- Email: [email protected]
The impact of intestinal microflora on serum bilirubin levels
Journal of Hepatology 42 (2005) 238–243 www.elsevier.com/locate/jhep
The impact of intestinal microflora on serum bilirubin levels Libor Vı´tek1,2,*, Jaroslav Zelenka1, Marie Zadinova´3, Jirˇ´ı Malina4 1
Institute of Clinical Biochemistry and Laboratory Diagnostics, 1st Medical Faculty, Charles University of Prague, Prague, Czech Republic 2 4th Department of Internal Medicine, 1st Medical Faculty, Charles University of Prague, Prague, Czech Republic 3 Institute of Medical Biophysics, 1st Medical Faculty, Charles University of Prague, Prague, Czech Republic 4 Department of Clinical Microbiology, Barrandov Outpatient Center, Prague, Czech Republic
See Editorial, pages 170–172 Background/Aims: Intestinal microflora plays an important role in the pathogenesis of neonatal jaundice by inhibiting enterosystemic circulation of bilirubin. The present study aimed to investigate the influence of intestinal microflora on serum bilirubin levels in hyperbilirubinemic Gunn rats. Methods: After a baseline phase Gunn rats received oral antibiotics (either clindamycin/neomycine or cotrimethoxazole for four days, phase II). Intestinal colonization was carried out either with a bilirubin-reducing strain of C. perfringens or C. pasteurianum incapable of reducing bilirubin (phase III). Serum bilirubin and fecal bile pigments were determined at the end of each phase. Results: Oral administration of clindamycin/neomycine resulted in the disappearance of fecal urobilinoids. Simultaneously, serum bilirubin increased dramatically (186G31 vs. 289G35 mmol/l, PZ0.004). Intestinal colonization with C. perfringens led to reappearance of fecal urobilinoid production accompanied with a partial decrease of serum bilirubin (289G35 vs. 239G17 mmol/l, PZ0.013), whereas the effect of C. pasteurianum on bile pigment metabolism was negligible. Co-trimethoxazole therapy had no effect on serum and intestinal metabolism of bilirubin. Conclusions: Intestinal microflora greatly affects intravascular metabolism of bilirubin. Prolonged use of certain antibiotics in man may lead to an increase in serum bilirubin levels, while the enhancement of intestinal catabolism may have an opposite effect. q 2004 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. Keywords: Antibiotic; Bilirubin; Bilirubin catabolism; Enterohepatic circulation; Hyperbilirubinemia; Intestinal microflora; Neonatal jaundice; Urobilinoids 1. Introduction Microbial catabolism of bilirubin in gut lumen is believed to contribute to serum bilirubin homeostasis, in particular during neonatal age [1,2]. In the early newborn period bilirubin-reducing microflora is absent due to Received 28 May 2004; received in revised form 22 September 2004; accepted 4 October 2004; available online 2 November 2004 * Corresponding author. Address: 4th Department of Internal Medicine, 1st Medical Faculty, Charles University of Prague, U Nemocnice 2, Praha 2, 128 08, Prague, Czech Republic. Tel.: C420 224 962 534; fax: C420 224 923 524. E-mail address: [email protected] (L. Vı´tek). Abbreviations ALT, alanine aminotransferase; AST, aspartate aminotransferase; ATB, antibiotic; B., Bacteroides; C., Clostridium.
insufficient and delayed colonization of neonatal gastrointestinal tract and unconjugated bilirubin (UCB) accumulated in the gut lumen may undergo substantial enterohepatic and enterosystemic circulation. It is known that only negligible amounts of fecal urobilinoids (for Review and nomenclature see Ref. [3]) are present in the intestinal lumen of infants during the first months of life due to undeveloped intestinal microflora capable of reducing bilirubin [2,4–6]. On contrary, the urobilinoid production in adults is highly efficient. Under normal conditions only small amounts of bilirubin can be found in stools of adults (5–20 mg/day) while urobilinoids are predominant bile pigments (50–250 mg/day) [4]. Nevertheless, only four bacterial strains capable of bilirubin conversion have been isolated so far with certainty:
0168-8278/$30.00 q 2004 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jhep.2004.10.012
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243
Clostridium (C.) ramosum [7], C. perfringens and C. difficile [2] and Bacteroides (B.) fragilis [8] despite the fact that hundreds of different indigenous microbial strains inhabit human gastrointestinal tract [9]. These organisms are capable of carrying out many enzymatic reactions that the host human cells cannot catalyze. It was estimated that there may be more than 400 different bacterial species in the human gastrointestinal tract at any given time with considerable variations in the fecal flora from one individual to another [10]. It is known that more than 99% of all intestinal bacteria are obligate anaerobes [11] with B. fragilis being the commonest microorganism [12]. Among others Eubacteria, Enterobacteria, Enterococci, Lactobacilli and Clostridia belong to the most prevalent species [13]. The importance of intestinal microflora for bilirubin catabolism in the gut lumen was established in studies with germfree rats. These animals do not excrete any urobilinoids in their feces or urine [7], whereas conjugated bilirubins represent major fecal bile pigments [14]. Colonization of their gastrointestinal tract with indigenous microflora of conventional animals results in prompt production of fecal urobilinoids [15]. Another evidence for the role and importance of bacteria in bilirubin catabolism in the intestinal lumen comes from studies of subjects treated with antibiotics. It was demonstrated that oral antibiotics eliminate fecal production of urobilinoids as a result of eradication of intestinal microflora reducing bilirubin [16–18]. Despite the importance of intestinal catabolic pathway of bilirubin only little is known about its relation to bilirubin metabolism in the intravascular compartment in adult age and the effect of oral antibiotics on bilirubin-reducing
Fig. 1. Intestinal metabolism of bilirubin in adults under physiologic conditions. [This figure appears in colour on the web.]
239
Fig. 2. Intestinal metabolism of bilirubin in adults after eradication of bilirubin-reducing microflora. [This figure appears in colour on the web.]
microflora has never been related to the serum bilirubinemia. In the present study we investigated the influence of the intestinal microflora on serum bilirubin levels in hyperbilirubinemic Gunn rats (Figs. 1 and 2).
2. Material and methods 2.1. Animal experiments 1. The effect of eradication of bilirubin-reducing microflora with subsequent colonization of gastrointestinal tract with a strain of C. perfringens reducing bilirubin. Eleven male Gunn rats (RHA/jj) with congenital deficiency of bilirubin UDP-glucuronosyltransferase (weight range 250–270 g) were used in this study. Each animal was placed in a separate metabolic cage enabling collection of stools and preventing coprophagy. After an initial baseline phase I (1 week of feeding a normal diet) the rats received antibiotics (clindamycin (Dalacin C, Pharmacia/Upjohn, Belgium) 4 mg/day plus neomycine (Bykomycin, Byk Gulden, Germany) 30 mg/day)) via a gastric tube for subsequent 4 days (phase II). Finally, intestinal colonization was carried out with a single, bilirubin-reducing strain of C. perfringens isolated from neonatal feces (2) (1.5 ml of suspension of bacteria (5.5K7.0!106 cells in logarithmic phase of growth per ml of 2% yeast extract) administered via gastric tube for further consecutive four days (phase III). Antibiotic susceptibility of this strain was tested in vitro with Antibiogram Disks (Biorad, USA) for following antibiotics: clindamycin, co-trimethoxazole, gentamycine, amikacine, netilmycine, neomycine and tobramycine. Serum bilirubin and fecal excretion of bile pigments were determined at the end of each phase. Blood sampling was carried out from the ocular vein during ongoing experiment, whereas heart puncture was used for blood collection at the end of each experiment. Food intake, body weights and fecal excretion were recorded every day during the whole experimental periods. 2. The effect of eradication of bilirubin-reducing microflora with subsequent colonization of gastrointestinal tract with a strain of C. pasteurianum incapable of reducing bilirubin. Six male Gunn rats (RHA/jj) with similar characteristics and treated with clindamycin plus neomycine in a similar way as in the first study were used in this study. In contrast to the first study, intestinal colonization
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243
240
(phase III) was carried out with a single strain of C. pasteurianum (DSM 525, DSMZ, Germany) incapable of reducing bilirubin as demonstrated in previous studies (L. Vı´tek, unpublished results). Serum and fecal bile pigments were determined as described above. 3. The effect of oral administration of ATB not influencing bilirubinreducing microflora Six male Gunn rats (RHA/jj) with similar characteristics as in the first study were used in this study. In contrast to the first study, clindamycin/neomycine treatment was exchanged for co-trimethoxazole (Biseptol, Polfa, Poland, 27 mg/kg/day) not affecting aerobic microflora. Serum and fecal bile pigments were determined as described above. All aspects of the study met the accepted criteria for the humane care and experimental use of laboratory animals and all protocols were approved by the Animal Research Committee of the 1st Medical Faculty, Charles University of Prague.
