Chloroquine-induced autophagic vacuole accumulation and cell death in glioma cells is p53 independent

Abstract

Glioblastoma (GBM) is a high-grade central nervous system malignancy and despite aggressive treatment strategies, GBM patients have a median survival time of just 1 year. Chloroquine (CQ), an antimalarial lysosomotropic agent, has been identified as a potential adjuvant in the treatment regimen of GBMs. However, the mechanism of CQ-induced tumor cell death is poorly defined. We and others have shown that CQ-mediated cell death may be p53-dependent and at least in part due to the intrinsic apoptotic death pathway. Here, we investigated the effects of CQ on 5 established human GBM lines, differing in their p53 gene status. CQ was found to induce a concentration-dependent death in each of these cell lines. Although CQ treatment increased caspase-3-like enzymatic activity in all 5 cell lines, a broad-spectrum caspase inhibitor did not significantly attenuate death. Moreover, CQ caused an accumulation of autophagic vacuoles in all cell lines and was found to affect the levels and subcellular distribution of cathepsin D, suggesting that altered lysosomal function may also play a role in CQ-induced cell death. Thus, CQ can induce p53-independent death in gliomas that do not require caspase-mediated apoptosis. To potentially identify more potent chemotherapeutics, various CQ derivatives and lysosomotropic compounds were tested on the GBM cells. Quinacrine and mefloquine were found to be more potent than CQ in killing GBM cells in vitro and given their superior blood-brain barrier penetration compared with CQ may prove more efficacious as chemotherapeutic agents for GBM patients.

Fig. 1.

CQ induces cell death in GBM cells irrespective of their p53 status. (A) Exposure to CQ for 24 hours induced a concentration-dependent decrease in viability in the U87 as well as the LN308 cells. Bright-field images demonstrating the morphological changes in U251 cells following CQ (50 µM) treatment for 24 hours. Data points represent mean ± standard deviation with n = 6 (*P < .05 by 1-way analysis of variance/Bonferroni post-test compared with untreated [UT] group). (B) Elevated levels of caspase-3–like activity were detected in the LN308 cells 24 hours following exposure to CQ. Levels of caspase-3–like activity were normalized to the UT group. Data points represent mean ± standard deviation with n = 6 (*P < .05 by 1-way analysis of variance/Bonferroni post-test compared with UT group). (C) Western blot analysis of lysates collected from U87 cells treated for 24 hours with 75 µM CQ revealed an increase in levels of the activated form of caspase-3. β-Tubulin served as loading control. (D) CQ-induced death was not significantly attenuated by caspase inhibition. The broad-spectrum caspase inhibitor BAF (100 µM) was able to prevent caspase activation (24 hours) but had no effect on CQ-induced death in U251 cells (48 hours). In contrast, BAF attenuated STS-mediated death in these cells (48 hours). Data points represent mean ± standard deviation with n = 6 (*P < .05 by 2-way analysis of variance/Bonferroni post-test compared with the group with CQ/STS treatment only).

Fig. 2.

CQ promotes AV accumulation in GBM cells. (A) LN308 cells exposed to 50 µM CQ for 24 hours demonstrated significantly higher levels of AV-associated LC3 immunoreactivity (red) relative to untreated (UT) cells. Cell nuclei were counterstained with bisbenzimide (blue). Scale bar equals 20 µm. (B) Western blot analysis of U87 whole cell lysates collected at 3, 6, and 24 hours following treatment with 50 µM CQ revealed a time-dependent increase in levels of LC3-II. β-Tubulin served as loading control. (C) Addition of 3-MA (5 mM) exacerbated cell death in response to STS treatment but provided significant protection against CQ (50 µM)-induced death as assessed 36 hours following treatment. Data points represent mean ± standard deviation with n = 6 (*P < .05 by 1-way analysis of variance/Bonferroni post-test compared with UT group).

Fig. 3.

CQ affects the subcellular localization and processing of cathepsin D. (A) Untreated (UT) LN308 cells showed discrete lysosome-associated cathepsin D immunoreactivity (red), whereas CQ-treated cells (50 µM, 24 hours) exhibited diffuse cytoplasmic immunoreactivity. Cell nuclei were counterstained with bisbenzimide (blue). Scale bar equals 20 µm. (B) Western blot analysis of lysates collected from U87 cells treated for 24 hours with 75 µM CQ revealed a time-dependent decrease in levels of the activated form of cathepsin D. β-Tubulin served as loading control.

Fig. 4.

Effect of CQ-related compounds on GBM cell lines. (A) Quinacrine and mefloquine induced a concentration-dependent decrease in cell viability and (B) increase in levels of caspase-3–like enzymatic activity in U87 cells. Data points represent mean ± standard deviation with n = 6 (*P < .05 by 1-way analysis of variance/Bonferroni post-test compared with the untreated [UT] group). (C) Relative to UT cells, quinacrine-treated U251 cells (6 µM, 48 hours) demonstrated an increase in “clumped”, AV-associated LC3 immunoreactivity (red) and also exhibited (D) diffuse cathepsin D immunoreactivity (red). Cell nuclei were counterstained with bisbenzimide (blue). Scale bars equal 20 µm.