The popliteal lymph node assay (PLNA) was proposed for the preclinical prediction of xenobiotics-induced autoimmune reactions in humans. Among the substances so far tested in this model, procainamide and isoniazid gave negative PLNA responses despite reports of lupus syndromes in man. To confirm the hypothesis that a metabolite instead of the parent molecule is involved, rats were pretreated with phenobarbital or beta-naphthoflavone, then injected with procainamide or isoniazid. In additional groups of animals, procainamide or isoniazid were injected together with S9-mix following various incubation times. Pretreated rats had a positive PLNA response when injected with procainamide, whereas preincubation with S9-mix resulted in a positive response to isoniazid. These results further support the validity of the PLNA.