Low-density Lipoprotein Receptor-related Protein-1 (LRP1) Mediates Autophagy and Apoptosis Caused by Helicobacter pylori VacA*

Background: Helicobacter pylori VacA receptor(s) responsible for apoptotic cell death and autophagy has not been identified. Results: VacA-induced autophagy via low-density lipoprotein receptor-related protein-1 (LRP-1) binding precedes apoptosis. Conclusion: LRP1 mediates VacA-induced autophagy and apoptosis. Significance: This study identified LRP1 as a VacA receptor associated with toxin-induced autophagy and apoptosis and demonstrated its importance in the processes. In Helicobacter pylori infection, vacuolating cytotoxin (VacA)-induced mitochondrial damage leading to apoptosis is believed to be a major cause of cell death. It has also been proposed that VacA-induced autophagy serves as a host mechanism to limit toxin-induced cellular damage. Apoptosis and autophagy are two dynamic and opposing processes that must be balanced to regulate cell death and survival. Here we identify the low-density lipoprotein receptor-related protein-1 (LRP1) as the VacA receptor for toxin-induced autophagy in the gastric epithelial cell line AZ-521, and show that VacA internalization through binding to LRP1 regulates the autophagic process including generation of LC3-II from LC3-I, which is involved in formation of autophagosomes and autolysosomes. Knockdown of LRP1 and Atg5 inhibited generation of LC3-II as well as cleavage of PARP, a marker of apoptosis, in response to VacA, whereas caspase inhibitor, benzyloxycarbonyl-VAD-fluoromethylketone (Z-VAD-fmk), and necroptosis inhibitor, Necrostatin-1, did not inhibit VacA-induced autophagy, suggesting that VacA-induced autophagy via LRP1 binding precedes apoptosis. Other VacA receptors such as RPTPα, RPTPβ, and fibronectin did not affect VacA-induced autophagy or apoptosis. Therefore, we propose that the cell surface receptor, LRP1, mediates VacA-induced autophagy and apoptosis.

In Helicobacter pylori infection, vacuolating cytotoxin (VacA)-induced mitochondrial damage leading to apoptosis is believed to be a major cause of cell death. It has also been proposed that VacA-induced autophagy serves as a host mechanism to limit toxin-induced cellular damage. Apoptosis and autophagy are two dynamic and opposing processes that must be balanced to regulate cell death and survival. Here we identify the low-density lipoprotein receptor-related protein-1 (LRP1) as the VacA receptor for toxin-induced autophagy in the gastric epithelial cell line AZ-521, and show that VacA internalization through binding to LRP1 regulates the autophagic process including generation of LC3-II from LC3-I, which is involved in formation of autophagosomes and autolysosomes. Knockdown of LRP1 and Atg5 inhibited generation of LC3-II as well as cleavage of PARP, a marker of apoptosis, in response to VacA, whereas caspase inhibitor, benzyloxycarbonyl-VAD-fluoromethylketone (Z-VAD-fmk), and necroptosis inhibitor, Necrostatin-1, did not inhibit VacA-induced autophagy, suggesting that VacA-induced autophagy via LRP1 binding precedes apoptosis. Other VacA receptors such as RPTP␣, RPTP␤, and fibronectin did not affect VacA-induced autophagy or apoptosis. Therefore, we propose that the cell surface receptor, LRP1, mediates VacA-induced autophagy and apoptosis.
Helicobacter pylori colonizes more than half the world's population. Although persistent infection by H. pylori is accepted as a major cause of gastroduodenal diseases (e.g. peptic ulcer disease, gastric lymphoma, gastric adenocarcinoma), the responsible cellular pathways have not been defined. Variation in manifestations of H. pylori infection in different populations suggests differences in virulence of strains, host genetic susceptibility, and responses to environmental factors. Many H. pylori strains isolated from patients contain the cagA gene (cytotoxin-associated gene A) as well as produce the vacuolating cytotoxin, VacA. Additional H. pylori products, including urease, OipA, adhesins, heat-shock protein, and lipopolysaccharide appear to be involved in virulence (1,2).
