methotrexate, harmine, harmaline, harmol and norharman. None of the harmine analogs was able to inhibit AF NAT activity significantly and were not studied ...
DUNDEE MEETING
87
Acetylation pharmacogenetics: profiles of arylamine pharmacology and toxicology in a genetic mouse model W. W. WEBER, T. A. HULTIN and D. A. MORGOTT Department of Pharmacology, University of Michigan, Ann Arbor, MI 48109, U.S.A. Acetylation, a major pathway for conjugation and excretion of foreign compounds, is a common route of arylamine biotransformation (Williams, 1959)and an important factor in the expression of mutagenic, carcinogenic and more acute effects of these substances (King & Weber, 1981). An hereditary polymorphism in acetylation, due to genetic variation in the activity of liver and gut mucosal NAT activity, is an interesting feature of this reaction which is well known both in man and rabbit enabling individuals in each of these species to be classified as 'rapid' and 'slow' acetylators of many arylamines and hydrazines in the environment (Weber & Glowinski, 1980). Human genetic epidemiological studies have shown that slow acetylators are predisposed to toxicity from arylamine and hydrazine drugs (Weber & Hein, 1984) and to urinary bladder cancer resulting from arylamine carcinogens (Cartwright et al., 1982). These studies are essential but in some instances they have led to controversy about the relationship of acetylator status to these effects, perhaps because they are based on epidemiological data from selected patient groups rather than on more precise experimental studies in animals (Weber et al., 1983). Thus we have concentrated on developing and characterizing experimental models for the human acetylation polymorphism as tools for elucidation of pharmacogenetic factors which can modify toxicity or reveal other peculiar responses to arylamine drugs and carcinogens. Recently we have identified hereditary acetylation polymorphisms in the mouse (Glowinski & Weber, 1982a, b and hamster (Hein et al., 1982). It was apparent from a survey of some 20 inbred mouse strains that most strains were rapid acetylators as typified by C57BL/6J (B6) mice, but a few strains (A/J, AHe/J, XG/f) were slow acetylators of certain carcinogens (Glowinski et al., 1982~). Genetic analysis of the mouse trait in B6 and A mice, which were selected for further study, indicated that the NAT activity in these mice is primarily controlled by a single gene wiht two major alleles. B6 mice have approximately ten times more NAT activity for benzidine, and for the alternative substrate, AF, than A/J (A) mice. Moreover, the levels of liver NAT activity were similar to levels that have been reported for human liver NAT in rapid and slow acetylators (Glowinski et al., 1978). An apparent K,,, difference of some 20-fold between A and B6 mouse liver AF NAT suggests that the inter-strain difference in B6 and A N-acetylating activity is probably accounted for by a structural difference in the NAT gene product (Glowinski & Weber, 1982b). We are currently investigating the pharmacologic and toxicdogic profiles of AF and MDA in relation to acetylator status in further characterizing the genetic mouse model. We have investigated the dose-response characteristics and pharmacokinetics of AF both in intact A, B6, B6AF1, and A.B6-NAT' animals and in hepatocytes isolated from A and B6 animals, and have obtained preliminary results on AFinduced DNA damage in A and B6 hepatocytes. Relatively little is known about the biological effects of MDA even though in man acute exposure produces a distinct picture of Abbreviations used : AAF, acetylaminofluorene; AF, 2-aminofluorene; MDA, methylene dianiline; NAT, N-acetyltransferase (EC 2.3.1.5); NT, nitrofluorene; UDS, unscheduled DNA synthesis.
