- Email: [email protected]
Caffeine demethylation measured by breath analysis in experimental liver injury in the rat
Journal of Hepatology 1995; 22: 82-87 Printed in Denmark AN rights reserved Munksgaard Copenhagen
Journalof Hepatology ISSN 0168-8278
Caffeine demethylation measured by breath analysis in experimental liver injury in the rat Heinz J. Schaad’,
Eberhard
L. Renner’,
‘Department of Clinical Pharmacology,
Hubertus
Wietholtz’,
Maurice J. Arnaud2 and Rudolf Preisig’
University of Berne, Berne and ‘Nestle Research Center, Nestec Ltd., Vevey, Switzerland
To assess the effects of experimental liver injury on caffeine metabolism, 1 &i/kg hw. of [3-methyP4C]caffeine (together with 5 mg/kg bw. of the cold compound) was injected i.p to four different experimental groups and respective controls of unanesthetized male Sprague-Dawley rats. Exhaled 14C02 was completely collected during 4 h and peak exhalation rate and fraction of dose recovered were calculated. l/3 hepatectomy affected 14C02 exhalation to a limited extent, decreasing solely peak exhalation rate (~~0.05 compared to sham-operated controls). 2/3 hepatectomy, on the other hand, resulted in significant reduction QKO.01) in both peak exhalation rate (by 59%) and fraction of dose recovered (by 47%), that were proportionate to the loss of liver mass (59%). End-to-side portocaval shunt led to the well-documented hepatic “atrophy”, liver weight being dhuinished on average to 50% within 2 weeks of surgery;
however, reductions in peak exhalation rate (by 75%) and fraction of dose recovered (by 64%) were even more pronounced. Finally, 48 h bile duct ligation was equivalent to “functional 2/3 hepatectomy”, peak exhalation rate (by 65%) and fraction of dose recovered (by 56%) being markedly diminished despite increased liver weight. These results indicate that 14C02 exhalation curves following administration of specifically labelled caffeine are quantitative indicators of acute or chronic loss of functioning liver mass In addition, the 3-demethylation pathway appears to be particularly sensitive to the inhibitory effects of cholestasis on microsomal function.
B
exclusively hepatic metabolism (1,2), 1,3,7-tri-methylxanthine (caffeine) has been proposed as a model compound for measuring liver function in man (3). Indeed, comparison of caffeine clearance, galactose elimination capacity and aminopyrine breath test in patients with various forms of liver disease has suggested that caffeine may represent an almost ideal test compound for assessing hepatic microsomal function quantitatively (3,4). In the past a series of animal studies explored the metabolic rate of caffeine within the body (2,5,6) and the effect(s) of inducing agents on hepatic caffeine metabolism (7,8). The specific influence on hepatic caffeine disposition of common features of liver disease,
such as reduced functioning liver cell mass or cholestasis, however, has thus far not been systematically determined, despite the fact that such knowledge may help interpretation of altered caffeine metabolism in human liver disease. We therefore decided to investigate the effects of partial hepatecotomy, bile duct ligation and end-to-side portocaval shunt on hepatic caffeine metabolism in the rat. Based on previous studies from this laboratory (8), we determined demethylation of caffeine in position 3, the predominant metabolic pathway in the rat (9), by quantitating 14C02 exhalation following application of specifically labelled [3-methyl-i4C]-caffeine.
Received 27 January 1994
Material
Correspondence
Animals
ECAUSE of its almost
of Medicine, Switzerland.
82
to:
Heinz J. Schaad, M.D., Department
University
Hospital,
3010 Berne-Inselspital,
Key words: Bile duct ligation; Breath analysis; Caffeine demethylation; Hepatectomy; Liver disease models; Liver function; Portocaval anastomosis; Rat. 0 Journal of Hepatology.
and Methods
All studies were performed in male Sprague-Dawley rats (Deutsche Versuchstierfarm, Tuttlingen, Ger-
Caffeine test in model liver diseases
many) weighing 150 to 350 g. Animals were kept at constant temperature (22°C) and humidity (48?2%), under an artificial 12-h light-dark cycle. They were housed in groups of 5-6 per cage, except for the portocaval shunt rats (and their respective sham-operated controls) which were kept in individual cages to prevent mutilation caused by aggressive behavior. All animals were allowed free access to standard rat chow (Altromin@ 300, Altromin GmbH, Lage, Germany) and tap water until 12 h prior to the experiment. “C-labelled caffeine [3-methyl-14C]-caffeine was prepared at the Nestle Research Center according to methods described previously (8). The final preparation contained less than 0.1% radiochemical impurity. Labelled caffeine was dissolved in 0.9% NaCl to a final concentration of 1 &i/ml, and stored at 4°C until use. Experimental models Partial hepatectomy was carried out via a midline abdominal incision under light ether anesthesia. The vessels to the median lobe (l/3 hepatectomy) and the left lobe (2/3 hepatectomy) of the liver were ligated en bloc and the lobe(s) was excised and weighed. Simple opening and closing of the abdomen using a midline incision served as a sham operation for the respective controls; care was taken to assure that the duration of anesthesia was comparable to the hepatectomy procedure (about 10 min). Bile duct ligation was performed via midline abdominal incision under light ether anesthesia. Following ligation of the common bile duct proximally and distally, the intervening small piece (about 1 cm) was excised. Sham operation consisted of a simple midline laparotomy. End-to-side portocaval anastomosis was performed under ether anesthesia after ligation of the gastroduodenal vein and the proximal vena cava, as previously described (10). As corresponding sham operation the gastroduodenal vein was ligated and the portal vein clamped for 15 min.
ate, Dr. G. Gattiker, Pharmazeutische Praeparate, Zurich, Switzerland). Exhaled CO2 was completely collected into a series of interchangeable counting vials connected to each individual restraining cage. The air was continuously sucked out of the cages by a vacuum pump through the collection vials where CO2 was trapped with ethanolamine, as described previously (498). At the end of the experiment, animals were anesthetized with i.p. pentobarbital (50 mgkg). A blood sample was taken for determination of serum levels of liver enzymes, bilirubin and bile acids, followed by sacrifice of the animals with an overdose of i.p. pentobarbital. The liver (and testes of portocaval anastomosis and sham portocaval anastomosis rats) was excised and weighed, and portocaval anastomosis shunt patency verified by probing. Analytical procedures Radioactivity of 14C in breath samples was counted following the addition of a scintillator (8) in a Packard liquid scintillation counter (TRI-CARB 2660). Total serum bilirubin and serum enzymes were measured in the Center Hospital Laboratory using routine methods, total serum bile acids with a commercially available radioimmunoassay (RIA)-kit (BectonDickinson, Orangeburg, NY, USA). Calculations 14C02 exhalation rate was expressed as percent of the dose per minute, the highest value being taken as peak elimination rate. The fraction of the dose recovered was obtained by calculating the area under the exhalation-time curve (AUC) between 0 and 240 min using the trapezoidal rule. A single value, namely the exhalation rate at 60 min (ER& was arbitrarily chosen for comparison with fraction of the dose recovered. All results are presented as mean2S.D. Group comparisons are based on Student’s t-test with Welch’s correction in the case of uneven distribution. PcO.05 was regarded as statistically significant.
