Acetaldehyde suppresses HBV-MHC class I complex presentation on hepatocytes via induction of ER stress and Golgi fragmentation

Abstract

Alcohol consumption worsens hepatitis B virus (HBV) infection pathogenesis. We have recently reported that acetaldehyde suppressed HBV peptide-major histocompatibility complex I (MHC class I) complex display on hepatocytes, limiting recognition and subsequent removal of the infected hepatocytes by HBV-specific cytotoxic T lymphocytes (CTLs). This suppression was attributed to impaired processing of antigenic peptides by the proteasome. However, in addition to proteasome dysfunction, alcohol may induce endoplasmic reticulum (ER) stress and Golgi fragmentation in HBV-infected liver cells to reduce uploading of viral peptides to MHC class I and/or trafficking of this complex to the hepatocyte surface. Hence, the aim of this study was to elucidate whether alcohol-induced ER stress and Golgi fragmentation affect HBV peptide-MHC class I complex presentation on HBV+ hepatocytes. Here, we demonstrate that, while both acetaldehyde and HBV independently cause ER stress and Golgi fragmentation, the combined exposure provided an additive effect. Thus we observed an activation of the inositol-requiring enzyme 1α-X-box binding protein 1 and activation transcription factor (ATF)6α, but not the phospho PKR-like ER kinase-phospho eukaryotic initiation factor 2α-ATF4-C/EBP homologous protein arms of ER stress in HBV-transfected cells treated with acetaldehyde-generating system (AGS). In addition, Golgi proteins trans-Golgi network 46, GM130, and Giantin revealed punctate distribution, indicating Golgi fragmentation upon AGS exposure. Furthermore, the effects of acetaldehyde were reproduced by treatment with ER stress inducers, thapsigargin and tunicamycin, which also decreased the display of this complex and MHC class I turnover in HepG2.2.15 cells and HBV-infected primary human hepatocytes. Taken together, alcohol-induced ER stress and Golgi fragmentation contribute to the suppression of HBV peptide-MHC class I complex presentation on HBV+ hepatocytes, which may diminish their recognition by CTLs and promote persistence of HBV infection in hepatocytes.

NEW & NOTEWORTHY Our current findings show that acetaldehyde accelerates endoplasmic reticulum (ER) stress by activating the unfolded protein response arms inositol-requiring enzyme 1α-X-box binding protein 1 and activation transcription factor (ATF)6α but not phospho PKR-like ER kinase-p eukaryotic initiation factor 2α-ATF4-C/EBP homologous protein in hepatitis B virus (HBV)-transfected HepG2.2.15 cells. It also potentiates Golgi fragmentation, as evident by punctate distribution of Golgi proteins, GM130, trans-Golgi network 46, and Giantin. While concomitantly increasing HBV DNA and HBV surface antigen titers, acetaldehyde-induced ER stress suppresses the presentation of HBV peptide-major histocompatibility complex I complexes on hepatocyte surfaces, thereby promoting the persistence of HBV infection in the liver.

INTRODUCTION

Hepatitis B virus (HBV) infection is a global public health problem. It is estimated that one-third of the world's population has been infected with HBV and that around 290 million are chronic carriers (60). The pathogenesis and clinical manifestations of hepatitis B are controlled by the interaction of this noncytopathogenic virus with a host immune system, which leads to liver injury as a result of immune-mediated HBV-infected hepatocyte clearance. In ~80% of cases, HBV causes acute hepatitis, but some patients develop chronic hepatitis and end-stage liver diseases. Both innate and adaptive immune responses contribute to the immune control of HBV infection although the clearance of HBV and disease pathogenesis are mainly performed by the adaptive immune response (9). The outcome of HBV infection is influenced by several viral and host factors and can be aggravated by alcohol consumption to prolong its persistence (1, 14, 56). Importantly, the combination of alcohol and HBV enhances the progression to steatosis, fibrosis, cirrhosis, and hepatocellular carcinoma (HCC) (13, 54). Previous studies reported that alcohol consumption and HBV infection operate synergistically (12, 29, 65) and share mechanisms for HCC development (66). However, the exact role of alcohol in HBV infection-associated liver injury has not been clearly elucidated.

It is known that HBV is not cytotoxic to hepatocytes, and the clearance of infected cells depends on the ability of cytotoxic T lymphocytes (CTLs) to recognize and eliminate HBV-expressing hepatocytes (4, 44, 52). Recognition of the HBV peptide-major histocompatibility complex I (MHC class I) complex on hepatocyte surfaces is a necessary step of effective hepatocyte killing by CTLs. Previously, we have shown that the ethanol metabolite, acetaldehyde, suppresses the presentation of HBV peptide-MHC class I complex on the hepatocyte surface, and this occurs due to suppressing effects of alcohol on the peptide processing by the proteasome (19). The impairment of immunoproteasome function and protein-loading complex (PLC) components, TAP and tapasin, are IFN-γ dependent, and their downregulation after acetaldehyde treatment could be explained by attenuation of IFN-γ signaling via the JAK-STAT1 pathway (19). However, these findings could not fully explain some IFN-γ-independent events contributing to ethanol-induced decline in HBV peptide-MHC class display on hepatocytes, which may be attributed to impaired trafficking of this complex to cell surface. In fact, after proteasomal degradation of HBV peptides followed by their loading to human leukocyte antigen (HLA) in endoplasmic reticulum (ER), these HBV peptide-MHC class I complexes are trafficked via the trans-Golgi to the plasma membrane, where they are recognized by T lymphocyte cell receptors (TCR) on CTLs (11, 15). To this end, ER stress (observed either in HBV infection or under ethanol exposure) can suppress the trafficking of peptide-MHC class I complex and contribute to Golgi fragmentation (2, 7, 11, 27, 28, 36). However, the combined effects of alcohol and HBV in ER stress and Golgi fragmentation in terms of MHC class I-restricted antigen presentation have not been investigated before. Hence, the focus of this study was to investigate whether ethanol metabolism affects the presentation of HBV peptide-MHC class I complex on infected hepatocytes by promoting ER stress and Golgi fragmentation.

