BAG3-dependent noncanonical autophagy induced by proteasome inhibition in HepG2 cells

Abstract

Emerging lines of evidence have shown that blockade of ubiquitin-proteasome system (UPS) activates autophagy. The molecular players that regulate the relationship between them remain to be elucidated. Bcl-2 associated athanogene 3 (BAG3) is a member of the BAG co-chaperone family that regulates the ATPase activity of heat shock protein 70 (HSP70) chaperone family. Studies on BAG3 have demonstrated that it plays multiple roles in physiological and pathological processes, including antiapoptotic activity, signal transduction, regulatory role in virus infection, cell adhesion and migration. Recent studies have attracted much attention on its role in initiation of autophagy. The current study, for the first time, demonstrates that proteasome inhibitors elicit noncanonical autophagy, which was not suppressed by inhibitors of class III phosphatidylinositol 3-kinase (PtdIns3K) or shRNA against Beclin 1 (BECN1). In addition, we demonstrate that BAG3 is ascribed to activation of autophagy elicited by proteasome inhibitors and MAPK8/9/10 (also known as JNK1/2/3 respectively) activation is also implicated via upregulation of BAG3. Moreover, we found that noncanonical autophagy mediated by BAG3 suppresses responsiveness of HepG2 cells to proteasome inhibitors.

Introduction

The ubiquitin-proteasome system and autophagy-lysosome system are two major intracellular protein degradation pathways responsible for the clearance proteins and organelles, which are critical in the maintenance of cellular homeostasis. The UPS is the primary proteolytic route for thousands of short-lived proteins and regulates a variety of functions, including cell cycle progression, cell survival, proliferation, apoptosis and other critical cellular functions.¹ The UPS also serves as a mechanism of protein quality control to clear out misfolded or aggregated proteins.^2,3 By contrast, autophagy describes a catabolic process in which cell constituents such as long-lived proteins and organelles are delivered to the lysosomal compartment for degradation. Traditionally, autophagy has been considered as a less selective degradative pathway than the UPS and is frequently described as bulky removal of defective organelles and large portions of cytoplasm by lysosomal degradation. Macroautophagy (hereafter referred to as autophagy) is the regulated catabolic pathway by which cytoplasmic constituents are sequestered in vacuoles known as autophagosomes or autophagic vacuoles (AVs), where they are targeted to lysosomes for degradation.^4-6 Autophagy is classically elicited by metabolic stress, such as nutrient deprivation, and much is known about starvation-induced autophagy.⁶

For a long time, the UPS and autophagy were generally assumed to be independent proteolytic pathways with few or no points of intersection.⁷ However, this view was challenged recently and accumulating lines of evidence support that the two cellular degradation systems are functionally interrelated catabolic processes.^8-18 Inadequate function of UPS can activate autophagy, which has been suggested to compensate for the reduced proteasomal-mediated degradation of misfolded and/or damaged proteins.^9-17 On the other hand, reduced autophagy also results in enhancement of protein degradation by UPS.^4,18,19 However, the mechanism by which autophagy and UPS functions are coordinated is little understood.

Autophagy is initiated by the formation and elongation of a double-layered phagophore to form autophagosomes. The formation of autophagosomes is a multistep process controlled by a set of evolutionarily conserved proteins termed autophagy-related (ATG) proteins.²⁰ Of these ATG proteins, the BECN1 (mammalian homolog of yeast Vps30/Atg6) complex and two ubiquitin-like conjugation systems: the ATG12 and LC3/MAP1LC3 (microtubule-associated protein 1 light chain 3, the mammalian homolog of yeast Atg8) systems act sequentially during the nucleation and elongation of the phagophore membrane.^21,22 BECN1 is initially isolated as a B-cell CLL/lymphoma 2 (BCL2)-interacting tumor suppressor.²³ As knockdown of BECN1 inhibits autophagy, therefore it is considered as a key initiator of autophagy in mammalian cells.²¹ BECN1 is part of the class III phosphatidylinositol 3-kinase (PtdIns3K) complex, which is essential for recruitment of other ATG proteins to the phagophore assembly site and is required for the formation of AVs.²¹ However, recently it has been revealed that autophagy can also occur in a BECN1- or PtdIns3K-independent manner, so named as noncanonical autophagy.^24-31 In contrast to classical or canonical autophagy, noncanonical autophagy is a process that is neither blocked by the knockdown of BECN1 or its binding partner PIK3C3, nor inhibited by pharmacological inhibitors wortmannin (WM) or 3-methyladenine (3-MA), which interfere with the activity of the PtdIns3K.^24-31 The autophagy involved in vacuole biogenesis in plants also appears to be noncanonical, as it is also resistant to PtdIns3K inhibitors.²⁵ A better understanding of the role of noncanonical autophagy calls for further investigations to elucidate the molecular mechanism by which autophagy occurs bypassing BECN1 and PIK3C3.

