miR-21 Targets ASPP2 to Inhibit Apoptosis via CHOP-Mediated Signaling in Helicobacter pylori-Infected Gastric Cancer Cells

Helicobacter pylori (H. pylori) infection affects cell survival pathways, including apoptosis and proliferation in host cells, and disruption of this balance is the key event in the development of H. pylori-induced gastric cancer (HPGC). H. pylori infection induces alterations in microRNAs expression that may be involved in GC development. Bioinformatic analysis showed that microRNA-21 (miR-21) is significantly upregulated in HPGC. Furthermore, quantitative proteomics and in silico prediction were employed to identify potential targets of miR-21. Following functional enrichment and clustered interaction network analyses, five candidates of miR-21 targets, PDCD4, ASPP2, DAXX, PIK3R1, and MAP3K1, were found across three functional clusters in association with cell death and survival, cellular movement, and cellular growth and proliferation. ASPP2 is inhibited by H. pylori-induced miR-21 overexpression. Moreover, ASPP2 levels are inversely correlated with miR-21 levels in HPGC tumor tissues. Thus, ASPP2 was identified as a miR-21 target in HPGC. Here, we observed that H. pylori-induced ASPP2 suppression enhances resistance to apoptosis in GC cells using apoptosis assays. Using protein interaction network and coimmunoprecipitation assay, we identified CHOP as a direct mediator of the ASPP2 proapoptotic activity in H. pylori-infected GC cells. Mechanistically, ASPP2 suppression promotes p300-mediated CHOP degradation, in turn inhibiting CHOP-mediated transcription of Noxa, Bak, and suppression of Bcl-2 to enact antiapoptosis in the GC cells after H. pylori infection. Clinicopathological analysis revealed correlations between decreased ASPP2 expression and higher HPGC risk and poor prognosis. In summary, the discovery of H. pylori-induced antiapoptosis via miR-21-mediated suppression of ASPP2/CHOP-mediated signaling provides a novel perspective for developing HPGC management and treatment.


Introduction
Gastric cancer (GC) is ranked as the ffth most common malignancy worldwide [1]. In 2023, there were 26,500 newly diagnosed cases of GC and 11,130 deaths attributed to GC in the United States [2]. Furthermore, East Asia exhibits the highest incidence and mortality rates of GC [1]. Te development of GC is infuenced by various host factors, including gene polymorphisms, diabetes, and metabolic syndrome, and environmental factors, such as smoking, alcohol consumption, high-salt and nitrite-rich foods, EB virus infection [3], and notably, Helicobacter pylori (H. pylori) infection [4,5]. H. pylori infection accounts for over 60% of the GC cases [6], and it is classifed as a class I carcinogen, representing a signifcant etiologic factor in the development of GC.
H. pylori is a Gram-negative spiral-shaped bacterium that colonizes the human gastric mucosa, estimated to have infected approximately half the world's population. Chronic gastritis develops in virtually all persistent H. pylori-colonized individuals, among which the majority are asymptomatic and only 2-5% develop H. pylori-induced GC (HPGC) [7]. Te most remarkable feature of persistent H. pylori infection is the chronic infammatory response, which increases the risk of gastric malignancy [8,9]. H. pylori infection-induced gastritis is mediated by multiple infammatory cytokines [3]. Te heterogeneity in H. pylori strain-specifc virulence factors contributes to the variety of outcomes after infection [10]. Te diverse outcomes of H. pylori-infected patients have additionally been attributed to diferent host responses [5].
H. pylori-induced infammation disrupts the balance between cell proliferation and apoptosis in host cells, which represents a key event in the HPGC pathogenesis [11]. Apoptosis is regarded as a natural barrier to tumorigenesis and malignancy progression. Cancer cells engage various mechanisms to evade apoptosis that facilitate their survival. Several studies have reported that H. pylori-mediated antiapoptosis facilitates HPGC development [12][13][14]. For instance, H. pylori upregulates BIRC3 eRNA, which is essential for transcription of the apoptosis inhibitor, cIAP2, and confers resistance to apoptosis in the GC cells [12]. H. pylori-mediated DARPP-32 overexpression plays key roles in EGFR stabilization and AKT activation, in turn, preventing apoptosis in the GC cells [13]. Moreover, H. pylori-induced VCP transactivation is involved in proteasome degradation of various cellular factors and promotes the antiapoptotic response in the GC cells [14].
In the present study, miR-21 was identifed as the most signifcantly upregulated miRNA in HPGC with the aid of bioinformatic approaches. miR-21 functions as an oncogenic miRNA involved in tumorigenesis by targeting tumor suppressive genes. For example, miR-21 downregulates PTEN to promote proliferation and antiapoptosis processes in acute myeloid leukemia cells [28]. Additionally, miR-21 targets PTEN, RECK, and Bcl-2 to enhance migration and invasion in lung cancer cells [29]. To elucidate the mechanisms underlying the involvement of miR-21 in HPGC pathogenesis, the stable isotope labeling with amino acids in cell culture (SILAC)-based quantitative proteomic approach was adopted to identify important downstream targets in the AGS cell model. Five potential miR-21-regulated target proteins were identifed with diferent oncogenic functions in the GC cells. Among these, ASPP2 was identifed as a miR-21 target in the HPGC. Subsequent experiments showed that ASPP2 downregulation induced by H. pylori infection promotes antiapoptosis through disruption of the CHOP-modulated pathway in the GC cells. Further clinical fndings validated ASPP2 as a key determinant of HPGC pathogenesis. Te collective results of this study provide novel insights into the mechanisms underlying HPGC malignancy. Moreover, the newly discovered prognostic potential of ASPP2 in HPGC progression may efectively assist in GC management and treatment.

