Background: Recent data suggest that NK cell mediated antibody dependent cellular cytotoxicity (ADCC) is a major mechanism of action of the anti-CD20 monoclonal antibody (mAb) rituximab and the anti-CD52 mAb alemtuzumab, which are frequently applied in patients with non-Hodgkin's lymphoma and chronic lymphocytic leukemia. However, the exact mechanisms leading to NK cell activation are not completely understood and the cytotoxic subpopulation of peripheral blood NK cells mediating ADCC remains to be defined. In order to quantify and characterize the NK cells mediating ADCC, we used a novel flow cytometric assay, which detects the lytic granule membrane protein CD107a as a marker for NK cell degranulation.Methods: PBMCs from healthy individuals were coincubated at 37°C for 3 h with different human leukemia and lymphoma cell lines. In each tube, containing 200μl effector/target suspension (4x105 cells), 15μl of PE-Cy5 conjugated anti-CD107a monoclonal antibody was added prior to incubation. To assess antibody dependent cellular cytotoxicity (ADCC) saturating concentrations (10μg/ml) of rituximab or alemtuzumab were used. After the first 1 h 5μl of the secretion inhibitor 2 mM monensin was added. At the end of coincubation cells were stained with mAbs (CD56, CD3, NKG2D, CD69, CD94, NKp30, NK46) for flow cytometry. NK cell-mediated cytotoxicity (specific lysis) was analyzed by flow cytometric detection of propidium iodide uptake.Results: After coincubation with NK sensitive K562 cells up to 6% of CD56+ cells expressed CD107a, indicating that a subpopulation of NK cells releases cytotoxic granules after contact with these target cells. In contrast, coincubation with NK-resistant leukemia cells (ML2, EHEB, DAUDI, RAJI, AM0-1, YT-1) was not followed by an increased surface expression of CD107a. However, when rituximab was added to CD20+ lymphoma or leukemia cells (EHEB, DAUDI, RAJI) we observed that up to 15% of NK cells expressed CD107a after coincubation. In contrast no increased CD107a surface expression was observed when rituximab was added to the CD20− cell lines AMO-0 and YT-1, which excludes unspecific NK cell activation. When alemtuzumab was added to the CD52+ cell lines AMO-1, DAUDI, EHEB, RAJI and YT-1, surface expression of CD107a on NK cells was increased considerably. The majority of degranulating NK cells had the phenotype: CD56dim, CD69+, NKG2D+, NKp30−, NKp46− and CD94−. Furthermore we found that the CD107a assay can also visualize ADCC under clinical conditions as we observed increased numbers of NK cells degranulating in response to CD20+ lymphoma cell lines in patients with non-Hodgkin's lymphoma treated with rituximab. The number of degranulating NK cells was closely related to the concentration of rituximab and the effector:target ratio, showing a maximum at a ratio of 1:1 and concentrations above 5μg/ml. CD107a surface expression and specific lysis demonstrated a strong positive correlation (r2 = 0.99), confirming that NK cell cytotoxicity can be assessed by this method.Conclusion: The CD107a assay represents a promising new method not only for assessment of natural cytotoxicity on a single cell level but also for determination of ADCC in vitro and in patients treated with mAb. In clinical practice, it may help to find optimal doses and time schedules for the treatment with different mAbs.