Here, we hypothesize that HBV synergizes with acetaldehyde to induce ER stress, as well as disrupt Golgi to ultimately decrease the delivery of HBV peptide-MHC class I complex to hepatocyte surface. To mimic ethanol metabolism in ethanol-nonmetabolizing HBV-transfected HepG2.2.15 cells (which are HLA-A2 positive and express 18-27 HBV core peptide), we exposed the cells to an acetaldehyde-generating system (AGS). AGS has been well characterized and successfully used in our previous studies (19–22, 24). Major end-point results were confirmed on HBV-infected human primary hepatocytes (HLA-A2+) treated with 50 mM ethanol. We demonstrated here that both acetaldehyde and ethanol cause ER stress by activating the inositol-requiring enzyme 1α (IRE1α) and activation transcription factor (ATF)6 pathway and Golgi fragmentation, thereby decreasing the HBV core 18-27 peptide-MHC class I complex presentation on hepatocytes.

MATERIALS AND METHODS

Reagents and media.

High-glucose Dulbecco’s modified Eagle medium (DMEM) and fetal bovine serum were purchased from Invitrogen (Carlsbad, CA). Williams E media and hepatocyte plating and maintenance supplements were from Gibco by Thermo Fisher Scientific (Foster City, CA). TRIzol from Life Technologies (Carlsbad, CA) was used. PCR and RT-PCR reagents, such as High-Capacity cDNA Reverse Transcription Kit and TaqMan Universal Master Mix II, with UNG, were from Applied Biosystems by Thermo Fisher Scientific. For flow cytometry studies, human Fc receptor binding inhibitor, anti-human HLA-A2-APC, and mouse isotype control were from Affymetrix-eBioscience (San Diego, CA). Expression of HBV core peptide FLPSDEFPSV-HLA-A2 was measured by flow cytometry using antibody to 18-27 HBV peptide-HLA-A2 purified from supernatant of hybridoma cells at Department of Internal Medicine, University of Nebraska Medical Center (Omaha, NE). Hybridoma cells were obtained from PHARMEXA (San Diego, CA). Anti-IRE1, anti-nuclear factor erythroid 2-related factor 2 (Nrf2), and anti-ATF6 antibodies were from Novus Biologicals (Centennial, CO); anti-PKR-like ER kinase (PERK), anti-glutathione peroxidase 1(GPx1), anti-ATF, anti-GM130, anti-trans-Golgi network 46 (TGN46), and anti-Giantin were obtained from Abcam (Cambridge, MA); anti-C/EBP homologous protein (CHOP), cleaved caspase-3, and anti-eukaryotic initiation factor 2α (eIF2α) antibodies were from Cell Signaling (Beverly, MA). Anti-β-actin was from Santa Cruz Biotechnology (Santa Cruz, CA).

Cells and treatments.

In this study, we used parent HepG2 cells and HepG2.2.15 cells stably transfected with HBV and able to replicate virus and produce viral particles (HBV genotype D) (67). To mimic ethanol metabolism, we employed an AGS, which does not require transfection for generation of acetaldehyde (Ach). We believe that Ach is a major acting ethanol metabolite in HBV-infected cells since HBx protein downregulates cytochrome P4502E1 (CYP2E1) expression (38). Cells were treated or not with AGS for 72 h. We have characterized AGS in our recent publications (20, 22–24). Briefly, AGS contains yeast alcohol dehydrogenase (ADH), 50 mM ethanol as a substrate, and nicotinamide adenine dinucleotide (NAD⁺) as a cofactor and provides continuous enzymatic generation of physiologically relevant amounts of Ach for at least 72 h. This generated Ach easily enters into the cells and induces biological effects without causing significant cell toxicity (24). To mimic ER stress, we used two inducers, thapsigargin and tunicamycin, that were procured from Fisher Scientific (Atlanta, GA). Primary human hepatocytes (HLA-A2+) were from Triangle Research Laboratories (Durham, NC). They were infected with HBV (50 GEq/cell) and after 5 days of infection were treated with 50 mM EtOH for 2 days. The overall infection period was 7 days. Since hepatocytes plated on collagen undergo fast dedifferentiation and lose CYP2E1 and ADH expression after 24 h (25) and because the sustained expression of these ethanol-metabolizing enzymes is necessary for the successful ethanol treatment, cells were plated on custom soft gels [polyelectrolyte multilayer (PEM) film coating on top of the polydimethyl siloxane surface, 2D culture] to support long-term cell functionality (16).

