Early and late genome-wide gastric epithelial transcriptome response during infection with the human carcinogen Helicobacterpylori

Infection of the stomach by Helicobacter pylori is a major risk factor for the development of gastric cancer. Colonization of the gastric epithelium leads to the activation of multiple disease-related signaling pathways. Serine protease HtrA represents an important secreted virulence factor that mediates cleavage of cellular junctions. However, its potential role in nuclear responses is unknown. Here, we performed a genome-wide RNA-seq analysis of polarized gastric epithelial cells infected by wild-type (wt) and ΔhtrA mutant bacteria. Fluorescence microscopy showed that H. pylori wt, but not ΔhtrA bacteria, preferably localized at cellular junctions. Our results pinpointed early (2 h) and late (6 h) transcriptional responses, with most differentially expressed genes at 6 h post infection. The transcriptomes revealed HtrA-dependent targeting of genes associated with inflammation and apoptosis (e.g. IL8, ZFP36, TNF). Accordingly, infection with the ΔhtrA mutant induced increased apoptosis rates in host cells, which was associated with reduced H. pylori CagA expression. In contrast, transcription of various carcinogenesis-associated genes (e.g. DKK1, DOCK8) was affected by H. pylori independent of HtrA. These findings suggest that H. pylori disturbs previously unknown molecular pathways in an HtrA-dependent and HtrA-independent manner, and provide valuable new insights of this significant pathogen in humans and thus potential targets for better controlling the risk of malignant transformation.

H. pylori virulence factor HtrA enables the bacteria to localize at cellular junctions. Transcriptional changes are specific between early and late infection time points. HtrA impacts the transcription of host genes associated with inflammation and apoptosis. H. pylori HtrA promotes CagA expression upon infection. HtrA deficiency in H. pylori leads to higher apoptosis rates of host cells.

Keywords: Helicobacter
Cancer RNA-seq HtrA Apoptosis Inflammation A B S T R A C T Infection of the stomach by Helicobacter pylori is a major risk factor for the development of gastric cancer. Colonization of the gastric epithelium leads to the activation of multiple disease-related signaling pathways. Serine protease HtrA represents an important secreted virulence factor that mediates cleavage of cellular junctions. However, its potential role in nuclear responses is unknown. Here, we performed a genome-wide RNA-seq analysis of polarized gastric epithelial cells infected by wild-type (wt) and ΔhtrA mutant bacteria. Fluorescence microscopy showed that H. pylori wt, but not ΔhtrA bacteria, preferably localized at cellular junctions. Our results pinpointed early (2 h) and late (6 h) transcriptional responses, with most differentially expressed genes at 6 h post infection. The transcriptomes revealed HtrA-dependent targeting of genes associated with inflammation and apoptosis (e.g. IL8, ZFP36, TNF). Accordingly, infection with the ΔhtrA mutant induced increased apoptosis rates in host cells, which was associated with reduced H. pylori CagA expression. In contrast, transcription of various carcinogenesis-associated genes (e.g. DKK1, DOCK8) was affected by H. pylori independent of HtrA. These findings suggest that H. pylori disturbs previously unknown molecular pathways in an HtrA-dependent and HtrAindependent manner, and provide valuable new insights of this significant pathogen in humans and thus potential targets for better controlling the risk of malignant transformation.