Human Subjects and Tissue Specimens.
In total, 790 patients with GC were enrolled from November 1998 to December 2012 in the National Taiwan University Hospital. Our study was approved by the Institutional Review Board of the Research Ethical Committee of the National Taiwan University Hospital (NTUH IRB number: 201903018RINA). Patient biopsies were reviewed by GI pathologists. H. pylori infection was detected via serologic enzyme-linked immunosorbent assay and histological analysis as described previously [4]. A total of 92 HPGC patients older than 20 years with surgical resection were included in this retrospective case-control study. Endoscopic biopsies of 89 non-GC subjects were additionally included (29 H. pylori-induced intestinal metaplasia (HPIM) cases, 30 H. pylori-induced gastritis (HPGS) cases, and 30 H. pylori-negative (HP (−)) healthy controls with no cancer in the digestive tract). Parafn-embedded sections of HPGC tumors and paired adjacent normal and non-GC samples were prepared for further analyses. Demographic and clinicopathological features of the subjects were recorded for subsequent analyses. All procedures were conducted in accordance with the approved guidelines and regulations.

Bioinformatic Analysis.
To identify potential dysregulated miRNAs in HPGC, we integrated data from two datasets. Te frst dataset, obtained from Gene Expression Omnibus (accession number: GSE62503), showed that 301 out of 718 miRNAs exhibited dysregulation in gastric tissues of HPGC compared to their counterparts in HPGS in a mouse model (p < 0.05). Te second dataset, obtained from Te Cancer Genome Atlas (data ID: TCGA.STAD.sampleMap/ miRNA_HiSeq_gene), showed that 843 out of 2179 miRNAs displayed dysregulation in human GC tumors compared to adjacent normal control tissues (p < 0.05). Distribution of signifcance versus fold changes of a total of 80 common dysregulated miRNAs from the two datasets was further analyzed using the volcano plot. Among these miRNAs, miR-21 was the most signifcantly upregulated in HPGC (log 2 (fold change) >2.0 and -log (p value) >1.3). For the prediction of miR-21 targets, we utilized the miRSystem (http://mirsystem. cgm.ntu.edu.tw), which integrates seven diferent programs (DIANA-microT, miRanda, mirBridge, PicTa, PITA, rna22, and TargetScan), and conducted a BLAST search against two experimentally validated databases, TarBase and miRecords [30]. Biological function enrichment and clustered interaction network analyses of miR-21 candidate targets were performed via ingenuity pathway analysis (IPA, Ingenuity Systems, Redwood City, CA, USA; http://www.ingenuity.com). Te ranking and signifcance of biological functions was based on the p values.

In Situ Hybridization and Immunohistochemistry (IHC).
Histological analysis of miR-21 was performed via in situ hybridization with a miRCURY LNA ™ microRNA ISH Optimization Kit (Exiqon, Vedbaek, Denmark) as per the manufacturer's instructions. Briefy, sample slides were deparafnized and stained with miR-21 probes. Cell nuclei were counterstained with Nuclear Fast Red ™ (Termo Fisher Scientifc, Waltham, MA, USA). Expression of miR-21 was ranked based on H score ranging from 0 to 3, formulated as ((percentage of 0) × 0) + ((percentage of 1+) × 1) + ((percentage of 2+) × 2) + ((percentage of 3+) × 3) (0, nil; 1+, low or weak; 2+, moderate; and 3+, high or strong) [31]. For histological determination of ASPP2 and cleaved caspase 3 expression, IHC was performed with a Novocastra ™ Manual Detection System kit (Leica Biosystems, Newcastle, UK) as described previously [31]. High expression was defned as positive staining in ≥50% of the cells. Culture medium was supplemented with 10% fetal bovine serum (HyClone), penicillin (100 U/L), and streptomycin (10 mg/L). Te H. pylori strain (ATCC 43504) was obtained from the American Type Culture Collection. H. pylori was inoculated on Columbia agar containing 5% sheep blood (BD Biosciences, San Jose, CA, USA) and grown at 37°C in a microaerophilic chamber under 10% CO 2 , 5% O 2 , and 85% N 2 . Te study on H. pylori was approved by the Biosafety Committee of National Taiwan University. All procedures were conducted in compliance with the approved guidelines and regulations. AGS and MKN45 cells were infected with H. pylori at multiplicity of infection (MOI) of 30 and 90, respectively, for the indicated times after starvation in serum-free medium for 1.5 h.

