979 PHARMACOGENETIC STUDIES ON THE DRUG- RELATED LUPUS SYNDROME Differences in Antinuclear Antibody Development and Drug-Induced DNA Damage in Rapid and Slow Acetylator Animal Models W. W. WEBER and R. H. TANNEN Pharmacogenetic study of an inbred mouse acetylators may also be found in abnormally high num- model system derived from A/J (slow acetylator) and bers in certain groups of patients with spontaneous C57BL/6J (rapid acetylator) parental strains shows lupus erythematosus (8). These correlations have sug- that spontaneous occurrence of antinuclear antibodies gested more than a simple pharmacologic relation be- is associated with the slow acetylator phenotype al- tween drug-induced lupus and the slow acetylator though the development of spontaneous and pro- phenotype and we have devised new experimental ap- cainamide-induced antinuclear antibodies is a dis- proaches to seek such possibilities. sociable process. In another study using primary We have found from pharmacogenetic studies in cultures of intact hepatocytes obtained from slow and an inbred mouse model system that spontaneous occur- rapid acetylator rabbits, observations indicate that the rence of antinuclear antibodies (ANA) is associated amount of DNA damage induced by exposure to hydra- with the slow acetylator phenotype (9, lo), although the zine and arylamine containing foreign compounds de- development of spontaneous and procainamide-induced pends on the concentration of the foreign compound ANA is a dissociable process. Also, we have found that used as well as on the acetylator phenotype. Exposure the amount of drug-induced DNA damage in primary to hydralazine induced greater DNA damage in slow cultures of intact hepatocytes obtained from rapid and acetylator hepatocytes whereas exposure to the aryla- slow acetylator rabbits is dependent on the specific mine carcinogen, 2-aminofluorene, induced greater hydrazine drug or arylamine to which the hepatocytes DNA damage in rapid acetylator hepatocytes. are exposed as well as the acetylator phenotype. A lupus erythematosus-like syndrome is some- times seen in patients receiving therapeutic doses of Acetylator phenotype and ANA hydralazine, procainamide (PA), or isoniazid (1). Each Twenty inbred strains of mice, including NZB, of these drugs contains either a hydrazine or an aryla- NZW, and NZB/W F, female hybrids, were screened mine group which undergoes N-acetylation, an impor- for their ability to N-acetylate arylamine substrates. tant early step in their elimination from the body (2-6). The NZB/W F, female hybrid mice are characterized Persons with a low capacity for N-acetylation, such as by spontaneous appearance of ANA and several other phenotypic slow acetylators, are more likely to develop complications of human systemic lupus erythematosus drug-induced lupus or some manifestations of this dis- (1 1). One parental strain, the NZB, is characterized by a order from these drugs than rapid acetylators (7). Slow severe Coombs positive hemolytic anemia while the other parent, NZW, is afiparently normal immunologi- From the Department of Pharmacology, University of Mich- cally. Although the autoimmune disease of these mice is igan, AM Arbor, Michigan 48109. not worsened by lupus-inducing drugs (12,13), the dis- Address reprint requests to Dr. W. W. Weber, Department of Pharmacology, Medical Science Building 1, Room 6322, University of Of another autoimmune strain, A/J, is exacerbated Michigan, AM Arbor MI 48109. by these agents (14-16). Arthritis and Rheumatism, Vol. 24, No. 8 (August 1981) 980 WEBER AND TANNEN combinant inbred lines developed from A/J and A/J C57BL/6J strains (22). These observations suggested that A/J mice C 57B L/6 J could be a useful model for studying the human lupus ee4.s 0 diathesis since this strain is characterized by a pre- disposition to spontaneous and drug-induced ANA and t R by the slow acetylator phenotype. To evaluate the im- portance of the slow acetylator phenotype in the lupus diathesis, A/J mice were mated with C57BL/6J mice : fiffiittl 0 and associations were sought between slow acetylator phenotype and both spontaneous and procainamide-in- S 3 duced ANA among offspring from F,, F,, and backcross 2 matings. Since procainamide was the provocative drug in ,t this study, it was important to demonstrate that metabo- 0 50 100 150 200 250 300 350 lism of procainamide reflects the acetylator polymor- BLOOD PA-NAT ACTIVITY phism in mice. We showed that in both parental strains and animals in the F, and backcross offspring slow or Figure 1. Blood PA-NAT activity in rapid and slow acetylators @mol/min/mg protein). C57BL/6J and A/J mice can be distin- rapid p-aminobenzoic acid acetylators were also slow or guished on the basis of PA as well as PABA acetylation. Offspring in rapid acetylators of procainamide (Figure 1). However, the F2 and