Introduction
About half of the world's population is carrying the gastric bacterium Helicobacter pylori that represents a high risk factor for developing gastric diseases including malignancies. This Gram-negative spiral-shaped bacterium naturally infects the human stomach, mostly asymptomatically; though in a subset of patients gastric disorders can develop. H. pyloridriven pathologies range from chronic active gastritis, peptic ulcer disease to gastric cancer and mucosa-associated lymphoid tissue (MALT) lymphoma (Kusters et al., 2006). Gastric disease outcome depends on the complex interaction of the host with the bacterium. Specific host genetic polymorphisms and gastric acid production control the colonization of the stomach, and antibiotics against H. pylori in patients suffering from gastritis and peptic ulcer markedly reduced the risk of recurring disease (Kuo et al., 2019;Treiber and Lambert, 1998). Individual H. pylori strains significantly vary in terms of their virulence potential and are divided into more virulent and less virulent groups (Covacci et al., 1997). Indeed, the severity of H. pylori-raised disorders widely depends on a range of bacterial factors. The most comprehensively studied H. pylori virulence factors associated with severe disorders are cytotoxin-associated gene A (CagA) and vacuolating cytotoxin A (VacA) (Atherton et al., 1995;Blaser et al., 1995;Tegtmeyer et al., 2017a). CagA is an effector protein injected into the host cell by a type IV secretion system (T4SS). This T4SS is encoded by the cag pathogenicity island (cagPAI) Olbermann et al., 2010). Upon delivery into host cells, CagA disturbs host molecular signaling, which leads to cytoskeletal rearrangements, loss of cell polarity, and epithelial barrier disruption (Knorr et al., 2019;Naumann et al., 2017;Sharafutdinov et al., 2020). Another H. pylori virulence factor, the cytotoxin VacA, triggers cell vacuolization, pore formation, and apoptosis in infected cells (Chauhan et al., 2019;McClain et al., 2017). Other H. pylori virulence factors that have been extensively studied in the last decade are the high temperature requirement A (HtrA) serine protease, blood group antigen binding protein A (BabA), sialic acid binding protein A (SabA), Helicobacter outer membrane protein Q (HopQ), and outer-inflammatory protein A (OipA) (Ansari and Yamaoka, 2019;Backert et al., 2016;Javaheri et al., 2016). Among those, special interest was focused on HtrA as this protease is secreted into the supernatant by H. pylori in order to disrupt cellular junctions (Tegtmeyer et al., 2017b).
HtrA serine proteases are widely distributed in both prokaryotic and eukaryotic organisms. Human HtrAs are involved in the maintenance of cellular homeostasis, in stress response and in cell death, and disturbance of their function can result in neurodegenerative diseases, musculoskeletal disorders or tumorigenesis (Clausen et al., 2011). In bacteria, HtrA family members combine chaperone and proteolytic activities by proper folding or degradation of misfolded proteins, respectively (Clausen et al., 2011). HtrA activity is commonly found in the periplasm and is necessary for bacterial survival under stress conditions, as was shown, for instance, in Escherichia coli (Skorko-Glonek et al., 1999), Listeria monocytogenes (Wonderling et al., 2004) and Streptococcus mutans (Biswas and Biswas, 2005). Furthermore, bacterial HtrAs can play a major role in pathogenesis by either directly conflicting damage to host tissue or by maintaining other bacterial virulence factors. For example, a mutant of the human pathogen Streptococcus pyogenes with impaired HtrA expression showed reduced amounts of mature streptococcal pyrogenic exotoxin B (SpeB) (Cole et al., 2007). Campylobacter jejuni, H. pylori and Bacillus anthracis and probably more bacteria secrete HtrA into the extracellular space and therefore exert proteolytic activity on host proteins .
In H. pylori and other bacteria, protease HtrA exhibits both chaperone and proteolytic activities, providing bacteria with cell viability and contributing to bacterial virulence. Recently, the chaperone activity of HtrA was shown to play an essential role in H. pylori survival under thermal, pH and osmotic stress conditions (Zarzecka et al., 2019a(Zarzecka et al., , 2019b. Furthermore, the proteolytic activity of HtrA was required for its efficient secretion by H. pylori. Interestingly, genetic inactivation of htrA was associated with mutations in SecA, which is involved in protein translocation from the cytoplasm into the periplasm, suggesting a functional relationship between HtrA and the Sec translocation system in H. pylori (Zawilak-Pawlik et al., 2019). Biochemical analyses showed that H. pylori's HtrA has a very high thermal stability in vitro and can restore its active structure after exposure to denaturing conditions (Zarzecka et al., 2019a(Zarzecka et al., , 2019b. These observations imply that HtrA is well adapted to both protein quality control in the bacterial periplasm and to pathogenesis when secreted. Secreted H. pylori HtrA was initially shown to target the host adherens junction protein E-cadherin via cleavage of its ectodomain (Hoy et al., 2010). Further analysis showed that the preferential HtrA cleavage sites in E-cadherin contain a [VITA]-[VITA]-x-x-D-[DN] sequence pattern . Finally, the tight junction proteins occludin and claudin-8 were identified as two additional substrates cleaved by HtrA during H. pylori paracellular transmigration (Tegtmeyer et al., 2017b). Remarkably, the htrA gene locus is highly conserved among H. pylori strains worldwide, implying a pivotal role of this protease for the bacterium . However, how HtrA affects the host cellular signaling upon H. pylori infection of gastric cells remains widely unclear. To elucidate the affected by HtrA upstream regulators and pathways, we performed an RNA-seq analysis of MKN-28 gastric epithelial cells after 2or 6-h infection with either H. pylori wt bacteria or with an isogenic ΔhtrA mutant. The RNA-seq analysis revealed new host cellular targets affected by H. pylori in an HtrA-dependent manner, comprising inflammatory, carcinogenic and apoptotic processes, which is emphasizing the significance of the HtrA in the pathogenesis of H. pylori.