Results Experimental design Caffeine breath tests were performed 1 h after recovery from anesthesia and surgery in partial hepatectomy animals, 48 h after bile duct ligation in bile duct ligated rats and 14 days after surgery in porto-caval anastomosis animals (and respective sham-operated controls), respectively. Animals were fasted overnight and placed in restraining cages. Following i.p. injection of 1 ,&i/ kg b.w. of [3-methyl-14C]-caffeine together with 5 mg/ kg b.w. of the cold compound (caffeine sodium benzo-
Characterization of experimental models By design, body weights of experimental groups and respective sham-operated controls (C) are almost identical, with the exception of portocaval anastomosis animals (Table 1). Such reduction (on the average 15%) in body weight 2 weeks after portocaval anastomosis is well known (11). Liver weights in partially hepatectomized animals were reduced to 4.82kO.42 g (versus 7.1120.87 g in C) by l/3 hepatectomy and to 3.1OkO.29 g (versus
83
H. J. Schadd et al.
7.4820.69 g in C) by 213 hepatectomy, respectively (Table 1). In bile duct ligated animals, total serum bilirubin was increased to 106.2236.0 pmol/l (versus 4.3~3.6 PmoVl in C), total serum bile acids to 168?7 1 ,umolll (versus 2.72 1.3 pmol/l) aspartate aminotransferase to 784t290 IU/l (versus 81? 16 IU/l) and alkaline phosphatase to 851298 IU/l (versus 455-+ 17 IU/l), respectively, thus documenting severe cholestasis. Functioning of the shunt in portocaval anastomosis rats is indicated by the expected increase in total serum bile acid levels (40246 versus 6.6k6.8 pmol/l in C) and the decrease in liver weight (4.7820.81 versus 9.30-+ 1.23 g) and in testicular weight (1.75kO.13 versus 2.7120.47 g), respectively (10,12). In addition, all shunts were patent at autopsy.
14C02 exhalation
curves
The 14C02 breath curves in l/3- and 2/3-hepatectomized animals, compared with their respective C, are depicted in Fig. 1. Whereas the curves of both C groups are superimposable with peak exhalation rate reached after 70 to 80 min, reduction in liver mass due to hepatectomy results in plateauing at a lower level of the 14C02 exhalation rate, which is much more pronounced following 2/3 hepatectomy. Thus, following 2/ 3 hepatectomy, peak exhalation rate - once reached is practically maintained throughout the period of observation. A virtual plateau at an even lower level, also represents the main effect of bile duct ligation and portocaval anastomosis on the 14C02 exhalation rate (Fig. 2), while the breath curves of all experimental groups are characterized by reduction in the initial specific 14C02 activity, evident already in the breath sample
‘4co*
O.WO-
Exhatatlon rate
/
Sham operated
Sham 0p.rat.d
0.100. o.oso% dosmlmtn l/S
Howt*ctomy
O.OlO_
I
-
0
120
240
0
-2/a
H.p*t~ctomy
120
240
NlRYt..
Fig. 1. Exhalation rates of 14C02, plotted against time, in rats following IN-hepatectomy (n=6, left panel) and 2/3hepatectomy (n=6, right panel) and in respective shamoperated controls (n=9 and 6, respectively). 84
Caffeine test in model liver diseases “Co,
0.500
Sham operatad
Shpm opwrtod 0.100 0.050 % do**/mln
611. duct Ilpptlon 0.010 0.005
0
120
240
0
120
240
Yhlutes
Fig. 2. Exhalation rates of “C02, plotted against time, in rats following porto-caval anastomosis (n=6, left panel) and bile-duct ligation (n=8, right panel), and in respective sham-operated controls (n=6 and 6, respectively).
collected during the first 10 min following 14C-caffeine injection. Pharmacokinetic
calculations
The results of peak exhalation rate, fraction of dose recovered, and ERm for all experimental groups are given in Table 1. There is no significant difference in any of these parameters between the various C groups. This attests to the reproducibility of the 14C2 caffeine breath test as performed. Whereas the effects of l/3-hepatectomy are limited to a decrease in peak exhalation rate of caffeine-3-demethylation, acute (2/3-hepatectomy) or chronic (portocaval anastomosis) reduction in liver mass exceeding 33% results in markedly diminished caffeine metabolism. Thus, following 2/3-hepatectomy peak exhalation rate and fraction of dose recovered are decreased in parallel with the change in liver weight (59%), on average by 59% and 47%, respectively. In animals with portocaval anastomosis, the reduction in peak exhalation rate (75%) and fraction of dose recovered (64%), in fact, slightly exceeds that in liver weight (49%) pointing to metabolic changes over and above those in hepatocyte mass (11). The divergence of liver weight and metabolic capacity, however, is most pronounced in animals with bile duct ligation. Despite a “gain” in liver weight of over 40%, peak exhalation rate is diminished by two thirds and fraction of dose recovered by 56%, respectively, presumably reflecting tissue edema, functional alteration(s) in and loss of hepatocytes. On the basis of an inspection of all 14C02 exhalation curves, it appeared reasonable to choose ERso as a potentially useful single point measurement. In general, its values reflect fraction of dose recovered (r2=0.972),
but tend to more closely parallel peak exhalation rate (r2=0.987) (Table 1). Since fraction of dose recovered reflects an integral of the 14C02 exhalation rate over time, this parameter was used for direct comparison with liver weight (Fig. 3). Whereas fraction of dose recovered is decreased roughly in parallel with surgical reduction of liver mass (l/3- and 2/3-hepatectomy, portocaval anastomosis), cholestasis (bile duct ligation) produces a functional defect approximately equivalent to 2/3_hepatectomy, in face of an increased liver weight.