RNA, DNA isolation, real-time PCR, and ddPCR.

Reagents for RNA isolation, cDNA synthesis, and real-time PCR were from Life Technologies and Applied Biosystems by Thermo Fisher Scientific (Carlsbad, CA and Foster City, CA). Total RNA was isolated from cells by TRIzol reagent. In a two-step procedure, 200 ng RNA was reverse transcribed to cDNA using the high-capacity reverse transcription kit. cDNA was amplified using TaqMan Universal Master Mix-II with fluorescent-labeled primers (TaqMan gene expression systems). After incubation in a model 7500 qRT-PCR thermal cycler, the relative quantity of each RNA transcript was calculated by its threshold cycle (Ct) after subtracting the reference cDNA (GAPDH). Data were expressed as the quantity of transcript (RQ). All primers and probes (single-vial primer probe) were obtained from Applied Biosystems (Foster City, CA).

HBV DNA.

HBV DNA levels were quantified by ddPCR method as we previously published (19). Total DNA was prepared using the DNeasy Kit (Qiagen, Germany) according to the manufacturer’s protocol. The concentrations of DNA were quantified using the QX200 Droplet Digital PCR System (Bio-Rad, Hercules, CA) according to the manufacturer’s instructions. Briefly, the 20 μL ddPCR reaction composed of 2× ddPCR Supermix (5 μL), reverse transcriptase (2 μL), and 300 mM DTT (1 μL) (Bio-Rad, Pleasanton, CA), 900 nmol/HBV sense (5′-CGA CGT GCA GAG GTG AAG-3′) and antisense (5′-CAC CTC TCT TTA CGC GGA CT-3′) primers, 250 nmol/ HBV probe (5′-/56-FAM/ATC TGC CGG /ZEN/ACC GTG TGC AC/3IABkFQ/-3′), and 5 μL adjusted DNA sample in RNase-free water. Primers and probes were from Integrated DNA Technologies, Inc. (Coralville, IA). Prepared droplets were transferred to corresponding wells of a Bio-Rad 96-well PCR plate using an Automated Droplet Generator as described in the instruction manual (no. 10043138). The PCR plate was subsequently heat-sealed with pierceable foil using the PX1 PCR plate sealer (Bio-Rad, Hercules, CA) and then amplified in the C1000 Touch deep-well thermal cycler (Bio-Rad). The cycling conditions were as follows: an initial denaturation cycle of 10 min at 95°C, followed by 45 cycles of denaturation for 30 s at 94°C, annealing for 60 s at 57°C (ramping rate set to 2°C/s), and a final incubation for 10 min at 98°C, ending at 4°C. After amplification, the 96-well plate was fixed in a plate holder and placed into the QX200 Droplet Reader (Bio-Rad). The ddPCR data were analyzed using the QuantaSoft analysis software (Bio-Rad). Fluorescent signals of droplets were manipulated with the QuantaSoft analysis software version 1.8 (Bio-Rad). Positive droplets with higher fluorescent signals and negative droplets with lower fluorescent signals were divided by applying a fluorescence amplitude threshold. The absolute concentration of each sample was automatically reported by the ddPCR software by calculating the ratio of the positive droplets over the total droplets combined with Poisson distribution. Thus the final concentration of template was equal to the results, as calculated by the software, multiplied by the dilution factor of the template in the reaction system.

HBV surface antigen Sandwich ELISA.

HBV surface antigen (HBsAg) levels were measured in cell culture media by ELISA using LSBio Kit (LifeSpan Biosciences, no. LS-F37979, Seattle, WA).

Immunofluorescence.

GM130, TGN46, and Giantin staining were performed by Immunofluorescence. HepG2 and HepG2.2.15 cells were plated on coverslips. After 72 h of AGS treatment, the cells were processed (including washing with PBS, fixation, permeabilization, and blocking) for staining and incubated with GM130, TGN46, and Giantin primary antibodies (1:50 dilution) for 2 h, followed by 1 h with secondary antibody (Alexa Fluor 594 donkey-anti-goat and goat-anti-mouse; 1:200 dilution, Life Technologies). Nuclei were stained with DAPI. The presence of GM130, TGN46, and Giantin were visualized using a ×63 lens in LSM 710 confocal microscope (Carl Zeiss, Peabody, MA).

Immunoblotting (Western blot).

Cell lysates prepared in 0.5 M EDTA, 2 M Tris, 20 mM Na₃VO₄, 200 mM Na₄P₂O₇, 100 mM PMSF, 1 M NaF, 20% Triton X-100, and aprotinin, pH 7 were separated and subjected to immunoblotting technique as previously described (23). Blots were developed using Odyssey infrared imaging system, and the protein bands were quantified using Li-Cor software (Li-Cor Bioscience, Lincoln, NE).

Flow cytometry analysis.