2.2. Biochemical analyses Serum bilirubin and transaminase activities were determined on the automatic analyzer (model 717; Hitachi, Tokyo, Japan). Bilirubin was analyzed by the diazo reaction according to Jendrassik and Gro´f. Serum transaminase activities were determined enzymatically with a-ketoglutarate and alanine or aspartate as substrates for ALT and AST, respectively. Fecal bile pigments and pH were determined in fecal suspensions in saline (1:10 w/v). Fecal urobilinoids were determined by spectrophotometry as a zinc complexes of total urobilins [19]. Fecal bilirubin concentrations were determined by spectrophotometry as p-iodoanilin azopigments [20].
2.3. Statistical analysis Data are presented as meanGSD. The statistical significance of differences between variables was evaluated by paired t-test or Wilcoxon Signed Rank test when data were not normally distributed. Differences were considered statistically significant when P values were less than 0.05.
3. Results All animals remained in good health during the whole experimental period without substantial changes in animal body weights and/or deterioration of food intake. No signs of diarrhoea were reported during all experiments. Oral administration of clindamycin plus neomycine to hyperbilirubinemic Gunn rats resulted in disappearance of fecal urobilinoids indicative of the eradication of intestinal microflora capable of reducing bilirubin (148G48 vs. 0 nmol/d/100 g b.wt, PZ0.001, Table 1). Simultaneously, serum bilirubin increased dramatically (186G31 vs. 289G Table 1 Serum bilirubin and fecal excretion of bile pigments in Gunn rats after eradication of bilirubin-reducing microflora and subsequent colonization with C. perfringens reducing bilirubin
Control phase ATB treatment Colonization phase
Serum bilirubin (mmol/L)
Fecal bilirubin (nmol/d/100 g b.wt)
Fecal urobilinoids (nmol/d/ 100 g b.wt)
186G31 289G35* 239G17**
65G17 103G47*** 120G72
148G48 0**** 39G28*****
*PZ0.004, compared to control phase; **PZ0.013, compared to ATB treatment;***PZ0.04,comparedtocontrolphase;****PZ0.001,compared to control phase; *****PZ0.001, compared to ATB treatment.
Table 2 Bile pigment balance in Gunn rats after eradication of bilirubinreducing microflora and subsequent colonization with C. perfringens reducing bilirubin
Control phase ATB treatment Colonization phase
Serum bilirubin poola (nmol)
Fecal bile pigment pool (nmol)
558G93 867G105 717G51
554G121 277G128* 447G234**
*PZ0.002, compared to control phase; **PZ0.102, compared to ATB treatment. a Serum pools calculated according to Belcher and Harris [36].
35 mmol/l, PZ0.004) (Table 1). Intestinal colonization with a strain of C. perfringens reducing bilirubin led to reappearance of urobilinoids in feces to 26% of the initial values (0 vs. 39G28 nmol/d/100 g b.wt, PZ0.001, Table 1) and to a partial decrease of serum bilirubin (by 17.3%; from 289G35 to 239G17 mmol/l, PZ0.013) (Table 1). Disappearance of urobilinoid production by intestinal microflora eradicated during antibiotic (ATB) therapy resulted in accumulation of bilirubin within intestinal lumen as evidenced by increase of fecal bilirubin output (65G17 vs. 103G47, PZ0.04, Table 1). However, total fecal bile pigment output decreased significantly after ATB treatment (554G121 vs. 277G128 nmol/d, PZ0.002) (Table 2). Subsequently, decrease of serum bilirubin levels after colonization of intestinal tract with bacteria reducing bilirubin was accompanied with an increase of fecal bile pigment excretion (Table 2). However, when the intestinal tract of Gunn rats was colonized with a strain of C. pasteurianum incapable of reducing bilirubin, increased serum bilirubin levels decreased only insignificantly (by 4.6%; from 281G58 to 269G34 mmol/l, PZ0.625) (Table 3). Simultaneously, colonization with this strain of C. pasteurianum did not lead to reappearance of fecal urobilinoids (Table 3). Direct link between bilirubin-reducing microflora and serum bilirubin homeostasis was further corroborated by the results of another study when co-trimethoxazole was used as an antibiotic not affecting bilirubin-reducing microflora. In this study both serum bilirubin and fecal urobilnoids were not markedly influenced by this antibiotic (Table 4). Serum bilirubin levels increased by 13% (from 138G59 to Table 3 Serum bilirubin and fecal excretion of urobilinoids in Gunn rats after eradication of bilirubin-reducing microflora and subsequent colonization with C. pasteurianum incapable of reducing bilirubin
Control phase ATB treatment Colonization phase
Serum bilirubin (mmol/l)
Fecal urobilinoids (nmol/d/100 g b.wt)
153G20 281G58* 269G34**
132G45 0*** 0***
*PZ0.01, compared to control phase; **PZ0.625, compared to ATB treatment; ***PZ0.001, compared to control phase.
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243 Table 4 Serum bilirubin and fecal excretion of urobilinoids in Gunn rats after treatment of ATB not influencing bilirubin-reducing microflora
Control phase ATB treatment
Serum bilirubin (mmol/l)
Fecal urobilinoids (nmol/d/100 g b.wt)
138G59 156G69*
127G85 121G52**
*PZ0.55, compared to control phase; **PZ0.91, compared to control phase.
156G69 mmol/l, PZ0.55, Table 4), which is within a physiological variation of serum bilirubin levels in Gunn rats as evidenced by results of a separate study (data not shown). On contrary, highly significant increases by 55% (153G20 vs. 281G58 mmol/l, PZ0.01) and 84% (186G31 vs. 289G35 mmol/l, PZ0.004) were demonstrated in both clindamycin/neomycine studies, respectively (Tables 1 and 3). In in vitro experiments a strain of C. perfringens used for colonization was found to be susceptible only for clindamycin, aminoglycosides (with the exception of slight suppressing effect of netilmycine), as well as co-trimethoxazole were ineffective. Fecal pH did not change markedly during all the study phases (a range between 6.6 and 7.2 without substantial differences attributable to any individual study phase). No deterioration of serum transaminase activities was detected during antibiotic treatment (ALT: 1.6G0.3 vs. 2.0G0.3 vs. 1.7G0.2 mkat/l, PO0.05, for Phases I, II, and III, respectively; AST: 3.5G0.5 vs. 3.5G0.5 vs. 2.2G0.4 mkat/l, PO0.05, for Phases I, II, and III, respectively).
