An insight into the mechanistic role of p53-mediated autophagy ...

[Autophagy 5:5, 663-675; 1 July 2009]; ©2009 Landes Bioscience

Research Paper

An insight into the mechanistic role of p53-mediated autophagy induction in response to proteasomal inhibition-induced neurotoxicity Yunlan Du,1 Dehua Yang,2 Liang Li,2 Guangrui Luo,2 Ting Li,2 Xiaolan Fan,2 Qian Wang,1 Xin Zhang,1 Yi Wang2 and Weidong Le1-3,* 1Institute of Neurology; Ruijin Hospital; Shanghai JiaoTong University School of Medicine; Shanghai, China; 2Institute of Health Science; Shanghai Institutes for Biological Sciences; Chinese Academy of Sciences; & Shanghai JiaoTong University School of Medicine; Shanghai, China; 3Department of Neurology; Baylor College of Medicine; Houston, TX USA

Key words: p53, autophagy, ubiquitin-proteasome system, Parkinson disease, rapamycin

The ubiquitin-proteasome system (UPS) and the autophagylysosomal pathway (ALP) are the two most important components of cellular mechanisms for protein degradation. In the present study we investigated the functional relationship of the two systems and the interactional role of p53 in vitro. Our study showed that the proteasome inhibitor lactacystin induced an increase in p53 level and autophagy activity, whereas inhibition of p53 by pifithrin-α or small interference RNA (siRNA) of p53 attenuated the autophagy induction and increased protein aggregation. Furthermore, we found that pretreatment with the autophagy inhibitor 3-methyladenine or beclin 1 siRNA further activated p53 and its downstream apoptotic pathways, while the autophagy inducer rapamycin showed the opposite effects. Moreover, we demonstrated that rapamycin pretreatment increased tyrosine hydroxylase (TH) protein level in dopamine (DA) neurons, which was associated with its induction of autophagy to degrade aggregated proteins. Our results suggest that p53 can mediate proteasomal inhibition-induced autophagy enhancement which in turn can partially block p53 or its downstream mitochondria-dependent apoptotic pathways. Further autophagy induction with rapamycin protects DA neurons from lactacystin-mediated cell death by downregulating p53 and its related apoptotic pathways and by inducing autophagy to degrade aggregated proteins. Therefore, rapamycin may be a promising drug for protection against neuronal injury relevant to Parkinson disease (PD). Our studies thus provide a mechanistic insight into the functional link between the two protein degradation systems.

*Correspondence to: Weidong Le; Neurology Department; NB 205; Baylor College of Medicine; Houston, TX 77030 USA; Tel.: 713.798.5142; Fax: 713.798.8307; Email: [email protected] Submitted: 01/15/09; Accepted: 03/10/09 Previously published online as an Autophagy E-publication: http://www.landesbioscience.com/journals/autophagy/article/8377

www.landesbioscience.com

Introduction Parkinson disease (PD) is characterized by the degeneration of dopamine (DA) neurons in the substantia nigra (SN) and aggregation of misfolded proteins in susceptible neurons. A variety of cellular changes indicate that proteolytic impairment leading to protein aggregation is centrally involved in the pathogenesis of PD.1,2 There are two main proteolytic systems responsible for protein degradation: the ubiquitin-proteasome system (UPS) and the autophagy-lysosomal pathway (ALP). UPS contributes to a highly selective degradation of short-lived intracellular proteins. Failure of the UPS to clear misfolded proteins has been found to play a major role in the etiopathogenesis of PD.3-5 ALP is a less selective multistep process involving the formation of double membrane structures known as autophagic vacuoles (AVs) and fusing with lysosomes to form autophagolysosomes inside which their contents are degraded.6-8 Several studies have shown that proteasome inhibitor MG 132 can upregulate autophagy to enhance the degradation of aggregated polyubiquitinated proteins in murine embryonic fibroblasts (MEFs) and prostate cancer cell line DU145.9,10 However, whether UPS inhibition can also induce the autophagy enhancement in primary ventral mesencephalic (VM) neurons remains unclear. p53 is a sequence-specific transcription factor that is dramatically increased in response to a variety of cellular stresses and activation of it may cause cell death by directly inducing mitochondrial permeability and apoptosis.11,12 p53 is also known to be primarily cleared away by Mdm2-mediated ubiquitination and subsequent proteasome degradation.13,14 It has been reported that proteasomal inhibition in PD related models can increase the level of p53 which may contribute to the degeneration of DA neurons.15,16 Furthermore, it has been demonstrated that p53 activation is able to inhibit the activity of mammalian target of rapamycin (mTOR) and thus activate autophagy in MEFs.17,18 p53 itself can also induce autophagy in a DRAM (damage-regulated autophagy modulator)-dependent manner in human osteosarcoma saos-2 cell line.19,20 In accordance with those studies, we hypothesize that proteasomal inhibition can upregulate autophagy in VM neurons