Vector Construction, Transfection, and Viral
Transduction. For adenovirus-based miR-21 overexpression, adenovirus particles with pAdTrack-CMV Mock vectors (AdMock) or vectors coding for primary miR-21 sequence (AdmiR-21) were prepared using the HEK 293 cells as described previously [14]. For adenoviral transduction, AGS cells were infected with adenovirus particles at MOI of 2 for 36 h to establish AdMock or AdmiR-21 AGS cells. MiR-21 overexpression efciency was assessed with quantitative polymerase chain reaction (qPCR). For small nucleotide transfection or siRNAmediated gene knockdown, 500 pmol control siRNA (siCtrl), miR-21, or anti-miR-21 oligonucleotide and siRNA against CHOP (siCHOP) or p300 (sip300, Gene-Pharma, Shanghai, China) were transfected into AGS or MKN45 cells for 48 h with Lipofectamine (Termo Fisher Scientifc). For lentivirus-mediated ASPP2 knockdown, lentiviruses expressing control shRNA (shCtrl) or shRNA against ASPP2 (shASPP2) were prepared using HEK293 cells according to the supplier's instructions (RNAiCore, Academia Sinica, Taipei, Taiwan). For lentiviral transduction, medium containing shCtrl or shASPP2 virus was incubated with AGS and MKN45 cells for 72 h and the efciency of ASPP2 knockdown was confrmed via qPCR and immunoblotting. For ASPP2 overexpression, the ASPP2 gene was cloned into p3xFLAG-CMV-14 vector using a cDNA library generated from AGS cells. Mock control or ASPP2-overexpressing vectors (5 μg) were transfected into AGS and MKN45 cells with Lipofectamine for 48 h and the efciency of ASPP2 overexpression was confrmed via qPCR and immunoblotting. Te small nucleotide, shRNA, and siRNA target sequences are listed in Supplementary Table 1.

qPCR, Luciferase Reporter Assay, and Chromatin Immunoprecipitation (ChIP) Assays.
In the qPCR assay, the total RNA in cell lysates or plasma specimens was isolated with TRIzol reagent (Termo Fisher Scientifc) and reverse-transcribed using a TaqMan MicroRNA Reverse Transcription Kit (Termo Fisher Scientifc). qPCR was conducted using a SensiFastTM SYBR Hi-ROX kit (BIOLINE, London, UK) and the transcripts were detected with the StepOne Plus system (Applied Biosystems, Cheshire, UK). Gene expression levels were calculated via the comparative 2 −△△CT method. In the luciferase reporter assay, AGS cells were transfected with control or miR-21 expression vector and pmirGLO vectors expressing wildtype or mutant ASPP2 3′UTR as indicated. Te binding interactions between miR-21 and ASPP2 3′ UTR were evaluated at 24 h post-transfection using the Dual-Luciferase reporter gene kit (Promega). In the ChIP assay, AGS cell lysates were harvested and DNA-protein complexes were immunoprecipitated with control IgG, p65, or CHOP antibodies bound on Protein G or A beads (G-Biosciences, St Louis, MO, USA) at 4°C overnight. Immunoprecipitated DNA was retrieved and analyzed for promoters of miR-21, Noxa, Bak, or Bcl-2 using qPCR. Te conserved motif PuPuPuTGCAAT(A/C)CCC for CHOP binding to Noxa and Bak promoter [32] and C/EBP binding site for CHOP binding to suppress Bcl-2 transcription [33,34] were searched using ensembl (http:// www.ensembl.org/index.html). Te primer sequences are listed in Supplementary Table 1.

Immunoblotting and Coimmunoprecipitation (Co-IP)
Assays. In the immunoblot assay, cells were lysed and separated via sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to PVDF membranes (Merck Millipore, Burlington, MA, USA). Immunoblotting was performed with the indicated antibodies (listed in Supplementary Table 1) as described previously [35]. In the Co-IP assay, AGS cell lysates were harvested and incubated with antibodies against ASPP2, CHOP, or p300 bound on Protein G or A beads (G-Biosciences) at 4°C overnight. Immunoprecipitated protein complex was eluted and analyzed via SDS-PAGE and immunoblotting.

Terminal Deoxynucleotidyl Transferase dUTP Nick End Labeling (TUNEL) and Annexin V/Propidium Iodide (PI)
Staining Apoptosis Assays. In the TUNEL assay, apoptotic cells were determined by detecting DNA double strand breaks using the APO-BRDU ™ kit (Enzo Life Sciences, Farmingdale, NY, USA) according to the manufacturer's instructions. For the Annexin V/PI assay, cells were collected and stained with FITC-conjugated antibodies against Annexin V (BioLegend, San Diego, CA, USA) and PI (BioLegend). Te percentages of apoptotic cells in both TUNEL and Annexin V/PI staining assays were measured with a Cytomics FC500 fow cytometer (Beckman Coulter, Brea, CA, USA) as described previously [14].

