Characterization of tumour necrosis factor α release by human granulocytes in response to procainamide challenge

Abstract

The role of human granulocytes in the promotion of procainamide (PA) toxicity in vitro has been studied and one of the agents responsible for DNA strand scission and cell death in human target cells has been characterized. Crude peripheral blood mononuclear cells (cPBMNs) isolated by density centrifugation, and the lymphocyte cell lines—CCRF-HSB2 and WIL-2NS—were exposed to PA, and DNA strand breaks were quantified by fluorescent analysis of DNA unwinding. Therapeutic plasma concentrations of PA (0–50 μM) caused dose-dependent cytotoxicity, determined by dye exclusion, and strand breaks in cPBMNs incubated for 3 and 1.5 hr at 37 °, respectively. Using 50 μM PA a five-fold increase in DNA strand breaks was observed after 1.5 hr, with significant induction of strand breaks also being observed for 10 and 25 μM concentrations. Toxicity was much reduced in lymphocyte cell lines (maximal killing = 3.0% at 50 μM PA compared with 13.2% in cPBMNs). A similar decrease in toxicity was observed where N-acetyl procainamide (NAPA) was substituted for PA (less than 50% of strand breaks at all concentrations). Further investigations showed that the presence of a contaminating granulocyte population in the cPBMN fraction was responsible for the induction of PA toxicity. Incubation of a highly enriched granulocyte population with PA for 1 hr prior to exposure to purified peripheral blood mononuclear cells (pPBMNs) led to the complete restoration of the toxic effects. The resulting cyto- and genotoxicity were not significantly different to levels observed in cPBMNs. Significantly, incubation of granulocytes with NAPA did not induce toxicity in target pPBMNs. Ultrafiltration of granulocyte supernatants led to the identification of two toxic fractions of < 3000 and > 30,000 Da. Temporal studies showed that the toxicity associated with the < 3000 Da fraction appeared during the first 10–15 min incubation with PA whereas the > 30,000 Da fraction did not display significant toxicity until the 40–60 min period. Further assessment of the nature of these agents indicated that the 30,000 Da fraction was a protein. SDS-PAGE analysis showed an inducible 17,800 Da species appearing in granulocyte supernatants after 40 min incubation with PA. Dot blot analysis indicated that tumour necrosis factor α (TNFα) was present in the > 30,000 Da fraction. Evidence that TNFα was the high-molecular weight species responsible for PA-induced toxicity was obtained from neutralization assays employing an anti-TNFα antibody. Treatment of granulocytederived supernatants with this antibody led to the specific removal of a significant portion of both the cyto- and genotoxicity (> 50% in both cases). Our data suggest a highly effective mechanism for the promotion of drug-induced lupus by PA. The specific induction of TNFα represents a highly effective mechanism for the induction of the inflammatory response commonly associated with this syndrome and provides a potential route for the specific induction of severe adverse drug reactions. Furthermore, the biochemical mechanisms of action of lupus-inducing drugs may provide insight into the molecular mechanisms leading to development of idiopathic lupus.