We read the study by Inoue et al. [1] with great interest. In this study, a GFP-LC3 transgenic mouse model was used to monitor autophagy in kidney tissues during cisplatin treatment. It was shown that cisplatin induced autophagy in proximal tubular cells in kidneys prior to apoptosis. Cisplatin also induced autophagy in NRK-52E cells and, interestingly, inhibition of autophagy could block apoptosis in these cells. It was suggested that autophagy may contribute to tubular cell apoptosis during cisplatin nephrotoxicity. The in vivo study using the GFP-LC3 transgenic mouse model is very impressive. It shows that cisplatin can rapidly induce autophagy in kidneys. Importantly, the cells undergoing autophagy are positively stained for aquaporin1, a marker of proximal tubular cells. These observations are consistent with our previous examination by electron microscopy, which demonstrated autophagy in proximal tubular cells following cisplatin treatment of C57BL/6 mice [2]. In that study, autophagy in kidney tissues was also suggested by the accumulation of LC3-II, a lipidated form of LC3 that accumulates in autophagosomes [2]. In vitro, cisplatin was shown to induce autophagy in rat and pig proximal tubular cell lines [2, 3]. Inoue et al. have now further extended the findings to NRK-52E cells, another rat proximal tubular cell line. In all three cell lines, cisplatin induces autophagy in a treatment time-dependent manner, as shown by autophagosome formation (puncta in GFP-LC3-transfected cells) and LC3-II accumulation. Notably, the autophagic changes occur hours before apoptosis detected by cell morphology, caspase activation, and annexin V staining [1–3]. Clearly, autophagy is an early response of the cells to cisplatin-induced stress and not a result of apoptosis. The relationship between autophagy and cell death or apoptosis is, however, less clear. Whereas previous studies suggest a cytoprotective and prosurvival role for autophagy during cisplatin treatment of renal proximal tubular cells [2, 3], the study by Inoue et al. apparently reached an opposite conclusion [1]. Based on the observation that pharmacological or genetic inhibition of autophagy suppresses cisplatin-induced apoptosis in NRK-52E cells, Inoue et al. concluded that autophagy is a cell-damage index and is involved in cell death. However, in a rat proximal tubular cell line used in our previous study [2], inhibition of autophagy sensitized the cells to cisplatininduced apoptosis, suggesting that autophagy is not a cellkilling but a prosurvival mechanism. Similar results were shown by Yang et al. [3] in LLC-PK1, a pig proximal tubular cell line. The cause of the obvious discrepancy between the Inoue study and previous studies is unclear, although it is generally believed that, depending on the injury or stress, autophagy can be either cytoprotective or detrimental. Nevertheless, it is noteworthy that autophagy has been shown to be protective in relevant renal injury models [4]. For example, using both in vitro and in vivo experimental models, our recent study suggests a renoprotective role for autophagy in renal ischemia–reperfusion [5]. Is autophagy indeed protective or destructive in kidneys? We do not have a clear-cut answer at the moment. M. Jiang Z. Dong (&) Department of Cellular Biology and Anatomy, Medical College of Georgia, 1459 Laney Walker Blvd., Augusta, GA 30912, USA e-mail: zdong@mail.mcg.edu