Cycloheximide promotes paraptosis induced by inhibition of cyclophilins in glioblastoma multiforme

Abstract

Cancer is the second leading cause of death worldwide. Current treatment strategies based on multi-agent chemotherapy and/or radiation regimens have improved overall survival in some cases. However, resistance to apoptosis often develops in cancer cells, and its occurrence is thought to contribute to treatment failure. Non-apoptotic cell death mechanisms have become of great interest, therefore, in hopes that they would bypass tumor cell resistance. Glioblastoma multiforme (GBM), a grade IV astrocytic tumor is the most frequent brain tumor in adults, and has a high rate of mortality. We report that NIM811, a small molecule cyclophilin-binding inhibitor, induces catastrophic vacuolization and cell death in GBM cells. These unique features are distinct from many known cell death pathways, and are associated with an incompletely defined cell death mechanism known as paraptosis. We found that NIM811-induced paraptosis is due to unresolved ER stress. The abnormal upregulation of protein translation was responsible for the build-up of misfolded or unfolded proteins in ER, whereas pro-survival autophagy and UPR signals were shutdown during prolonged treatment with NIM811. Although cycloheximide has been claimed to suppress paraptosis, instead we find that it only temporarily delayed vacuole formation, but actually enhanced paraptotic cell death in the long term. On the other hand, mTOR inhibitors rescued cells from NIM811-induced paraptosis by sustaining autophagy and the UPR, while specifically restraining cap-dependent translation. These findings not only provide new insights into the mechanisms underlying paraptosis, but also shed light on a potential approach to enhance GBM treatment.

Figure 1

NIM811 induces ER vacuolization and paraptosis cell death in GBM cells. (a) Pan-caspase inhibitor Qvd-oph was not able to rescue cell death induced by 10 μM NIM811 at 24 h. (b) Electron microscopic images revealing the morphological structures of vacuoles, scale bar=10 μm. (c) Labeling cells with GFP-ER signal sequence of calreticulin-KDEL (1–2) and TagRFP-Leader sequence of E1 alpha pyruvate dehydrogenase (3-4) or RFP-lamp1 (5-6) indicate that the vacuoles originate from the ER. Scale bar=25 μm

Figure 2

NIM811 inhibits tumor growth in vivo (a) 15 μM of NIM811 caused vacuolization in G22VF cells at 24 h, and substantial cell death at 48 h. Scale bar=50 μm. (b) Nude mice (n=30) were injected subcutaneously with 2E⁶ G22VF cells. When the mice developed tumors larger than 100 mm³, they were randomly divided into three groups to receive treatments: vehicle (1% ethanol+9% cremophor EL+ 90% normal saline), NIM811 intraperitoneal (25 mg/kg), NIM811 oral gavage (25 mg/kg). Using 1 cm³ as cut-off, NIM811-treated mice had slower tumor growth rate than the mice in vehicle group. Curves were significant different by the log-rank (Mantel–Cox) test, P=0.0081. (c) Using tumor size of 2.5 cm³ as killing criteria, NIM811-treated mice survived longer than vehicle-treated mice. Curves were significantly different by log-rank (Mantel–Cox) test, P=0.0198

Figure 3

NIM811-induced vacuolization is sensitive to protein synthesis inhibition. (a) U251 cells were incubated with 10 μM NIM811 containing media for 9 h, then media were replaced with fresh DMEM (10% FBS). Cells were then tracked by live microscopy (scale bar=25 μm) for 24 h. (b) U251 cells were pretreated with 20 μM cycloheximide (CHX) or 10 μg/ml of blasticidin or 20 μM of U0126 respectively for 2 h followed by 24 h 10 μM NIM811 incubation (scale bar=25 μm). (c) Twenty hours of NIM811 treatment leads to accumulation of ubiquitinated proteins, pretreatment with 20 μM cycloheximide for 2 h prevented this build-up. (d) After 24 h or 48 h of 10 μM NIM811 treatment, cells were collected for 20S proteasome activity assay. Cells treated with 1 or 5 μM MG132 served as positive control for this assay

