We investigated the in vivo effects of rufloxacin and ciprofloxacin on T-cell subsets and tumor necrosis factor production in mice infected with Bacteroidesfiagiis. These quinolones did not alter the helper/suppressor ratio but did modulate the kinetics of tumor necrosis factor production in infected animals. This result correlated with the observed therapeutic efficacies of the quinolones.Several recent reports indicate that quinolone antibiotics at clinically achievable levels modulate the in vitro growth of T cells and the production of cytokines by activated lymphoid cells (1,2,5,(7)(8)(9). The purpose of this investigation was to determine whether quinolones in vivo alter the T-cell response and cytokine production in an infection model. Intra-abdominal abscesses produced by Bacteroides fragilis are a good model to study the in vivo effects of quinolones because, first, abscess formation and elimination of abscesses require T-cell function (10, 11) and, second, the cytokines play an important role in elimination of infection (3,11).Five-to six-week-old female C3HIHeN mice were infected with an abscess-forming mixture of B. fragilis with sterile cecal contents and BaSO4 as previously described (3). Six hours after the inoculation, the mice were treated with rufloxacin (50 mg/kg of body weight) once a day or with ciprofloxacin (40 mg/kg) twice daily for 10 days. Each group contained 11 animals. The antibiotics were given intraperitoneally. Rufloxacin was given once a day because of its long half-life (6). As a control, a group of mice were treated with saline. In addition, a group of uninfected animals used as controls received either ciprofloxacin or rufloxacin so as to determine the effect of the quinolones on T-cell subsets without the associated effects of infection with B. fragilis. On day 11 posttreatment, mice were sacrificed by cervical dislocation under anesthesia, the spleens were removed, and a single cell suspension was prepared as previously described (4). Cells were stained with fluorescein isothiocyanate-labeled Thy 1.2 (total T-cell), L3T4 (helper), or LYT2 (suppressor) antibody (Becton Dickinson, San Jose, Calif.). The binding of antibodies to T lymphocytes was analyzed with a FACScan flow cytometer (Becton Dickinson). Mice treated with or without drugs were bled at 2, 24, 48, and 72 h posttreatment, sera were pooled, and tumor necrosis factor (TNF) levels in the serum were determined by enzyme-linked immunosorbent assay as suggested by the manufacturer (Genzyme, Cambridge, Mass.). Mice were killed 1 day after completion of therapy; their abdomens were opened and examined for the presence or absence of ab-* Corresponding author.scesses. If abscesses were present, then pus was cultured to confirm the presence of B. fragilis. B. fragilis was recovered from the abdominal abscess pus of all control untreated animals. The absence of B. fragilis was considered to indicate that the animals had been cured. The results were analyzed for statistical significance by chi-square analysis, and P values of ...