4. Discussion The relationship between the intestinal catabolism of bilirubin and serum bilirubin levels has been described only in newborn infants [1,2]. In the present study we demonstrate a profound effect of oral clindamycin/neomycine therapy administered to hyperbilirubinemic Gunn rats on eradication of intestinal microflora reducing bilirubin resulting in inhibition of bilirubin catabolism in the intestinal lumen. As a consequence, the levels of serum bilirubin rose remarkably. Increase in serum bilirubin after oral ATB treatment eradicating bilirubin-reducing microflora was accompanied by a drop of fecal bile pigment excretion with accumulation of bilirubin within intestinal lumen (Table 1). Furthermore, colonization of intestinal tract of Gunn rats with a strain of C. perfringens reducing bilirubin had a prompt reverse effect on serum bilirubin levels as well as fecal urobilinoid production, while no such effect could be demonstrated after colonization of gastrointestinal tract with C. pasteurianum incapable of reducing bilirubin. Similar lacking effect on bilirubin metabolism was observed also after oral administration of antibiotic not affecting bilirubin-reducing microflora further supporting
241
the importance of bilirubin-reducing microflora on intravascular bilirubin homeostasis. Under physiological conditions of bilirubin disposition via hepatocytes bilirubin might be reabsorbed into portal circulation and undergo enterohepatic circulation. A substantial portion of bilirubin reabsorbed from the gut lumen can spill over into the systemic circulation because of its low z30% single-pass hepatic extraction [21] (Figs. 1 and 2). However, it is necessary to note that under conditions of deficient bilirubin conjugation in the liver tissue such as in Gunn rats, where the major disposition route for bilirubin is direct permeation through the intestinal wall [22], the effect of inhibition of bilirubin degradation resulting in its accumulation in the intestinal lumen might lead in even more pronounced elevation of serum bilirubin levels. In fact, inverse relationship between serum bilirubin levels and fecal bile pigment excretion is in support of direct link between intestinal and intravascular metabolism of bilirubin. As can be calculated from Table 2, clindamycin/ neomycine treatment increased serum bilirubin pool by 309 nmol, while daily fecal bile pigment output decreased by 277 nmol. Similar balanced state can be seen between treatment and colonization phases (decrease of serum bilirubin pool by 150 vs. increase of intestinal bile pigment pool by 170 nmol). These data indicate that intestinal disposition of bile pigments was markedly disturbed by clindamycin/neomycine treatment. Since it was suggested that there is a dynamic equilibrium between serum bilirubin levels and bilirubin concentrations within intestinal lumen [22], based on our results it seems that treatment with antibiotics affecting bilirubin-reducing microflora impairs bilirubin disposition in Gunn rats. This is supported by the fact that intestinal colonization with a single strain of C. perfringens reducing bilirubin led to a reappearance of urobilinoids in feces and to a partial, but significant decrease of serum bilirubin demonstrating thus the direct role of intestinal microflora on bilirubin homeostasis, while no such effect could be detected in animals colonized with a single strain of C. pasteurianum incapable of reducing bilirubin. It is highly likely that treatment with antibiotics affecting bilirubin-reducing microflora might influence serum bilirubin levels also under physiological conditions of bilirubin disposition via hepatocytes by enhancement of enterohepatic and enterosystemic circulation of bilirubin, i.e. by a principle similar to that described above for Gunn rats. Our data are consistent with results of previous studies performed on germfree and conventional rats. As stated above, germfree rats do not excrete any urobilinoids in their feces or urine [7], whereas conjugated bilirubins represent major fecal bile pigments [14]. However, one day after conventialization fecal urobilinoids became positive reaching physiologic values 7 days later [15]. Interestingly, after monocontamination of germfree rats with C. ramosum isolated from intestinal contents of conventionalized rats approximately 10% of the bilirubin was excreted in the feces as urobilinoids. The fecal output of
242
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243
urobilins increased to 30% after superinfection of monocontaminated ex-germfree rats with a strain of E. coli, but never reached the values found in the conventional rats (i.e. almost 80% of bilirubin excreted as fecal urobilinoids). Interestingly, after contamination of germfree rats with feces from conventional rats, the ex-germfree rats produced urobilins to the same extent as conventional rats [7]. It is important to note that a strain of E. coli used in this study was not able to produce urobilinoids alone and only potentiated production by C. ramosum indicating that complex interactions between several specific microorganisms might be responsible for high production of urobilinoids seen in conventional rats as well as adult healthy humans. These data are consistent with our findings that colonization of sterile gastrointestinal tract of Gunn rats with a strain of C. perfringens reducing bilirubin resulted only in partial restoration of urobilinoid production. Interestingly, no urobilinoid production was observed after colonization of intestinal tract with strains of Peptostreptococcus species [23], Lactobacillus acidophilus, Bifidobacterium bifidum [24], C. welchii type A, Streptococcus faecalis, Proteus vulgaris, Escherichia coli, Bacillus subtilis, Sarcina and Mucor [7]. In our in vitro studies no bilirubin-reducing activity was detected in various clostridial strains including C. pasteurianum, C. acetobutylicum, C. novyii, C. paraputrificum, C. sporogenes, C. bifermentans, C. scindens, C. hiranonis and C. sordelii (L. Vı´tek, unpublished results). Based on these limited data it seems that bilirubin reduction within gastrointestinal tract is confined to bacterial species of C. perfringens, C. difficile, C. ramosum and possibly B. fragilis [2,7,8]. In a study by Tally et al. 86% of 49 strains of C. ramosum tested was found to be sensitive to clindamycin [25]. In another study Clostridia were reported to be sensitive to clindamycin, ampicillin and erythromycin [26]. Oral treatment with clindamycin was shown to cause fecal bile pigment pattern similar to that of germfree animals [16] with return of urobilinoid production to conventional values within 4–6 weeks [27]. It was demonstrated by Heimdahl and Nord that clindamycin strongly disturbs microbial colonization causing at least a 100-fold increase in indigenous facultative anaerobic flora [28]. The effect of antibiotics given orally for 6 days to healthy humans was investigated also in studies by Saxerholt et al. and Midtvedt [16,17]. Intake of bacitracin, vancomycin, clindamycin, erythromycin and ampicillin resulted in a pronounced suppression of fecal urobilinoid excretion, whereas intake of doxycycline, metronidazole, nalidixic acid, ofloxacin and co-trimethoxazole had no significant effect. Intake of clindamycin also led to a marked increase of conjugated bilirubin in the feces and this drug was the only one present in high concentrations in the feces of subjects tested [16]. Similar suppressive effect on urobilinoid producing microflora was demonstrated in earlier human studies for aureomycin [17], chloromycetin [29] or sulphasalazine [30] as well as for a number of other antibiotics including
penicillins, tetracyclins or aminoglycosides tested in animal studies [31]. These findings correlate well with available data that only Clostridia are capable of reducing bilirubin [2,7,32–35]. It can be implied from the list of antibiotics listed above, that many of the antimicrobial agents used in human medicine may have potential to affect serum bilirubin levels (Fig. 2). Although this effect will be negligible in otherwise healthy subjects, it might be of importance in patients with other underlying clinical conditions influencing metabolism of bilirubin. According to our knowledge this is the first report demonstrating direct link between eradication of intestinal microflora capable of reducing bilirubin and serum bilirubin levels. On the basis of our data it may be concluded that the intestinal microflora may substantially affect serum metabolism of bilirubin not only in newborn infants but also in adult age. Prolonged use of antibiotics in man may lead to an increase of serum bilirubin levels, while enhancement of its intestinal catabolism may have an opposite effect.