VOl. 12
severe hepatotoxic injury (Kopelman et af., 1966a,b Kopelman, 1968; Brooks et af., 1974; McGill & Motto, 1974; Williams et al., 1974). The low incidence (1-2%) of persons affected among persons exposed suggests that a genetic component may contribute to the susceptibility of MDAinduced cholestatic jaundice (NIOSH, 1976). These findings and the structural resemblance of MDA to benzidine and to methylene bis-chloroaniline, both of which are distinctly polymorphic substrates for hepatic NAT (Glowinski et al., 1978), has led us to conduct a series of acute toxicity studies in the genetic mouse acetylation model to determine whether acetylator status alters individual susceptibility to acute cholestatic jaundice resulting from MDA. Some of the highlights of these studies are presented below. AF studies AF elimination rates were measured in the B6 and A mice after a single intraperitoneal injection of AF at each of three different doses. Elimination of AF from the blood was dosedependent but appeared to be a first-order process suggesting the presence of a limiting factor. AF elimination rates were 2.5-4 times faster in B6 than in A mice at 30mg/kg (0.83 f0.04h-l vs 0.35 & 0.07h-I), 50mg/kg (0.86 & 0.11 h-' vs 0.21 f0.04h-I) and 100 mg/kg (0.36 f 0.02h-l vs 0.12 & 0.02h-') (Fig. 1). The 100mg/kg dose of AF was chosen for detailed pharmacokinetic study in F, offspring and in the congenic mouse line, A.B6-NAT'. Since the A.BdNAT' mouse has the rapid NAT gene on the A background, the specific influence of N-acetylation on differences in AF elimination and in V, can be determined by comparison of congenic and A mice. AF elimination rates in F, mice (0.24&0.02h-l) were intermediate between the parental values (Fig. 1).This provides additional evidence for the co-dominant inheritance of the response resulting from the two major alleles of the NAT gene. In A.B6-NATr mice, AF elimination rates were also intermediate and had an average value of 0.27 f 0.05h-' (data not shown on Fig. 1). This is significantly lower than that found in B6 mice (0.36 f 0.02h-l). When prior studies of AF NAT activity in liver and blood with recombinant inbred lines from B6 and A mice are taken into account, the presence of modifier genes in the A mouse background which suppress the B6 gene is the most likely explanation for the low value of the elimination rate constant observed in this congenic mouse line. We also determined the effect of genetic differences in Nacetylation of AF disappearance in hepatocytes isolated from B6 and A mice (Williams et a f . , 1977). Rates of AF disappearance and of AAF appearance were quantitated by h.p.1.c. methods and compared in hepatocytes isolated from the two strains exposed to various concentrations of AF. Since previous studies in our laboratory had shown that there was a 2-fold difference in de-acetylation of AAF to k f ; in these strains (1.45 nmol/min per mg of protein for B6 vs 0.76 for A), paraoxon, an AAF de-acetylase inhibitor, was included in the incubation medium. In hepatocytes from both strains the disappearance of AF and the appearance of AAF were dependent on time and dose. Initially, AF disappearance followed pseudo zero-order kinetics but later it followed first-order kinetics when the concentration of AF fell to approximately 8pM. Under first-order conditions, AF disappearance was 2 times faster in B6 hepatocytes than in A hepatocytes (Table 1). Under saturating conditions, the
'I
88
BIOCHEMICAL SOCIETY TRANSACTIONS
50
Table 1. AF disappearance and AF-induced DNA damage in hepatocytes isolated from rapid (B6) and slow ( A ) acetylator mice K values (mean +s.E.M.) determined under first-order conditions in the presence of paraoxon; number of animals given in parentheses. UDS values are duplicate determinations in hepatocytes from a single animal from each phenotype..
0 C57BU6J
0 NJ
*
B6AFi/J
Mouse strain B6
A
K (h-') 0.59f0.06 (5)
AF concentration (molil) 10-3
UDS (net grains/nucleus) Toxic
10-4 10-5 10-6
22.1 19.6 18.8 Toxic 11.6 7.3 6.1
0.34k0.04 (4)
10-3 10-4
10-5 10-6
1 '
I
I
I
1
2
3
I
I
4
5
r-
6
7
8
Time (h)
Fig. 1. Blood AF elimination in C57BL/6J,AIJ, and B6AFIIJ mice Animals were injected intraperitoneally with lOOmg/kg of AF and 5Opl serial blood samples were taken by orbital sinus puncture. Blood was lysed in 1ml of water, 10- M-NF was added as an internal standard and samples were extracted 2 times with 1Ovols. of ether. Samples were dried under N2, redissolved in 70% acetonitrile and loop1 aliquots were injected on to a Whatman C-18 Partisil lOpm ODS-3 h.p.1.c. column. AF and N F were eluted with a linear gradient of 60% trifluoroacetic acid/40% acetonitrile at 0 time to 100% acetonitrile at l0min. Retention times were approximately 3.7min for AF and 12.5min for NF. Concentrations of AF were calculated by reference to an AF/NF vs AF standard curve.
