To THE EmlOR-Tuberculosis is still an important health problem, especially in developing countries [1]. The AIDS pandemic has exacerbated the problem, as immunosuppressed individuals are highly susceptible to infections with mycobacteria [2]. As intracellular pathogens, mycobacteria infect and replicate within cells of the mononuclear phagocyte series [3]. Although in mice the classic macrophage-activating interferon-v (IFN-'Y) renders macrophages/monocytes resistant to mycobacterial growth [4], it is not clear if a similar activation can be attained in human cells [4]. Interest has been increasing regarding the role that neutrophils may play in mycobacterial infections [5-7]. Indeed, neutrophil recruitment in the early lesions of tuberculosis has been observed in early studies [8]. The possibility that these cells, activated or not by cytokines, could kill Mycobacterium tuberculosis efficiently is important and deserves careful examination. A virulent strain of M. tuberculosis H37Rv was grown in 7H9 broth (Difco, Detroit) and 0.05%(vol/vol) Tween 80. Mycobacteria were harvested by centrifugation at 1500 g for 10 min, washed, and resuspended in Hanks' balanced salt solution (HBSS) with 0.01% gelatin and om% Tween 80. The concentration (colony-forming units, cfu) was determined by spectrophotometry. Staphylococcus aureus 38F was cultured overnight in nutrient broth no. 2 (Oxoid, London) at 37 C, washed in PBS, and suspended in HBSS-gelatin-Tween 80 at 107 cfu/ml. Neutrophils were isolated from purified protein derivative (Pl'Dj-negative individuals by layering whole blood on Ficoll-Isopaque solution (Pharmacia, Uppsala, Sweden). After lysis of erythrocytes, cells were adjusted at 107 granulocytes/ml in HBSS-gelatin-Tween 80. Cells were 97% neutrophils, at a viability of> 95%. To determine killing of microbes, serum was prepared from blood of healthy donors with AB blood group that was clotted for 1 h. Equal volumes of H37Rv or S. aureus (107 or 106 cfu) and 107 neutrophils/ml in HBSS-gelatin-Tween were mixed with 10% AB serum in a plastic tube at 37 C under slow rotation (4 rpm). At predetermined intervals, samples were taken, 0.1 ml was added to 0.9 ml of ice cold distilled water with 0.01% bovine serum albumin and 0.01% Tween 80, and clumps were disrupted by a brief (30-s) sonication (power output, 2.5 WIs). The number of viable mycobacteria was determined by plating dilutions on 7H1O agar plates