The assay was done as we previously published (19). HepG2.2.15 cells either were exposed or not to ER stress inducers, thapsigargin and tunicamycin, overnight, while human hepatocytes were treated with 50 mM EtOH and ER stress inducers. The cells were washed with PBS and detached by accutase, Invitrogen by Thermo Fisher Scientific (Carlsbad, CA). Cells were collected by centrifugation, then incubated with human Fc receptor-binding inhibitor for 20 min, followed by exposure to HBV peptide 18-27-HLA-A2 primary antibody for 60 min. After the washing, Alexa Fluor 647 secondary antibody was added for 30 min, and the cells were fixed in 2% paraformaldehyde for flow cytometry analysis. All incubations were on ice. Data were collected on a BD LSR2 flow cytometer and analyzed by using BD FACSDiva Software v6.0.

Low-acid wash.

Cells were either treated or not overnight with thapsigargin and tunicamycin overnight. Before the end of treatment, cells were washed with citrate buffer (0.062 M Na₂HPO₄, 0.132 M citric acid, 0.5% BSA, pH 3) for 2 min as described by Sugawara et al. (57). The citrate buffer wash was then neutralized by 3× wash with medium. Cells were incubated in fresh medium at 37°C for up to 4 h and analyzed by flow cytometry for expression of HLA-A2 as described in Flow cytometry analysis.

Statistical analyses.

Data from at least three duplicate independent experiments were expressed as mean values ± standard error. Comparisons among multiple groups were determined by one-way ANOVA, using a Tukey’s post hoc test. For comparisons between two groups, we used Student’s t test. A probability value of 0.05 or less was considered significant.

RESULTS

We provide here the mechanistic evidence that the acetaldehyde-induced ER stress and Golgi fragmentation may also affect the trafficking of the HBV core peptide 18-27-HLA-A2 (FLPSDFFPSV-HLA-A2) complex, thereby contributing to the impaired presentation of this complex on hepatocyte surface.

Effect of AGS on unfolded protein response mRNA markers in HepG2.2.15 cells.

ER stress activates a complex signaling network referred to as the unfolded protein response (UPR) to reduce ER stress and restore homeostasis. Here, we studied how the combination of HBV expression and exposure to AGS affects the UPR in liver cells. We treated HepG2 and HBV-transfected HepG2.2.15 cells with AGS for 72 h and then measured mRNA expression of master regulators of UPR signaling candidates as previously reported (46). The differential regulation of UPR genes in AGS-treated HepG2 and HBV-transfected cells is presented in Fig. 1. We found that the first arm of UPR, IRE1-related genes X-box binding protein 1 [spliced (s) XBP1; Fig. 1A] and ER degradation-enhancing-α-mannosidase-like protein (EDEM1), were elevated in HepG2 and HepG2.2.15 (HBV+) cells exposed to AGS, but the magnitude of response to AGS was higher in HBV+ cells; furthermore, HBV expression in HepG2.2.15 cells increased sXBP1 and EDEM1 levels (Fig. 1, A and C). We found no increase in uXBP1 gene activation in response to AGS in both HepG2 and HepG2.2.15 cells; HBV expression in HepG2.2.15 elevated uXBP1 mRNA levels compared with parental HepG2 cells (Fig. 1B). In addition, expression of ER chaperone, immunoglobin-binding protein (BiP), as a central regulator of the UPR stress sensors and ER chaperone in assisting protein folding was not changed in all groups (Fig. 1D).

Fig. 1.

Effects of acetaldehyde-generating system (AGS) and hepatitis B virus (HBV) on unfolded protein response (UPR) mRNA markers in HepG2 and HepG2.2.15 cells. Cells were either treated or not treated with AGS for 72 h and then harvested in TRIzol for RNA isolation. Real-time PCR (RT-PCR) analysis was done for spliced X-box binding protein 1 (sXBP-1) (A), uXBP-1 (B), ER degradation-enhancing-α-mannosidase-like protein 1 (EDEM1) (C), immunoglobin-binding protein (BiP) (D), and activation transcription factor (ATF)4 (E), and C/EBP homologous protein (CHOP) (F) GAPDH was used as an internal control. Data are from 3 independent experiments presented as means ± SE. Bars marked with the same letter are not significantly different from each other; bars with different letters are significantly different (P ≤ 0.05).

Next, we measured the genes related to another arm of UPR, the PERK pathway, and found that the level of ATF4 mRNA expression was not changed in all treatment groups (Fig. 1E). Also, the CHOP mRNA levels were only slightly enhanced by AGS exposure in HepG2 cells, while, in HepG2.2.15 cells, we observed threefold increase in response to AGS, and CHOP expression in untreated HepG2.2.15 exceeded one in HepG2 cells (Fig. 1F).

AGS affects the IRE1 and ATF6 arm of the UPR but not the PERK pathway.

Next, we measured the expression of ER stress pathway proteins. We found that, while total IRE1α protein was not affected, AGS elevated the expression of phosphorylated IRE1 protein (pIRE1) in both HepG2 and HepG2.2.15 cells (P < 0.05); in addition, pIRE1 protein expression was higher in untreated HBV+ cells vs. HepG2 cells (Fig. 2A). Thus pIRE1 response to AGS was more robust in HepG2.2.15 cells compared with HepG2 cells, and IRE1 phosphorylation was also increased by HBV in the absence of AGS (Fig. 2B).

Fig. 2.