Interestingly, VacA causes epithelial damage in mouse models both when given orally as a single agent (3) and when delivered by a toxigenic strain of H. pylori during gastric infection (4,5). In vitro, VacA is internalized by endocytosis (6), which is inhibited by CagA (7,8), and exerts multiple effects on susceptible cells, including vacuolation and mitochondrial damage, leading eventually to apoptosis (9 -13). In addition, VacA forms hexametric pores, followed by endocytosis and processing into late-endosomal compartments (14), which then undergo osmotic swelling to become large acidic vacuoles. Although vacuolation is the most obvious effect of VacA in vitro, it is not as obvious in vivo. The pleiotropic effects of VacA appear to result from activation of different signal transduction pathways through binding to several epithelial cell receptors, e.g. receptor protein-tyrosine phosphatase (RPTP) 3 ␤ and ␣ (15,16), fibronectin (FN) (17), sphingomyelin (18).
VacA enhanced tyrosine phosphorylation of the G proteincoupled receptor kinase-interactor 1 (Git1) as did pleiotrophin, an endogenous ligand of RPTP␤ (19). Oral administration of VacA to wild-type mice, but not to RPTP␤ knock-out mice, resulted in gastric ulcer. However, cells lacking RPTP␤ were able to internalize VacA and undergo vacuolation (20), suggesting that other VacA receptors were responsible for vacuolation. Recent interest has focused on the immunosuppressive effects of VacA, i.e. VacA inhibited proliferation of T cells due to down-regulation of interleukin-2 (IL-2) transcription (21,22). Through interactions with the ␤2-integrin subunit CD18 of the leukocyte-specific integrin LFA-1 (23), VacA plays an important role in inhibition of interleukin-2 (IL-2) gene expression after clathrin-independent endocytosis via PKC-dependent phosphorylation of the cytoplasmic tail of CD18 (24). Thus, VacA has effects on both epithelial cells (25) as well as inflammatory cells (26).
Over the last 10 years, studies have focused on the mechanism of cell death resulting from mitochondrial damage caused by VacA (10,12,13,27). Additional recent studies have shown that VacA induces autophagy, but the pathway has not been identified (28,29). Autophagy can promote the survival of dying cells (30). However, increased autophagic activity can also lead to cell death (31)(32)(33)(34)(35), suggesting that autophagy can be responsible for both cytoprotective and cytotoxic activities, depending on the specific cellular conditions.
Here we purified from AZ-521 cells, a human gastric epithelial cell line, a surface membrane protein, p500, which binds VacA, and identified it as low-density lipoprotein receptor-related protein-1 (LRP1). LRP1 binding of VacA was shown to be specifically responsible for VacA-induced autophagy and apoptosis. Similar to RPTP␣ and RPTP␤, LRP1 mediates VacA internalization in AZ-521 cells, but in contrast to RPTP␣ and RPTP␤, LRP1 targeted downstream pathways leading to autophagy and apoptosis.
Purification of VacA-The toxin-producing H. pylori strain ATCC 49503 was the source of VacA for purification as previously described (36).
Assay for Vacuolating Activity-Vacuolating activity was assessed using AZ-521 cells as previously described (36). Briefly, cells (1 ϫ 10 4 cells/well, 100 l) were grown as monolayers in 96-well culture plates for 24 h in a 5% CO 2 atmosphere at 37°C. VacA was added, and cells were incubated at 37°C for the indicated times. To quantify vacuolating activity, the uptake of neutral red into vacuoles was determined.
Preparation of Alexa 555-labeled VacA-To investigate VacA binding to cells and co-localization with other proteins in cells, VacA was labeled using the Alexa Fluor 555 Protein Labeling Kit (Molecular Probes), according to instructions provided by the manufacturer. In brief, 50 l of 1 M sodium bicarbonate buffer (pH 8.5) were added to 500 l (500 g in phosphatebuffered saline (PBS)) of VacA, followed by incubation with the reactive dye in the vial for 15 min at room temperature. To remove excess dye, the reaction mixture was applied to a PD-10 column (Amersham Biosciences). Alexa 555-labeled VacA (100 g/ml) was stored at Ϫ20°C.