rate of AAF formation was also approximately twice as fast in B6 hepatocytes as in A hepatocytes. DNA damage as measured by UDS (Williams, 1977) was assessed in hepatocytes from both strains to determine the influence of acetylator status on AF-induced DNA damage. Preliminary evidence suggests that UDS is dose-dependent in both strains and that more DNA damage is induced in B6 hepatocytes than in A hepatocytes (Table 1). The capacity of various chemicals to inhibit one or more of the metabolic steps in the activation of arylamine carcinogens was also studied. Competitive-type inhibitors of the enzyme(s) involved are most interesting inasmuch as they may need to recover their activity to carry out essential cellular functions. Because N-acetylation of AF to AAF appears to play a role in the susceptibility to AF-induced cancer, it may be a good candidate for inhibition. Recently, studies of human blood showed that p-aminobenzoic acid, folic acid and methotrexate reversibly inhibited human blood p-aminobenzoic acid N AT (Mandelbaum-Shavit & Blondheim, 1981). Another study demonstrated that the beta-carboline derivatives, harmine and harmaline, were potent inhibitors of NAT partially purified from hamster and rat liver (Wright et al., 1979). In our studies, cytosolic
Table 2. Kivalues for inhibition of AF NAT activity in liver cytosols from rapid (B6) and slow ( A ) acetylator mice Kivalues are the mean +s.E. Number of animals studied are in parentheses. Inhibitor (mM) Mouse strain €36 (3) A (3)
L
p-Aminobenzoic acid 0.16 fO.01 0.18 f0.01
Folic acid 1.06f0.08 2.80 f0.73
\
Methotrexate 0.92k0.16 2.36k0.17
preparations of NAT from B6 and A mice were incubated with concentrations of AF ranging from to 2 x 1 0 - 3 ~ in the presence of p-aminobenzoic acid, folic acid, methotrexate, harmine, harmaline, harmol and norharman. None of the harmine analogs was able to inhibit AF NAT activity significantly and were not studied further. pAminobenzoic acid is a competitive inhibitor of mouse AF NAT activity while both folk acid and methotrexate are non-competitive inhibitors. K, values calculated from Dixon plots are shown in Table 2. The Kivalues for paminobenzoic acid are the same in B6 and A mice and about 6-16 times lower than the Ki values for folic acid and methotrexate. The Kivalues for folk acid and methotrexate are about 2.5 times lower in B6 than in A mice, with those for methotrexate being slightly less than for folic acid. The appearance of AAF in hepatocytes from B6 mice was almost completely inhibited by 10- M-p-aminobenzoicacid while 5 x 10-4~-folicacid inhibited AAF appearance only a small amount (Table 3). In addition, p-aminobenzoic acid inhibited the disappearance of AF almost completely while folic acid inhibited AF disappearnce only slightly. In A mice, p-aminobenzoic acid and folic acid inhibition of AAF appearance followed the same pattern as seen in B6 mice. Conversely, p-aminobenzoic acid inhibited AF disappearance only moderately; folic acid inhibition of AF disappearance was similar to that seen in B6 mice (Table 3). Thus disappearance of AF in B6 mice appears to result almost entirely from N-acetylation of AF to AFF while AF disappearance in A mice seems to be influenced by processes in addition to N-acetylation. Thus we find in this genetic mouse model for the hereditary acetylation polymorphism that there are differ1984
DUNDEE MEETING
89
Table 3. Inhibition of AF disappearance and AAF appearance in hepatocytesfrom rapid (B6) and slow ( A ) acetylator mice % inhibition = % of control. B6 values are means of two hepatocyte preparations done in duplicate. % inhibition f
Mouse strain B6 (2) A (2)
h
I
p-Aminobenzoic acid
r
AF disappearance 83 50
A
Folic acid \
AAF appearance 91 98
10.0