RNA-seq of gastric epithelial cells infected with H. pylori
The present study was designed to identify the genome-wide affected genes as well as associated biological processes and molecular interaction networks in gastric MKN-28 cells upon infection with either the H. pylori N6 wild-type (wt) strain or an isogenic H. pylori N6ΔhtrA mutant in which the protease gene was deleted (Fig. 1A). Upon infection, H. pylori wt and ΔhtrA mutant showed similar bacterial loads on MKN-28 cells after 2 and 6 h of infection (Fig. 1B). In agreement with previous studies (Tegtmeyer et al., 2017b), wt bacteria tended to localize in the cell-to-cell junctions area, particularly after 6 h infection, in contrast to the ΔhtrA mutant that clustered significantly less near the cell junctions ( Fig. 1C and D). To determine the "early" and "late" transcriptomic response of polarized MKN-28 gastric epithelial cells during infection with H. pylori, samples for RNA-seq were collected after 2 and 6 h, respectively. Non-infected "mock" MKN-28 cells served as control. All experiments were performed in quadruple for statistical significance of the data. After quality filtering of the raw reads, on average 47,619,289 high-quality reads per sample were obtained, ranging from 40,628,500 to 56,376, 713 reads (Table S1). On average, 43,084,974 reads mapped uniquely against the human reference genome (Ensembl GRCh37), of which 40, 094,004 reads were counted over exons (84% of the raw reads). A sample-to-sample heatmap of DESeq2-normalized and log 2 transformed counts across replicates revealed strong homogeneity among the replicates and absence of outlier samples (Fig. S1). Furthermore, a principal component analysis (PCA) of log 2 transformed read counts showed a distinct difference between the uninfected control and cells infected for 6 h along PC 1, with 88% of the total variance explained along this axis (Fig. 1E). In contrast, the uninfected controls and the cells infected for 2 h separated along PC 2, which explained only 8% of the total variance, indicating that most transcriptomic changes occurred by 6 h post infection. Interestingly, the groups of MKN-28 cells infected with H. pylori N6 wt or with the N6ΔhtrA deletion mutant clustered together at either time point, indicating only minor differences between the wt and ΔhtrA groups in the host response to the bacterial infection.