Autophagy

663

The mechanistic role of autophagy induction by proteasome inhibitor

and p53 may play a role in this process of induction. Whether autophagy is a cell defense mechanism or a regulated cell-killing pathway (referred as to type II programmed cell death) in response to cellular stress conditions is still a much-debated question.21,22 Several recent studies have revealed that autophagy may have a dual role in cell survival, depending on the actual cellular environment. On the one hand, the initial activation of autophagy may be able to protect cells by degrading intracellular macromolecules and organelles to provide energy for cell functioning and reduce further cell injury by facilitating removal of damaged organelles and misfolded proteins. On the other hand, prolonged or vigorous activation of autophagy may lead to an excessive elimination of the affected cells.23-25 To further reveal the role of enhanced autophagy in response to proteasomal inhibition and its interaction with p53 or its downstream targets, we used the proteasome inhibitor lactacystin in cultures of primary VM neurons and human dopaminergic neuroblastoma SH-SY5Y cells as cell models to investigate the mechanisms of p53 mediated autophagy induction and mitochondriadependent apoptotic pathways.

Results

Figure 1. Proteasomal inhibition induces autophagy enhancement. The primary VM neurons were treated with lactacystin at a concentration of 5 μm for various times. The LC3 (A), Beclin 1 (B) and p62 protein expression (C) were determined using immunoblotting assay. The density of the bands was evaluated by densitometric analysis. The ratio of LC3-II, Beclin 1 or p62 density to that of β-actin was calculated, respectively, and data were expressed as folds of control. The beclin 1 mRNA expression was detected by real-time PCR and the average CT values were calculated. The data were expressed as folds of control (D). Data were the mean ± SD values. *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with control. n = 5, each.

Proteasomal inhibition induces autophagy enhancement in a time-dependent manner. To investigate whether inhibition of proteasome can activate autophagy in the primary VM neurons, we used 5 μM lactacystin (a specific inhibitor of the 20S proteasome) in the cultures. We found that incubation with lactacystin elicited a time-dependent increase of LC3-II, a lipidated form of LC3 localized on the AVs, and Beclin 1, a positive regulator involved in the formation of autophagosomes and initiation of autophagy (Fig. 1A and B), while p62 protein, the ubiquitin and LC3-binding protein, which regulates the formation of protein aggregates to be removed by autophagy26,27 was decreased after treatment with lactacystin (Fig. 1C). However, LC3-II and Beclin 1 levels were increased at different rates after exposure to lactacystin for various times. LC3-II was increased 3-fold at 5 h (p < 0.01; Fig. 1A) and went up to 7-fold at 36 h (p < 0.001; Fig. 1A). Beclin 1 was increased 5-fold at 5–12 h (p < 0.01; Fig. 1B) and 10-fold at 36 h (p < 0.001; Fig. 1B). We further examined beclin 1 mRNA expression in VM neurons and found that it began to elevate significantly after exposure to lactacystin for 12 h (p < 0.05; Fig. 1D). Proteasomal inhibition leads to an increase in phosphorylated p53. We then examined the effects of proteasomal inhibition on 664