Statistical Analysis.
Student's t-test was used for the general evaluation of the experimental signifcance and quantifed values expressed as mean ± SD. Te diferences in expression and correlations between ASPP2 and miR-21 or cleaved caspase 3 in HPGC tumors and control counterparts were examined with the Mann-Whitney test and Spearman correlation analysis. ASPP2 levels among HPGC, HPIM, HPGS, and healthy control groups were evaluated using the Mann-Whitney test. Univariate and multivariate logistic regression analyses were utilized to compute the odds ratios (ORs), and 95% confdence interval (CI) of demographic features, and ASPP2 levels for HPGC. Univariate and multivariate Cox's proportional hazard regression models were used to compute the hazard ratios (HRs), and 95% CI of demographic and clinicopathological features, and ASPP2 levels in HPGC patients. OS curves based on ASPP2 expression were constructed using the Kaplan-Meier method and compared with the log-rank test. All statistical evaluations were two-tailed, with p values <0.05 considered statistically signifcant. All statistical analyses were performed using IBM SPSS Statistics (version 22.0; Chicago, IL, USA).

NF-κB-Mediated miR-21 Upregulation Is Involved in HPGC. H. pylori infection induces alterations in the ex-
pression of miRNAs that may be involved in the development of GC [26]. To clarify the involvement of miRNA dysregulation in HPGC pathogenesis, we aimed to identify potential miRNAs critical for HPGC using bioinformatic approaches. Using the GEO dataset, 301 signifcantly dysregulated miRNAs were identifed in the gastric tissues of HPGC mice relative to their HPGS counterparts (p < 0.05). Eighty commonly dysregulated miRNAs were further extracted upon comparison with 843 signifcantly dysregulated miRNAs in human GC tumors (relative to paired adjacent normal controls, p < 0.05) from TCGA dataset (Figure 1(a)). Te eighty miRNAs were further examined using a volcano scatter plot. Consequently, miR-21 was identifed as the most signifcantly upregulated miRNA potentially associated with H. pylori infection in human GC (log 2 (fold change) >2.0 and −log (p value) >1.3; Figure 1(b)). Consistently, our results showed marked upregulation of miR-21 in HPGC tumor tissues (n � 35) compared to normal controls (n � 35; Figure 1(c)).
In vitro, H. pylori infection promoted miR-21 expression in AGS cells (Supplementary Figure 1A). Te most remarkable feature of H. pylori infection is infammatory response, which is mediated by the activation of NF-κB [8,9]. NF-κB has been identifed as a major transcriptional activator of miR-21 [36,37]. Te phosphorylation of the NF-κB subunit p65 at Ser 536 facilitates the translocation of NF-κB to the nucleus and initiates transcription, which is a crucial event for NF-κB activation [38]. As shown in Supplementary Figure 1B, phosphorylation of p65 at Ser 536 in AGS cells increased after H. pylori infection, whereas treatment with MG132, a proteasome inhibitor of NF-κB, led to signifcant suppression of p65 phosphorylation and miR-21 expression in H. pylori-infected AGS cells compared to nontreated controls (Supplementary Figures 1C and D). Moreover, the binding of p65 to miR-21 promoter was signifcantly increased in cells infected with H. pylori compared to noninfected controls (Supplementary Figure 1E). Collectively, these results suggest that H. pyloriinduced miR-21 upregulation is mediated by NF-κB activation in GC cells.