Purification and Identification of p500-To purify p500 using affinity columns, AZ521 cells (5 ϫ 10 7 cells) were washed twice with PBS, and suspended in 10 ml of Sol buffer containing 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 10% glycerol, 1% Triton X-100, with protease inhibitor mixture (Roche Diagnostics)) for 15 min on ice. After centrifugation (20 min at 17,400 ϫ g), the supernatant was filtered (0.45 m, Millipore) and the filtrate (10 ml) applied to a Maackia amurensis (MAA)-agarose column (2 ml bed volume, Seikagaku Corporation). After washing the column, Sol buffer containing 50 mM ethylenediamine was used to elute the carbohydrate-containing proteins in 1-ml fractions. To confirm the presence of p500 in the eluted fractions, proteins in effluents were detected by lectin blotting using MAA as described previously (15,16). To identify p500, proteins in effluents were precipitated with chloroform/methanol, then heated at 100°C for 10 min in 1ϫ SDS-PAGE sample buffer, separated in 6% gels, and transferred to PVDF membranes, which were stained with Coomassie Brilliant Blue. The stained bands were used for LC-MS/MS analysis.
Immunoprecipitation-Immunoprecipitation of VacA-binding proteins from AZ521 cells was performed as described previously. In brief, biotinylated AZ521 cell lysates (100 g/200 l) were incubated at 4°C for 1 h with 1 g of native VacA or heat-inactivated VacA (100°C, 10 min), followed by incubation overnight at 4°C with 1 l of rabbit anti-VacA antibodies. Antibody-bound proteins were collected after addition of 20 l of rProtein G-agarose (Invitrogen), 50% (v/v) in Sol buffer, and incubated at 4°C for 1.5 h. After the beads were washed three times with Sol buffer, proteins were solubilized in SDS-PAGE sample buffer, resolved by SDS-PAGE, and transferred to PVDF membranes (Millipore; Immobilon-P membranes), which were incubated with streptavidin-HRP (Amersham Biosciences). Biotinylated proteins were detected using the enhanced chemiluminescence system (Pierce).
Immunofluorescence Confocal Microscopy-For immunofluorescence analysis of VacA co-localization with LRP1, RPTP␣, RPTP␤, or LC3B, AZ-521 cells (1 ϫ 10 5 cells) on coverglass (Matsunami) were incubated with 120 nM Alexa 555-labeled VacA for the indicated times, cells were fixed with 4% paraformaldehyde (PFA) at room temperature for 15 min, washed with PBS twice, and then immediately permeabilized with ice-cold 100% methanol for 10 min at Ϫ20°C. The cells are then rinsed three times with PBS and incubated with blocking buffer (5% goat serum, 0.3% Triton X-100 in PBS) at room temperature for 1 h. To visualize LRP1 (8G1 antibody, 1:50), RPTP␣ (antibody provided by Jan Sap, 1:100), RPTP␤ (polyclonal, 1:250), or LC3B (D11, 1:200), cells were further incubated with the primary antibodies in 1% BSA/PBS buffer overnight at 4°C, washed twice with PBS and incubated with anti-rabbit Alexa 488 (Molecular Probes), anti-mouse 488 (Molecular Probes), or anti-mouse Cy5 (Jackson ImmunoResearch Laboratories Inc.) antibodies at room temperature for 1 h in the dark. After washing with PBS three times, cells were mounted on glass slides using Prolong Gold Antifade reagent with DAPI. For staining the lysosomal compartment in VacA-treated cells, cells were incubated with 100 nM LysoTracker Red DND-99 (Molecular Probes) according to the instruction manual, before fixation with 4% paraformaldehyde. Colocalization of VacA and the indicated proteins was analyzed by FV10i-LIV confocal microscopy (Olympus). The images were arranged with Adobe Photoshop CS4.
Statistics-Densitometric analysis on the immunoblots was done by Image Gauge software (FUJI FILM). The p values for densitometric analysis and vacuolating assay were determined by Student's t test with GraphPad Prism software (GraphPad, San Diego, CA). p values of Ͻ0.05 were considered statistically significant.

RESULTS
Purification and Identification of p500-Our analysis of membrane proteins that bind VacA revealed three proteins, i.e. RPTP␣, RPTP␤, and an unidentified p500. The latter protein had a molecular mass higher than RPTP␤ and reacted with MAA lectin (15,16). In the present study, we purified p500 using MAA-agarose column chromatography and identified it by LC-MS/MS as LRP1 (Fig. 1). We confirmed its association with native VacA by immunoprecipitation (Fig. 1).