Effects of acetaldehyde-generating system (AGS) and hepatitis B virus (HBV) on unfolded protein response (UPR) protein expression in HepG2 and HepG2.2.15 cells. Cells were either treated or not with AGS for 72 h. Then protein expressions were detected by immunoblotting in cell lysates. A: phospho- and total inositol-requiring enzyme 1α (IRE1α). C: cleaved activation transcription factor (ATF)6α. B and D: quantification of immunoblotting bands. E: phospho-PKR-like ER kinase (PERK), phospho eukaryotic initiation factor 2α (peIF2α), ATF4, and C/EBP homologous protein (CHOP). F: immunoglobin-binding protein (BiP). G: cleaved caspase-3. H: nuclear factor erythroid 2-related factor 2 (Nrf2) and glutathione peroxidase 1(GPx1). Equal (20 µg) amounts of protein were loaded in each lane. β-Actin was used as an internal control. Data are from 3 independent experiments presented as means ± SE. Bars marked with the same letter are not significantly different from each other; bars with different letters are significantly different (P ≤ 0.05).

Furthermore, HBV expression in HepG2.2.15 cells increased the level of ATF6α protein, while the magnitude of upregulation of this protein by AGS exposure was similar in HepG2 and HepG2.2.15 cells (Fig. 2, C and D). There were no changes in the PERK pathway protein markers, pPERK, peIFα, ATF4, and BiP/GRP78 induced by AGS or HBV; however, in HepG2 cells, but not HepG2.2.15 cells, AGS exposure increased CHOP expression (Fig. 2, E and F). The latter corresponded to increased caspase-3 cleavage in AGS-treated HepG2 cells (Fig. 2G), indicating that AGS exposure can induce apoptosis only in HBV nonexpressing cells. To link this event to differential expression of antioxidative proteins, Nrf2 and GPx1, in HBV+ and HBV− liver cells, we measured these protein levels in HepG2 and HepG2.2.15 cells. Surprisingly, neither Nrf2 nor GPx1 proteins were elevated in HBV-transfected cells in the presence or absence of AGS (Fig. 2H).

AGS induces Golgi fragmentation in HBV-transfected HepG2.2.15 cells.

To study whether HBV and AGS affected the Golgi morphology to impair HBV core 18-27 peptide-MHC class I complex trafficking through Golgi, we visualized Golgi morphology by immunofluorescence staining of GM130 (trans-Golgi marker, a peripheral cytoplasmic protein located on the cis side of the Golgi) and TGN46 (trans-Golgi marker, a transmembrane protein, localized in the trans-Golgi network). GM130 and TGN46 staining revealed that AGS and HBV independently cause Golgi fragmentation, which was further increased by AGS in HBV− transfected cells compared with both noninfected HepG2 cells and untreated HepG2.2.15 cells (Fig. 3, A and B). The percentage of cells with fragmented Golgi (well punctuated) was even more in AGS-treated cells. In addition, we confirmed these changes in Golgi morphology by immunofluorescent staining with another Golgi protein, Giantin, a cis-medial-Golgi marker. Based on Giantin staining, there was a fragmentation of Golgi only in AGS-treated-HBV-transfected cells, but we did not see changes in both noninfected and HBV-transfected cells in the absence of AGS. (Fig. 3C). These findings indicate that acetaldehyde mediates ethanol-induced Golgi fragmentation.

Fig. 3.

Acetaldehyde-generating system (AGS) modulates Golgi fragmentation in HepG2.2.15 and HepG2 cells. Cells were either treated or not with AGS for 72 h and processed for staining. A–C: immunofluorescence staining of GM130, trans-Golgi network 46 (TGN46), and Giantin expression. Staining was visualized using a ×63 lens in LSM 710 confocal microscope. Pictures are shown as the representative from 3 independent experiments with similar results.

Role of ER stress on HBV peptide-MHC class I complex presentation and surface MHC class I restoration.

We characterized the input of AGS-HBV-induced ER stress in the presentation of HBV peptide-MHC class I complex on hepatocyte surface. To mimic the effects of acetaldehyde, we used thapsigargin (100 nM for overnight) and tunicamycin (10 µg for overnight) to induce ER stress in HBV-transfected cells followed by measuring the HBV peptide-MHC class I complex (FLPSDFFPSV-HLA-A2) by flow cytometry using a specific antibody to the HBV core peptide-MHC class I complex as we previously published (19). While AGS treatment suppressed this presentation for 50% (Fig. 4A), both thapsigargin and tunicamycin decreased the HBV core 18-27 peptide-MHC class I complex display by ~20% and 23%, respectively, compared with untreated HBV-transfected cells (Fig. 4B). Unlike alterations in proteasome/immunoproteasome activity, these changes were not IFN-γ dependent (data not shown).

Fig. 4.

Acetaldehyde-generating system (AGS) and endoplasmic reticulum (ER) stress inducers suppress expression of hepatitis B virus (HBV) peptide-human leukocyte antigen-A2 (HLA-A2) complex on the surface of HepG2.2.15 cells. A: HepG2.2.15 cells were treated or not with AGS for 72 h, and expression of HBV core peptide (18-27) FLPSDEFPSV-HLA-A2 was measured by flow cytometry with antibody to HBV peptide-HLA-A2 followed by exposure to Alexa Fluor 647 secondary antibody. Mouse IgG2b K isotype control was used. B: expression of HBV core peptide (18-27) FLPSDEFPSV-HLA-A2 measured by flow cytometry in thapsigargin (100 nM)- and tunicamycin (10 µg)-treated HepG2.2.15 cells. Data were processed using BD LSR2 flow cytometer, analyzed by BD FACSDiva Software v6.0 D, and shown as the representative expression from 3 independent experiments with similar results. C: quantification of flow data. Bars marked with the same letter are not significantly different from each other; bars with different letters are significantly different (P ≤ 0.05).