Identifcation of ASPP2 as a miR-21 Target in HPGC.
In view of the functional importance of miR-21 in GC, we further focused on the potential mechanisms underlying miR-21-mediated GC progression. We assessed the functional efects of miR-21 upregulation on GC cells with the aid of miR-21-overexpressing AGS cells. Upregulation of miR-21 clearly promoted proliferation, migration, and invasion and inhibited apoptosis of AGS cells ( Supplementary  Figures 2A-F). To further identify potential miR-21 targets that could contribute to the cancer phenotype, SILAC-based  Upper, representative images were captured at 200x magnifcation. Scale bar, 50 μm. Lower, relative quantitation of miR-21 levels by H score ranking. Te signifcance of diferences in H scores between two groups was compared using the nonparametric Mann-Whitney U test. * * p < 0.01. quantitative proteomic analysis and bioinformatic integration were performed. A schematic description of the approach is illustrated in Figure 2(a). Diferential proteomes were compared between AGS cells infected with AdMock and AdmiR-21. A total of 47 downregulated candidate targets of miR-21 (AdmiR-21/AdMock ratio <0.7) were further identifed via in silico validation by miRSystem.
To clarify the functional characteristics of the 47 candidate miR-21 targets, these proteins were subjected to the biological function analysis using IPA. We found eight main functional groups that strongly enriched (p value <0.05) in miR-21regulated proteins (Figure 2(b)). Interestingly, we observed top three enriched functions, cell death and survival, cellular movement, and cellular growth and proliferation, corresponding to the results of miR-21 functional assays (Supplementary Figures 2A-2F). For identifying potential targets under miR-21 regulation during oncogenesis, the interacting linkage and functional clustering of molecules among the top three enriched biological functions were performed via IPA interaction network analysis. As shown in Figure 2 We further investigated the regulation of miR-21 on fve candidate targets in GC cells using miR-21 overexpression and suppression assays. As shown in Supplementary Figure 3, we observed the most signifcant post-transcriptional regulation between miR-21 and ASPP2 in AGS cells, with PDCD4 serving as a positive control [39]. Terefore, ASPP2, playing key roles in modulation of cell growth and apoptosis and reported to exert tumor suppressor efects [40], was selected for further analysis. A quantitative mass spectrum of the representative peptide of ASPP2 was shown in Supplementary  Figure 4. Moreover, protein levels of ASPP2 were consistently modulated by miR-21 in AGS and MKN45 cells (Figure 3(a)), further confrming the direct association between miR-21 and ASPP2. In the TargetScan analysis, miR-21 was predicted to bind the 3′UTR of ASPP2 (Figure 3(b), upper). Te binding of miR-21 to the ASPP2 3′UTR was subsequently validated using the luciferase reporter assay. Overexpression of miR-21 signifcantly inhibited the luciferase reporter activity of the wild-type ASPP2 3′UTR, but not the mutant ASPP2 3′UTR. Tis provides evidence that ASPP2 is a direct target of miR-21 in the GC cells (Figure 3(b), lower).
We additionally examined ASPP2 expression under H. pylori-induced miR-21 overexpression in both GC cells. ASPP2 was downregulated in AGS and MKN45 cells after H. pylori infection for 6 h, while it was recovered under miR-21 inhibition by the treatment of anti-miR-21 (Figure 3(c)). Te results suggest that H. pylori-induced ASPP2 suppression is mediated by miR-21. To validate the clinical signifcance of ASPP2 in HPGC, we analyzed its expression in tumor and surrounding normal tissues using IHC. As shown in Figure 3(d), left, ASPP2 expression was signifcantly lower in HPGC tumor tissues (n = 35) compared to adjacent normal counterparts (n = 35). Moreover, the evaluation of ASPP2 and miR-21 levels indicated a negative correlation in HPGC (Figure 3(d), right). Our data collectively indicate that ASPP2 is a miR-21 target in HPGC.

H. pylori-Induced ASPP2 Suppression Promotes Antiapoptosis of HPGC.
In view of the implication of miR-21 upregulation in association with antiapoptosis from functional assays and proteomic analyses, we next clarify the antiapoptotic role of ASPP2 suppression in GC cells. shASPP2 and shCtrl lentivirus were generated for infection of AGS and MKN45 cells. ASPP2 mRNA and protein levels were efciently suppressed in both shASPP2 GC cells compared to shCtrl cells (Supplementary Figure 5A). Results from TUNEL and Annexin V/PI assays indicated that knockdown of ASPP2 attenuated 5-fuorouracil (5-FU)induced apoptosis in both the AGS and MKN45 cells, providing evidence that ASPP2 plays a role in apoptosis in GC cells ( Supplementary Figures 5B and C).
To study the role of H. pylori-induced ASPP2 suppression in the apoptosis of GC cells, we subsequently examined the efects of H. pylori infection on apoptosis by time-dependent experiments in AGS and MKN45 cells. As shown in Supplementary Figures 6A and B, after H. pylori infection in AGS and MKN45 cells, induction of apoptosis in the GC cells was observed from 8 to 24 h, whereas antiapoptosis was stimulated rapidly at 32 h post-infection. Next, we observed the efect of H. pylori-induced ASPP2 suppression on the apoptosis in GC cells. As shown in Figure 4(a), H. pylori induced ASPP2 suppression in AGS and MKN45 cells at 32 h post-infection, which was rescued following ectopic overexpression of ASPP2. Data from the TUNEL assay showed that H. pylori-induced ASPP2 downregulation inhibited apoptosis in mock GC cells, which was recovered under conditions of ASPP2 overexpression (Figure 4(b)). Consistently, Annexin V/PI assays confrmed induction of antiapoptosis through ASPP2 suppression in H. pylori-infected GC cells (Figure 4(c)). In clinical patient samples, apoptotic maker cleaved caspase 3 was signifcantly downregulated in HPGC tumor tissues (n = 35) compared to adjacent normal counterparts (n = 35, Figure 4(d), left). Correlation analysis additionally indicated a positive correlation of ASPP2 attenuation with cleaved caspase 3 downregulation in HPGC tumors (Figure 4(d), right). Tese fndings support a key antiapoptotic role of H. pylori-induced ASPP2 suppression in the progression of HPGC.