LRP1 Mediates VacA Binding and Internalization in AZ-521 Cells-Confocal microscopy analysis revealed that in AZ-521 cells VacA colocalized with LRP1 on cell membranes, and was internalized, whereas heat-inactivated VacA did not show colocalization and internalization with LRP1 (data not shown) ( Fig.  2A). Furthermore, AZ-521 cells transfected with siRNA of LRP1 did not show significant toxin binding resulting in internalization, suggesting that LRP1 mediates VacA binding to the cell surface and facilitates its internalization. In agreement with these data, silencing of the p500 gene inhibited vacuole formation caused by VacA (Fig. 2B). These results suggest that LRP1 is associated with toxin internalization.
VacA Induced Generation of LC3-II in an LRP1-dependent Manner-Based on the prior reports (28,29) that VacA induced autophagy in AGS cells, we determined whether VacA induced LC3-II generation from LC3-I in AZ-521 cells. Consistent with previous findings, Western blot analysis showed that VacA induced LC3-II generation from LC3-I in a time-dependent manner (Fig. 3a). As expected, immunoblots of VacA-treated cells transfected with control siRNA indicated a progressive conversion over 10 h of LC3-I to LC3-II. In LRP1 siRNA-transfected cells, LRP1 expression was down-regulated after 4 h with VacA and conversion of LC3-II from LC3-I was suppressed ( Fig. 3b and supplemental Fig. S1). These data suggest an important role of LRP1 in mediating autophagy in AZ-521 cells in response to VacA.
VacA Induced Formation of Autophagosomes and Autolysosomes in AZ-521 Cells-To determine whether VacA induces autophagic vacuoles, AZ-521 cells were incubated with 120 nM VacA. We microscopically observed that active VacA (A) is sufficient to trigger autophagic vacuoles such as autophagosomes containing LC3-II after a 4-h incubation, followed after by 12 h incubation by formation of autolysosomes as detected by LysoTracker (Fig. 4A). Cells incubated with heat-inactivated VacA (IA) showed low or undetectable levels of these autophagic vacuoles after 12 h incubation. Furthermore, confocal microscopy analysis showed that intracellular VacA partially co-localized with LC3-II and LRP1, consistent with the conclusion that LRP1 plays an important role in VacA-induced autophagosome formation. However, LRP1 knockdown with siRNA suppressed VacA co-localization with LC3-II, suggesting that LRP1 is essential for formation of autophagosomes in response to VacA (Fig. 4B).
Vacuoles Caused by VacA Are Characterized as Autophagosomes and Autophagolysosomes-Confocal microscope visualization of LC3-II, VacA, and LRP1 revealed that vacuoles caused by VacA are of at least two different types; one type consists of autophagic vacuoles such as autophagosomes and autophagolysosomes and the second type lacks LC3-II (Fig.  5A). These observations support previous findings that VacAdependent autophagosomes and large vacuoles are distinct intracellular compartments and autophagy is independent of the formation of large vacuoles by VacA (29). Interestingly, some vacuoles observed with RPTP␤ revealed small light vacuoles without LC3-II (Fig. 5B) and dense vacuoles with RPTP␣ were devoid of LC3-II (Fig. 5C). Although little is known about the physiological importance of the autophagy-dependent degradation of mitochondria (mitophagy) (39), several studies have suggested that PINK1/parkin-dependent mitophagy selectively degrades mitochondria (40), implying that mitophagy contributes to mitochondrial quality control. As shown in Fig. 5D, after 10 h incubation mitochondria were not observed in vacuoles with LC3-II. Furthermore, recent studies revealed that p62 binds to LC3 on the autophagosome membrane to target aggregates to autophagosomes for degradation (41). After 24 h incubation, VacA, not heat-inactivated VacA, induced formation of puncta, which were colocalized with LC3-II and p62 (Fig. 5E).