Discussion The aim of the present study ‘was to explore the effects of a reduced liver cell mass and of cholestasis on caffeine metabolism in three well-established rat liver disease models (11-17). Our results can be summarized as follows: Acute, partial hepatectomy results in a reduction of 14C02 exhalation, which is proportionate to the amount of functioning liver mass removed. Peak rate expiration appears to be primarily affected, while fraction of dose recovered is only significantly reduced after 2/3-PH. Chronic liver “atrophy” due to porto-caval shunt, as expected, decreases 14C02 exhalation rate. The fact that the functional change tends to exceed the loss of liver weight is in keeping with previous studies (13,14). Indeed, the presence of mild cholestasis suggested by increased serum bile salt levels may have additionally inhibited microsomal function. Markedly diminished 14C02 exhalation despite in-
25-
14
co
sham opommd
Exhalation
20-
%do55/4h mte
,5
113ltqwwtomy w 2/9l+wpl=@my
10 -
4
s,
/
’
/
/
’
’
‘P Porlomv8I~hunt
Oh7777 bar walght %b.w.
Fig. 3. Relationship between the cumulative 14C0, exhalation rate over 4 h (FD) and liver weight (expressed as % of body weight) in the various experimental groups studied. The dotted line (- - -) represents an imaginary line the
fraction of dose recovered would follow if changes in fraction of dose recovered were in parallel with changes in liver weight. 85
H. J. Schadd et al.
creased liver weight characterizes bile duct ligated animals. Although structural damage (leading to hepatocyte loss) undoubtedly occurred, deterioration of microsomal function in this model is primarily a consequence of cholestasis (16,17), an effect documented also in previous investigations using aminopyrine (11). The complexities of both the breath test and the experimental models of liver injury employed merit emphasis. Thus, although demethylation in position 3 yielding paraxanthine is the predominant primary pathway of caffeine metabolism in the rat (9), exhaled i4C02 following injection of [3-methyl-‘4C]-caffeine also originates from 3-demethylation of other caffeine demethylation products, i.e. theobromine and theophylline, being converted to 7-methylxanthine and lmethylxanthine, respectively (9,18). The observed breath curves, therefore, reflect a cascade of metabolic events, rather than a single process. Nevertheless, the flatness of exhalation curves in 2/3_hepatectomized, portocaval anastomosis and bile-duct ligated animals suggests that saturation of caffeine 3-demethylation was reached with the dose of caffeine used in these models. It should be recognized that the time interval chosen for determination of fraction of dose recovered (4 h) is somewhat arbitrary and reflects the sum of xanthine 3-demethylation over this time-period. Since fraction of dose recovered represents an AUC, it is a measure of i4C02 kinetics independent of pharmacokinetic models and observer bias. Also, it is recognized that theoretically fraction of dose recovered will become equal in all experimental groups, provided it is measured over a long enough period. In addition to fraction of dose recovered values, peak exhalation rate and exhalation rate at 60 min were therefore determined. Finally, conversion of methyl groups to CO2 via the intermediates formaldehyde and formate (19) although considered not to be rate-limiting - may be affected by various, as yet ill-defined factors inherent in the liver disease models used. While surgical trauma per se might affect caffeine metabolism, performance of breath tests 1 h after partial hepatectomy appears sufficient to allow complete recovery of the animals from anesthesia. The lack of any significant difference in the breath curves of the four C groups further underscores this notion. Clearly, the time elapsed between hepatectomy and performance of breath tests was too short to account for metabolic effect of regenerative processes (20). Establishment of an end-to-side porto-caval shunt is associated with major systemic sequelae, as evidenced by weight loss, signs of “encephalopathy” and testicular atrophy (10,12). These factors, together with (in 86
part disparate) hepatic functional changes (IO, 13-l 5) presumably affected microsomal processing of caffeine over and above the loss of liver mass. Animals 48 h after bile duct ligation exhibit no weight loss and appear well despite cholemia. In this sense, bile duct ligation may represent the “purest model” of hepatic injury. Caffeine 3-demethylation is mediated by cytochrome P-450IA2 (21-23). It is well known that the overall content of cytochrome P-450 is reduced in the bile duct ligated rat (24,25). Certain P-450 isoforms have been shown to be selectively diminished in certain disease states (26). Our data would be entirely compatible with the caffeine-metabolizing P-450IA2 being overproportionately decreased in cholestasis. Since P-450 isoforms were not determined in the present study, however, this remains entirely speculative. Interestingly a more profound decrease in P-4501A in cholestatic compared to hepatocellular cirrhosis in man has been described (27). Despite these limitations and uncertainties, our studies do allow us to link the effects of experimental hepatic injury on caffeine disposition in rats with those of liver disease in man. In man, specific activity of 14C02 in breath following injection of 14C-caffeine is highly correlated with caffeine plasma clearance over a wide range of values (3,28). In the rat, 14C02 output was continuously and completely collected. The resulting, integrated breath curves should, therefore, represent caffeine metabolism even more fully. It seems reasonable to assume that in the rat cumulative 14C02 excretion in breath closely reflects hepatic caffeine clearance. The clearance of a low extraction compound such as caffeine, exhibiting a constant volume of distribution and being eliminated almost exclusively by hepatic metabolism, is expected to be diminished primarily by loss of hepatocyte mass (29). This was confirmed in the present study in animal models leading to acute (partial hepatectomy) and chronic reduction in liver weight (portocaval anastomosis). It is, therefore, suggested that reduction in caffeine clearance previously observed in patients with cirrhosis (3) primarily reflects partial functional hepatectomy. Finally, the profound effect of cholestasis on caffeine metabolism may render caffeine clearance measurements a sensitive indicator of functional deterioration in acute or chronic liver disease in man which, besides a reduction in functioning liver cell mass, are often associated with cholestatic features. Since cytochrome P-450IA2 plays an important role in the genesis of carcinogenic compounds (21), caffeine has gained interest as a test compound for the assessment of cytochrome P-450IA2 activity in man (30).
Caffeine test in model liver diseases
This study describing the influence of various models of liver disease on the 3-demethylation activity may be helpful in the interpretation of results from tests in humans affected by liver diseases.