ASPP2 Interaction with CHOP Regulates the Expression of CHOP-Modulated Apoptotic Factors in H. pylori-Infected GC
Cells. ASPP2 is involved in apoptosis through interactions with a variety of apoptosis-associated proteins [40]. To elucidate the mechanisms underlying ASPP2 suppression-mediated antiapoptosis in HPGC, an IPA mapping tool was utilized to identify key downstream molecules of ASPP2 by constructing a protein interaction network incorporating potential ASPP2 interactants and apoptosis-associated proteins. As shown in Supplementary Figure 7A, CHOP was identifed as a highly connected molecule within the network. Te signifcance of CHOP as a potential downstream mediator of ASPP2-modulated apoptosis was further studied. In      Journal of Oncology of CHOP in H. pylori-infected AGS cells, suggesting that ASPP2 is involved in regulation of CHOP degradation. Earlier reports suggest that p300 E4 ligase is a crucial factor in degradation of CHOP [41]. Knockdown of p300 suppressed CHOP ubiquitination and enhanced its protein expression in AGS cells after H. pylori infection for 8 h, suggesting p300 modulates the ubiquitination and degradation of CHOP (Supplementary Figure 7c). We further investigated the role of ASPP2 in p300-mediated CHOP degradation. As shown in Figure 5(b), ASPP2 knockdown under treatment with MG132 promoted interactions of CHOP with p300, leading to increase in CHOP ubiquitination in AGS cells after H. pylori infection for 8 h, indicating the importance of ASPP2 in the regulation of CHOP/p300 interaction as well as CHOP activity under H. pylori infection.
CHOP was proposed to act as a transcription factor potentially regulating the expression of both proapoptotic (Noxa, Bak, Bim, Bax, PUMA, and DR5) and antiapoptotic (Bcl-2, Bcl-xL, and Mcl-1) genes [42,43]. After screening several CHOP-regulated transcription factors, knockdown of CHOP signifcantly reduced Noxa and Bak, while enhanced Bcl-2 at the transcription level in AGS cells after H. pylori infection for 8 h (Supplementary Figure 7D). We further examined the role of ASPP2 in the regulation of CHOP-mediated transcription of apoptotic factors using ChIP assay. As shown in Figure 5(c), ASPP2 knockdown led to reduced binding of Noxa and Bak and enhanced binding of Bcl-2 promoter regions by CHOP in AGS cells after H. pylori infection for 8h. Additionally, ASPP2 knockdown decreased CHOP, Noxa, and Bak and increased Bcl-2 protein levels in H. pylori-infected AGS cells, where the expressions of cleaved caspase 9, caspase 3, and PARP were also attenuated ( Figure 5(d)). Taken together, interaction with ASPP2 prevents p300-mediated degradation of CHOP, in turn promoting transcription of Noxa, Bak, and suppression of Bcl-2 expression in GC cells after H. pylori infection.

H. pylori-Induced ASPP2 Downregulation Promotes Antiapoptosis through Attenuation of CHOP-Modulated
Apoptotic Pathway in GC Cells. Next, we examined the efect of H. pylori infection on the molecular changes of ASPP2/ CHOP-mediated apoptotic pathway in a time-dependent manner. As shown in Figure 6

Associations of ASPP2 and Clinicopathological Characteristics with Overall Survival (OS) in HPGC.
To establish the ASPP2 level in gastric tissues, we examined histological patterns at diferent stages of Correa's cascade, including HP (−) healthy control (n = 30), HPGS (n = 30), HPIM (n = 29), and HPGC (n = 57) samples. As presented in Figure 7(a), expression of ASPP2 was markedly lower in gastric tissues from HPGC and HPIM stages relative to HPGS. No differences were observed between HPGS and HP (−) healthy control groups. Te demographic and clinicopathological features and ASPP2 levels of all subjects analyzed in this study are shown in Supplementary Table 4. We consistently observed signifcantly lower ASPP2 expression in HPGC and HPIM gastric tissues compared to HPGS counterpart tissues. Moreover, ORs of demographic features and ASPP2 levels for HPGC were evaluated via logistic regression analysis. As shown in Table 1, univariate analysis revealed correlations of cigarette smoking, alcohol consumption, and ASPP2 levels with risk of HPGC. Multivariate logistic regression analysis confrmed that reduced expression of ASPP2 is associated with the higher risk of HPGC. We further assessed the HRs of demographic and clinicopathological features and ASPP2 levels in HPGC patients via Cox's proportional hazards analysis. As shown in Table 2, univariate analysis showed that tumor stage, invasive depth, nodal metastasis, metastasis, and ASPP2 levels were correlated with OS of HPGC. After adjustment for age, gender, tumor stage, invasive depth, nodal metastasis, and metastasis in multivariate analysis, metastasis and ASPP2 levels were correlated with OS of HPGC patients. Te OS of HPGC patients in relation to ASPP2 expression was further examined using Kaplan-Meier survival curves. As shown in Figure 7(b), low expression of ASPP2 was associated with poor OS in HPGC patients. Based on the collective results, we proposed that reduced ASSP2 expression could serve as a novel independent prognostic predictor of HPGC. All the data were presented as mean ± SD of 3 replicates. * * p < 0.01 and * * * p < 0.01.