Among VacA-binding Proteins, LRP1, but Not RPTPs and FN, Mediates VacA-dependent Autophagy-To assess which VacAbinding proteins were responsible for VacA-induced autophagy, we examined the effect of silencing and knockout of the genes for RPTP␤, RPTP␣, and fibronectin. Although LRP1 silencing blocked VacA-stimulated generation of LC3-II as shown in Fig.  3b, silencing these other genes did not show a similar effect, suggesting that only LRP1 may be critical for VacA-induced autophagy (Fig. 6).

LRP-1, but Not RPTPs, Mediates Cleavage of Caspase-7 and PARP Caused by
VacA-Excessive autophagy can cause cell death (34,42). Furthermore, VacA-induced cell death may occur through a programmed necrosis pathway in a caspase-independent process in AZ-521 cells (27). Therefore, we examined whether VacA-induced cell death resulted from autophagy via an LRP1-dependent pathway. Western blot analysis showed that LRP1 silencing blocked VacA-induced generation of LC3-II as well as cleavages of effector caspase-7 and PARP, suggesting that VacA binding to LRP-1 is responsible for not only autophagy but also for apoptosis in AZ-521 cells (Fig. 7).

Effects of Atg5 Silencing, Z-VAD-fmk and Necrostatin-1 on VacA-induced LC3-II Production and Cleavage of PARP-To
further examine the link between autophagy and apoptosis, the effects of Atg5 silencing with siRNA, general caspase inhibitor (Z-VAD-fmk) and RIPK inhibitor (Necrostatin-1) on LC3-II generation, and PARP cleavage was evaluated. Silencing of the Atg5 gene inhibited generation of LC3-II as well as PARP cleavages in response to VacA (Fig. 8), whereas both inhibitors, Z-VAD-fmk and Necrostatin-1, which interfere with apoptosis (43), did not inhibit VacA-induced autophagy, suggesting that VacA-induced autophagy precedes apoptosis in AZ-521 cells. Necrostatin-1, which inhibits necroptosis (44), did not interfere with VacA-induced generation of LC3-II and PARP cleavage.
Effect of Anion Channel Blockers, NPPB and DIDS, on VacAinduced LC3-II Production-To assess whether membrane channels formed by VacA may also be involved in autophagy (29), we tested the effects of pretreating AZ-521 or AGS cells with chloride channel blockers, NPPB and DIDS, which are known to block both VacA-mediated channel activity and cellular vacuolation (45). AZ-521 cells were pretreated for 30 min with 100 M NPPB or 100 M DIDS prior to incubation with VacA for 6 h. Both NPPB and DIDS inhibited VacA-induced LC3-II generation in AZ-521 cells (Fig. 9a), but not in AGS cells under these conditions (Fig. 9b). " Immunocomplexes were separated by SDS-PAGE in 6% gels and transferred to PVDF membranes. VacAbinding proteins were detected with streptavidin-HRP. b, proteins immunoprecipitated (IP) with heat-inactivated or wild-type VacA were separated by SDS-PAGE in 6% gels and transferred to PVDF membranes, which were incubated with MAA-lectin conjugated to digoxigenin and then with anti-digoxigenin Fab fragments conjugated to alkaline phosphatase, followed by reaction with 4-nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl phosphate. c, biotinylated AZ-521 cell lysates were incubated overnight with a MAA-agarose column (2 ml bed volume), which was washed with 20 ml of Sol buffer. Bound proteins were eluted, concentrated, and separated by SDS-PAGE as described under "Experimental Procedures." MAA-lectin blotting is shown in the left panel and Coomassie Brilliant Blue (CBB) staining in the right panel. The stained p500 protein band was hydrolyzed with trypsin and subjected to LC-MS/MS analysis. The procedures described in a-c were repeated at least three times with similar results. IB, immunoblot.