Acknowledgements The authors thank Mrs. Anna Fliegerova´ and Jirˇ´ı Oborˇil for excellent technical assistance. This work was supported by the research grant GACˇR No. 310/021436 (Granting Agency of the Czech Republic).
References [1] Poland RL, Odell GB. Physiologic jaundice: the EHC of bilirubin. New Engl J Med 1971;284:1–6. ˇ erna´ M, Chmelarˇ D, et al. [2] Vı´tek L, Kotal P, Jirsa M, Malina J, C Intestinal colonization in neonates leading to fecal urobilinoid excretion may play a role in the pathogenesis of neonatal hyperbilirubinemia. J Pediatr Gastroenterol Nutr 2000;30:294–298. [3] Moscowitz A, Weimer M, Lightner A, Petryka ZJ, Davis E, Watson CJ. The in vitro conversion of bile pigments to the urobilinoids by a rat Clostridia species as compared with human fecal flora III. Natural d-urobilin, synthetic i-urobilin, and synthetic iurobilinogen. Biochem Med 1970;4:149–170. [4] With TK. Bile pigment in feces. Bile pigments. Chemical, biological and clinical aspects. New York: Academic Press; 1968 p. 585–613. [5] Norin KE, Gustafsson BE, Lindblad BS, Midtvedt T. The establishment of some microflora associated biochemical characteristics in feces from children during first years of life. Acta Paediatr Scand 1985;74:207–212. [6] Midtvedt AC, Duke BC, Norin KE, Saxerholt H, Midtvedt T. Development of five metabolic activities associated with the intestinal microflora of healthy infants. J Pediatr Gastroenterol Nutr 1988;7: 559–567. [7] Gustafsson BE, Lanke LS. Bilirubin and urobilins in germfree, exgermfree and conventional rats. J Exp Med 1960;112:975–981. [8] Fahmy K, Gray CH, Nicholson DC. The reduction of bile pigments by faecal and intestinal bacteria. Biochim Biophys Acta 1972;264:85–97. [9] Moore WEC, Holdenman LV. Human fecal flora: the normal flora of 20 Japanese-Hawaiians. Appl Microbiol 1974;27:961–979.
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243 [10] Holdeman LV, Good IJ, Moore WEC. Human fecal flora: Variation in bacterial composition within individual and a possible effect of emotional stress. Appl Environ Microbiol 1976;31:359–375. [11] Hylemon PB. Matebolism of bile acids in intestinal microflora. In: Danielsson H, Sjo¨vall J, editors. Sterols and bile acids. Amsterdam: Elsevier Science; 1985. p. 331–343. [12] Hayakawa S. Microbial transformation of bile acids. Adv Lipid Res 1973;11:143–192. [13] Hill MJ, Drasar BS. The normal colonic bacterial flora. Gut 1975;16: 318–323. [14] Saxerholt H, Skar V, Midtvedt T. HPLC separation and quantification of bilirubin and its glucuronide conjugates in faeces and intestinal contents of germ-free rats. Scand J Clin Lab Invest 1990; 50:487–495. [15] Saxerholt H, Midtvedt T, Gustafsson BE. Deconjugation of bilirubin conjugates and urobilin formation by conventionalized germ-free rats. Scand J Clin Lab Invest 1984;44:573–577. [16] Saxerholt H, Carlstedt-Duke B, Hoverstadt T, Lingaas E, Norin KE, Steinbakk M, et al. Influence of antibiotics on the faecal excretion of bile pigments in healthy subjects. Scand J Gastroenterol 1986;21: 991–996. [17] Midtvedt T. Effects of antimicrobial agents upon functional part of the intestinal flora. Scand J Infect Dis Suppl 1986;49:85–88. [18] Sborov VM, Jay AR, Watson CJ. The effect of aureomycin on urobilinogen formation and the fecal flora. J Lab Clin Med 1951;37: 52–59. [19] Kotal P, Fevery J. Quantitation of urobilinogen in faeces, urine, bile, and serum by direct spectrophotometry of zinc complex. Clin Chim Acta 1991;202:1–10. [20] Heirwegh KPM, Fevery J, Meuwissen J, de Groote J. Recent advances in the separation and analysis od diazo-positive bile pigments. Methods Biochem Anal 1974;22:205–250. [21] Ga¨rtner U, Goeser T, Wolkoff AW. Effect of fasting on the uptake of bilirubin and sulfobromophtalein by isolated perfused rat liver. Gastroenterology 1997;113:1707–1713. [22] Kotal P, van der Veere C, Sinaasappel M, Elferink RO, Vı´tek L, Brodanova´ M, et al. Intestinal excretion of unconjugated bilirubin in man and rats with inherited unconjugated hyperbilirubinemia. Pediatr Res 1997;42:195–200.
243
[23] Saxerholt H. The normal microflora and bilirubin. In: Grubb R, Midtvedt T, Norin E, editors. The regulatory and protective role of the normal microflora. New York: Stockton Press; 1988. p. 169–201. [24] Norin KE, Persson AK, Saxerholt H, Midtvedt T. Establishment of Lactobacillus and Bifidobacterium species in germfree mice and their influence on some microflora-associated characteristics. Appl Environ Microbiol 1991;57:1850–1852. [25] Tally FP, Armfield AY, Dowell VR, Kwok YY, Sutter VI, et al. Suseptibility of Clostridium ramosum to antimicrobial agents. Antimicrob Agents Chemother 1974;5:589–593. [26] Appelbaum PC, Chatterton SA. Susceptibility of anaerobic bacteria to ten antimicrobial agents. Antimicrob Agents Chemother 1978;14: 371–376. [27] Midtvedt T. Clindamycin-associated alterations in intestinal microfloraassociated characteristics in rats. Scand J Gastroenterol 1985;20:92–98. [28] Heimdahl A, Nord CE. Effect of erythromycin and clindamycin on the indigeneous human anaerobic flora and new colonization of the intestinal tract. Eur J Clin Microbiol 1982;1:38–48. [29] Kolb FO, Oppenheim DJ. Disappearance of urobilinogen during antibiotic therapy. Bull New Engl Med Center 1950;12:31. [30] Leijonmarck CE, Benno P, Carlstedt-Duke B, Monsen U, Norin E, Poppen B, et al. The function of the intestinal microflora in patients with ulcerative colitis before and after colectomy. Scand J Gastroenterol 1990;25:585–593. [31] Gustafsson BE, Norin KE. Development of germfree animal characteristics in conventional rats by antibiotics. Acta Pathol Microbiol Scand Sect B 1977;85:1–8. [32] Passini F, Czakes J. Ueber Urobilinbildung durch Reinkulturen anaerober Darmbakterien. Wien Klin Wochnschr 1923;36:657–659. [33] Ka¨mmerer H, Miller R. Ueber die Umwandlung der Gallenfarbstoffe durch fa¨ulniserregende Darmbakterien. Wien Klin Wochnschr 1922; 35:639–640. [34] Ka¨mmerer H, Miller K. Zur enterogenen urobilinbildung. Deutsche Arch Klin Med 1923;141:318–347. [35] Watson CJ. The urobilinoids: Milestones in their history and some recent developments. In: Berk PD, Berlin NI, editors. The chemistry and physiology of bile pigments. Washington, DC: US Govern Print Off; 1977. p. 469–480. [36] Belcher EH, Harris EB. Studies of plasma volume, red cell volume and total blood volume in young growing rats. J Physiol 1957;139:64–78.
