GzmH kills independent of caspases. (A) Using cell-penetrating caspase substrates we analyzed caspase activities in living cells 10 hours after granzyme treatment. Data were normalized to GzmB and pooled from 3 (left) and 2 (right) independent experiments ± SD. In contrast to GzmB cell death mediated by sublytic concentrations of perforin (right) or SLO (left), GzmH did not induce cleavage of Ac-DEVD-AMC, which monitors the activities of caspases 3, 6, 7, 8, and 10. Addition of zVAD-fmk completely blocked the activation of caspases by GzmB. The same observations were seen with the caspase 9 substrate, Ac-LEHD-AMC, here only monitored using SLO as a translocator. Cleavage of the caspase 2 substrate, Ac-VDVAD-AFC, was observed in both GzmB- and GzmH-SLO treated K562 cells. In comparison with GzmB, the effect of GzmH was weaker but clearly above the background levels caused by sublytic SLO alone. (B) Western blot analysis (representative of 2 independent experiments) reveals that cytochrome c release is not a hallmark of GzmH-induced cell death. K562 cells were exposed to GzmH or GzmB and sublytic SLO for 6 hours. After cytosolic and membrane supernatant preparation, cytochrome c, in the case of GzmH treatment, could only be detected in the membrane fraction. The membrane fractions prepared 1, 2, and 6 hours after exposure to GzmB show a gradual decline of cytochrome c and gradually reappear in the cytosolic fractions. (C) GzmH kills independently of caspases in K562 and HL60 cells. Shown is the percentage of surviving cells (AV^-/PI^-) that had been treated with SLO and GzmB (i, iii) or GzmH (ii, iv) for 12 hours. Values represent the average of 3 experiments ± SD. The caspase inhibitors zVAD-fmk and zDEVD-fmk rescued cells from GzmB-induced apoptosis in a dose-dependent fashion, whereas killing by GzmH was minimally affected. Notably, the caspase 2 inhibitor zVDVAD-fmk did not protect the cells against the actions of GzmH. n.d indicates not determined.