The increasing variance in gene expression over the mean (Fig. S2) suggested that the Negative Binomial distribution fits the count dataset best, while the Poisson distribution would result in an increased number of differentially expressed genes (DEGs) that are false-positive (Soneson and Delorenzi, 2013). Therefore, in order to assess the role of HtrA in the host response to H. pylori infection, a differential gene expression analysis was performed using the DESeq2 tool (Love et al., 2014). The final dispersion estimates clustered around the fitted trend line, which implied that data are well distributed and suitable for differential gene expression analysis (Fig. S3). To address the problem of false positive DEGs and reduce the false discovery rate (FDR), we used the DESeq2 default shrinkage estimator to identify and remove weakly expressed genes that showed relatively high variability between the replicates. The number of DEGs amounted to 918 genes (669 upregulated and 249 downregulated) and 1708 genes (1069 upregulated and 639 downregulated) in the groups infected for 2 h with the H. pylori wt strain or the ΔhtrA mutant, respectively ( Fig. 2A). In contrast, the number of DEGs after 6 h of infection was much higher, showing 7681 (3960 upregulated and 3721 downregulated) and 7880 (4111 upregulated and 3769 downregulated) DEGs, respectively. 2.2. RNA-seq reveals distinct gene expression profiles in gastric cells after 6 h of infection with H. pylori wt or ΔhtrA Differentially expressed genes that displayed a 1.5-fold change and thus a log 2 fold change of !0.58 (upregulated genes) or À0.58 (downregulated genes) with a p value < 0.05 were selected for further analysis. This reduced the number of DEGs in the 2-h infection wt and ΔhtrA groups to 106 genes (98 upregulated and 8 downregulated) and 130 genes (125 upregulated and 5 downregulated), respectively ( Fig. 2A). The DEG numbers in the 6-h infection wt and ΔhtrA groups were reduced to 2228 genes (913 upregulated and 1315 downregulated) and 2329 genes (997 upregulated and 1332 downregulated), respectively, suggesting progressive downstream signaling after infection. The total of 2722 different DEGs, which were present in at least one of the groups, were clustered and visualized in a heatmap (Fig. 2B). Similar to the PCA in Fig. 1E, the DEGs clustered into three groups; (I) uninfected mock control; (II) infected for 2 h; and (III) infected for 6 h (Fig. 2B). Interestingly, the groups infected with the H. pylori wt strain or the ΔhtrA mutant sub-clustered into minimally distinct groups only after 6 h of infection, suggesting an increasing effect of the bacterial HtrA protease on the transcription of MKN-28 cells over time.

HtrA deficiency in H. pylori results in altered expression of host genes involved in immune responses and transcription activity
To unravel the biologically most relevant genes in MKN-28 cells affected by H. pylori, we created a volcano plot, which plots the DEGs by their statistical significance versus the magnitude of change. For visualization, we selected a threshold of an absolute log 2 fold change of !1.0 with a p value 10 À5 . After 2 h, infection with H. pylori wt or with the ΔhtrA mutant upregulated a similar set of genes ( Fig. 3A and B). As expected, the infection stimulated a strong immune response in MKN-28 cells, characterized by an over 2-fold increased transcription of genes IL1A, IL8, and IL24 encoding inflammatory cytokines, and of chemokine receptor ligand genes CXCL2, CXCL3 and CCL20. In addition, the bacterial infection activated transcription of EREG and EPGN, whose products are the ligands of the Epidermal Growth Factor Receptor (EGFR), as well as of STC2, TIPARP, TNFAIP3, CYP1A1, CYP1B1, PTGS2, LHX4, among others. Interestingly, the transcription profile of MKN-28 cells changed drastically after 6 h of infection by either H. pylori strain ( Fig. 3C and D) with many more genes significantly upregulated or downregulated. As indicated above, the number of DEGs increased to 2228 and 2329 in the H. pylori wt and ΔhtrA infection groups, respectively. Among the highly upregulated genes in both infection groups were ESPL1, DOCK8, ARID3A, PFKFB4, ANKZF1, BANP and GRB7, all of which exhibited a log 2 fold change of !1.5. Significantly downregulated genes in both groups with a log 2 fold change of À1.5 comprised CTGF, H1F0, CYR61, CTH, RGS2, DKK1, DNAJB9, ARG2, CDH5, and others.