Figure 4

NIM811 induced a transient upregulation of UPR and autophagy. (a and b) In all, 0 μM NIM811 transiently induced UPR signaling (P-EIF2a, Bip, ATF4, CHOP upregulation) at 6 h, but except for Bip, the activation of UPR mediators disappeared at 9 h. Tunicamycin (Tm) and thapsigargin (Tg) were used as positive controls for the UPR. (c) In total, 10 μM NIM811 was added to GFP-LC3B transfected U251 cells and imaged by fluorescence microscopy at 0 h and 6 h. Scale bar=25 μm. LC3 puncta were detected at 3–6 h. LC3 puncta formation occurred before the appearance of vacuoles. (d) LC3-I and -II conversion occurred at 4–6 h of 10 μM NIM811 treatment accompanied by p62 degradation. (e and f) In all, 10 μM NIM811 incubation caused p62 accumulation at 24 h, which persisted at later time points (40–46 h). (g) Immunofluorescence staining of SQSTM1/p62 (red) and nucleus (blue) after 24 h vehicle (left) or 10 μM NIM811-treated (right) U251 cells. Scale bar=25 μm

Figure 5

mTOR inhibitors block NIM811-induced vacuolization. (a and b) U251 cells were treated with 10 μM NIM811 in combination with 100 μM chloroquine (CQ) or 100 nM bafilomycin A1 (Baf-A) for 24 h. Scale bar=50 μm. (c) Four-hour pretreatment with 100 nM rapamycin or 150 nM torin-2 blocked vacuolization in U251 cells at 48 h. Scale bar=50 μm. (d and e) Western blotting of NIM81-treated cells demonstrated transient stimulation of mTOR as well as AKT phosphorylation at 2-6 h. Bip remained upregulated after 24 h. (f) Cell lysates were collected from 2 h to 10 h of 10 μM NIM811-treated U251 cells, and analyzed for protein ubiquitination

Figure 6

NIM811 activates both CD and CI translation. (a-c) CD and CI translation were measured by the dual renilla/firefly luciferase assay after 24 or 48 h of indicated treatments, and relative light units were normalized to cell numbers. NIM811 significantly increased the CI and CD translation at 10 and 15 μM, *P<0.05. (d-f) Pretreatment of rapamycin, torin-2 effectively decreased CD translation during NIM811 incubation, whereas brief pretreatment with cycloheximide substantially decreased both CD and CI translation. *P<0.05

Figure 7

Cycloheximide inhibits autophagy and UPR signaling, whereas activating downstream mTOR substrate phosphorylation. (a) Brief treatment with cycloheximide led to increased p62 accumulation at 20 h of NIM811 treatment (quantitation with the actin loading control on the right). (b) Two-hour cycloheximide pretreatment decreased LC3-I and II levels even after prolonged NIM811 treatment (44 h), (c) 10 μM NIM811 stimulated phosphorylation of P70S6K at 2 h and p-EIF2a at 4 h, while 2 h of cycloheximide increased the P-P70S6K level but decreased eIF2a phosphorylation. (d) Two-hour cycloheximide pre-incubation boosted the p-p70S6K level after 24 h, whereas, with rapamycin pretreatment, P70S6K phosphorylation decreased and p62 was degraded more efficiently

Figure 8

Cycloheximide pretreatment exacerbates cell death induced by prolonged NIM811 treatment. (a) Cycloheximide pretreatment did not improve U251 live cell numbers after 24 h of 10 μM NIM811 incubation. (b) Colony number counts on day 13 indicates that pretreatment with rapamycin significantly increased colony numbers over NIM811 alone treatment (from 0.5 to 5 μM), *P<0.05, whereas, cycloheximide pretreatment substantially decreased colony numbers by 0.5 μM NIM811 incubation, *P<0.05. (c) Microscopic images were obtained on day 13 to demonstrating the morphology of the colony cells. Rapamycin also blocked the vacuolization triggered by NIM811. Scale bar=50 μm