ences in AF elimination rates in intact B6 and A mice and in hepatocytes isolated from these mice. These differences are correlated directly with AF N-acetylation rates seen in liver cytosol. In addition, there appears to be a differential susceptibility to AF-induced DNA damage which correlates with differences in their acetylating capacity. Sincep aminobenzoic acid is a reasonably potent inhibitor of AF acetylation in hepatocytes as well as in liver cytosol, it will be of interest to determine whether it will also inhibit AFinduced D N A damage in the intact cell. MDA studies MDA dose-response characteristics were determined in male B6 and A mice 24h after a single intraperitoneal injection of MDA at each of five different doses. Dosedependent elevations of serum bilirubin, alkaline phosphatase and serum glutamate-pyruvate transaminase were observed with maximum responses occurring between doses of 60-100mg/kg body wt. The degree of hepatic damage assessed by these biochemical indices of liver function was significantly higher in the rapid B6 males than in the slow A males, suggesting that acetylation was an important step in the toxic pathway. Further study revealed, however, that the hepatic damage induced by MDA was more complex than originally supposed, and profoundly influenced by sex. When further data were obtained on females of both strains and plotted according to sex for each strain (Fig. 2), it was obvious that
VOl. 12
L
r
AF disappearance 26 26
~
\
AAF appearance 25 30
Ib) C57BU6J
A mice of both sexes responded to almost the same degree at a level intermediate to B6 males and females: in addition, B6 males were more responsive than B6 females. The possibility that acetylator status might still be influential in the toxic response but was being masked by strain (B6)specific sex differences suggested that recombinant inbred lines of mice derived from B6 and A lines might serve to isolate any single gene effect(s) contributing to the interstrain differences within each sex. Thus males and females of each of several (nine) recombinant lines available were given a 60mg/kg dose of MDA intraperitoneally and responses were measured as described above (Table 4). The strain distribution patterns within each sex did not indicate that the response was controlled by a single gene. Reexamination of the strain distribution patterns across sexes showed that two recombinant inbred lines (BXA15, AXB17) and possibly a third (BXA1) responded in a B6-like manner (i.e. with a lower response in females than males) while all other recombinant lines responded in an A-like manner (i,e, without a sex-related response). Additional comparisons showed no correlation of these strain distribution patterns with the strain distribution pattern for NAT in these recombinant lines. These findings would appear to suggest that acetylation does not play an especially important role in the hepatotoxic damage acutely produced by MDA. But this possibility is not entirely precluded by the information obtained so far because the presence of the strain-specificsex difference in
90
BIOCHEMICAL SOCIETY TRANSACTIONS
Table 4. Total serum bilirubin in rapid (B6) and slow ( A ) acetylator mice 24 h after a single intraperitoneal dose (6Omgl kg) of methylenedianiline Values represent the mean&s.E.; the number of treated animals are in parentheses. Total serum bilirubin (mg/100ml) Mouse strain A/J B6 F1 BXA 1 BX.46 BXAll BXA I4 BXA 15 AXB2 AXB13 AXB 15 AXB 17
Acetylator phenotype A B6 -
B6 A B A A B6 B6 A B6
r
A
Female 4.67k0.16 (4) 1.24k0.28 ( 5 ) 0.79k0.48 (5) 0.28+0.04 ( 5 ) 4.65k0.64 ( 5 ) 3.31 k0.25 (4) 9.59k0.83 (5) 5.70+ 1.01 (5) 6.95 k 0.66 (4) 11.11k0.71 (5) 8.24+ 1.06 (4) 1.35k0.45 (5)
>
Male 4.96k0.29 (4) 7.48k0.55 (4) 10.74k1.38 (4) 0.50k0.07 (4) 5.03k0.56 (5) 3.67k0.28 (4) 7.20k 1.03 (4) 11.872 1.33 (4) 7.45 f0.86 (4) 9.70k0.34 ( 5 ) 9.34k0.90 (5) 4.64k0.29 (4)
B6 mice may merely be masking the true relationship between acetylator phenotype and hepatotoxicity. Since this sex difference has not been observed in any other mouse strain (or in any other susceptible laboratory animal species studied), it would be useful to test other rapid acetylator mouse strains for an effect of acetylator status on MDAinduced cholestatic jaundice. This work was supported in part by USPHS Grant G M 27028. Brooks, L. J., Neale, J. M. & Pieroni, D. R. (1974) JAMA 242, 1527-1528