BAG3, a cochaperone of HSP70, is a stress- and survival-related protein, which is induced upon various types of cellular stress.^32,33 Studies of BAG3 demonstrate that this protein is involved in a number of physiological and pathological processes, including cell proliferation, apoptosis, adhesion and migration and viral infection.³⁴ Recently, the involvement of BAG3 in selective degradation of misfolded proteins by autophagy has attracted a large amount of attention.^35-41 For the first time, we demonstrated that proteasome inhibition elicits BECN1-independent noncanonical autophagy. In addition, for the first time, we report that BAG3 is a candidate as a specific “molecular player” implicated in noncanonical autophagy elicited by proteasome inhibition.

Results

Increase in autophagosomes and activation of autophagy by proteasome inhibitors in HepG2 cells

As obvious punctate GFP-LC3B/GFP-MAP1LC3B distribution is a marker for autophagosome formation,⁴² HepG2 cells were stably transfected with a plasmid encoding EGFP-LC3B and examined by immunofluorescence microscopy (Fig. 1A). EGFP-LC3B was mainly distributed evenly throughout the cell in vehicle-treated control cells, while MG132-treated HepG2 cells displayed obviously punctate distribution of EGFP-LC3B (Fig. 1A). Since both autophagosome formation and impaired autophagosomes degradation ascribes to increased AVs,⁶ the effect of inhibiting lysosomal turnover of autophagosome contents by cloroquine (CQ) or ammonium chloride (NH₄Cl) were also examined. Both CQ and NH₄Cl markedly increased the numbers of AVs elicited by MG132 exposure (Fig. 1A). Since the transition from cytoplasmic (LC3-I) to the cleaved, lipidated and membrane-bound (LC3-II) form is accompanied by a shift to a faster migrating band detectable by western blot,⁴² we further investigated the LC3-II transition using western blot and found that MG132 treatment elicited significant increase in LC3-II levels, and CQ or NH₄Cl further augmented such increase mediated by MG132 (Fig. 1B). Ultrastructural observation using transmission electron microscopy (TEM) demonstrated that similar to Earle’s balanced salt solution (EBSS), MG132 caused reduction of cellular organelles (asterisks) and accumulation of small vacuoles (arrows) in the cytoplasm (Fig. 1C). Immunofluorescence microscopy also demonstrated that MG132 treatment increased autophagic vacuoles in HepG2 cells as assessed by acridine orange (AO) staining (Fig. S1A). In addition, the accumulation of acidic vacuoles was further enhanced by cotreatment with CQ or NH₄Cl (Fig. S1A). We then investigated whether other proteasome inhibitors including bortezomib (BZ), epoxomicin (Epox) and lactacystin (Lacta) could activate autophagy in HepG2 cells. Both AO staining (Fig. S1B) and distribution of EGFP-LC3B (Fig. 1D) confirmed that all the proteasome inhibitors tested increased vacuole numbers in HepG2 cells. LC3-II generation, detected by western blot confirmed that all these tested proteasome inhibitors elicited a significant increase in LC3-II levels, and such an increase was further augmented by CQ (Fig. 1E). Real-time RT-PCR demonstrated that these proteasome inhibitors had no effects on LC3 mRNA expression (Fig. 1F).

Figure 1. Activation of autophagy by proteasome inhibitors in HepG2 cells. (A) HepG2 cells stably overexpressing EGFP-LC3B were treated with vehicle or MG132 in the absence or presence of cloroquine (CQ) or ammonia chloride (NH₄Cl), the punctate distribution of EGFP-LC3B was visualized under the fluorescence microscopy. (B) HepG2 cells were treated with MG132 alone or in combination with CQ or NH₄Cl, and western blot analysis was performed using the indicated antibodies. (C) HepG2 cells were treated with vehicle, MG132 or EBSS, and ultrastructure was analyzed using transmission electron microscopy. Asterisks point to intracellular organelles, arrows point to vacuoles. (D) HepG2 cells stably overexpressing EGFP-LC3B were treated with vehicle, bortezomib (BZ), epoxomicin (Epox), or lactacystin (Lacta), the punctate distribution of EGFP-LC3B was visualized under the fluorescence microscopy. (E) HepG2 cells were treated with BZ, Epox, Lacta or MG132 in the absence or presence of CQ, LC3 production was analyzed using western blot analysis. (F) HepG2 cells were treated with vehicle, BZ, Epox, Lacta or MG132, and LC3 mRNA was measured using real-time RT-PCR. N.S., not significant.