The impact of intestinal microflora on serum bilirubin levels Libor Vı´tek1,2,*, Jaroslav Zelenka1, Marie Zadinova´3, Jirˇ´ı Malina4 1
Institute of Clinical Biochemistry and Laboratory Diagnostics, 1st Medical Faculty, Charles University of Prague, Prague, Czech Republic 2 4th Department of Internal Medicine, 1st Medical Faculty, Charles University of Prague, Prague, Czech Republic 3 Institute of Medical Biophysics, 1st Medical Faculty, Charles University of Prague, Prague, Czech Republic 4 Department of Clinical Microbiology, Barrandov Outpatient Center, Prague, Czech Republic
See Editorial, pages 170–172 Background/Aims: Intestinal microflora plays an important role in the pathogenesis of neonatal jaundice by inhibiting enterosystemic circulation of bilirubin. The present study aimed to investigate the influence of intestinal microflora on serum bilirubin levels in hyperbilirubinemic Gunn rats. Methods: After a baseline phase Gunn rats received oral antibiotics (either clindamycin/neomycine or cotrimethoxazole for four days, phase II). Intestinal colonization was carried out either with a bilirubin-reducing strain of C. perfringens or C. pasteurianum incapable of reducing bilirubin (phase III). Serum bilirubin and fecal bile pigments were determined at the end of each phase. Results: Oral administration of clindamycin/neomycine resulted in the disappearance of fecal urobilinoids. Simultaneously, serum bilirubin increased dramatically (186G31 vs. 289G35 mmol/l, PZ0.004). Intestinal colonization with C. perfringens led to reappearance of fecal urobilinoid production accompanied with a partial decrease of serum bilirubin (289G35 vs. 239G17 mmol/l, PZ0.013), whereas the effect of C. pasteurianum on bile pigment metabolism was negligible. Co-trimethoxazole therapy had no effect on serum and intestinal metabolism of bilirubin. Conclusions: Intestinal microflora greatly affects intravascular metabolism of bilirubin. Prolonged use of certain antibiotics in man may lead to an increase in serum bilirubin levels, while the enhancement of intestinal catabolism may have an opposite effect. q 2004 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. Keywords: Antibiotic; Bilirubin; Bilirubin catabolism; Enterohepatic circulation; Hyperbilirubinemia; Intestinal microflora; Neonatal jaundice; Urobilinoids 1. Introduction Microbial catabolism of bilirubin in gut lumen is believed to contribute to serum bilirubin homeostasis, in particular during neonatal age [1,2]. In the early newborn period bilirubin-reducing microflora is absent due to Received 28 May 2004; received in revised form 22 September 2004; accepted 4 October 2004; available online 2 November 2004 * Corresponding author. Address: 4th Department of Internal Medicine, 1st Medical Faculty, Charles University of Prague, U Nemocnice 2, Praha 2, 128 08, Prague, Czech Republic. Tel.: C420 224 962 534; fax: C420 224 923 524. E-mail address: [email protected] (L. Vı´tek). Abbreviations ALT, alanine aminotransferase; AST, aspartate aminotransferase; ATB, antibiotic; B., Bacteroides; C., Clostridium.
insufficient and delayed colonization of neonatal gastrointestinal tract and unconjugated bilirubin (UCB) accumulated in the gut lumen may undergo substantial enterohepatic and enterosystemic circulation. It is known that only negligible amounts of fecal urobilinoids (for Review and nomenclature see Ref. [3]) are present in the intestinal lumen of infants during the first months of life due to undeveloped intestinal microflora capable of reducing bilirubin [2,4–6]. On contrary, the urobilinoid production in adults is highly efficient. Under normal conditions only small amounts of bilirubin can be found in stools of adults (5–20 mg/day) while urobilinoids are predominant bile pigments (50–250 mg/day) [4]. Nevertheless, only four bacterial strains capable of bilirubin conversion have been isolated so far with certainty:
0168-8278/$30.00 q 2004 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.jhep.2004.10.012
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243
Clostridium (C.) ramosum [7], C. perfringens and C. difficile [2] and Bacteroides (B.) fragilis [8] despite the fact that hundreds of different indigenous microbial strains inhabit human gastrointestinal tract [9]. These organisms are capable of carrying out many enzymatic reactions that the host human cells cannot catalyze. It was estimated that there may be more than 400 different bacterial species in the human gastrointestinal tract at any given time with considerable variations in the fecal flora from one individual to another [10]. It is known that more than 99% of all intestinal bacteria are obligate anaerobes [11] with B. fragilis being the commonest microorganism [12]. Among others Eubacteria, Enterobacteria, Enterococci, Lactobacilli and Clostridia belong to the most prevalent species [13]. The importance of intestinal microflora for bilirubin catabolism in the gut lumen was established in studies with germfree rats. These animals do not excrete any urobilinoids in their feces or urine [7], whereas conjugated bilirubins represent major fecal bile pigments [14]. Colonization of their gastrointestinal tract with indigenous microflora of conventional animals results in prompt production of fecal urobilinoids [15]. Another evidence for the role and importance of bacteria in bilirubin catabolism in the intestinal lumen comes from studies of subjects treated with antibiotics. It was demonstrated that oral antibiotics eliminate fecal production of urobilinoids as a result of eradication of intestinal microflora reducing bilirubin [16–18]. Despite the importance of intestinal catabolic pathway of bilirubin only little is known about its relation to bilirubin metabolism in the intravascular compartment in adult age and the effect of oral antibiotics on bilirubin-reducing
Fig. 1. Intestinal metabolism of bilirubin in adults under physiologic conditions. [This figure appears in colour on the web.]
239
Fig. 2. Intestinal metabolism of bilirubin in adults after eradication of bilirubin-reducing microflora. [This figure appears in colour on the web.]
microflora has never been related to the serum bilirubinemia. In the present study we investigated the influence of the intestinal microflora on serum bilirubin levels in hyperbilirubinemic Gunn rats (Figs. 1 and 2).
2. Material and methods 2.1. Animal experiments 1. The effect of eradication of bilirubin-reducing microflora with subsequent colonization of gastrointestinal tract with a strain of C. perfringens reducing bilirubin. Eleven male Gunn rats (RHA/jj) with congenital deficiency of bilirubin UDP-glucuronosyltransferase (weight range 250–270 g) were used in this study. Each animal was placed in a separate metabolic cage enabling collection of stools and preventing coprophagy. After an initial baseline phase I (1 week of feeding a normal diet) the rats received antibiotics (clindamycin (Dalacin C, Pharmacia/Upjohn, Belgium) 4 mg/day plus neomycine (Bykomycin, Byk Gulden, Germany) 30 mg/day)) via a gastric tube for subsequent 4 days (phase II). Finally, intestinal colonization was carried out with a single, bilirubin-reducing strain of C. perfringens isolated from neonatal feces (2) (1.5 ml of suspension of bacteria (5.5K7.0!106 cells in logarithmic phase of growth per ml of 2% yeast extract) administered via gastric tube for further consecutive four days (phase III). Antibiotic susceptibility of this strain was tested in vitro with Antibiogram Disks (Biorad, USA) for following antibiotics: clindamycin, co-trimethoxazole, gentamycine, amikacine, netilmycine, neomycine and tobramycine. Serum bilirubin and fecal excretion of bile pigments were determined at the end of each phase. Blood sampling was carried out from the ocular vein during ongoing experiment, whereas heart puncture was used for blood collection at the end of each experiment. Food intake, body weights and fecal excretion were recorded every day during the whole experimental periods. 2. The effect of eradication of bilirubin-reducing microflora with subsequent colonization of gastrointestinal tract with a strain of C. pasteurianum incapable of reducing bilirubin. Six male Gunn rats (RHA/jj) with similar characteristics and treated with clindamycin plus neomycine in a similar way as in the first study were used in this study. In contrast to the first study, intestinal colonization
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243
240
(phase III) was carried out with a single strain of C. pasteurianum (DSM 525, DSMZ, Germany) incapable of reducing bilirubin as demonstrated in previous studies (L. Vı´tek, unpublished results). Serum and fecal bile pigments were determined as described above. 3. The effect of oral administration of ATB not influencing bilirubinreducing microflora Six male Gunn rats (RHA/jj) with similar characteristics as in the first study were used in this study. In contrast to the first study, clindamycin/neomycine treatment was exchanged for co-trimethoxazole (Biseptol, Polfa, Poland, 27 mg/kg/day) not affecting aerobic microflora. Serum and fecal bile pigments were determined as described above. All aspects of the study met the accepted criteria for the humane care and experimental use of laboratory animals and all protocols were approved by the Animal Research Committee of the 1st Medical Faculty, Charles University of Prague.
