bile pigments in rats in RA/jj, RHA/jj, TR-/jj, TR-/JJ, and Wistar(RA/JJ) strains was collected by opening the intestine immediately after exsanguination. Intestinal contents were removed, homogenized, and processed as described in “Methods.” Statistical differences of the intestinal content of bile pigments between Wistar and jaundiced strains(RA/jj, RHA/jj, TR/jj) were insignificant (p = 0.49, 0.80, and 0.53, respectively).

Full size image

Fecal excretion of bile pigments measured during 5 consecutive days was also comparable in Gunn and Wistar rats and amounted to 363.2 ± 70.2 and 288.7 ± 29.2 nmol/d/100 g of body weight (n = 5), respectively (p = 0.06). UCB represented 47.8 ± 11.1% of excreted bile pigments in Gunn and 10.6 ± 2.5% in Wistar rats(p = 1.8 × 10-4). Urobilinogen excretion averaged 184.6± 31.2 nmol/d/100 g of body weight in jaundiced RA/jj rats. This is far in excess of rates previously measured(28); the difference is presumably due to the previous use of a less sensitive method.

To locate the direct transfer of UCB, two bile-drained Gunn rats with isolated intestinal segments (duodenum, jejunum, ileum, cecum, and colon) were injected with 14C-bilirubin. Radioactivity was found 2 h after i.v. injection in all parts of the intestine in both rats. About 2% of injected dose of radioactivity was recovered. The recovery is comparable with the fractional turnover of bilirubin in bile-drained Gunn rats determined previously(3, 29). Subsequent HPLC analysis of extracts of intestinal content proved the presence of labeled bilirubin.

To investigate a possible intestinal transfer of conjugated bilirubins, the common bile duct of Wistar rats was ligated. This resulted in an increase of serum level of bilirubins from 0.92 ± 0.08 in control rats to 225± 60 nmol/L in ligated ones 5 d after ligation (p = 6.9× 10-5, Table 5). The fecal excretion of bile pigments was 20 ± 10 nmol/d/100 g of body weight compared with 290± 30 in nonobstructed rats (p = 1.4 × 10-7, mean of 5 subsequent days). The intestinal content of bile pigments in rats with a common bile duct ligation dropped to 26% (245 ± 49.2 nmol/100 g of body weight versus 62.9 ± 39.6, p = 4.1 × 10-4) of the control group after 5 d of obstruction(Table 5). For comparison, data from a neonate with biliary atresia are given; again only traces of Ugen and UCB were detected in stools (Table 6).

Table 5 Intestinal content and serum level of bile pigments in intact and bile duct ligated Wistar rats
Full size table
Table 6 Serum levels of bilirubin and concentration of bile pigments in stools of a neonate with biliary atresia
Full size table

The content of bile pigments was also determined in stools of patients with CN disease and GS, and from normal persons collected during three subsequent 24-h periods. The amount of total pigment in stools of CN and GS patients was comparable to that of control subjects (1660 ± 750 nmol/24h/kg of body weight versus 1710 ± 980 versus 1660 ± 870,p = 0.99, Table 7), but the amount of UCB excreted in stools of CN patients tended to be higher (330 ± 210 nmol/24 h/kg of body weight versus 60 ± 5, p = 0.03, Table 7).

Table 7 Serum levels and fecal excretion of bile pigments in CN and GS patients and controls
Full size table

DISCUSSION

In bilirubin UDP-glucuronosyltransferase deficiency states, UCB cannot be efficiently excreted in bile and is thus retained in the body. Because bilirubin production is not decreased and inasmuch as the serum and tissue bilirubin levels are stable, alternative routes have to compensate for the nonfunctional biliary excretion of bilirubin conjugates. Schmid et al.(3–5, 28) reported the presence of polar diazo-negative bilirubin derivatives in bile of Gunn rats, which remain unidentified. The same authors also suggested that UCB could be directly transferred across the gut wall(3). Recently, Ugen was determined as a major bilirubin derivative in Gunn rat bile(13). Other pathways have also been proposed to enhance bilirubin catabolism in Gunn rats; they include bilirubin oxidation(6–10) possibly mediated by induced cytochrome P450. However, proof that oxidation represents a physiologically significant pathway of bilirubin disposal was not obtained(11). In the present study, we have compared the biliary and fecal excretion and intestinal content of bile pigments in man and rats.