PtdIns3K-independent autophagic response induced by proteasome inhibitors in HepG2 cells

Pharmacological inhibitors of PtdIns3K, including 3-MA and WM, are effective at inhibiting starvation-induced autophgy.^6,43 However, neither 3-MA nor WM could suppress the increases in AVs elicited by MG132 as measured using punctate distribution of EGFP-LC3B (Fig. 2A) and AO staining (Fig. S2A). Western blot confirmed that neither 3-MA nor WM suppressed LC3-II production elicited by MG132 treatment (Fig. 2B). On the contrary, both 3-MA and WM significantly reduced LC3-II generation elicited by EBSS (Fig. 2C), indicating that starvation-induced autophagy was intact in HepG2 cells. To further confirm the effectiveness of 3-MA or WM on lipid kinase activity of PtdIns3K, we further transfected HepG2 cells with a p40(phox)PX-EGFP plasmid, whose dot distribution and density indicate the lipid kinase activity of PtdIns3K.^44,45 EBSS significantly increased punctate distribution and density of PX-EGFP, as well as AV numbers as assessed by LysoTracker Red staining (Fig. 2D and E). Both 3-MA and WM significantly suppressed EBSS-induced increase in PX-EGFP dot density and accumulation of AVs (Fig. 2D and E). Different from EBSS, MG132 significantly increased AV numbers, while demonstrated no obvious effects on dot distribution and density of PX-EGFP (Fig. 2F and G). Both 3-MA and WM significantly suppressed PX-EGFP dot density, while neither 3-MA nor WM demonstrated obvious effects on increase in AVs elicited by MG132 (Fig. 2F and G). To test whether other proteasome inhibitors also cause PtdIns3K-independent activation of autophagy, we treated HepG2 cells with different proteasome inhibitors in the absence or presence of 3-MA or WM. Western blot analysis demonstrated that neither 3-MA nor WM had effects on LC3-II production elicited by these proteasome inhibitors (Fig. 2H). We also treated p40(phox)PX-EGFP transfected HepG2 with BZ (Fig. S2B), Epox (Fig. S2C), or Lacta (Fig. S2D) in the absence or presence of PtdIns3K inhibitors, and AVs were measured using LysoTracker Red staining. Similar to MG132, BZ, Epox and Lacta significantly increased AV numbers without obvious effects on punctate distribution of PX-EGFP (Fig. S2B–S2E). Cotreatment with 3-MA or WM significantly reduced punctate distribution of PX-EGFP, while had no obvious effects on accumulation of AVs elicited by BZ, Epox or Lacta (Fig. S2B–S2E). We also found that MG132 caused PtdIns3K-independent autophagy in other cell types including HEK293, FRO, KTC1, OVCAR3 cells (data not shown). These data indicated that proteasome inhibitors generally induced PtdIns3K-independent autophagy.

Figure 2A–E.

General activation of PtdIns3K-independent autophagy by proteasome inhibitors in HepG2 cells. (A) HepG2 cells stably overexpressing EGFP-LC3B were treated with vehicle or MG132 in the absence or presence of 3-methyladenine (3-MA) or wortmannin (WM), the punctate distribution of EGFP-LC3B was visualized under the fluorescence microscopy. (B) HepG2 cells were treated with MG132 alone or in combination with 3-MA or WM, and western blot analysis was performed to detect LC3-II generation. (C) HepG2 cells were treated with EBSS alone or in combination with 3-MA or WM, and LC3-II generation was analyzed using western blot. (D) HepG2 cells stably overexpressing PX-EGFP were cultured in EBSS medium in the absence or presence of 3-MA or WM, acidic vacuoles were stained with LysoTracker Red and images were acquired using the fluorescence microscopy. (E) Light-microscopy quantitation of PX-EGFP dots in HepG2 cells treated with EBSS in the absence or presence of 3-MA or WM. Results shown represent mean ± SD from five representative microscopic fields and three independent experiments were performed.

Figure 2F–H.