2.2. Biochemical analyses Serum bilirubin and transaminase activities were determined on the automatic analyzer (model 717; Hitachi, Tokyo, Japan). Bilirubin was analyzed by the diazo reaction according to Jendrassik and Gro´f. Serum transaminase activities were determined enzymatically with a-ketoglutarate and alanine or aspartate as substrates for ALT and AST, respectively. Fecal bile pigments and pH were determined in fecal suspensions in saline (1:10 w/v). Fecal urobilinoids were determined by spectrophotometry as a zinc complexes of total urobilins [19]. Fecal bilirubin concentrations were determined by spectrophotometry as p-iodoanilin azopigments [20].
2.3. Statistical analysis Data are presented as meanGSD. The statistical significance of differences between variables was evaluated by paired t-test or Wilcoxon Signed Rank test when data were not normally distributed. Differences were considered statistically significant when P values were less than 0.05.
3. Results All animals remained in good health during the whole experimental period without substantial changes in animal body weights and/or deterioration of food intake. No signs of diarrhoea were reported during all experiments. Oral administration of clindamycin plus neomycine to hyperbilirubinemic Gunn rats resulted in disappearance of fecal urobilinoids indicative of the eradication of intestinal microflora capable of reducing bilirubin (148G48 vs. 0 nmol/d/100 g b.wt, PZ0.001, Table 1). Simultaneously, serum bilirubin increased dramatically (186G31 vs. 289G Table 1 Serum bilirubin and fecal excretion of bile pigments in Gunn rats after eradication of bilirubin-reducing microflora and subsequent colonization with C. perfringens reducing bilirubin
Control phase ATB treatment Colonization phase
Serum bilirubin (mmol/L)
Fecal bilirubin (nmol/d/100 g b.wt)
Fecal urobilinoids (nmol/d/ 100 g b.wt)
186G31 289G35* 239G17**
65G17 103G47*** 120G72
148G48 0**** 39G28*****
*PZ0.004, compared to control phase; **PZ0.013, compared to ATB treatment;***PZ0.04,comparedtocontrolphase;****PZ0.001,compared to control phase; *****PZ0.001, compared to ATB treatment.
Table 2 Bile pigment balance in Gunn rats after eradication of bilirubinreducing microflora and subsequent colonization with C. perfringens reducing bilirubin
Control phase ATB treatment Colonization phase
Serum bilirubin poola (nmol)
Fecal bile pigment pool (nmol)
558G93 867G105 717G51
554G121 277G128* 447G234**
*PZ0.002, compared to control phase; **PZ0.102, compared to ATB treatment. a Serum pools calculated according to Belcher and Harris [36].
35 mmol/l, PZ0.004) (Table 1). Intestinal colonization with a strain of C. perfringens reducing bilirubin led to reappearance of urobilinoids in feces to 26% of the initial values (0 vs. 39G28 nmol/d/100 g b.wt, PZ0.001, Table 1) and to a partial decrease of serum bilirubin (by 17.3%; from 289G35 to 239G17 mmol/l, PZ0.013) (Table 1). Disappearance of urobilinoid production by intestinal microflora eradicated during antibiotic (ATB) therapy resulted in accumulation of bilirubin within intestinal lumen as evidenced by increase of fecal bilirubin output (65G17 vs. 103G47, PZ0.04, Table 1). However, total fecal bile pigment output decreased significantly after ATB treatment (554G121 vs. 277G128 nmol/d, PZ0.002) (Table 2). Subsequently, decrease of serum bilirubin levels after colonization of intestinal tract with bacteria reducing bilirubin was accompanied with an increase of fecal bile pigment excretion (Table 2). However, when the intestinal tract of Gunn rats was colonized with a strain of C. pasteurianum incapable of reducing bilirubin, increased serum bilirubin levels decreased only insignificantly (by 4.6%; from 281G58 to 269G34 mmol/l, PZ0.625) (Table 3). Simultaneously, colonization with this strain of C. pasteurianum did not lead to reappearance of fecal urobilinoids (Table 3). Direct link between bilirubin-reducing microflora and serum bilirubin homeostasis was further corroborated by the results of another study when co-trimethoxazole was used as an antibiotic not affecting bilirubin-reducing microflora. In this study both serum bilirubin and fecal urobilnoids were not markedly influenced by this antibiotic (Table 4). Serum bilirubin levels increased by 13% (from 138G59 to Table 3 Serum bilirubin and fecal excretion of urobilinoids in Gunn rats after eradication of bilirubin-reducing microflora and subsequent colonization with C. pasteurianum incapable of reducing bilirubin
Control phase ATB treatment Colonization phase
Serum bilirubin (mmol/l)
Fecal urobilinoids (nmol/d/100 g b.wt)
153G20 281G58* 269G34**
132G45 0*** 0***
*PZ0.01, compared to control phase; **PZ0.625, compared to ATB treatment; ***PZ0.001, compared to control phase.
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243 Table 4 Serum bilirubin and fecal excretion of urobilinoids in Gunn rats after treatment of ATB not influencing bilirubin-reducing microflora
Control phase ATB treatment
Serum bilirubin (mmol/l)
Fecal urobilinoids (nmol/d/100 g b.wt)
138G59 156G69*
127G85 121G52**
*PZ0.55, compared to control phase; **PZ0.91, compared to control phase.
156G69 mmol/l, PZ0.55, Table 4), which is within a physiological variation of serum bilirubin levels in Gunn rats as evidenced by results of a separate study (data not shown). On contrary, highly significant increases by 55% (153G20 vs. 281G58 mmol/l, PZ0.01) and 84% (186G31 vs. 289G35 mmol/l, PZ0.004) were demonstrated in both clindamycin/neomycine studies, respectively (Tables 1 and 3). In in vitro experiments a strain of C. perfringens used for colonization was found to be susceptible only for clindamycin, aminoglycosides (with the exception of slight suppressing effect of netilmycine), as well as co-trimethoxazole were ineffective. Fecal pH did not change markedly during all the study phases (a range between 6.6 and 7.2 without substantial differences attributable to any individual study phase). No deterioration of serum transaminase activities was detected during antibiotic treatment (ALT: 1.6G0.3 vs. 2.0G0.3 vs. 1.7G0.2 mkat/l, PO0.05, for Phases I, II, and III, respectively; AST: 3.5G0.5 vs. 3.5G0.5 vs. 2.2G0.4 mkat/l, PO0.05, for Phases I, II, and III, respectively).