Acknowledgements The authors wish to thank Ms. M. Kappeler for the art work and Ms. M. Sommer for secretarial help in preparing the manuscript. E.L.R. is the recipient of a SCORE award from the Swiss National Foundation for Scientific Research. Supported by the Swiss National Foundation for Scientific Research.
References 1. Arnaud MJ, Welsh C. Theophylline and caffeine metabolism in man. In: Rietbrock N, Woodcock BG, Staib AH, eds. Theophylline and Other Methylxanthines. Braunschweigi Wiesbaden, Germany: Vieweg and Son, 1981: 13548. 2. Bonati M, Garattini S. Interspecies comparison of caffeine disposition. In: Dews PB, ed. Caffeine: Perspectives from Recent Research. Berlin, Germany: Springer Verlag, 1984: 4856. 3. Renner E, Wietholtz H, Huguenin P et al. Caffeine: a model compound for measuring liver function. Hepatology 1984; 4: 3846. 4. Jost G, Wahllaender A, von Mandach U, et al. Overnight salivary caffeine clearance: a liver function test suitable for routine use. Hepatology 1987; 7: 33844. 5. Aldridge A, Neims AH. Relationship beetween the clearance of caffeine and its 7-N-demethylation in developing beagle puppies. Biochem Pharmacol 1980; 29: 1909-14. 6. Ferrer0 JL, Neims AH. Metabolism of caffeine by mouse liver microsomes: GSH or cytosol causes a shift in products from 1,3,7_trimethylurate to substituted uracil. Life Sci 1983; 33: 1173-8. 7. Aldridge A, Parsons WD, Neims AH. Stimulation of caffeine metabolism in the rat by 3-methylcholantrene. Life Sci 1977; 21: 967-74. 8. Wietholtz H, Voegelin M, Arnaud MJ, et al. Assessment of the cytochrome p-448 dependent liver enzyme system by caffeine breath test. Eur J Clin Pharmacol 1981; 21: 53-9. 9. Arnaud MJ. Identification, kinetic and quantitative study of [2-i4C] and [1-Me-i4C]caffeine metabolites in rat’s urine by chromographic separations. Biochem Med 1976; 16: 67-76. 10. Lauterburg BH, Sautter V, Preisig R, et al. Hepatic functional deterioration after portocaval shunt in the rat. Gastroenterology 1976;71:221-7. 11. Herz R, Sautter V, Robert F, et al. The Eck fistula rat: definition of an experimental model. Eur J Clin Invest 1972; 2: 390-7. 12. Lauterburg B, Bircher J. Expiratory measurement of maximal aminopyrine demethylation in vivo. Effects of phenobarbital, partial hepatectomy, portocaval shunt and bile duct ligation in the rat. J Pharmacol Exp Ther 1976; 196: 501-9. 13. Herz R, Paumgartner G, Preisig R. Bile salt metabolism and
bile formation in the rat with a portocaval shunt. Eur J Clin Invest 1974; 4: 223-8. 14. Fisher B, Fisher ER, Lee S. Experimental evaluation of liver atrophy and portocaval shunt. Surg Gynecol Obstet 1967; 125: 1253-8. 15. Schroeder R, Mueller 0, Bircher J. The portocaval and splenorenal shunt in the normal rat. J Hepatol 1985; 1: 10723. 16. Schacter BA, Joseph E, Firneisz G. Effects of cholestasis produced by bile duct ligation on hepatic heme and hemoprotein metabolism in rats. Gastroenterology 1983; 84: 227-35. 17. Babany G, Descatoire V, Corbic M, et al. Regulation of renal cytochrome P-450. Effects of two-thirds hepatectomy, cholestasis, biliary cirrhosis and post-necrotic cirrhosis on hepatic and renal microsomal enzymes. Biochem Pharmacol 1985; 34: 31 l-20. 18. Arnaud MJ, Welsh C. Metabolic pathway of theobromine in the rat and identification of two new metabolites in human urine. J Agric Food Chem 1979; 27: 524-7. 19. Protwich F, Aebi H. Katalaseaktivitaet und Oxidation der Ein-Kohlenstoff-Fragmente in der Leber des Menschen. Klin Wochenschr 1960; 38: 753-7. 20. Chamuleau RAFM, Bosman DK. Liver regeneration. Hepatogastroenterol 1988; 35: 309-12. 21. Butler MA, Iwasaki M, Guengerich FP Kadlubar FE Human cytochrome P-450p, (P-450IA2), the phenacetin-o-deethylase, is primarily responsible for the hepatic 3-demethylation of caffeine and N-oxidation of carcinogenic arylamines. Proc Nat1 Acad Sci USA 1989; 86: 7696700. 22. Fuhr U, Doehmer J, Battula N, et al. Biotransformation of caffeine and theophylline in mammalian cell lines genetically engineered for expression of single cytochrome P450 isoforms. Biochem Pharmacol 1992; 43: 225-35. 23. Berthou F, Flinois J-P, Ratanasavanh D, Beaune P Riche C, Guillouzo A. Evidence for the involvement of several cytochromes P-450 in the first steps of caffeine metabolism by human liver microsomes. Drug Metab Disp 1991; 19: 561-7. 24. MacKinnon AM, Simon FR. Reduced synthesis of hepatic microsomal cytochrome P-450 in the bile duct ligated rat. Biochem Biophys Res Commun 1974; 56: 43743. 25. Dueland S, Reichen J, Everson GT, Davis RA. Regulation of cholesterol and bile acid homeostasis in bile-obstructed rats. Biochem J 1991; 280: 373-7. 26. Murray M, Cantrill E, Ishita M, Farrell GC. Impaired expression of microsomal cytochrome P450 2Cll in cholinedeficient rat liver during the development of cirrhosis. J Pharmacol Exp Ther 1992; 261: 373-80. 27. George J, Murray M, Farrell GC. Selective reductions of individual cytochrome P450 isoforms in different types of human cirrhosis. Hepatology 1993; 18: 315A. 28. Kotake AN, Schoeller DA, Lambert GH, et al. The caffeine CO2 breath test: dose response and route of N-demethylation in smokers and nonsmokers. Clin Pharmacol Ther 1982; 32: 261-9. 29. Larrey D, Branch RA. Clearance by the liver: current concepts in understanding the hepatic disposition of drugs. Semin Liver Dis 1983; 3: 285-97. 30. Kalow W, Tang B-K. The use of caffeine for enzyme assays: a critical appraisal. Clin Pharmacol Ther 1993; 53: 503-14.