General activation of PtdIns3K-independent autophagy by proteasome inhibitors in HepG2 cells. (F) HepG2 stably overexpressing PX-EGFP cells were treated with MG132 in the absence or presence of 3-MA or WM, acidic vacuoles were stained with LysoTracker Red and images were acquired using the fluorescence microscopy. (G) Light-microscopy quantitation of PX-EGFP dots in HepG2 cells treated with MG132 in the absence or presence of 3-MA or WM. Results shown represent mean ± SD from five representative microscopic fields and three independent experiments were performed. (H) HepG2 cells were treated with BZ, Epox, Lacta or MG132 alone or in combination with 3-MA or WM, LC3-II production was analyzed using western blot analysis. *p < 0.01, N.S. not significant.

Activation of autophagy in a BECN1-independent manner by proteasome inhibitors in HepG2 cells

As BECN1 associates with PtdIns3K to induce autophagy,⁴⁶ we further investigated the effect of shRNA against BECN1 (shBECN1) on autophagy induced by proteasome inhibitors. First, we analyzed BECN1 expression upon treatment with different proteasome inhibitors in HepG2 cells. To different extents, reproducible reduction of BECN1 was observed in all these tested proteasome inhibitors (Fig. 3A). In addition, MG132 caused reproducible downregulation of BECN1 in several other cells, including HEK293, FRO, KTC1 and OVCAR3 cells (data not shown). To confirm the role of BECN1 in the autophagy elicited by proteasome inhibitors, we transfected HepG2 cells with shRNA against BECN1 (shBECN1), western blot analysis confirmed that shBECN1 significantly reduced BECN1 expression, as well as LC3-II generation elicited by EBSS, while a scrambled shRNA demonstrated no obvious effects on BECN1 expression and LC3-II production elicited by EBSS (Fig. 3B). shBECN1-transfected HepG2 cells were treated with the panel of proteasome inhibitors and western blotting demonstrated that LC3-II production elicited by proteasome inhibitors was unaffected by shBECN1 transfection (Fig. 3C). Western blots also demonstrated that proteasome inhibitors reduced the phosphorylated form of RPS6KB, while shBECN1 had no obvious effects on decreased RPS6KB phosphorylation elicited by proteasome inhibitors (Fig. 3C). It should be noted that Epox and MG132 treatment reproducibly resulted in a faster-migrating band, which was recognized by RPS6KB antibody (Figs. 3C and 4C). LysoTracker Red staining also confirmed that shBECN1 was not able to block AVs formation elicited by these proteasome inhibitors, but significantly blocked AVs formation in cells cultured under EBSS (Fig. S3A). Immunostaining of endogenous LC3 confirmed that shBECN1 significantly blocked puncta distribution of LC3 mediated by EBSS, while no obvious effects on LC3 puncta distribution elicited by proteasome inhibitors (Fig. S3B and S3C).

Figure 3. Activation of BECN1-independent autophagy by proteasome inhibitors in HepG2 cells. (A) HepG2 cells were treated with vehicle, BZ, Epox, Lacta or MG132, and BECN1 expression levels were analyzed using western blot analysis. (B) HepG2 cells were transfected with control (scrambled) or shRNA against BECN1 (shBECN1), then treated with EBSS. LC3-II production was analyzed using western blot analysis. (C) HepG2 cells were transfected with scrambled shRNA or shBECN1 and treated with EBSS. Western blotting was performed using the indicated antibodies. “*” indicates a potential truncated form of RPS6KB.

Figure 4. Implication of BAG3 in proteasome inhibitor-mediated autophagy in HepG2 cells. (A) HepG2 cells were treated with vehicle, BZ, Epox, Lacta or MG132, and BAG3 mRNA expression was measured using real-time RT-PCR. (B) HepG2 cells were treated with the indicated proteasome inhibitor and western blot analysis was performed using the indicated antibodies. (C) HepG2 cells were transfected with scrambled shRNA or shRNA against BAG3 (shBAG3), then treated with the indicated proteasome inhibitor. Western blot analysis was performed using the indicated antibodies. (D) HepG2 cells were transfected with scrambled shRNA or shBAG3, then treated with the indicated proteasome inhibitor. Real-time RT-PCR was performed to measure LC3 mRNA levels. (E) HepG2 cells were transfected with scrambled or shBAG3, then treated with vehicle or EBSS. Western blot analysis was performed using the indicated antibodies.

