S164 DISCUSSION: Drug-Induced Lupus-like Syndromes Ziff: Up to this point, we have directed our primary attention to the interplay of genetic and virologic influences as determinants of susceptibility to systemic lupus. This is an appropriate moment to turn the discussion to the lupus-like disease states induced by certain drugs. For those not familiar with this area, there are certain drugs, such as hydralazine and procainamide, which when taken for a period of weeks or months, produce a reversible disease state that resembles systemic lupus erythematosus in many clinical and laboratory characteristics. The disease clearly differs from classic lupus in that it is less severe and rarely, if ever, has manifest nephritis. Moreover, the disease wanes and disappears after withdrawal of the drug. It is an unresolved question whether the drug related disease is qualitatively or quantitatively different from classic systemic lupus. Bluestein: One of the curious features about drug-induced lupus is that the prevalence of antinuclear antibodies in individuals taking the inducing drugs greatly exceeds the number of those who develop an actual clinical illness. I would like briefly to describe our finding of increased frequency and elevated levels of antilymphocyte antibody in patients with procainamide-induced lupus. I n contrast to the antinuclear antibody titers, the levels of these antilymphocyte antibodies correlate very well with the clinical symptoms in the patients. Lymphocyte antibody levels fell rapidly after the drug was discontinued. Serial serum samples were available from 5 patients. In 3 patients with initial 50% cytotoxicity titers of 1:4, 1:8, and 154, lymphocytotoxicity was not detectable 8 weeks after cessation of procainamide therapy. In 2 other patients with initial titers of 1:32, the titers fell to 1:2 and 1:4 in the same period. The occurrence of these lymphocyte antibodies led us to speculate that perhaps they are being generated by a new antigenic determinant created by a drug-cell membrane interaction. In searching for evidence of such an interaction, we’ve looked for direct effects of the drug on in vitro T-cell function. We found that at pharmacologic concentrations of procainamide, there is an enhancement of the vitro response to PHA. This is most marked at suboptimal stimulating concentrations of PHA where the response is 50% more than in cultures without pro- cainamide. With higher doses of drug, molar, which is about 100-fold higher than the concentration in the patients’ serum should be), we eliminate lymphocyte responsiveness to PHA without killing the cells. I would like to think that the drug-cell membrane interaction, either by creating a neoantigen or by altering lymphocyte regulation, or both, may play a central role in the generation of the drug-induced autoimmune syndrome. We have shown that Dilantin has the same in vitro effect on PHA stimulation that was described for procainamide. Warner: Were the patients with lymphocyte antibodies lymphopenic? Bluestein: The patients for whom we have the data all had peripheral lymphocyte counts greater than 1 OOO/ mm3. Hess: I think that drug induced lupus is a most intriguing model and we’ve really rather ignored it. I’d make a plea for lots more people to get into the act. We have about 25 patients we’ve followed now for up t o 3 years, doing the complete immunologic profile before they went on hydralazine and then following them along at stated periods to see what really happens while they are on this drug. A s you would expect, about 50% of them produced positive antinuclear reactions in undiluted sera. With dosages of 400 mg of hydralazine, which is about twice what is usually given, only one person in a group of 25 patients followed for 3 years evidenced the clinical syndrome. I t is intriguing that although 50% demonstrate A N A in low titers, their PHA and cellular immune responses are rather normal all the time that they are on the drug. Patients in these studies were matched for age and sex, and since they are all older people who have heart disease and other kinds of problems, they were matched for other antihypertensive drugs and for their basic heart disease. Dr. Michael Denman found that in 19 of 20 people before practolol exposure, the cell-mediated immune responses were normal; later they were diminished in some who developed the clinical syndrome. DISCUSSION Ziff: So it wasn’t primary. Hess: No, it wasn’t primary. I think that is an important point. Also, we’ve done lymphocytotoxic antibodies in o u r study and they were just what you would expect to find in the normal population. So I would like to know from Dr. Bluestein how many patients did he have that were positive. And were these the symptomatic patients? Bluestein: The lymphocytotoxic antibody studies were done on 18 patients with well developed clinical syndromes. Control groups were I5 patients taking procainamide who were free of SLE and 63 healthy middle-aged (45-65 years) men. The complement dependent, dye-exclusion microtoxicity assay was used. Sera from 10 of 18 procainamide-induced SLE patients killed at least 50% of normal human peripheral blood lymphocytes at serum dilution of 1:2 or greater. Five of these sera had 50% cytotoxicity titers of 1 :8 or more. In contrast, only 3 of 15 sera from patients without SLE on the drug were cytotoxicnone of those at titers greater than 1: 2. Lymphocytotoxicity was found in 4% of sera from the middleaged men-again, none with titers above 1 : 2 Hess: One wonders if it isn’t an uncovering mechanism because there really is nothing there until the clinical syndrome appears. I think that is an important point to establish. It may not be the drug itself. Winfield: My comment may be relevant to the statements Dr. Kunkel made concerning the induction of these autoantibodies and also to Dr. Bluestein’s statement that procainamide may work at the level of the membrane. In New York we studied patients with acute myocardial infarction and found that in a large number of these patients there was release of ribonucleoprotein antigen into the circulation. If these patients had an arrhythmia and were treated with proclaimaide, this antigen was not detectable. Antibodies to ribonucleoprotein appeared in serum in a very short time. The thesis we made was that the procainamide combined with ribonucleoprotein to create a new antigen. One of the unusual aspects of this study was the S165 rapidity with which these anti-ribonucleoprotein antibodies developed; one wondered whether procainamide was perhaps interfering with suppressor cell function in the development of this very rapid autoimmune response. McDuffie: To follow up the comment that Dr. Hess made, one idea about the drug-induced lupus is that there may be some kind of immunologic cross-reaction between the drug and the nuclear antigen. Several years ago Dr. Hahn showed that several patients who developed drug-induced lupus made antibodies to hydralazine as well as to DNA. Dr. Hess and her group in Cincinnati showed that rabbits immunized with hydralazine conjugates make antibodies to hydralazine that cross-react with DNA. Like Dr. Hess, we have also been following a number of patients on hydralazine, and a number of them have developed antibodies to DNA. But, whereas we find that the guinea pig antibodies that bind to DNA produced by immunizing with hydralazine conjugates can be readily inhibited by hydralazine, we cannot inhibit the human antibodies in the same way. So there doesn’t seem to be a cross-reaction. Agnello: John Oates has looked at the metabolism of procainamide in those individuals who develop positive ANAs and those who don’t, and he finds that the ones w h o develop positive ANAs eliminate the drug much more slowly. So that is another factor involved here. Hess: With respect to the race studies Dr. Christian mentioned in which lupus is three times more prevalent in blacks than i n whites, it is of interest to point out that in the drug-related lupus, this is absolutely not so. Mostly white patients get the drug-induced lupus; it is extremely rare in blacks. Whether this is related to the acetylation rate isn’t entirely clear. There are more slow acetylators in whites than in blacks, and this finding also holds true in natural lupus. Different acetylation rates seem to be a genetic trait. How the various metabolites of drugs and other agents relate to this genetic trait is intriguing to think about.