87
Journalof Hepatology ISSN 0168-8278
Caffeine demethylation measured by breath analysis in experimental liver injury in the rat Heinz J. Schaad’,
Eberhard
L. Renner’,
‘Department of Clinical Pharmacology,
Hubertus
Wietholtz’,
Maurice J. Arnaud2 and Rudolf Preisig’
University of Berne, Berne and ‘Nestle Research Center, Nestec Ltd., Vevey, Switzerland
To assess the effects of experimental liver injury on caffeine metabolism, 1 &i/kg hw. of [3-methyP4C]caffeine (together with 5 mg/kg bw. of the cold compound) was injected i.p to four different experimental groups and respective controls of unanesthetized male Sprague-Dawley rats. Exhaled 14C02 was completely collected during 4 h and peak exhalation rate and fraction of dose recovered were calculated. l/3 hepatectomy affected 14C02 exhalation to a limited extent, decreasing solely peak exhalation rate (~~0.05 compared to sham-operated controls). 2/3 hepatectomy, on the other hand, resulted in significant reduction QKO.01) in both peak exhalation rate (by 59%) and fraction of dose recovered (by 47%), that were proportionate to the loss of liver mass (59%). End-to-side portocaval shunt led to the well-documented hepatic “atrophy”, liver weight being dhuinished on average to 50% within 2 weeks of surgery;
however, reductions in peak exhalation rate (by 75%) and fraction of dose recovered (by 64%) were even more pronounced. Finally, 48 h bile duct ligation was equivalent to “functional 2/3 hepatectomy”, peak exhalation rate (by 65%) and fraction of dose recovered (by 56%) being markedly diminished despite increased liver weight. These results indicate that 14C02 exhalation curves following administration of specifically labelled caffeine are quantitative indicators of acute or chronic loss of functioning liver mass In addition, the 3-demethylation pathway appears to be particularly sensitive to the inhibitory effects of cholestasis on microsomal function.
B
exclusively hepatic metabolism (1,2), 1,3,7-tri-methylxanthine (caffeine) has been proposed as a model compound for measuring liver function in man (3). Indeed, comparison of caffeine clearance, galactose elimination capacity and aminopyrine breath test in patients with various forms of liver disease has suggested that caffeine may represent an almost ideal test compound for assessing hepatic microsomal function quantitatively (3,4). In the past a series of animal studies explored the metabolic rate of caffeine within the body (2,5,6) and the effect(s) of inducing agents on hepatic caffeine metabolism (7,8). The specific influence on hepatic caffeine disposition of common features of liver disease,
such as reduced functioning liver cell mass or cholestasis, however, has thus far not been systematically determined, despite the fact that such knowledge may help interpretation of altered caffeine metabolism in human liver disease. We therefore decided to investigate the effects of partial hepatecotomy, bile duct ligation and end-to-side portocaval shunt on hepatic caffeine metabolism in the rat. Based on previous studies from this laboratory (8), we determined demethylation of caffeine in position 3, the predominant metabolic pathway in the rat (9), by quantitating 14C02 exhalation following application of specifically labelled [3-methyl-i4C]-caffeine.
Received 27 January 1994
Material
Correspondence
Animals
ECAUSE of its almost
of Medicine, Switzerland.
82
to:
Heinz J. Schaad, M.D., Department
University
Hospital,
3010 Berne-Inselspital,
Key words: Bile duct ligation; Breath analysis; Caffeine demethylation; Hepatectomy; Liver disease models; Liver function; Portocaval anastomosis; Rat. 0 Journal of Hepatology.
and Methods
All studies were performed in male Sprague-Dawley rats (Deutsche Versuchstierfarm, Tuttlingen, Ger-
Caffeine test in model liver diseases
many) weighing 150 to 350 g. Animals were kept at constant temperature (22°C) and humidity (48?2%), under an artificial 12-h light-dark cycle. They were housed in groups of 5-6 per cage, except for the portocaval shunt rats (and their respective sham-operated controls) which were kept in individual cages to prevent mutilation caused by aggressive behavior. All animals were allowed free access to standard rat chow (Altromin@ 300, Altromin GmbH, Lage, Germany) and tap water until 12 h prior to the experiment. “C-labelled caffeine [3-methyl-14C]-caffeine was prepared at the Nestle Research Center according to methods described previously (8). The final preparation contained less than 0.1% radiochemical impurity. Labelled caffeine was dissolved in 0.9% NaCl to a final concentration of 1 &i/ml, and stored at 4°C until use. Experimental models Partial hepatectomy was carried out via a midline abdominal incision under light ether anesthesia. The vessels to the median lobe (l/3 hepatectomy) and the left lobe (2/3 hepatectomy) of the liver were ligated en bloc and the lobe(s) was excised and weighed. Simple opening and closing of the abdomen using a midline incision served as a sham operation for the respective controls; care was taken to assure that the duration of anesthesia was comparable to the hepatectomy procedure (about 10 min). Bile duct ligation was performed via midline abdominal incision under light ether anesthesia. Following ligation of the common bile duct proximally and distally, the intervening small piece (about 1 cm) was excised. Sham operation consisted of a simple midline laparotomy. End-to-side portocaval anastomosis was performed under ether anesthesia after ligation of the gastroduodenal vein and the proximal vena cava, as previously described (10). As corresponding sham operation the gastroduodenal vein was ligated and the portal vein clamped for 15 min.
ate, Dr. G. Gattiker, Pharmazeutische Praeparate, Zurich, Switzerland). Exhaled CO2 was completely collected into a series of interchangeable counting vials connected to each individual restraining cage. The air was continuously sucked out of the cages by a vacuum pump through the collection vials where CO2 was trapped with ethanolamine, as described previously (498). At the end of the experiment, animals were anesthetized with i.p. pentobarbital (50 mgkg). A blood sample was taken for determination of serum levels of liver enzymes, bilirubin and bile acids, followed by sacrifice of the animals with an overdose of i.p. pentobarbital. The liver (and testes of portocaval anastomosis and sham portocaval anastomosis rats) was excised and weighed, and portocaval anastomosis shunt patency verified by probing. Analytical procedures Radioactivity of 14C in breath samples was counted following the addition of a scintillator (8) in a Packard liquid scintillation counter (TRI-CARB 2660). Total serum bilirubin and serum enzymes were measured in the Center Hospital Laboratory using routine methods, total serum bile acids with a commercially available radioimmunoassay (RIA)-kit (BectonDickinson, Orangeburg, NY, USA). Calculations 14C02 exhalation rate was expressed as percent of the dose per minute, the highest value being taken as peak elimination rate. The fraction of the dose recovered was obtained by calculating the area under the exhalation-time curve (AUC) between 0 and 240 min using the trapezoidal rule. A single value, namely the exhalation rate at 60 min (ER& was arbitrarily chosen for comparison with fraction of the dose recovered. All results are presented as mean2S.D. Group comparisons are based on Student’s t-test with Welch’s correction in the case of uneven distribution. PcO.05 was regarded as statistically significant.