Implication of BAG3 in autophagy elicited by proteasome inhibitors

We and others have reported that BAG3 is upregulated upon proteasome inhibitors in various cancer cells.^47,48 Considering the increasing number of reports on induction of autophagy by BAG3,^{36-38,40,41,49-52} we therefore investigated whether BAG3 was implicated in autophagy induced by proteasome inhibitors. Consistent with previous reports,^47,48 real-time RT-PCR and western blots confirmed that all these proteasome inhibitors increased BAG3 mRNA (Fig. 4A) and protein levels (Fig. 4B) in HepG2 cells. HepG2 cells were then transfected with shRNA against BAG3 (shBAG3), western blot analysis confirmed that shBAG3 successfully reduced basal and proteasome inhibition-induced BAG3 expression (Fig. 4C). LC3-II generation elicited by proteasome inhibitors was markedly blocked by shBAG3 transfection (Fig. 4C). shBAG3 had little effect on RPS6KB phosphorylation elicited by proteasome inhibitors (Fig. 4C). No alteration of LC3 mRNA expression was observed by shBAG3 transfection as assessed by real-time RT-PCR (Fig. 4D). We also investigated the potential involvement of BAG3 on traditional autophagy induced by EBSS (Fig. 4E). EBSS reduced BAG3 expression, and further downregulation of BAG3 by shBAG3 produced no obvious effect on LC3-II generation elicited by EBSS (Fig. 4E). LysoTracker Red staining also demonstrated that AVs formation elicited by these proteasome inhibitors was significantly reduced in shBAG3 cells (Fig. S4). However, shBAG3 produced no obvious effects on AVs formation elicited by EBSS (Fig. S4).

Regulation of autophagy elicited by proteasome inhibitors by MAPK8/9 via BAG3

The MAPK8/9 signaling pathway has been shown to regulate autophagy in response to different stimuli, such as starvation, ER stress, T cell receptor activation, growth factor withdrawal and caspase inhibition.^21,53-56 In addition, we have reported that MAPK8/9 is involved in BAG3 induction mediated by proteasome inhibitors.⁵⁷ MAPK8/9/10 inhibitor SP600125 demonstrated a dose-dependent reduction of LC3-II production elicited by MG132 (Fig. 5A). We then investigated whether MAPK8/9/10 was implicated in autophagy induced by other proteasome inhibitors. Fifty micrometers of SP600125 significantly suppressed punctate distribution of EGFP-LC3 mediated by proteasome inhibitors (Fig. 5B). Western blotting confirmed that all these proteasome inhibitors activated MAPK8/9 and increased BAG3 levels (Fig. 5C). 50 μM of SP600125 significantly reduced LC3-II generation, simultaneously blocked induction of BAG3, while had little effects on RPS6KB phosphorylation elicited by proteasome inhibitors (Fig. 5C). AO staining demonstrated that, similar to starvation-induced autophagy, MAPK8/9/10 inhibitor SP600125 significantly blocked AVs formation mediated by proteasome inhibitors (Fig. S5A).

Figure 5. Involvement of MAPK8/9/10 in proteasome inhibitor-elicited autophagy in HepG2 cells. (A) HepG2 cells were treated with vehicle or MG132 in combination with different concentrations of SP600125, and western blot analysis was performed using the indicated antibodies. (B) HepG2 cells stably expressing EGFP-LC3B were treated with BZ, Epox, Lacta, or MG132 in the absence or presence of 50 μM of SP600125, the punctate distribution of was visualized under the fluorescence microscopy. (C) HepG2 cells were treated with the indicated proteasome inhibitor alone or in combination with 50 μM of SP600125, and western blot analysis was performed using the indicated antibodies. (D) HepG2 cells were transfected with a mock or BAG3 eukaryotic expression vector, and treated with MG132 and SP600125 alone or in combination. Western blot analysis was performed using the indicated antibodies. (E) HepG2 cells were transfected with a mock or BAG3 eukaryotic expression vector, then treated with MG132 and SP600125 alone or in combination. LC3 mRNA expression was analyzed using real-time RT-PCR.

To confirm the role of BAG3, EGFP empty (mock) or EGFP-tagged BAG3 (EGFP-BAG3) eukaryotic expression vector transfected HepG2 cells were treated with 50 μM of SP600125 and 2 μM of MG132 alone or in combination. Western blots demonstrated that BAG3 overexpression significantly blocked the inhibitory effects of SP600125 on LC3-II generation elicited by MG132 (Fig. 5D). BAG3 overexpression increased phosphorylation of MAPK8 under basal condition, while demonstrated little impact on suppressive effects of SP600125 on MAPK8/9 phosphorylation induced by MG132 (Fig. 5D). Real-time RT-PCR demonstrated that neither 50 μM of SP600125 nor BAG3 transfection affected LC3 mRNA expression (Fig. 5E). LysoTracker Red staining also demonstrated that 50 μM of SP600125 significantly blocked accumulation of AVs in untransfected cells, while BAG3 overexpression markedly rescued the suppressive effects of SP600125 on AVs formation elicited by MG132 (Fig. S5B).