In Wistar rats, approximately 10% of biliary bilirubin efflux is reabsorbed in the intestine after transformation to Ugen and is reexcreted in bile(30). In UDP-glucuronosyltransferase-deficient rats, the biliary excretion of Ugen is comparable to that of the Wistar strain despite the markedly reduced biliary secretion of bilirubins. Because Ugen is formed from bilirubin in the intestinal lumen, the extrabiliary supply of bilirubin to the intestinal lumen in Gunn rats must be comparable to the biliary excretion rate of conjugated bilirubin in Wistar rats. A prolonged external bile diversion led to a complete elimination of bile pigments from the intestinal lumen of Wistar rats. In contrast, a considerable amount of bile pigments remained present in the intestinal lumen of bile-drained Gunn rats. In accordance with previous observations(12), Ugen continued to be excreted in bile of drained Gunn rats at a constant level. These results suggest the presence of an efficient direct transfer of unconjugated bilirubin from plasma across the intestinal wall.

The intestinal content and fecal excretion of bile pigments during a 5-d period in Gunn and Wistar rats was similar, despite the low biliary secretion of bilirubin in Gunn rats. This indicates that the extrabiliary intestinal excretion of bilirubin is responsible for 85-98% of the bilirubin present in the intestinal lumen and that the intestinal content and fecal excretion of bile pigments is independent of the biliary efflux of bilirubin in bilirubin UDP-glucuronosyltransferase deficient rat strains. Again, permeation across the intestinal wall seems the most likely mechanism involved.

The direct transfer of UCB across the intestinal wall was demonstrated by determination of radioactive bile pigments in the lumen of isolated intestinal segments after i.v. injection of labeled bilirubin to bile-diverted Gunn rats. The mechanism involved remains unidentified. Recent studies indicate that passive diffusion from blood to intestinal lumen may represent an important route for elimination of lipophilic poorly metabolized compounds(31, 32). Unconjugated bilirubin in conditions of bilirubin glucuronosyltransferase deficiency seems such a compound.

Despite high levels of conjugated bilirubin in serum, rats with a common bile duct ligation and a neonate with biliary atresia excreted only traces of bile pigments in stools. Hence, in contrast to UCB, conjugated bilirubin does not appear to permeate across the intestinal wall to a major extent.

CN patients as well as Gunn rats retain high levels of UCB in serum and tissues and excrete only traces of bilirubin in bile. Biliary excretion of bilirubin is also significantly decreased in GS patients(33, 34). However, feces of CN and GS patients have a normal color and contain normal amounts of bile pigments.

The direct permeation of UCB is further substantiated by the higher amount of UCB found in feces of Gunn rats and CN patients in comparison with control subjects. However, part of the bilirubin found in the intestine of CN patients treated with phototherapy could be derived from phototherapy-induced biliary secretion.

The extrabiliary elimination route is not very efficient, because it seems to depend on equilibration of unbound UCB, which will represent only a small fraction of total UCB. Conversion or trapping of UCB in the intestinal lumen by binding to adsorbents, such as undigested food fibers, can lead to final elimination. A shift of the equilibrium in favor of the intestinal bilirubin pool might be therapeutically useful also in other forms of unconjugated hyperbilirubinemia, such as neonatal jaundice and GS. A few attempts to trap UCB in the lumen and to enhance its fecal elimination have been made, however, with conflicting results(35–37). This elimination route could possibly be exploited by using adsorbents that more efficiently trap UCB in the intestinal lumen(38).