Results Experimental design Caffeine breath tests were performed 1 h after recovery from anesthesia and surgery in partial hepatectomy animals, 48 h after bile duct ligation in bile duct ligated rats and 14 days after surgery in porto-caval anastomosis animals (and respective sham-operated controls), respectively. Animals were fasted overnight and placed in restraining cages. Following i.p. injection of 1 ,&i/ kg b.w. of [3-methyl-14C]-caffeine together with 5 mg/ kg b.w. of the cold compound (caffeine sodium benzo-
Characterization of experimental models By design, body weights of experimental groups and respective sham-operated controls (C) are almost identical, with the exception of portocaval anastomosis animals (Table 1). Such reduction (on the average 15%) in body weight 2 weeks after portocaval anastomosis is well known (11). Liver weights in partially hepatectomized animals were reduced to 4.82kO.42 g (versus 7.1120.87 g in C) by l/3 hepatectomy and to 3.1OkO.29 g (versus
83
H. J. Schadd et al.
7.4820.69 g in C) by 213 hepatectomy, respectively (Table 1). In bile duct ligated animals, total serum bilirubin was increased to 106.2236.0 pmol/l (versus 4.3~3.6 PmoVl in C), total serum bile acids to 168?7 1 ,umolll (versus 2.72 1.3 pmol/l) aspartate aminotransferase to 784t290 IU/l (versus 81? 16 IU/l) and alkaline phosphatase to 851298 IU/l (versus 455-+ 17 IU/l), respectively, thus documenting severe cholestasis. Functioning of the shunt in portocaval anastomosis rats is indicated by the expected increase in total serum bile acid levels (40246 versus 6.6k6.8 pmol/l in C) and the decrease in liver weight (4.7820.81 versus 9.30-+ 1.23 g) and in testicular weight (1.75kO.13 versus 2.7120.47 g), respectively (10,12). In addition, all shunts were patent at autopsy.
14C02 exhalation
curves
The 14C02 breath curves in l/3- and 2/3-hepatectomized animals, compared with their respective C, are depicted in Fig. 1. Whereas the curves of both C groups are superimposable with peak exhalation rate reached after 70 to 80 min, reduction in liver mass due to hepatectomy results in plateauing at a lower level of the 14C02 exhalation rate, which is much more pronounced following 2/3 hepatectomy. Thus, following 2/ 3 hepatectomy, peak exhalation rate - once reached is practically maintained throughout the period of observation. A virtual plateau at an even lower level, also represents the main effect of bile duct ligation and portocaval anastomosis on the 14C02 exhalation rate (Fig. 2), while the breath curves of all experimental groups are characterized by reduction in the initial specific 14C02 activity, evident already in the breath sample
‘4co*
O.WO-
Exhatatlon rate
/
Sham operated
Sham 0p.rat.d
0.100. o.oso% dosmlmtn l/S
Howt*ctomy
O.OlO_
I
-
0
120
240
0
-2/a
H.p*t~ctomy
120
240
NlRYt..
Fig. 1. Exhalation rates of 14C02, plotted against time, in rats following IN-hepatectomy (n=6, left panel) and 2/3hepatectomy (n=6, right panel) and in respective shamoperated controls (n=9 and 6, respectively). 84
Caffeine test in model liver diseases “Co,
0.500
Sham operatad
Shpm opwrtod 0.100 0.050 % do**/mln
611. duct Ilpptlon 0.010 0.005
0
120
240
0
120
240
Yhlutes
Fig. 2. Exhalation rates of “C02, plotted against time, in rats following porto-caval anastomosis (n=6, left panel) and bile-duct ligation (n=8, right panel), and in respective sham-operated controls (n=6 and 6, respectively).
collected during the first 10 min following 14C-caffeine injection. Pharmacokinetic
calculations
The results of peak exhalation rate, fraction of dose recovered, and ERm for all experimental groups are given in Table 1. There is no significant difference in any of these parameters between the various C groups. This attests to the reproducibility of the 14C2 caffeine breath test as performed. Whereas the effects of l/3-hepatectomy are limited to a decrease in peak exhalation rate of caffeine-3-demethylation, acute (2/3-hepatectomy) or chronic (portocaval anastomosis) reduction in liver mass exceeding 33% results in markedly diminished caffeine metabolism. Thus, following 2/3-hepatectomy peak exhalation rate and fraction of dose recovered are decreased in parallel with the change in liver weight (59%), on average by 59% and 47%, respectively. In animals with portocaval anastomosis, the reduction in peak exhalation rate (75%) and fraction of dose recovered (64%), in fact, slightly exceeds that in liver weight (49%) pointing to metabolic changes over and above those in hepatocyte mass (11). The divergence of liver weight and metabolic capacity, however, is most pronounced in animals with bile duct ligation. Despite a “gain” in liver weight of over 40%, peak exhalation rate is diminished by two thirds and fraction of dose recovered by 56%, respectively, presumably reflecting tissue edema, functional alteration(s) in and loss of hepatocytes. On the basis of an inspection of all 14C02 exhalation curves, it appeared reasonable to choose ERso as a potentially useful single point measurement. In general, its values reflect fraction of dose recovered (r2=0.972),
but tend to more closely parallel peak exhalation rate (r2=0.987) (Table 1). Since fraction of dose recovered reflects an integral of the 14C02 exhalation rate over time, this parameter was used for direct comparison with liver weight (Fig. 3). Whereas fraction of dose recovered is decreased roughly in parallel with surgical reduction of liver mass (l/3- and 2/3-hepatectomy, portocaval anastomosis), cholestasis (bile duct ligation) produces a functional defect approximately equivalent to 2/3_hepatectomy, in face of an increased liver weight.