Suppression of proteasome inhibitors-elicited cytotoxicity by noncanonical autophagy

MTT assay first was performed to investigate the potential role of noncanonical autophagy in the responsiveness of HepG2 to proteasome inhibitors, and found that cell viability was significantly decreased in cells treated with proteasome inhibitor in combination with the MAPK8/9/10 inhibitor SP600125, compared respectively with the cells treated with the proteasome inhibitor alone (Fig. 6A). Trypan blue exclusion demonstrated that 50 μM of SP600125 significantly increased HepG2 cell death mediated by proteasome inhibitors (Fig. 6B). Hoechst 33258 staining confirmed that SP600125 cotreatment significantly enhanced apoptotic cells induced by proteasome inhibitors (Fig. 6C). To confirm the role of BAG3, EGFP-tagged shBAG3 eukaryotic expression vector-transfected HepG2 cells were treated with these proteasome inhibitors. MTT assay demonstrated that shBAG3 significantly enhanced growth inhibition induced by these proteasome inhibitors, while a scrambled shRNA demonstrated no obvious effects (Fig. 6D). To confirm the involvement of downregulation of BAG3 in the enhancing effects of SP600125 in proteasome inhibitors-mediated cytotoxicity, we further transfected HepG2 cells with EGFP- tagged BAG3 eukaryotic expression vector. MTT assay demonstrated that BAG3 overexpression significantly suppressed cytotoxicity induced by BZ, Epox, Lacta and MG132 (Fig. 6E). Moreover, the enhancing effects of SP600125 on proteasome inhibitor-mediated cytotoxicity were significantly blocked by BAG3 overexpression (Fig. 6E).

Figure 6. Protection of HepG2 from proteasome inhibitors-mediated cytotoxicity by MAPK8/9/10 and BAG3-mediated noncanonical autophagy. (A) HepG2 cells were treated with MG132, Lacta, Epox or BZ in the absence or presence of 50 μM of SP600125, cell viability was analyzed using MTT assay. (B) HepG2 cells were treated with MG132, Lacta, Epox or BZ in the absence or presence of 50 μM of SP600125, dead cells were counted using trypan blue staining. (C) HepG2 cells were treated with MG132, Lacta, Epox or BZ in the absence or presence of 50 μM of SP600125, nuclei morphology was analyzed using Hoechst 33258 staining. Apoptotic nuclei were counted in six representative microscopic fields and three independent experiments were performed. Percentage of dead cells were plotted in the graph. (D) HepG2 cells were transfected with a scrambled control shRNA or EGFP-tagged shRNA against BAG3 (shBAG3), then treated with BZ, Epox, Lacta, or MG132 in the absence or present of 50 μM of SP600125. Cell viability was analyzed using MTT assay. (E) HepG2 cells were transfected with a mock or BAG3 eukaryotic expression vector, then treated with BZ, Epox, Lacta or MG132 in the absence or presence of 50 μM of SP600125. Cell viability was analyzed using MTT assay. *p < 0.01, N.S. not significant.

Discussion

Protein homeostasis, which maintains the balance between protein synthesis, folding and clearance, is central to cell survival. The UPS and autophagy are two major intracellular protein degradation pathways, and regarded as independent, parallel degradation systems with few or no points of interaction for a long time. However, the last few years have led to substantial insight into the interactions of proteasomes and autophagy in protein clearance. It has become apparent that impairment of the UPS leads to increased autophagic function, and in some contexts this upregulation of autophagy is considered to be a compensatory mechanism for impaired UPS function.^8,9,58-61 While there is a general consensus about a compensatory role of autophagy upon proteasomal inhibition, there is little consensus on the exact mechanism(s) of this crosstalk, as a number of potential explanations have been proposed to rationalize the interaction between UPS and autophagy.¹⁰