Discussion
Limited knowledge exists regarding aberrantly expressed miRNAs and their targeted genes in GC [44]. MiR-21 is upregulated in many cancer types and promotes cancer progression by targeting several tumor suppressors [45]. Our experiments corroborated previous fndings [17], showing that miR-21 promotes cell proliferation and metastasis and inhibits apoptosis to a signifcant extent in AGS cells. Moreover, miR-21 expression was remarkably increased following H. pylori infection in both the GC cells in vitro and gastric tumor tissues in vivo, analogous to earlier results [17,46]. Te activation of NF-κB, increased IL-6 secretion, and enhanced expression of AP-1 and STAT3 by H. pylori are proposed mechanism for inducing miR-21 transcription [26]. NF-κB has also been identifed as an activating transcription factor for miR-21 in the GC cells [37,47]. Our experiments demonstrated that NF-κB regulates miR-21 expression by binding to its promoter region in H. pylori-infected cells, supporting a key role of miR-21 in HPGC development. MiR-21 regulates tumor suppressors to promote cancer progression [15]. Tis study used proteomic approaches to identify 1896 potential downregulated proteins in miR-21overexpressing AGS cells. Te TargetScan program in miRSystem predicted 47 potential miR-21 targets [30].
Functional enrichment analysis revealed that these downregulated proteins were involved in cell death, survival, and cellular movement and growth. Previous studies demonstrated that miR-21 targets LZTFL1 in breast cancer cells, RASA1 in colon cancer cells, and KLF5 in hepatocellular carcinoma cells [48][49][50]. Tis suggests that miR-21 and its target genes could be the molecular targets of HPGC. To gain insight into the miR-21 target proteins, interaction network analysis was applied to identify protein clusters with similar functions [51]. As a result, 32 out of 47 miR-21 downregulated proteins related to enriched functions form the three clusters. Among these proteins, we identifed fve candidates of miR-21 targets, including PDCD4, ASPP2, DAXX, PIK3R1, and MAP3K1.
MiR-21 promotes growth and invasion by inhibiting PDCD4 in the GC cells [39]. It also induces cell migration and invasion in breast cancer cells by targeting PIK3R1 [52]. However, our study did not observe signifcant efects of miR-21 on PIK3R1 and MAP3K1 gene expression. In silico analyses suggest that DAXX contains miR-21 binding sites [53] and it promotes cell apoptosis and inhibits cell invasion in the GC cells [54]. Moreover, miR-21 suppresses ASPP2 to modulate cell apoptosis, proliferation, and metastasis in glioblastoma, hepatoblastoma, and lung cancer cells [53,55,56]. Among these candidates, ASPP2 was signifcantly upregulated following miR-21 inhibition in our (c) Cell lysates of AGS and MKN45 cells incubated with H. pylori at 30 and 90 MOI for the indicated time followed by 10 μM MG132 for 3 h were immunoprecipitated with antibodies against p300 (upper) or CHOP (middle) immobilized on protein A beads for 16 hours. In panels (b) and (c), the Co-IP complex and 20% input (lower) samples were subjected to immunoblotting. All the data were presented as mean ± SD of 3 replicates. * p < 0.05, * * p < 0.01, and * * * p < 0.01. experiments. Luciferase reporter assay further validated ASPP2 as a direct miR-21 target. In view of its identifcation as a miR-21 target and potential tumor suppressor activity, the functional importance of ASPP2 was further explored.
ASPP2 suppression is linked to cell growth and inhibition of apoptosis in hepatocellular carcinoma cells and leukemia cells [40,57]. Here, ASPP2 was suppressed by miR-21 in H. pylori-infected GC cells. Te H. pylori-derived cytotoxin-associated gene A disrupts the apoptotic function of ASPP2 in AGS cells [58,59]. Our results showed that H. pylori-induced ASPP2 suppression enhanced antiapoptosis in the GC cells, which provides valuable clues for further investigating the signaling pathways involved in HPGC carcinogenesis. ASPP2 promotes apoptosis by binding to p53 and enhancing the transcription of proapoptotic genes, such as PUMA, PIG3, and Bax [60]. In addition to regulation of p53-dependent apoptosis, ASPP2 is implicated in apoptosis independent of p53 [40]. In most human cancer types, ASPP2 acts as a tumor suppressor and is often downregulated. For instance, ASPP2 knockdown suppresses p85α expression, leading to increased AKT phosphorylation and antiapoptosis in triple-negative breast cancer cells [61]. Downregulation of ASPP2 also confers resistance to apoptosis induced by 5-FU-and etoposide by promoting BECN1-dependent autophagy in hepatocellular carcinoma cells [62]. In our study, reduced ASPP2 expression promoted resistance to apoptosis in the H. pyloriinfected GC cells via disruption of the CHOP-mediated apoptotic pathway. Based on protein interaction network and Co-IP assays, we identifed CHOP as a direct mediator Lower, relative quantitation of ASPP2 levels by H score ranking. Te signifcance of diferences in H score between two groups was compared using the nonparametric Mann-Whitney U test.