Discussion The aim of the present study ‘was to explore the effects of a reduced liver cell mass and of cholestasis on caffeine metabolism in three well-established rat liver disease models (11-17). Our results can be summarized as follows: Acute, partial hepatectomy results in a reduction of 14C02 exhalation, which is proportionate to the amount of functioning liver mass removed. Peak rate expiration appears to be primarily affected, while fraction of dose recovered is only significantly reduced after 2/3-PH. Chronic liver “atrophy” due to porto-caval shunt, as expected, decreases 14C02 exhalation rate. The fact that the functional change tends to exceed the loss of liver weight is in keeping with previous studies (13,14). Indeed, the presence of mild cholestasis suggested by increased serum bile salt levels may have additionally inhibited microsomal function. Markedly diminished 14C02 exhalation despite in-
25-
14
co
sham opommd
Exhalation
20-
%do55/4h mte
,5
113ltqwwtomy w 2/9l+wpl=@my
10 -
4
s,
/
’
/
/
’
’
‘P Porlomv8I~hunt
Oh7777 bar walght %b.w.
Fig. 3. Relationship between the cumulative 14C0, exhalation rate over 4 h (FD) and liver weight (expressed as % of body weight) in the various experimental groups studied. The dotted line (- - -) represents an imaginary line the
fraction of dose recovered would follow if changes in fraction of dose recovered were in parallel with changes in liver weight. 85
H. J. Schadd et al.
creased liver weight characterizes bile duct ligated animals. Although structural damage (leading to hepatocyte loss) undoubtedly occurred, deterioration of microsomal function in this model is primarily a consequence of cholestasis (16,17), an effect documented also in previous investigations using aminopyrine (11). The complexities of both the breath test and the experimental models of liver injury employed merit emphasis. Thus, although demethylation in position 3 yielding paraxanthine is the predominant primary pathway of caffeine metabolism in the rat (9), exhaled i4C02 following injection of [3-methyl-‘4C]-caffeine also originates from 3-demethylation of other caffeine demethylation products, i.e. theobromine and theophylline, being converted to 7-methylxanthine and lmethylxanthine, respectively (9,18). The observed breath curves, therefore, reflect a cascade of metabolic events, rather than a single process. Nevertheless, the flatness of exhalation curves in 2/3_hepatectomized, portocaval anastomosis and bile-duct ligated animals suggests that saturation of caffeine 3-demethylation was reached with the dose of caffeine used in these models. It should be recognized that the time interval chosen for determination of fraction of dose recovered (4 h) is somewhat arbitrary and reflects the sum of xanthine 3-demethylation over this time-period. Since fraction of dose recovered represents an AUC, it is a measure of i4C02 kinetics independent of pharmacokinetic models and observer bias. Also, it is recognized that theoretically fraction of dose recovered will become equal in all experimental groups, provided it is measured over a long enough period. In addition to fraction of dose recovered values, peak exhalation rate and exhalation rate at 60 min were therefore determined. Finally, conversion of methyl groups to CO2 via the intermediates formaldehyde and formate (19) although considered not to be rate-limiting - may be affected by various, as yet ill-defined factors inherent in the liver disease models used. While surgical trauma per se might affect caffeine metabolism, performance of breath tests 1 h after partial hepatectomy appears sufficient to allow complete recovery of the animals from anesthesia. The lack of any significant difference in the breath curves of the four C groups further underscores this notion. Clearly, the time elapsed between hepatectomy and performance of breath tests was too short to account for metabolic effect of regenerative processes (20). Establishment of an end-to-side porto-caval shunt is associated with major systemic sequelae, as evidenced by weight loss, signs of “encephalopathy” and testicular atrophy (10,12). These factors, together with (in 86
part disparate) hepatic functional changes (IO, 13-l 5) presumably affected microsomal processing of caffeine over and above the loss of liver mass. Animals 48 h after bile duct ligation exhibit no weight loss and appear well despite cholemia. In this sense, bile duct ligation may represent the “purest model” of hepatic injury. Caffeine 3-demethylation is mediated by cytochrome P-450IA2 (21-23). It is well known that the overall content of cytochrome P-450 is reduced in the bile duct ligated rat (24,25). Certain P-450 isoforms have been shown to be selectively diminished in certain disease states (26). Our data would be entirely compatible with the caffeine-metabolizing P-450IA2 being overproportionately decreased in cholestasis. Since P-450 isoforms were not determined in the present study, however, this remains entirely speculative. Interestingly a more profound decrease in P-4501A in cholestatic compared to hepatocellular cirrhosis in man has been described (27). Despite these limitations and uncertainties, our studies do allow us to link the effects of experimental hepatic injury on caffeine disposition in rats with those of liver disease in man. In man, specific activity of 14C02 in breath following injection of 14C-caffeine is highly correlated with caffeine plasma clearance over a wide range of values (3,28). In the rat, 14C02 output was continuously and completely collected. The resulting, integrated breath curves should, therefore, represent caffeine metabolism even more fully. It seems reasonable to assume that in the rat cumulative 14C02 excretion in breath closely reflects hepatic caffeine clearance. The clearance of a low extraction compound such as caffeine, exhibiting a constant volume of distribution and being eliminated almost exclusively by hepatic metabolism, is expected to be diminished primarily by loss of hepatocyte mass (29). This was confirmed in the present study in animal models leading to acute (partial hepatectomy) and chronic reduction in liver weight (portocaval anastomosis). It is, therefore, suggested that reduction in caffeine clearance previously observed in patients with cirrhosis (3) primarily reflects partial functional hepatectomy. Finally, the profound effect of cholestasis on caffeine metabolism may render caffeine clearance measurements a sensitive indicator of functional deterioration in acute or chronic liver disease in man which, besides a reduction in functioning liver cell mass, are often associated with cholestatic features. Since cytochrome P-450IA2 plays an important role in the genesis of carcinogenic compounds (21), caffeine has gained interest as a test compound for the assessment of cytochrome P-450IA2 activity in man (30).
Caffeine test in model liver diseases
This study describing the influence of various models of liver disease on the 3-demethylation activity may be helpful in the interpretation of results from tests in humans affected by liver diseases.