The next step was to elucidate whether ER stress inhibited the delivery of recycled MHC I molecules (HLA-A2) to the hepatocyte surface. To this end, surface MHC class I molecules were removed from the HepG2.2.15 cell surface by a low-acid wash of thapsigargin or tunicamycin-exposed HepG2.2.15 cells, and cells were incubated for 6–8 h in fresh complete medium at 37°C. Surface MHC class I molecule presentation was measured by flow cytometry using anti-human-HLA-A2-APC antibody. We observed that the treatment with thapsigargin and tunicamycin delayed the HLA-A2 restoration up to 23% and 38%, respectively. These results are in concordance with acetaldehyde-induced delay in HLA-A2 molecule expression on HepG2.2.15 cells (Fig. 5, A and B), suggesting that ER stress contributes to the AGS-induced delay in surface MHC class I recycling.

Fig. 5.

Acetaldehyde-generating system (AGS) and endoplasmic reticulum (ER) stress inducers decrease on surface major histocompatibility complex I (MHC class I) turnover. A: HepG2.2.15 cells were treated or not with AGS for 72 h. B: HepG2.2.15 cells were treated or not with thapsigargin (100 nM) and tunicamycin (10 µg). Then surface MHC class I was removed from cell surface by low-acid wash, followed by measurement of restoration of human MHC class I by flow cytometry using anti-human leukocyte antigen-A2 (HLA-A2) antibody. Mouse IgG2b K isotype control APC was used. A representative experiment (1 out of 3 with similar results) was performed on detection of MHC class I measured by flow cytometry. C: quantification of flow data. Data are from 3 independent experiments presented as means ± SE. Bars marked with the same letter are not significantly different from each other; bars with different letters are significantly different (P ≤ 0 0.05).

As a proof of concept, we confirmed our in vitro findings using HepG2.2.15 cells by conducting selected experiments on primary human hepatocytes. Thus HLA-A2+ human hepatocytes were infected with HBV for 5 days and for the last 2 days were exposed to 50 mM EtOH. To mimic ethanol-induced ER stress, we also treated HBV-infected primary human hepatocytes with thapsigargin or tunicamycin overnight, then measured HBV core 18-27 peptide-MHC class I complex expression by flow cytometry. We found that EtOH treatment decreased the complex presentation by ~50%, whereas thapsigargin and tunicamycin reduced it ~30% and 29%, respectively (Fig. 6A). These data also confirm our in vitro findings that ethanol metabolism-induced ER stress plays a partial role in decreased HBV core 18-27 peptide-HLA-A2 presentation on hepatocyte surface. This decrease was not related to ethanol-induced suppression of HBV infection since HBV DNA and HBsAg levels were elevated (2-fold) in HBV-infected EtOH-treated vs. EtOH-untreated hepatocytes (Fig. 6, C and D).

Fig. 6.

Ethanol and endoplasmic reticulum (ER) stress inducers diminish expression of hepatitis B virus (HBV) peptide-human leukocyte antigen-A2 (HLA-A2) complex on the surface of hepatocytes. Here, we used primary human hepatocytes plated on polyelectrolyte multilayer (PEM) gels to prolong expression of alcohol-metabolizing enzymes, cytochrome s4502E1 (CYP2E1), and alcohol dehydrogenase (ADH). Hepatocytes were infected with HBV and exposed to 50 mM EtOH (as described in materials and methods). A: primary human hepatocytes were treated or not with either EtOH for 48 h or ER stress inducers overnight. Then expression of HBV core peptide (18-27) FLPSDEFPSV-HLA-A2 was measured by flow cytometry. Mouse IgG2b K isotype control was used. The results of representative (1 out of 3) experiment are shown. B: quantification of flow data. Tg, thapsigargin; Tn, tunicamycin. C: HBV DNA levels were measured by ddPCR, respectively. D: HBV surface antigen (HBsAg) levels were measured by Sandwich ELISA kit. Data are from 3 independent experiments presented as mean ± SE. Bars marked with the same letter are not significantly different from each other; bars with different letters are significantly different (P ≤ 0.05).