backcross populations which were phenotyped as rapid or no appreciable differences in either procainamide ace- slow acetylators of PABA are also appropriately rapid and slow ace- tylation in vitro by liver and certain other tissues or by tylators of PA. Correlation coefficient of blood PA-NAT versus procainamide acetylation in vivo was apparent between PABA-NAT activity in F,, F2, and backcross mice was 0.77 (P < 0.005). Each symbol represents 1 animal. (From J Pharmacol Exp the parental strains. Ther 213:485490, 1980. Reproduced by permission.) Mice 3 months of age were provided procain- amide in the drinking water at a concentration suf- ficient to supply 20 mg/kg/day. ANA determinations The principal substrates employed for screening were made in each mouse throughout the 37 weeks of were p-aminobenzoic acid and sulfamethazine since the study of untreated and procainamide-treated mice they show unimodal and bimodal populations of acety- as an indication of the autoimmune process occurring lation capacity respectively in human and rabbit popu- either spontaneously or as a result of procainamide ad- lations (17-19). In this survery, we found that the A/J ministration. An indirect slide test for ANA was em- strain is a slow acetylator phenotype (20). A/J mice can be differentiated from C57BL/6J mice, a rapid acety- lator autoimmune resistant inbred strain selected for these characteristics, by several measures of acetylation I C5 70L 16 J capacity: 1) by little or no detectable N-acetyl- transferase activities for p-aminobenzoic acid, pro- 0.. 0 cainamide (lo), aminofluorene, and benzidine in blood 1 (21); 2) by the excretion of less urinary acetyl- 0 I A/J a w sulfamethazine after oral administration of sulfametha- m I zine (11.4 f 4.4% for A/J versus 28.3 & 5.5% for 3 2 0 00 C57BL/6J) (10); and 3) by an 8-10 fold difference in N- 0 5 10 I5 20 acetyltransferase activity for aminofluorene or ben- zidine in liver (21). WHOLE LIFE ANA SCORE Genetic analysis of distributions for F,, F,, and Figure 2. Whole-life ANA scores of parental type animals. Whole- backcross progeny for both liver and blood reveals in- life ANA score is the sum of 8 determinations per mouse during the heritance patterns consistent with segregation at a single 37 weeks of the experiments. Individual tests were graded by using locus of 2 major codominant autosomal alleles for the scale previously described (10). A/J mice (both sexes) had higher whole-life ANA scores than C57BL/6J mice (2= 5.4; 2 > 3.8 is sig- aminofluorene N-acetyltransferase activity. This genetic nificant, P < 0.05). (From J Pharmacol Exp Ther 213:485490, 1980. hypothesis is further supported from analysis of re- Reproduced by permission.) PHARMACOGENETIC STUDIES 98 1 ployed which utilizes a peroxidase-conjugated rabbit A(J F YAfJ F1 XCS 7 c57 antimouse IgG to visualize ANA deposited in the nu- 1 'I I clear substrate (23). By using this test we codinned the findings of Ten Veen and Feltkamp (14) that A/J mice have a higher incidence of ANA-positive sera spontaneously than C57BL/6J mice (Figure 2). It was also noted that ~~ ~~ I ~ ANA responsiveness is probably a genetic trait since k2=88.9 backcrosses with the A/J parent produced offspring Figure 3. ANA responsiveness in parental lines and their back- with greater ANA positivity than offspring from back- crosses. a, Average percentage of positive ANA tests for all animals of a given type; 8 determinations per animal. Backcross animals were crosses with C57BL/6J parents (Figure 3). Whole-life pooled on the basis of negative findings for sex and drug treatment. b, ANA scores, the sum of 8 ANA determinations in each 2 values are computed on the basis of the average percentage of posi- mouse, also support the difference in ANA positivity tive ANA reactions and the total number of determinations (from 320 between strains (Figure 4). It is also evident from the to 636 per group). y > 3.8 is significant, P < 0.05. (From J Pharmacol whole-life ANA scores in the F, generation that an asso- Exp Ther 213:485-490, 1980. Reproduced by permission.) ciation still occurs after two generations between slow acetylator phenotype and high titers of spontaneous ANA. Effects of procainamide exposure on whole-life found that procainamide significantly suppressed ANA scores can be seen in Figure 5. Data in Figure 5 from titers in C57BL/6J mice as shown by the shift of the procainamide-treated mice (open-circles) show a shift to open circles to the left. Phenotypic slow acetylator off- the right indicating that ANA levels are induced in A/J spring, unlike their slow acetylator (A/J) parents, mice by procainamide as Ten Veen and Feltkamp (14) showed procainamide-suppression of ANA scores, observed. Contrary to their observations, however, we while ANA titers in rapid acetylator offspring were A/J I R I I : : 0 000 0.