Cartwright, R. A., Rogers, H. J., Barham-Hall, D., Kahn, M.A., Glashan, R. W., Ahmad, R. A. & Higgins, E. (1982) Lancet i, 842-846 Glowinski, I. B. & Weber, W. W. (1982a) J. Biol. Chem. 257,14241430 Glowinski, I. B. & Weber, W. W. (19826)J. Biol. Chem. 257,14311437 Glowinski, I. B., Radtke, H. E. & Weber, W. W. (1978) Mol. PharmacoL 14, 940-949 Hein, D. W., Omichinski, J. G., Brewer, J. A. & Weber, W. W. (1982) J. Pharm. Exp. Ther. 220, 8-15 King, C. M. & Weber, W. W. (1981) NCI Monograph 58, pp. 117122 Kopelman, H. (1968) Postgrad. Med. J. 44, 78-81 Kopelman, H., Robertson, M. H., Sanders, P. G. & Ash, I. (1966~) Brit. Med. J . 1, 514-516 Kopelman, H., Scheuer, P. J. &Williams, R. (1966b) Quart. Med. J. XXXV, 553-564 Mandelbaum-Shavit, F. & Blondheim, S. H. (1981) Biochem. Pharmacol. 30,6 5 4 9 McGill, D. B. & Motto, J. D. (1974) New Engl. J. Med. 291, 278282 NIOSH (1976) Current Intelligence Bulletin 4.4-Diaminodiphenylmethane (DDM), Department of Health, Education and Welfare, Public Health Service, Washington, D.C. Weber, W. W. & Glowinski, I. B. (1980) in Enzymatic Basis of Detoxicatwh (Jakoby, W. B., ed.), pp. 169-186, Academic Press, New York Weber, W. W. & Hein, D. W. (1984) Pharmacol. Rev. in the press Weber, W. W., Hein, D. W., Litwin, A. &Lower, Jr, G. M. (1983) Fed. Proc. 43,42-53 Williams, G. M. (1977) Cancer Res. 37, 1845-1851 Williams, G. M., Bermudez, E. & Scaramuzzino, D. (1977) In Vitro 13, 809-817 Williams, R. T . (1959) Detoxication Mechanisms, 2nd edn., Wiley New York Williams, S.V., Bryan, J. A,, Burk, J. R. & Wolf, F. S.(1974) New Eng. J. Med. 291, 1256 Wright, E. H., Bird, J. L. & Feldman, J. M. (1979) Res. Commun. Chem. Pathol. Pharmacol. 24, 259-272
Drug metabolism and tropical disease A. BRECKENRIDGE, K. AWADZI, D. J. BACK, G. EDWARDS, H. GILLES, M. ORME and S. WARD Unit of Tropical Clinical Pharmacology, University of Liverpool and Liverpool School of Tropical Medicine, Liverpool L69 3BX, U.K. Introduction Over the last 20 years, clinical pharmacology has made signal advances in three broad areas as follows. The scientific basis of inter-individual differences in drug response. Both pharmacokinetic and pharmacodynamic aspects have been investigated although much more effort has been expended on the former than the latter so that there are now relatively few commonly used drugs whose absorption, distribution, metabolism and elimination have not been studied in man. The clinical pharmacologist’s contribution to the measurement of drug effect has been less notable, perhaps reflecting the scientific training of the majority of research workers in the field, but there are now welcome signs that increasing attention is being paid to this important aspect. Two other determinants of drug response have received special attention. The first is drug interactions. While it is frequently thought that interactions are of more theoretical than practical importance, their clinical relevance now Abbreviations used: DEC, diethylcarbazine; M.I.C., mean inhibitory concentration.
receives much attention and it is possible to predict with some accuracy those situations where drug interactions may contribute beneficially or adversely to drug response. The second aspect is the role of active metabolites. As more detailed metabolic profiles of commonly used drugs have been built up, it has become apparent that metabolites may make a considerable contribution to both the therapeutic and toxic effects which drugs may manifest. The design of effective drug regimens. This is largely a consequence of understanding how subjects may vary in their response to administered drugs. To this end, pharmacokinetics has become a valuable clinical tool which the practising doctor must use at the bedside. Improvement in drug usage in common diseases. This is in part a consequence of the above two areas, but, in addition, socio-pharmacological factors play a part. Thus it is most important to know why patients do not comply with prescribed regimens, and what determines regional and national differences in drug usage. The clinical pharmacologist has rightly directed his attention quite extensively to this aspect of therapeutics. In many therapeutic areas, these three advances have improved patient care significantly. The use of anticonvulsant, antibacterial and antiarrhythmic agents can rightly be cited as examples. The aim of this paper is to demonstrate how the principles of clinical pharmacology are slowly being applied to the field of the drug treatment of tropical disease, to 1984