4. Discussion The relationship between the intestinal catabolism of bilirubin and serum bilirubin levels has been described only in newborn infants [1,2]. In the present study we demonstrate a profound effect of oral clindamycin/neomycine therapy administered to hyperbilirubinemic Gunn rats on eradication of intestinal microflora reducing bilirubin resulting in inhibition of bilirubin catabolism in the intestinal lumen. As a consequence, the levels of serum bilirubin rose remarkably. Increase in serum bilirubin after oral ATB treatment eradicating bilirubin-reducing microflora was accompanied by a drop of fecal bile pigment excretion with accumulation of bilirubin within intestinal lumen (Table 1). Furthermore, colonization of intestinal tract of Gunn rats with a strain of C. perfringens reducing bilirubin had a prompt reverse effect on serum bilirubin levels as well as fecal urobilinoid production, while no such effect could be demonstrated after colonization of gastrointestinal tract with C. pasteurianum incapable of reducing bilirubin. Similar lacking effect on bilirubin metabolism was observed also after oral administration of antibiotic not affecting bilirubin-reducing microflora further supporting
241
the importance of bilirubin-reducing microflora on intravascular bilirubin homeostasis. Under physiological conditions of bilirubin disposition via hepatocytes bilirubin might be reabsorbed into portal circulation and undergo enterohepatic circulation. A substantial portion of bilirubin reabsorbed from the gut lumen can spill over into the systemic circulation because of its low z30% single-pass hepatic extraction [21] (Figs. 1 and 2). However, it is necessary to note that under conditions of deficient bilirubin conjugation in the liver tissue such as in Gunn rats, where the major disposition route for bilirubin is direct permeation through the intestinal wall [22], the effect of inhibition of bilirubin degradation resulting in its accumulation in the intestinal lumen might lead in even more pronounced elevation of serum bilirubin levels. In fact, inverse relationship between serum bilirubin levels and fecal bile pigment excretion is in support of direct link between intestinal and intravascular metabolism of bilirubin. As can be calculated from Table 2, clindamycin/ neomycine treatment increased serum bilirubin pool by 309 nmol, while daily fecal bile pigment output decreased by 277 nmol. Similar balanced state can be seen between treatment and colonization phases (decrease of serum bilirubin pool by 150 vs. increase of intestinal bile pigment pool by 170 nmol). These data indicate that intestinal disposition of bile pigments was markedly disturbed by clindamycin/neomycine treatment. Since it was suggested that there is a dynamic equilibrium between serum bilirubin levels and bilirubin concentrations within intestinal lumen [22], based on our results it seems that treatment with antibiotics affecting bilirubin-reducing microflora impairs bilirubin disposition in Gunn rats. This is supported by the fact that intestinal colonization with a single strain of C. perfringens reducing bilirubin led to a reappearance of urobilinoids in feces and to a partial, but significant decrease of serum bilirubin demonstrating thus the direct role of intestinal microflora on bilirubin homeostasis, while no such effect could be detected in animals colonized with a single strain of C. pasteurianum incapable of reducing bilirubin. It is highly likely that treatment with antibiotics affecting bilirubin-reducing microflora might influence serum bilirubin levels also under physiological conditions of bilirubin disposition via hepatocytes by enhancement of enterohepatic and enterosystemic circulation of bilirubin, i.e. by a principle similar to that described above for Gunn rats. Our data are consistent with results of previous studies performed on germfree and conventional rats. As stated above, germfree rats do not excrete any urobilinoids in their feces or urine [7], whereas conjugated bilirubins represent major fecal bile pigments [14]. However, one day after conventialization fecal urobilinoids became positive reaching physiologic values 7 days later [15]. Interestingly, after monocontamination of germfree rats with C. ramosum isolated from intestinal contents of conventionalized rats approximately 10% of the bilirubin was excreted in the feces as urobilinoids. The fecal output of
242
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243
urobilins increased to 30% after superinfection of monocontaminated ex-germfree rats with a strain of E. coli, but never reached the values found in the conventional rats (i.e. almost 80% of bilirubin excreted as fecal urobilinoids). Interestingly, after contamination of germfree rats with feces from conventional rats, the ex-germfree rats produced urobilins to the same extent as conventional rats [7]. It is important to note that a strain of E. coli used in this study was not able to produce urobilinoids alone and only potentiated production by C. ramosum indicating that complex interactions between several specific microorganisms might be responsible for high production of urobilinoids seen in conventional rats as well as adult healthy humans. These data are consistent with our findings that colonization of sterile gastrointestinal tract of Gunn rats with a strain of C. perfringens reducing bilirubin resulted only in partial restoration of urobilinoid production. Interestingly, no urobilinoid production was observed after colonization of intestinal tract with strains of Peptostreptococcus species [23], Lactobacillus acidophilus, Bifidobacterium bifidum [24], C. welchii type A, Streptococcus faecalis, Proteus vulgaris, Escherichia coli, Bacillus subtilis, Sarcina and Mucor [7]. In our in vitro studies no bilirubin-reducing activity was detected in various clostridial strains including C. pasteurianum, C. acetobutylicum, C. novyii, C. paraputrificum, C. sporogenes, C. bifermentans, C. scindens, C. hiranonis and C. sordelii (L. Vı´tek, unpublished results). Based on these limited data it seems that bilirubin reduction within gastrointestinal tract is confined to bacterial species of C. perfringens, C. difficile, C. ramosum and possibly B. fragilis [2,7,8]. In a study by Tally et al. 86% of 49 strains of C. ramosum tested was found to be sensitive to clindamycin [25]. In another study Clostridia were reported to be sensitive to clindamycin, ampicillin and erythromycin [26]. Oral treatment with clindamycin was shown to cause fecal bile pigment pattern similar to that of germfree animals [16] with return of urobilinoid production to conventional values within 4–6 weeks [27]. It was demonstrated by Heimdahl and Nord that clindamycin strongly disturbs microbial colonization causing at least a 100-fold increase in indigenous facultative anaerobic flora [28]. The effect of antibiotics given orally for 6 days to healthy humans was investigated also in studies by Saxerholt et al. and Midtvedt [16,17]. Intake of bacitracin, vancomycin, clindamycin, erythromycin and ampicillin resulted in a pronounced suppression of fecal urobilinoid excretion, whereas intake of doxycycline, metronidazole, nalidixic acid, ofloxacin and co-trimethoxazole had no significant effect. Intake of clindamycin also led to a marked increase of conjugated bilirubin in the feces and this drug was the only one present in high concentrations in the feces of subjects tested [16]. Similar suppressive effect on urobilinoid producing microflora was demonstrated in earlier human studies for aureomycin [17], chloromycetin [29] or sulphasalazine [30] as well as for a number of other antibiotics including
penicillins, tetracyclins or aminoglycosides tested in animal studies [31]. These findings correlate well with available data that only Clostridia are capable of reducing bilirubin [2,7,32–35]. It can be implied from the list of antibiotics listed above, that many of the antimicrobial agents used in human medicine may have potential to affect serum bilirubin levels (Fig. 2). Although this effect will be negligible in otherwise healthy subjects, it might be of importance in patients with other underlying clinical conditions influencing metabolism of bilirubin. According to our knowledge this is the first report demonstrating direct link between eradication of intestinal microflora capable of reducing bilirubin and serum bilirubin levels. On the basis of our data it may be concluded that the intestinal microflora may substantially affect serum metabolism of bilirubin not only in newborn infants but also in adult age. Prolonged use of antibiotics in man may lead to an increase of serum bilirubin levels, while enhancement of its intestinal catabolism may have an opposite effect.