Next, we aimed to define, which DEGs (adjusted p value < 0.05, log 2 fold change of À1 or ! 1) were specific for the individual experimental groups (Fig. 4A, Table S2). Four genes were differentially expressed in all four infection groups, of which expression of PTGS2, IL8 and CYP1A1 were upregulated, while expression of C9orf169 was upregulated at 2 h post infection, but downregulated at the 6 h time point. At the 2 h infection time point, expression of 13 genes (IL1A, IL24, CXCL2, CXCL3, RND1, BMF, IER3, TNFAIP3, TIPARP, LHX4, STC2, EPGN, EREG) was significantly upregulated in MKN-28 cells after infection with either wt or ΔhtrA mutant H. pylori strains. Transcription of IL1B was elevated only in MKN-28 cells infected with H. pylori wt bacteria, while upregulation of HTR6, ZFP36, BCL3 and MN1 was specific to MKN-28 cells infected with the ΔhtrA mutant. Considerably more, 381 DEGs, were shared between the wt and ΔhtrA H. pylori infection groups at the 6 h time point, with the most significantly upregulated genes being GRB7, PFKFB4, DOCK8, and MEX3A, and the most significantly downregulated genes PTP4A3, DKK1, BHLHA15, MAP1B, HSPA6, CTGF and NFATC2. The number of strainspecific genes increased by the 6-h time point to 97 DEGs induced by the ΔhtrA mutant bacteria and 52 DEGs induced by the wt bacteria. However, the change in expression of those genes was lower than for most of the 381 DEGs that were shared between the two groups. Interestingly, only one gene, TNF, was significantly upregulated in MKN-28 cells infected with the ΔhtrA mutant at either time point, but not in cells infected with the H. pylori wt strain.
To validate the RNA-seq data, differential expression of a set of six genes that are involved in various biological processes as discussed below was checked by RT-qPCR of the same samples (Fig. 4B). These genes included ZFP36, TNF, and IL8, for all of which the RT-qPCR data confirmed elevated expression in ΔhtrA mutant-infected compared to wt H. pylori-infected epithelial cells. In addition, the RT-qPCR analysis validated downregulation of DKK1 expression in H. pylori-infected cells versus the uninfected control, upregulated NFKBIA expression compared to the uninfected control with slight differences between the 2 h and 6 h time points, and upregulation of DOCK8 at the latter time point (Fig. 4B). Taken together, we identified the most significant DEGs in MKN-28 cells affected by either H. pylori wt or ΔhtrA, as well as defined particular DEGs shared by the experimental groups.

Gene ontology (GO) enrichment analysis shows both unique and distinct biological pathways in MKN-28 cells affected by wt and ΔhtrA H. pylori
To assess the functional importance of H. pylori infection on the MKN-28 whole-genome transcriptome, we performed a functional analysis of significant DEGs by using an over-representation analysis. First, we enriched known Gene Ontology (GO) terms in ClusterProfiler (Yu et al., 2012) using the corresponding lists of significant DEGs (adjusted p value < 0.05, log 2 fold change < À0.58 or >0.58). The most enriched GO terms upon infection with H. pylori wt for 2 h included the biological processes "Response to bacterium" (GO:0009617) and "Positive regulation of protein kinase activity" (GO:0045860). "Response to bacterium" was characterised by enrichment for genes encoding cytokines (e.g. IL24, IL1A, IL1B, IL8) and other proinflammatory regulators (e.g. IRAK2, WNT5A, CXCL2, CXCL3, TNFAIP3, ZFP36, SASH1, ANKRD1, NFKBIA, TNFRSF11A, CEBPB, BCL3 and ICAM1) as well as metabolic enzymes involved in stress response (e.g. CYP1A1 and SOD2). The category "Positive regulation of protein kinase activity" was enriched with TGFA, IL1B, IRAK2, WNT5A, EPGN, EREG, EGR1, CDKN1A, SASH1, LPAR1, ERCC6, DKK1, TNFRSF11A, SDC4, GDF15 and CARD10 ( Fig. S4; Table S3). When MKN-28 cells were infected with the ΔhtrA mutant for 2 h, the enriched GO terms were similar to those in MKN-28 cells infected with the H. pylori wt strain ( Fig. S4; Table S4). However, a biological process was defined that was distinctly affected by the ΔhtrA mutant after 2 h, namely "cellular response to tumor necrosis factor" (GO:0071356), which was characterized by enrichment for DEGs ZC3H12A, NFE2L2, IL8, TNF, TNFAIP3, TRAF1, ANKRD1, KCNJ11, BIRC3, THBS1, NFKBIA, TNFRSF11A, ICAM1 and ZFP36. A further analysis of the transcriptionally upregulated genes showed major changes at the functional levels after 6 h of infection, with the major biological processes "Cell population proliferation" (GO:0008283), "Regulation of cell migration" (GO:0030334) and "Regulation of intracellular signal transduction" (GO:1902531) (Fig. 5A, Table S5). Similar to infection with wt bacteria, the enriched GO terms after 6 h infection with the ΔhtrA included "Regulation of cell migration" and "Cell population proliferation" (Fig. 5B, Table S6). However, the enriched GO terms representing "Cell death" or "Apoptotic process" (GO:0010941, GO:0008219, GO:0043067 and GO:0012501) were distinctly present after the extended infection with the ΔhtrA mutant, suggesting that H. pylori HtrA may be implicated in the inhibition of host cell apoptotic pathways.