the accumulation of p53 and phosphorylated p53 in primary VM neurons. Treatment with lactacystin at 5 μM for 5 h led to a significant increase in the level of p53 protein (p < 0.01; Fig. 2A and B). Accumulated p53 is phosphorylated at serine 15, which is known to enhance p53 stability.15,16 Similarly, the level of phosphorylated p53 was also markedly increased after treatment with lactacystin at 5 μM as determined by immunoblotting (p < 0.01; Fig. 2A and B). To further verify the p53 expression profile in individual DA neurons, we examined the subcellular localization of phosphorylated p53 using triple immunofluorescent stainings in primary VM neurons. The results showed that phosphorylated p53 immunostaining was significantly increased in most tyrosine hydroxylase (TH) positive neurons, but not in non-TH positive neurons (Fig. 2C and D), indicating that lactacystin can increase the phosphorylated p53 expression specifically in DA neurons. In addition, we measured the p53 mRNA level in primary VM neurons exposed to lactacystin for various times and found it moderately increased after 5–12 h exposure (p < 0.05, Fig. 2E)

Autophagy

2009; Vol. 5 Issue 5


autophagy activation in DA neurons. The results showed that lactacystin can significantly increase LC3 expression in TH-positive neurons (p < 0.01; Fig. 3D and E), whereas it has a moderate but significant effect on non-TH-positive neurons (p < 0.05; Fig. 3D and E). PFT-α pretreatment decreased the LC3 staining in most of TH-positive neurons (p < 0.01; Fig. 3D and E) but not in non-TH positive neurons after exposure to lactacystin (p > 0.05; Fig. 3D and E), indicating that the autophagy activation mediated by p53 mainly occurs in DA neurons. Using chloroquine, an inhibitor of lysosomal activity,29 we determined the effects of lactacystin and p53 on autophagic flux. The results showed that chloroquine further enhanced the LC3 staining induced by lactacystin in the primary VM cultures (p < 0.05; Fig. 3D and E), while it only enhanced the LC3 staining in TH-positive neurons after PFT-α treatment prior to lactacystin (p < 0.05; Fig. 3D and E). Therefore we cannot exclude the possibility that p53 may also regulate the lysosmal integrity and thus mediate the autophagy activation in DA neurons. PFT-α pretreatment also resulted in a moderate but significant decrease in the mRNA level of beclin 1 in VM neurons after exposure to lactacystin, as compared with the neurons Figure 2. Proteasomal inhibition leads to an increase in phosphorylated p53. Neurons were treated with lactacystin alone (p < 0.05; Fig. 3F). treated with lactacystin (5 μM) for different times. p53 protein and its phosphorylation levels Electronic microscopy (EM) images showed that were determined using immunoblotting assay (A). The ratio of p53 or phosphorylated p53 treatment with 5 μM lactacystin for 24 h induced density to that of β-actin was calculated, respectively, and data were expressed as folds of the buildup of AVs that include autophagosomes control (B). Twenty-four hours after lactacystin treatment at 5 μM, the cytoplasmic phosphorylated p53 was determined by immunostaining using mouse anti-phospho-Ser15 p53 antibody and multilamellar bodies, as compared with the (green) and rabbit anti-TH antibody (red) in both lactacystin-treated and control cultures. control (p < 0.01; Fig. 3G), and such effects Hoechst was used to stain the cell nuclei (blue) (C). The fluorescent density in each cell was seemed to be dose-dependent since higher lactaquantified and the data was expressed as percentage of the control (D). The p53 mRNA cystin dose (10 μM) induced greater AVs buildup expression was detected by real-time PCR and the average CT values were calculated. Data (p < 0.001; Fig. 3G). Pretreatment with PFT-α were the mean ± SD values (E). *p < 0.05, **p < 0.01, ***p < 0.001 as compared with control. Scale bar in (C) = 10 μm. n = 5, each. Con. = Control; pp53 = Phosphorylated reduced the buildup of AVs induced by lactacystin p53; Lac. = Lactacystin. exposure (p < 0.05; Fig. 3G). We then tested the possible link between and significantly increased after 24 h exposure (p < 0.01; Fig. PFT-α and lactacystin-induced protein aggregation by immu2E). These observations suggest that proteasomal inhibition can nofluorescent staining and found a strong aggregated α-synuclein increase both p53 and phosphorylated p53 levels in primary VM immunoreactivity in some TH-positive neurons after exposure to lactacystin (p < 0.05; Fig. 3H), which was enhanced by PFT-α neurons. p53 mediates autophagy activation in response to protea- pretreatment, as compared with the neurons treated with lactasomal inhibition. Pifithrin-α (PFT-α) is well known to inhibit cystin alone (p < 0.05; Fig. 3H), indicating that inhibition of p53 p53 activity.28 To determine whether p53-mediated pathways may promote α-synuclein protein aggregation through suppression are involved in the lactacystin-induced autophagy activation in of autophagy. PFT-α treatment alone did not significantly affect primary VM neurons, we pretreated these neurons with PFT-α at α-synuclein-positive protein aggregation. To further evaluate the role of p53 in the autophagy activation 10 μM for 30 min, followed by lactacystin exposure for another 24 h. The results showed that PFT-α pretreatment reduced the protein induced by proteasomal inhibition, we genetically suppressed the levels of LC3-II and Beclin 1 by 68% (p < 0.01; Fig. 3A) and 52% p53 expression in SH-SY5Y cells by transfecting the cells with p53 (p < 0.01; Fig. 3B), respectively, while it increased p62 protein level siRNA (Fig. 4A). We applied p53 siRNA for 24 h to inhibit p53 by 43% (p < 0.05; Fig. 3C), as compared with the neurons treated expression prior to lactacystin or chloroquine treatment for another with lactacystin alone. We then used the triple immunofluores- 24 h in the SH-SY5Y cells. Then we evaluated the autophagy level cent staining to observe the effects of lactacystin and PFT-α on by immunoblotting assay (Fig. 4B and C) and immunofluorescent www.landesbioscience.com