DISCUSSION

Liver disease is considered one of the leading causes of death worldwide (39, 64). Cell-to-cell interactions play either protective or pathogenic roles and are important for antiviral adaptive immune response in HBV infection (6). Cellular stress responses are crucial for hepatocytes because both alcohol metabolism and HBV replication occur in these cells (reviewed in 17). Hepatocytes are the predominant cell type in the liver and are responsible for producing large amounts of secretory proteins, and this can stress the liver (27). Experimental disease models and observational human data demonstrate that ER stress is a common feature observed in acute and chronic liver diseases caused by both HBV and alcohol (27, 30–32, 34, 35, 59, 62). However, the combined role of HBV and alcohol-induced ER stress and Golgi fragmentation in HBV-related liver injury has not been investigated in relation to HBV peptide-MHC class I presentation on hepatocyte surfaces. In this regard, we stress that the primary goal of this study is, not only to merely investigate the effects of acetaldehyde on ER stress and Golgi fragmentation in hepatoma cells, but also to elucidate the contribution of ethanol metabolism-triggered ER stress and Golgi fragmentation to impaired MHC class I-restricted presentation of HBV peptides on HBV+ cell surface (HepG2.2.15 cells). In fact, here, in most cases, we observed more robust increase in ER stress and Golgi fragmentation in HBV-expressing HepG2.2.15 cells exposed to acetaldehyde as a product of AGS compared with AGS-treated HepG2 cells. This indicates that HBV expression in the cells makes them more susceptible to ethanol metabolism-induced ER stress. While this observation is of interest, the primary focus of our study is the involvement of ER stress and Golgi fragmentation in suppressed ability of HBV-expressing liver cells to present HBV-MHC class I complex on hepatocyte surface, which is a prerequisite for elimination of infected hepatocytes by CTLs.

In the current study, we hypothesized that both alcohol and HBV cause ER stress and Golgi fragmentation, thereby impairing the CTL epitope, HBV core (18-27) peptide-MHC class I, presentation on hepatocytes, which finally leads to persistence of HBV infection and chronic liver disease. In this study, we used HepG2.2.15 cells, which are stably transfected with HBV and are HLA-A2+. All these features together make this cell line an excellent model to study HBV core 18-27 peptide-MHC class I complex presentation. HepG2.2.15 cells contain full HBV genome and express HBV proteins, such as HBsAg, HBeAg, and HBcAg, which are relevant to liver inflammation development. As we published earlier, HepG2.2.15 cells do not metabolize alcohol. Hence, to mimic alcohol metabolism, we used the AGS system, which provides a continuous generation of toxic alcohol metabolite acetaldehyde. Acetaldehyde produced by AGS system did not induce either apoptosis or necrosis in HepG2.2.15 cells (19). We measured HBV core peptide 18-27-HLA-A2 complex, a known CTL epitope (3, 5, 40), by using a specific antibody that recognizes HBV peptide-HLA-A2 complex on the surface of HepG2.2.15 cells. Therefore, HepG2.2.15 cells containing HBV core antigen are naturally degraded by proteasome machinery, then loaded to MHC class I (HLA-A2), and delivered to cell membranes by PLC transporters, TAPs to the ER. Thus ER stress contributes to impaired HBV peptide-MHC class I complex presentation on hepatocyte surface, which may potentially decrease the recognition of these epitopes by CTLs. Here, we studied the three arms of the UPR signaling pathways accompanying ER stress.

During ER stress, three ER stress sensors are activated via dissociation of the ER protein, chaperone BiP and/or direct association with unfolded/misfolded proteins. IRE1α dimerizes and autophosphorylates to activate its kinase and endoribonuclease activities. Activated IRE1α induces transcriptionally active XBP1s by an atypical splicing mechanism, and XBP1s translocate to the nucleus to regulate downstream target genes, including ER chaperones and genes involved in ER-associated degradation (ERAD), such as EDEM (38a, 39). Similarly, upon ER stress, ATF6α is activated by regulated intramembrane proteolysis in the Golgi to release the transcriptionally active 50-kDa cytosolic NH₂-terminal domain. Cleaved ATF6α heterodimerizes with spliced XBP1 (sXBP1) to transcriptionally induce several genes encoding ER chaperones and ERAD proteins like EDEM1 (41).

In our study, we observed that phosphorylation of IRE1α was increased in AGS-treated HepG2 and HBV-transfected cells, and this activated IRE1α increased sXBP1s, which finally led to increased ERAD mRNA expression EDEM1. Importantly, another arm of UPR branch, cleaved ATF6α, was increased in AGS-treated HepG2 cells, HBV+, and AGS-treated HBV-transfected cells compared with control noninfected cells. The observed additional effect of acetaldehyde on HBV-transfected cells in inducing ER stress and subsequent UPR activation is due to several reasons: 1) aldehyde(s)-protein adduct formation (45, 53), especially the generation of hybrid-aldehyde adducts (malondialdehyde-acetaldehyde adducts), which have been reported to play a role in triggering ER stress response (30); 2) acetaldehyde adducts affecting the ER Ca²⁺ handling (58), perturbing ER calcium homeostasis, and causing ER stress (43); 3) acetaldehyde-induced decrease in S-adenosylmethionine (SAM)-to-S-adenosylhomocysteine (SAH) ratio also considered as a causative factor for ER stress response (30); and 4) acetaldehyde increases levels of HBsAg involved in activation of the IRE1α and ATF6α pathways to upregulate EDEM family of proteins, which reduces the load of viral surface proteins, possibly contributing to HBV persistence and chronicity (33, 37), thereby causing the progression of liver diseases. In our recent study, we reported that acetaldehyde increases HBsAg levels approximately threefold in HBV-transfected cells (19). In the current study, we showed approximately twofold increase in HBsAg levels in EtOH-treated HBV-infected primary human hepatocytes. In addition, we also reported that ethanol metabolism increased acetaldehyde-induced adduct formation, such as 4-hydroxynonenal (4HNE) adduct formation and decreased SAM-to-SAH ratio (18, 20, 24). All these effects of AGS may contribute to promotion of ER stress involved in HBV pathogenesis.