Autophagy is a dynamic process used to sequester and degrade cytoplasm and entire organelles in double membrane-bound vesicles, referred to as autophagosomes, which ultimately fuse with lysosomes to degrade its autophagic cargo.^4-6 BECN1 is part of the PtdIns3K complex, which plays a critical role in the formation of autophagosomes.²¹ Nevertheless, the current study indicated that the regulation of autophagy activated by proteasome inhibitors differs from that of starvation-induced autophagy. We found that the autophagic process elicited by proteasome inhibitors was completely independent of PtdIns3K activity and BECN1. Consistent with the current study, several recent studies have reported the existence of BECN1-independent autophagy that is insensitive to PtdIns3K inhibitors, which has been named as noncanonical autophagy.²⁸ For example, resveratrol induces autophagy in a BECN1-independent manner in breast cancer MCF7 cells.²⁶ Parkinsonian neurotoxin MPP⁺ induces BECN1-independent autophagy in SHSY5Y neuroblastoma cells and primary dopaminergic midbrain neurons.³¹ In addition, Z18, a small compound targeting the BH3 binding groove of BCL2L1/Bcl-XL, activates BECN1-independent autophagy in HeLa cells.⁶² Collectively, coupled with the present study, these studies indicate that there is an uncharacterized mechanism for autophagosome formation that can bypass BECN1 and PtdIns3K.

BAG3 is a member of BAG domain-containing co-chaperone family, which has been characterized traditionally by their interaction and regulation of the ATPase activity of the HSP70 chaperone family. A plethora of functions has been assigned to BAG3, including an antiapoptotic activity,^63-66 signal transduction,⁶⁷ cell adhesion and migration^68-71 and a regulatory role in virus assembly.^72-74 Recent studies on BAG3 attract much attention on its role in stimulation of autophagy.^{37-40,49,75-77} In an aging model, it has been demonstrated that BAG3 is increased in aging cells and acts in concert with the ubiquitin-binding protein SQSTM1/p62 (sequestosome 1) to elicit autophagic degradation of polyubiquitinated proteins.³⁷ Therefore, BAG3-mediated recruitment of autophagic degradation pathway seems to be a compensatory mechanism of the protein clearance systems to maintain protein homeostasis in the presence of a increased pro-oxidant and aggregation-prone milieu characteristic of aging.³⁷ Consistent with previous reports,^47,48 the current study demonstrated that BAG3 was increased by proteasome inhibitors in HepG2 cells. Importantly, we also demonstrated that BAG3 was implicated in autophagic responses elicited by proteasome inhibitors, as downregulation of BAG3 by shRNA significantly suppressed autophagic activity elicited by proteasome inhibitors. Considering intracellular milieu characteristics, especially status of intracellular degradative systems, cells exposed to proteasome inhibition mirror aging cells very well. For example, both aging cells and cells exposed to proteasome inhibitors demonstrate increased autophagic activity, but decreased proteasomal activity. Both aging cells and cells exposed to proteasome inhibitors demonstrated increase in BAG3 expression, accumulation of damaged and misfolded proteins, as well as accumulation of reactive oxygen species.^37,47,78-80

MAPK8/9/10, which are stress-activated protein kinases, have previously been shown to be involved in autophagy induced by starvation and ceramide.^21,81 In the current study, we found that inhibition of MAPK8/9/10 activation by SP600125 suppressed induction of BAG3, simultaneously proteasome inhibitors-mediated autophagic response was also suppressed, which was rescued by BAG3 overexpression. As BAG3 overexpression demonstrated little effect on inactivation of MAPK8/9/10 by SP600125, BAG3 might exert its autophagy regulatory role downstream of MAPK8/9/10. Inhibition of MTORC1 is one of the hallmarks of starvation-induced autophagy,^82,83 the current study demonstrated that proteasome inhibitors also suppressed activity of MTORC1, as assessed by phosphorylation of RPS6KB. shBAG3 and MAPK8/9/10 inhibitor SP600125 suppressed LC3-II generation, while producing no obvious effect on inhibition of MTORC1 induced by proteasome inhibitors in HepG2 cells, suggesting that BAG3 and MAPK8/9/10 signaling might modulate proteasome inhibition-induced autophagy downstream of MTORC1, but upstream of conversion of LC3. The exact mechanism by which BAG3 regulates autophagy requires further investigation.

In future studies, it will be instructive to investigate whether BAG3-mediated initiation of autophagy is BECN1-independent in aging cells. It is also important to clarify whether BAG3 is ascribed to noncanonical autophagy elicited by resveratrol and MPP⁺ et al.^26,31 Further illumination of these questions could clarify the links between BECN1-dependent canonical autophagy and the alternative BAG3-dependent noncanonical autophagy.