Autophagy

665


Figure 3. For figure legend, see page 667.

666

Autophagy


The mechanistic role of autophagy induction by proteasome inhibitor Figure 3. p53 mediates autophagy activation in response to proteasomal inhibition in primary VM neurons. The neurons were pretreated with PFT-α at 10 μM for 30 min, followed by lactacystin exposure for another 24 h. The LC3, Beclin 1 and p62 expression were determined using immunoblotting assay and data were expressed as relative OD of β-actin (A–C). Twenty-four hours after lactacystin or chloroquine treatment in addition to PFT-α pretreatment, cytoplasmic LC3 or a-synuclein were determined by immunostaining using rabbit anti-LC3 antibody (green) and mouse anti-TH antibody (red) (D), or rabbit anti-a-synuclein antibody (green) and mouse anti-TH antibody (red) (H). Hoechst was used to stain the cell nuclei (blue). The fluorescent density in each neuron was quantified and the data were expressed as a percentage of the control (E). The beclin 1 mRNA expression was detected by real-time PCR after PFT-α pretreatment prior to lactacystin exposure (F). EM images showed formation of AVs including multilamellar bodies and autophagosomes (arrows) induced by lactacystin exposure. The number of AVs in each cell was counted at a final magnification of 5,000 (G). Data were the mean ± SD values. In (A–C and E–H), *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with control; +p < 0.05, ++p < 0.01 as compared with neurons treated with lactacystin. In (D), *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with control in TH-positive neurons; +p < 0.05, ++p < 0.01 as compared with TH-positive neurons treated with lactacystin; #p < 0.05 as compared with TH-positive neurons treated with lactacystin and chloroquine; &p < 0.05 as compared with TH-positive neurons treated with PFT-α and lactacystin; $p < 0.05, $$p < 0.01 as compared with control in non-TH-positive neurons; @p < 0.05 as compared with non-TH-positive neurons treated with lactacystin. Scale bar in (D) = 10 μm, in (G) = 50 nm, in (H) = 20 μm. n = 5, each. Con. = Control; Cq. = Chloroquine; Lac. = Lactacystin.