LRP1 Mediates VacA-induced Autophagy and Apoptosis
and channel forming by the N-terminal p33 domain alone drives pleiotropic cellular activities of VacA; i.e. vacuolation, mitochondria damage, apoptosis (10,47), autophagy (28,29), and programmed necrosis (27), suggesting that VacA may be characterized as a pore-forming toxin (47). Another study has indicated that both p33 and p55 are required to form a functional channel in the inner mitochondria membrane and trigger apoptosis (49). In addition, it is now widely accepted that the C-terminal p55 domain of VacA plays an essential role in its binding to target cells (50,51).  SEPTEMBER 7, 2012 • VOLUME 287 • NUMBER 37

JOURNAL OF BIOLOGICAL CHEMISTRY 31109
The present study defines a novel role for VacA signaling through LRP1 in AZ-521 cells, inducing autophagy and apoptosis (Fig. 7). LRP1 is a large endocytic receptor belonging to the LDL receptor family. This membrane protein consists of a 515-kDa heavy chain containing the extracellular ligand-binding domains and a noncovalently associated 85-kDa light chain, which consists of a transmembrane domain and a short cytoplasmic tail. LRP1 functions as a clearance receptor mediating

LRP1 Mediates VacA-induced Autophagy and Apoptosis
the uptake and catabolism of various ligands from the pericellular environment, e.g. LRP1 binds to apolipoprotein E-rich lipoproteins, lipoprotein lipase, ␣ 2 -macroglobulin, lactoferrin, and tissue plasminogen activator; it functions in lipoprotein metabolism, degradation of proteases and proteinase/inhibitor complexes, activation of lysosomal enzymes and cellular entry of viruses, and bacterial toxin such as Pseudomonas exotoxin A (52). LRP1 has also been shown to function in the turnover of fibronectin (53). This is the first study to provide evidence that LRP1 mediates autophagy. In AZ-521 cells, VacA triggered formation of autophagosomes, followed by autolysosome formation, consistent with the observations in AGS cells (29). Because LRP1 knockdown with siRNA resulted in inhibition of VacA-induced LC3-II generation and cleavage of both caspase 7 and PARP, induction by VacA of both autophagy and apoptosis occurred via, at least in part, association with LRP1. VacA also promoted formation of vacuoles containing RPTP␤ and RPTP␣, which

LRP1 Mediates VacA-induced Autophagy and Apoptosis
were characterized as light and dense vacuoles, respectively, by confocal microscopy, and large vacuoles (Fig. 5). We observed the presence in AZ-521 cells of large vesicles without autophagosome markers in wild-type and Atg12 knockdown AGS cells (29). In general, vacuole formation caused by VacA is required for VacA channel activity (54). Our studies using chloride channel blockers, NPPB and DIDS, to address the relationship between VacA-induced autophagy and channel activity of VacA in AZ-521 cells treated with VacA revealed that these channel blockers inhibited LC3-II generation in response to VacA (Fig. 9), suggesting that channel activity may be required for LRP1-dependent autophagy. More interestingly, VacA-induced autophagy was not blocked by caspase inhibitor and RIPK inhibitor, suggesting that VacA-induced autophagy via LRP1 binding precedes apoptosis.
Autophagy is a degradation process that involves formation of autophagosomes, which engulf cytoplasmic components, and fuse with the lysosome/vacuole for degradation of con-   4). b, the indicated siRNA-transfected AZ-521 cells were incubated with 120 nM heat-inactivated or wild-type VacA for 6 h at 37°C and the cell lysates were subjected to immunoblotting with anti-LC3B, anti-cleaved PARP, anti-RPTP␣, or anti-RPTP␤ antibodies. ␣-Tubulin served as a loading control. A blot representative of three separate experiments is shown.
tents. This process is considered cytoprotective but in certain settings excessive autophagy can cause cell death (34,42). Little information exists concerning the molecular mechanisms underlying the regulation of apoptosis by autophagy. Our data indicate that autophagy induced by VacA does not involve the canonical pathway in which Beclin-1 initiates the generation of autophagosomes by forming a multiprotein complex with class III PI3K, because 3MA, a class III PI3K inhibitor, and silencing of Beclin-1 did not inhibit autophagy induced by VacA (supplemental Fig. S2). The detailed mechanism by which VacA induces autophagy and apoptosis via LRP1 is not clear. Within VacA-intoxicated cells that provoke death signaling via mitochondrial damage, cells attempt to limit damage by seeking what catabolic benefits may be found in autophagy as indicated by Terebiznik et al. (29). Therefore, once the stress-provoking, death-signaling response to VacA is overwhelming, autophagy is futile, and it is beneficial to induce apoptosis. Future studies involving mouse models and human specimens will help to determine whether LRP1 plays a critical role in the pathobiology of H. pylori infection in vivo.