Abstract： Astrocytes maintain homeostasis of neuronal microenvironment, provide metabolic and trophic support to neurons and modulate neuronal responses to injury. Rotenone specifically inhibits mitochondrial complex I, and long exposure to rotenone may increase the risk for Parkinson's disease (PD) and cause Parkinsonism. However,little is known about the role of astrocytes in the process of rotenone-induced dopaminergic neuron injury. In order to investigate this issue, we used MN9D cells as a cell model of dopaminergic neurons and rotenone as a toxin to initiate mitochondrial deficiency. MN9D cells treated with the normal medium or astrocyte-conditioned medium (ACM) were exposed to different concentrations of rotenone for different time followed by cell viability measurement by MTT assay. Besides, various concentrations of ACM and temporally different treatments were devised to evaluate protective efficiency of ACM. Growth curve of cells in the normal medium or ACM was continuously assessed by cell counting for 8 d. The influence of rotenone and ACM on cellular oxidative stress was determined by DCFH-DA staining followed by flow cytometric analysis. Glutathione (GSH) content after treatment of ACM or rotenone was measured by GSH assay kit. Our results showed that rotenone decreased viability of MN9D cells in a dose-dependent manner and ACM treatment significantly attenuated rotenone toxicity at each concentration. No significant difference in growth rate was observed between the normal medium and ACM treatment. Four concentrations of ACM, namely 1/3ACM, 1/2ACM, 2/3ACM and pure ACM, all displayed protection, increasing cell viability to (124.15±0.79)%, (126.59±0.82)%, (125.84±0.61)% and (117.15±1.63)% of the cells exposed directly to rotenone, respectively. Treatment with ACM through the whole experiment except the initial 24 h, 24 h before or at the same time of rotenone addition all exerted protective effects, with cell viability being (110.11±2.52)%, (113.30±2.36)%, (114.42±2.00)% of the cells exposed directly to rotenone, respectively. Conversely, ACM treatment 12 h after rotenone addition had no protective effect, with cell viability being (102.54±1.36)% of the cells exposed directly to rotenone. Moreover, ACM treatment up-regulated GSH level in MN9D cells nearly twofold. Incubation with 100 nmol/L rotenone for 24 h depleted GSH level by nearly two thirds of the control, but ACM treatment mitigated the drop of GSH level, maintaining its content at (147.83±0.63)% of the control. Consistent with GSH change, rotenone administration resulted in a positive rate of 96.24% of DCF staining, implying a great extent of oxidative stress, whereas treatment with ACM reduced the extent of oxidative stress to a positive rate of 78.31%. Taken together, these findings suggest that astrocytes protect MN9D cells from oxidative stress caused by rotenone, and GSH partially accounts for the protection. Therefore, astrocytes may play a protective role in the process of PD.