Acknowledgements The authors wish to thank Ms. M. Kappeler for the art work and Ms. M. Sommer for secretarial help in preparing the manuscript. E.L.R. is the recipient of a SCORE award from the Swiss National Foundation for Scientific Research. Supported by the Swiss National Foundation for Scientific Research.
References 1. Arnaud MJ, Welsh C. Theophylline and caffeine metabolism in man. In: Rietbrock N, Woodcock BG, Staib AH, eds. Theophylline and Other Methylxanthines. Braunschweigi Wiesbaden, Germany: Vieweg and Son, 1981: 13548. 2. Bonati M, Garattini S. Interspecies comparison of caffeine disposition. In: Dews PB, ed. Caffeine: Perspectives from Recent Research. Berlin, Germany: Springer Verlag, 1984: 4856. 3. Renner E, Wietholtz H, Huguenin P et al. Caffeine: a model compound for measuring liver function. Hepatology 1984; 4: 3846. 4. Jost G, Wahllaender A, von Mandach U, et al. Overnight salivary caffeine clearance: a liver function test suitable for routine use. Hepatology 1987; 7: 33844. 5. Aldridge A, Neims AH. Relationship beetween the clearance of caffeine and its 7-N-demethylation in developing beagle puppies. Biochem Pharmacol 1980; 29: 1909-14. 6. Ferrer0 JL, Neims AH. Metabolism of caffeine by mouse liver microsomes: GSH or cytosol causes a shift in products from 1,3,7_trimethylurate to substituted uracil. Life Sci 1983; 33: 1173-8. 7. Aldridge A, Parsons WD, Neims AH. Stimulation of caffeine metabolism in the rat by 3-methylcholantrene. Life Sci 1977; 21: 967-74. 8. Wietholtz H, Voegelin M, Arnaud MJ, et al. Assessment of the cytochrome p-448 dependent liver enzyme system by caffeine breath test. Eur J Clin Pharmacol 1981; 21: 53-9. 9. Arnaud MJ. Identification, kinetic and quantitative study of [2-i4C] and [1-Me-i4C]caffeine metabolites in rat’s urine by chromographic separations. Biochem Med 1976; 16: 67-76. 10. Lauterburg BH, Sautter V, Preisig R, et al. Hepatic functional deterioration after portocaval shunt in the rat. Gastroenterology 1976;71:221-7. 11. Herz R, Sautter V, Robert F, et al. The Eck fistula rat: definition of an experimental model. Eur J Clin Invest 1972; 2: 390-7. 12. Lauterburg B, Bircher J. Expiratory measurement of maximal aminopyrine demethylation in vivo. Effects of phenobarbital, partial hepatectomy, portocaval shunt and bile duct ligation in the rat. J Pharmacol Exp Ther 1976; 196: 501-9. 13. Herz R, Paumgartner G, Preisig R. Bile salt metabolism and
bile formation in the rat with a portocaval shunt. Eur J Clin Invest 1974; 4: 223-8. 14. Fisher B, Fisher ER, Lee S. Experimental evaluation of liver atrophy and portocaval shunt. Surg Gynecol Obstet 1967; 125: 1253-8. 15. Schroeder R, Mueller 0, Bircher J. The portocaval and splenorenal shunt in the normal rat. J Hepatol 1985; 1: 10723. 16. Schacter BA, Joseph E, Firneisz G. Effects of cholestasis produced by bile duct ligation on hepatic heme and hemoprotein metabolism in rats. Gastroenterology 1983; 84: 227-35. 17. Babany G, Descatoire V, Corbic M, et al. Regulation of renal cytochrome P-450. Effects of two-thirds hepatectomy, cholestasis, biliary cirrhosis and post-necrotic cirrhosis on hepatic and renal microsomal enzymes. Biochem Pharmacol 1985; 34: 31 l-20. 18. Arnaud MJ, Welsh C. Metabolic pathway of theobromine in the rat and identification of two new metabolites in human urine. J Agric Food Chem 1979; 27: 524-7. 19. Protwich F, Aebi H. Katalaseaktivitaet und Oxidation der Ein-Kohlenstoff-Fragmente in der Leber des Menschen. Klin Wochenschr 1960; 38: 753-7. 20. Chamuleau RAFM, Bosman DK. Liver regeneration. Hepatogastroenterol 1988; 35: 309-12. 21. Butler MA, Iwasaki M, Guengerich FP Kadlubar FE Human cytochrome P-450p, (P-450IA2), the phenacetin-o-deethylase, is primarily responsible for the hepatic 3-demethylation of caffeine and N-oxidation of carcinogenic arylamines. Proc Nat1 Acad Sci USA 1989; 86: 7696700. 22. Fuhr U, Doehmer J, Battula N, et al. Biotransformation of caffeine and theophylline in mammalian cell lines genetically engineered for expression of single cytochrome P450 isoforms. Biochem Pharmacol 1992; 43: 225-35. 23. Berthou F, Flinois J-P, Ratanasavanh D, Beaune P Riche C, Guillouzo A. Evidence for the involvement of several cytochromes P-450 in the first steps of caffeine metabolism by human liver microsomes. Drug Metab Disp 1991; 19: 561-7. 24. MacKinnon AM, Simon FR. Reduced synthesis of hepatic microsomal cytochrome P-450 in the bile duct ligated rat. Biochem Biophys Res Commun 1974; 56: 43743. 25. Dueland S, Reichen J, Everson GT, Davis RA. Regulation of cholesterol and bile acid homeostasis in bile-obstructed rats. Biochem J 1991; 280: 373-7. 26. Murray M, Cantrill E, Ishita M, Farrell GC. Impaired expression of microsomal cytochrome P450 2Cll in cholinedeficient rat liver during the development of cirrhosis. J Pharmacol Exp Ther 1992; 261: 373-80. 27. George J, Murray M, Farrell GC. Selective reductions of individual cytochrome P450 isoforms in different types of human cirrhosis. Hepatology 1993; 18: 315A. 28. Kotake AN, Schoeller DA, Lambert GH, et al. The caffeine CO2 breath test: dose response and route of N-demethylation in smokers and nonsmokers. Clin Pharmacol Ther 1982; 32: 261-9. 29. Larrey D, Branch RA. Clearance by the liver: current concepts in understanding the hepatic disposition of drugs. Semin Liver Dis 1983; 3: 285-97. 30. Kalow W, Tang B-K. The use of caffeine for enzyme assays: a critical appraisal. Clin Pharmacol Ther 1993; 53: 503-14.
87