Autophagy

667

The mechanistic role of autophagy induction by proteasome inhibitor Figure 4. p53 mediates autophagy activation in response to proteasomal inhibition in SH-SY5Y cells. The cells were exposed to p53 siRNA for various times to detect its inhibiting effects (A). p53 expression was inhibited by RNA interference for 24 h, followed by lactacystin or chloroquine exposure for another 24 h before Beclin 1 and LC3 expression were determined using immunoblotting assay. The data were expressed as relative OD of β-actin (B and C). Cytoplasmic LC3 or p62 protein was determined by double immunostaining using rabbit anti-LC3 antibody (green) and Hoechst (blue) (D),or mouse anti-p62 antibody (red) and rabbit anti-p53 antibody (green) (E). The fluorescent density in each cell was quantified and the data was expressed as percentage of the control. Data were the mean ± SD values. *p < 0.05, **p < 0.01,***p < 0.001 as compared with control; +p < 0.05, ++p < 0.01 as compared with cells treated with lactacystin; #p < 0.05 as compared with the cells treated with lactacystin and chloroquine; &p < 0.05 as compared with the cells treated with PFT-α and lactacystin. Scale bar in (D and E) = 10 μm. n = 5, each. Con. = Control; Cq. = Chloroquine; Lac. = Lactacystin.

staining (Fig. 4D and E) in the cells. The quantitative analysis showed that 100 nM of p53 siRNA pretreatment attenuated the expression of LC3-II and Beclin 1 proteins by about 60% (p < 0.01; Fig. 4B–D), while increased the p62 protein level by 65% (p < 0.01; Fig. 4E), respectively, compared with the cells exposed to lactacystin and siRNA vehicle. p53 siRNA treatment alone did not significantly affect autophagy and its cargo levels. Using chloroquine to inhibit lysosomal activity, we further explored the effects of p53 inhibition on lysosomal flux in SH-SY5Y cultures after treatment with p53 siRNA and then lactacystin. Similar to the results using PFT-α, we found that chloroquine co-administration with lactacystin also enhanced the LC3 staining in the cells pretreated with p53 siRNA, as compared with the cells treated with p53 siRNA and lactacystin (p < 0.05; Fig. 4D). Proteasomal inhibition activates the downstream mitochondria-dependent apoptotic pathways of p53. It has been reported that mitochondria-dependent apoptotic machinery can mediate the injury of cortical neurons or sympathetic neurons induced by proteasomal inhibition.30,31 To further explore the role of autophagy following proteasomal inhibition in mitochondria-dependent apoptotic pathways, we then primarily tested the possible link between proteasomal inhibition and intrinsic mitochondrial apoptotic pathways in primary VM neurons. The results showed that lactacystin treatment at 5 μm increased the release of cytochrome c (Cyt C) starting at 5 h and reaching the maximal response at 24 h (Fig. 5A), as compared with Figure 5. Proteasomal inhibition activates mitochondrial injury-related apoptotic paththe neurons without lactacystin exposure. Exposure of ways. The primary VM neurons were treated with lactacystin treatment for different neurons to lactacystin also led to a dramatic increase time durations. The cytosolic Cyt C (A), cleaved caspase 3 (B), cleaved PARP (C) and in the cleaved caspase 9, cleaved caspase 3 and cleaved cleaved caspase 9 (D) levels were determined using immunoblotting assay and density PARP levels. The expression levels of cleaved caspase of the bands was evaluated by densitometric analysis. The ratio of their density to that of β-actin was calculated, respectively, and data were expressed as folds of control. Data 3 and cleaved PARP started at 12 h and reached a were the mean ± SD values. *p < 0.05, **p < 0.01 and ***p < 0.001 as compared maximal response at 24 h (Fig. 5B and C), while that with control. n = 5, each. Casp.9 = Caspase 9; Casp.3 = Caspase 3. of cleaved caspase 9 started at 12 h and increased significantly at 36 h (Fig. 5D). The results support the theory levels of phosphorylated p53, cleaved caspase 9, cleaved caspase that proteasomal inhibition can activate mitochondria-dependent 3, cleaved PARP and cytosolic Cyt C, as well as mitochondrial membrane potential (Fig. 6). The quantitative analysis showed apoptotic pathways in primary VM neurons. Autophagy is involved in lactacystin-induced alteration of that pretreatment of primary VM neurons with 3-MA can increase phosphorylated p53 and mitochondria impairment pathways. the levels of phosphorylated p53, cleaved PARP, cleaved caspase To confirm the effects of autophagy induced by lactacystin on 3, cleaved caspase 9 levels and Cyt C release by 51% ± 6.91 (p < phosphorylated p53 or its downstream mitochondria-dependent 0.05), 49% ± 7.23 (p < 0.05), 40% ± 2.10 (p < 0.01), 54% ± 3.03 apoptotic machinery in primary VM neurons or SH-SY5Y cells, we (p < 0.01), 51% ± 9.35 (p < 0.05), respectively, as compared with used the autophagy inhibitor 3-methyladenine (3-MA) or beclin 1 the neurons treated with lactacystin alone (Fig. 6A). Pretreatment siRNA prior to lactacystin exposure and detected the protein of SH-SY5Y cells with beclin 1 siRNA had similar effects and 668