Acknowledgements The authors thank Mrs. Anna Fliegerova´ and Jirˇ´ı Oborˇil for excellent technical assistance. This work was supported by the research grant GACˇR No. 310/021436 (Granting Agency of the Czech Republic).
References [1] Poland RL, Odell GB. Physiologic jaundice: the EHC of bilirubin. New Engl J Med 1971;284:1–6. ˇ erna´ M, Chmelarˇ D, et al. [2] Vı´tek L, Kotal P, Jirsa M, Malina J, C Intestinal colonization in neonates leading to fecal urobilinoid excretion may play a role in the pathogenesis of neonatal hyperbilirubinemia. J Pediatr Gastroenterol Nutr 2000;30:294–298. [3] Moscowitz A, Weimer M, Lightner A, Petryka ZJ, Davis E, Watson CJ. The in vitro conversion of bile pigments to the urobilinoids by a rat Clostridia species as compared with human fecal flora III. Natural d-urobilin, synthetic i-urobilin, and synthetic iurobilinogen. Biochem Med 1970;4:149–170. [4] With TK. Bile pigment in feces. Bile pigments. Chemical, biological and clinical aspects. New York: Academic Press; 1968 p. 585–613. [5] Norin KE, Gustafsson BE, Lindblad BS, Midtvedt T. The establishment of some microflora associated biochemical characteristics in feces from children during first years of life. Acta Paediatr Scand 1985;74:207–212. [6] Midtvedt AC, Duke BC, Norin KE, Saxerholt H, Midtvedt T. Development of five metabolic activities associated with the intestinal microflora of healthy infants. J Pediatr Gastroenterol Nutr 1988;7: 559–567. [7] Gustafsson BE, Lanke LS. Bilirubin and urobilins in germfree, exgermfree and conventional rats. J Exp Med 1960;112:975–981. [8] Fahmy K, Gray CH, Nicholson DC. The reduction of bile pigments by faecal and intestinal bacteria. Biochim Biophys Acta 1972;264:85–97. [9] Moore WEC, Holdenman LV. Human fecal flora: the normal flora of 20 Japanese-Hawaiians. Appl Microbiol 1974;27:961–979.
L. Vı´tek et al. / Journal of Hepatology 42 (2005) 238–243 [10] Holdeman LV, Good IJ, Moore WEC. Human fecal flora: Variation in bacterial composition within individual and a possible effect of emotional stress. Appl Environ Microbiol 1976;31:359–375. [11] Hylemon PB. Matebolism of bile acids in intestinal microflora. In: Danielsson H, Sjo¨vall J, editors. Sterols and bile acids. Amsterdam: Elsevier Science; 1985. p. 331–343. [12] Hayakawa S. Microbial transformation of bile acids. Adv Lipid Res 1973;11:143–192. [13] Hill MJ, Drasar BS. The normal colonic bacterial flora. Gut 1975;16: 318–323. [14] Saxerholt H, Skar V, Midtvedt T. HPLC separation and quantification of bilirubin and its glucuronide conjugates in faeces and intestinal contents of germ-free rats. Scand J Clin Lab Invest 1990; 50:487–495. [15] Saxerholt H, Midtvedt T, Gustafsson BE. Deconjugation of bilirubin conjugates and urobilin formation by conventionalized germ-free rats. Scand J Clin Lab Invest 1984;44:573–577. [16] Saxerholt H, Carlstedt-Duke B, Hoverstadt T, Lingaas E, Norin KE, Steinbakk M, et al. Influence of antibiotics on the faecal excretion of bile pigments in healthy subjects. Scand J Gastroenterol 1986;21: 991–996. [17] Midtvedt T. Effects of antimicrobial agents upon functional part of the intestinal flora. Scand J Infect Dis Suppl 1986;49:85–88. [18] Sborov VM, Jay AR, Watson CJ. The effect of aureomycin on urobilinogen formation and the fecal flora. J Lab Clin Med 1951;37: 52–59. [19] Kotal P, Fevery J. Quantitation of urobilinogen in faeces, urine, bile, and serum by direct spectrophotometry of zinc complex. Clin Chim Acta 1991;202:1–10. [20] Heirwegh KPM, Fevery J, Meuwissen J, de Groote J. Recent advances in the separation and analysis od diazo-positive bile pigments. Methods Biochem Anal 1974;22:205–250. [21] Ga¨rtner U, Goeser T, Wolkoff AW. Effect of fasting on the uptake of bilirubin and sulfobromophtalein by isolated perfused rat liver. Gastroenterology 1997;113:1707–1713. [22] Kotal P, van der Veere C, Sinaasappel M, Elferink RO, Vı´tek L, Brodanova´ M, et al. Intestinal excretion of unconjugated bilirubin in man and rats with inherited unconjugated hyperbilirubinemia. Pediatr Res 1997;42:195–200.
243
[23] Saxerholt H. The normal microflora and bilirubin. In: Grubb R, Midtvedt T, Norin E, editors. The regulatory and protective role of the normal microflora. New York: Stockton Press; 1988. p. 169–201. [24] Norin KE, Persson AK, Saxerholt H, Midtvedt T. Establishment of Lactobacillus and Bifidobacterium species in germfree mice and their influence on some microflora-associated characteristics. Appl Environ Microbiol 1991;57:1850–1852. [25] Tally FP, Armfield AY, Dowell VR, Kwok YY, Sutter VI, et al. Suseptibility of Clostridium ramosum to antimicrobial agents. Antimicrob Agents Chemother 1974;5:589–593. [26] Appelbaum PC, Chatterton SA. Susceptibility of anaerobic bacteria to ten antimicrobial agents. Antimicrob Agents Chemother 1978;14: 371–376. [27] Midtvedt T. Clindamycin-associated alterations in intestinal microfloraassociated characteristics in rats. Scand J Gastroenterol 1985;20:92–98. [28] Heimdahl A, Nord CE. Effect of erythromycin and clindamycin on the indigeneous human anaerobic flora and new colonization of the intestinal tract. Eur J Clin Microbiol 1982;1:38–48. [29] Kolb FO, Oppenheim DJ. Disappearance of urobilinogen during antibiotic therapy. Bull New Engl Med Center 1950;12:31. [30] Leijonmarck CE, Benno P, Carlstedt-Duke B, Monsen U, Norin E, Poppen B, et al. The function of the intestinal microflora in patients with ulcerative colitis before and after colectomy. Scand J Gastroenterol 1990;25:585–593. [31] Gustafsson BE, Norin KE. Development of germfree animal characteristics in conventional rats by antibiotics. Acta Pathol Microbiol Scand Sect B 1977;85:1–8. [32] Passini F, Czakes J. Ueber Urobilinbildung durch Reinkulturen anaerober Darmbakterien. Wien Klin Wochnschr 1923;36:657–659. [33] Ka¨mmerer H, Miller R. Ueber die Umwandlung der Gallenfarbstoffe durch fa¨ulniserregende Darmbakterien. Wien Klin Wochnschr 1922; 35:639–640. [34] Ka¨mmerer H, Miller K. Zur enterogenen urobilinbildung. Deutsche Arch Klin Med 1923;141:318–347. [35] Watson CJ. The urobilinoids: Milestones in their history and some recent developments. In: Berk PD, Berlin NI, editors. The chemistry and physiology of bile pigments. Washington, DC: US Govern Print Off; 1977. p. 469–480. [36] Belcher EH, Harris EB. Studies of plasma volume, red cell volume and total blood volume in young growing rats. J Physiol 1957;139:64–78.