(b) Kaplan-Meier OS curves for 92 HPGC patients stratifed by ASPP2 expression were constructed using the Kaplan-Meier method and compared with the log-rank test. * * * p < 0.001.   [63]. Here, ASPP2 silencing markedly reduced CHOP expression in the H. pylori-infected GC cells, suggesting ASPP2 involvement in CHOP degradation regulation. Te critical fnding indicated that loss of the ASPP2 and CHOP interaction promotes p300-mediated degradation of CHOP in the H. pylori-infected GC cells. CHOP is a transcriptional repressor via competing the promoter binding with C/EBP to inhibit C/EBP-mediated transcription of Bcl-2 and Bcl-xL [33,34]. CHOP upregulates Bim expression, activating Bak and Bax to induce mitochondrial apoptosis [42]. In addition, CHOP and ATF5 could activate Noxa gene expression to induce apoptosis during stress [33]. Noxa, a BH3-only member of the Bcl-2 family, is required for apoptosis activation in response to ER stress. Our results demonstrate that CHOP mediates ASPP2-induced apoptosis by upregulating Noxa and Bak and suppressing Bcl-2 expression in the GC cells. Moreover, our study found that H. pylori initially promotes apoptosis in the GC cells, but later induces antiapoptosis through miR-21-mediated ASPP2 suppression, which disrupts CHOP-mediated apoptotic signaling by facilitating p300 dependent CHOP degradation in the GC cells. Te overall data suggest that the H. pylori-infected host cells initiate alternative pathways to prevent apoptosis and promote GC development. ASPP2 expression in H. pylori-positive cancer or precancerous tissues is reported to be lower than that in H. pylori-negative tissues [64]. In this study, ASPP2 was frequently downregulated in HPGC and HPIM compared with HPGS tissues, suggesting that H. pylori infection suppresses ASPP2 expression, leading to inhibition of apoptosis and promotion of cell proliferation. Our data support attenuation of ASPP2 expression as a mechanism to promote resistance to apoptosis in HPGC. Moreover, univariate OR analysis disclosed that lower ASPP2 expression is a risk factor of HPGC. Similarly, cigarette smoking and alcohol consumption were associated with the risk of HPGC, analogous to previous fndings [4]. Although the incidence rate of GC varies by age and gender [65], ORs of these factors showed no signifcant associations in our study. Similarly, in multivariate OR analysis, lower ASPP2 level was signifcantly correlated with HPGC risk.
While ASPP2 shows prognostic value in human cancers [57,66], its role in GC remains to be established. Data from our univariate HR analysis showed that lower ASPP2 expression is an indicator of poor prognosis in HPGC patients. Moreover, tumor stage, invasive depth, nodal metastasis, and metastasis had a signifcant infuence on the OS of HPGC patients, which is consistent with previous results [67][68][69][70]. Although histological grade, tumor location, and Lauren classifcation are reported to be associated with GC prognosis [69,71], these characteristics showed no signifcant associations in our study. Furthermore, in multivariate HR analysis, only metastasis and ASPP2 expression were correlated with the OS of HPGC patients. Previously, the signifcance of several clinicopathological features in GC prognosis was compared in diferent studies, with divergent fndings [67][68][69][70][71]. Tis may be attributable to heterogeneous etiologies and mechanisms underlying GC development [4]. Lower ASPP2 expression was correlated with poor survival rate in an earlier study on esophageal squamous cell carcinoma patients [66]. Likewise, ASPP2 suppression was associated with high-risk disease and therapy failure in acute leukemia patients [57]. Here, we identifed ASPP2 as a potential prognostic biomarker in the HPGC population.
Te efectiveness of H. pylori eradication in HPGC prevention has achieved success. However, the efcacy is frequently compromised in those patients with precancerous lesions [72]. It indicates that other than H. pylori virulence factors, the aberrant host response underlying H. pylori infection is also important for developing new therapeutic approach of HPGC. In this study, we demonstrated dysregulation of ASPP2/CHOP pathway in response to H. pylori-induced miR-21 overexpression, which confers apoptosis resistance in host cells. Low expression of ASPP2 increases the occurrence of GC in the H. pylori-infected patients. Restoration of ASPP2 expression or activity may ofer a breakthrough in developing therapeutic intervention against HPGC. Moreover, clinicopathological analyses suggested a predictive potential of lower ASPP2 in HPGC poor prognosis. In conclusion, our study provides both therapeutic and prognostic insights into developing precision medicine against HPGC.

Conclusions
H. pylori-induced miR-21-mediated ASPP2 suppression confers resistance to apoptosis through inhibition of ASPP2/ CHOP-mediated transcription of Noxa and Bak and suppression of Bcl-2 in gastric cancer cells (Figure 8). In addition, lower ASPP2 expression is associated with higher risk and poor prognosis in the H. pylori-induced gastric cancer patients.

Data Availability
Te data presented in this study are available on request from the corresponding author.

Ethical Approval
Our study was approved by the Institutional Review Board of the Research Ethical Committee of National Taiwan University Hospital (NTUH IRB number: 201903018RINA). All procedures were conducted in accordance with the approved guidelines and regulations.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Supplementary Materials
Supplementary