Autophagy





Autophagy

669

The mechanistic role of autophagy induction by proteasome inhibitor Figure 6 (See previous page). Induced autophagy in response to lactacystin inhibits phosphorylated p53 or its downstream apoptotic pathways. Half an hour pretreatment of primary VM neurons with 3-MA, followed by lactacystin treatment for another 24 h, expression of phosphorylated p53, cleaved caspase 9, cleaved caspase 3 and cleaved PARP as well as Cyt C release levels were determined using immunoblotting assay and the ratio of their density to that of β-actin was calculated, respectively. Data was expressed as folds of control (A). The SH-SY5Y cells were exposed to beclin 1 siRNA for various times to detect its inhibiting effects. Twenty-four hours after pretreatment with beclin 1 siRNA prior to lactacystin exposure, cleaved caspase 9 and Cyt C release levels were determined using immunoblotting assay. The data were expressed as folds of control (B). To observe the mitochondrial membrane potential alteration under the fluorescence microscope, the VM neurons were incubated with JC-1. Red fluorescence (JC-1 PE) represents polarized mitochondrial membrane potential, while green fluorescence (JC-1 FITC) in the cytosol indicates mitochondrial membrane depolarization. Percentage of JC-1 FITC positive neurons was measured on a FACS Calibur instrument and analyzed with Cell Quest software. The data were quantified and expressed as percentage of JC-1 FITC positive neurons (C). Data were the mean ± SD values. *p < 0.05, **p < 0.01 and ***p < 0.001 as compared with control. +p < 0.05 as compared with cells treated with lactacystin. Scale bar in (C) = 15 μm. n = 5, each. RA = Rapamycin; Lac. = Lactacystin; Con. = Control; Casp.9 = Caspase 9; Casp.3 = Caspase 3.

increased the levels of cleaved caspase 9 levels and Cyt C release by 69% ± 5.23 (p < 0.01), 49% ± 3.42 (p