Materials and Methods

Western blot analysis

Cells were lysed in lysis buffer (20 mM TRIS-HCl, 150 mM NaCl, 2 mM EDTA, 1% Triton-X100 and protease inhibitor cocktail (Sigma-Aldrich, P2714). Cell extract protein amounts were quantified using the BCA^TM protein assay kit (PIERCE, 23225). Equivalent amounts of protein (25 μg) were separated using 12% SDS-PAGE and transferred to PVDF membrane (Millipore Corporation, IPVH00010).

Transmission electron microscopy

Cells grown in 35-mm dishes were fixed for at least 60 min in 2.5% glutaraldehyde at 4°C. After fixation, cell monolayers were washed three times in PBS and then postfixed in aqueous 1% OsO₄ and 1% K₃Fe(CN)₆ for 1 h. After 3 PBS washes, the cultures were dehydrated through a graded series of 30 to 100% ethanol, infiltrated, and then embedded in Polybed 812 epoxy resin (Polysciences, 02597-50). Ultrathin (60 nm) sections were collected on copper grids and stained with 2% uranyl acetate in 50% methanol for 10 min, followed by 1% lead citrate for 7 min. Sections were photographed using a JEOL JEM 1210 transmission electron microscope (JEOL) at 80 kV.

DNA construction and transfection

p40(phox)PX-EGFP plasmid was generously provided by Professor Jae U. Jung (University of Southern California). EGFP-LC3B and EGFP-BAG3 fusion plasmid were constructed by PCR and cloned into pcDNA3.1-EGFP vector, respectively. The constructs were verified by DNA sequencing. EGFP-tagged short hairpin RNA (shRNA) against BECN1 (shBECN1) or BAG3 (shBAG3) was purchased from Open Biosystems. Cells were transfected with Lipofectamine 2000 reagent (Invitrogen, 11668-027) as instructed by the supplier.

Fluorescence microscopy

Cells were stained with acridine orange (AO), monodansylcadaverine (MDC), LysoTracker Red at a final concentration of 1 μg/ml, 0.05 mmol/L and 50 nmol/L, respectively. For EGFP-LC3B translocation and PX-EGFP dot density, cells were stably transfected with pcDNA3.1-EGFP-LC3B and p40(phox)PX-EGFP plasmid, respectively. Images of live cells were taken using Olympus IX71 fluorescence microscopy (Olympus) equipped with digital epifluorescence imaging.

Cell viability assays

For cell viability assays, cells were plated in 96-well dishes (1 × 10⁴ cells per well) and the next day were treated with or without apoptosis inducing agents in 10% FBS-containing media and grown over a 24 h period. Cell viability was assessed using the 3-(4,5-dimethylthiazol-2-thiazolyl)-2,5-diphenyl tetrazolium bromide (MTT) assay (Chemicon, CT02) according to the manufacturer’s instruction.

RNA isolation and real-time reverse transcription-polymerase chain reaction (RT-PCR)

Total RNA was isolated using TRIZOL Reagent (Invitrogen, 15596-026) and first strand cDNA was synthesized from 2 μg of total RNA using SuperScript^TM II Rnase H^- Reverse Transcriptase (Invitrogen, 18064-014) according to the manufacturer’s instructions. Real-time RT-PCR analysis was performed in triplicate on the ABI prism 7000 sequence detection system (Applied Biosystems) using the SYBR Green PCR Master Mix (Applied Biosystems, 4367659). Results were normalized against those of GAPDH and presented as ratio vs. vehicle-treated control.

Statistics

The statistical significance of the difference was analyzed by ANOVA and post hoc Dunnett’s test. Statistical significance was defined as p < 0.05. All experiments were repeated three times, and data were expressed as the mean ± SD (standard deviation) from a representative experiment.

Glossary

Abbreviations:

ATG

autophagy-related

BAG3

Bcl-2 associated athanogene 3

BECN1

Beclin 1

BCL2

B-cell CLL/lymphoma 2

HSP70

heat shock protein 70

LC3/MAP1LC3

microtubule-associated protein 1 light chain 3

PtdIns3K

phosphatidylinositol 3-kinase

SQSTM1/p62

sequestosome 1

3-MA

3-methyladenine

BZ

bortezomib

CQ

cloroquine

EBSS

Earle’s balanced salt solution

Epox

epoxomicin

Lacta

lactacystin

NH₄Cl

ammonium chloride

WM

wortmannin

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplemental Materials

Supplemental materials may be found here: www.landesbioscience.com/journals/autophagy/article/24292