Interestingly, the third arm of UPR branch, pPERK-peIFα-ATF4-CHOP, was not activated in AGS-treated HepG2.2.15 cells compared with HBV-transfected or noninfected HepG2 cells. However, we observed a significant increase in CHOP mRNA levels, while there were no changes in protein levels. It has been reported that CHOP plays a critical role in ER stress-induced apoptosis (48). In fact, there was no cleaved caspase-3 activation in AGS-exposed HepG2.2.15 cells, indicating that, in agreement with unchanged CHOP protein levels, apoptosis was not detected. The lack of apoptosis was not attributed to enhanced antioxidant protection since the protective antioxidant enzymes, Nrf2 and GPx1, were not upregulated in these groups. Currently, we cannot link the unchanged CHOP protein levels to impaired protein degradation since CHOP is a proteasomal substrate (32a), and lowering of proteasome activity demonstrated before in response to AGS and HBV (19) would rather cause CHOP stabilization. In addition, it has been reported that the increased CHOP expression does not ultimately result in apoptosis although it does sensitize cells to apoptosis (42). In our previous studies, we have also demonstrated that AGS did not induce apoptosis (measured by annexin V staining) in HBV-transfected cells (19). Overall, acetaldehyde activates IRE1α and ATF6α pathways as a protective mechanism and inactivates the pPERK-peIFα-ATF4-CHOP arm, thereby preventing apoptosis of infected cells, which may finally lead to HBV persistence and chronicity.

An association between ER stress and Golgi fragmentation was recently reported (49). We have also shown that the stressed ER often corresponds to the stressed Golgi since proteins from the ER are trafficked to Golgi before they leave the cells (reviewed in Ref. 17). The immunofluorescence localization of three Golgi proteins, GM130, TGN46, and Giantin, revealed a fragmented Golgi in AGS-treated HBV-transfected cells. Our results were supported by the observations of others that alcohol metabolism causes Golgi fragmentation (8, 55). In this case, the fragmented Golgi cannot function properly, and along with a malfunctioning of ER impairs the secretory function in hepatocytes. It has also been reported that ethanol-induced Golgi fragmentation and disorganization of Golgi matrix proteins are among the main contributors of Golgi scattering. Importantly, Golgi fragmentation caused by alcohol may contribute to the alteration in Golgi-to-plasma membrane trafficking (8, 50, 51). HBV-induced fragmentation is not unique for this virus because fragmentation of Golgi by herpes simplex virus and HCV infection has been already reported (2, 28).

Finally, we examined how acetaldehyde-induced ER stress may affect the HBV core 18-27 peptide-MHC class I complex presentation in HepG2.2.15 cells and HBV-infected primary hepatocytes. We found that ER stress inducers thapsigargin and tunicamycin attenuated the HBV peptide-MHC class I complex presentation and surface MHC class I molecule (HLA-A2) turnover, mimicking the effects of acetaldehyde in HBV-transfected cells demonstrated in this study and before (19). Interestingly, the magnitude of ethanol-induced suppression of HBV core 18-27 peptide-MHC class I complex display by ER stress inducers was even more prominent in primary human hepatocytes than in AGS-treated, HBV-transfected cells. Thus acetaldehyde-induced UPR activation may be partially responsible for the suppression of both complex presentation and surface MHC class I turnover. Our data showing that ER stress affects MHC class I-restricted antigen presentation and the expression of surface MHC class I molecules are supported by others (10, 26, 47). However, these other studies were not performed in the context of the effects of acetaldehyde on HBV peptide presentation on hepatocytes.

In conclusion, our study shows that acetaldehyde-induced ER stress and Golgi fragmentation contribute to impaired ER-Golgi-to-plasma membrane trafficking of HBV peptide-MHC class I complex, which may potentially reduce the recognition of HBV-infected cells by CTLs, thereby causing HBV persistence (Fig. 7) and further worsening of hepatitis B pathogenesis.

Fig. 7.

Alcohol suppresses hepatitis B virus (HBV) peptide-major histocompatibility complex I (MHC class I) presentation on HBV-infected hepatocytes via induction of endoplasmic reticulum (ER) stress and Golgi fragmentation. Alcohol metabolites (acetaldehyde) potentiate ER stress and Golgi fragmentation in HBV-expressing hepatocytes. This impairs the delivery of HBV peptide-MHC class I complex and turnover of MHC class I on infected hepatocyte surface. Consequently, it may diminish recognition of infected hepatocytes by cytotoxic T lymphocytes (CTLs), leading to chronic persistence of HBV infection.

GRANTS

This work is supported by K01-Mentored Research Scientist Development Award 1K01AA026864 from National Institute on Alcohol Abuse and Alcoholism (NIAAA).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.G. conceived and designed research; M.G., S.M., E.M., and S.K. performed experiments; M.G., S.M., E.M., and A.P. analyzed data; M.G. interpreted results of experiments; M.G. prepared figures; M.G. drafted manuscript; A.P., K.K.K., L.I.P., C.A.C., and N.A.O. edited and revised manuscript; M.G. and N.A.O. approved final version of manuscript.