Since the data indicated a potential role of HtrA in apoptosis, we studied the induction of apoptosis in MKN-28 cells infected with either H. pylori wt or the ΔhtrA mutant. In agreement with the GO enrichment analysis, H. pylori ΔhtrA infection resulted in an higher apoptosis rate than infection with wt bacteria after both 2 h or 6 h infection (Fig. 5C and  D). Among the bacterial factors that could explain such an effect on apoptosis is H. pylori oncoprotein CagA which is known to exhibit antiapoptotic properties upon infection (Mimuro et al., 2007;Backert and Tegtmeyer, 2017). In our experiments, we found that infection of MKN-28 with H. pylori wt bacteria resulted in~2-fold higher CagA expression at both 2 and 6 h post infection compared to infection with the H. pylori ΔhtrA mutant (Fig. 5E), suggesting a crucial role of HtrA in CagA delivery. Therefore, it seems that HtrA mediates the downstream signaling in MKN-28 cells upon H. pylori infection at least partially through CagA.
2.5. Ingenuity Pathway analysis (IPA) reveals HtrA-dependent activation and inhibition of various upstream regulators The results of differential gene expression analysis and RT-qPCR along with the GO enrichment assay indicated a possible role of HtrA in TNF signaling in MKN-28 cells after 2 h of H. pylori infection. Therefore, we aimed to investigate the molecular interactions in MKN-28 cells upon infection with the ΔhtrA mutant in more detail. The most activated molecular functions in MKN-28 cells upon infection with H. pylori wt for 2 h after GO enrichment analysis appeared to be "Cytokine activity", "Cytokine receptor binding", "Receptor ligand activity", "Receptor regulator activity" and "Signaling receptor activator activity" (Fig. 6A), all of which correspond to the affected biological process "Response to bacterium" (Fig. S4). Thus, infection with H. pylori wt bacteria triggers a quick response of the host that results in activation of transcription factor genes JUN and NFKB1 (Fig. 6A). Activation of these transcription factors may lead to expression of various cytokines genes such as IL1B, CXCL2 and CXCL3, which in turn recruits immune cells to the site of infection. The ΔhtrA deletion mutant activated similar cellular functions in the "Response to bacterium" category (Fig. 6B). However, infection with the ΔhtrA mutant, but not with the wt bacteria, resulted in enrichment of genes involved in tumor necrosis factor (TNF) signaling, including TNF (encoding tumor necrosis factor), TNFAIP3 (tumor necrosis factor alphainduced protein 3), TNFRSF11A (tumor necrosis factor receptor superfamily member 11 A) and ZFP36 (anti-inflammatory modulator Zinc finger protein 36). We further assessed the TNF signaling pathway (hsa04668) in MKN-28 cells affected by ΔhtrA mutant infection using the KEGG enrichment pathway analysis. The TNF signaling pathway overview indicated that TNF activates transcription of downstream genes AP1, NFKB and PI3K and a set of chemokine genes including CXCL1, CXCL2, CXCL3, and IL1B, while expression of JNK and MMP14 was slightly inhibited (Fig. 6C). Thus, the analysis of DEGs using GO and KEGG enrichment assays indicated a likely role of the HtrA protease in the TNF-α response to infection with H. pylori.
In order to analyze the observed changes in MKN-28 gene expression upon H. pylori infection, we next attempted to pinpoint the upstream transcriptional regulators and their activation state using Ingenuity Pathway analysis (IPA) (Kramer et al., 2014). To identify the affected upstream regulators, IPA utilizes prior knowledge from its own database  to predict expected effects between transcriptional regulators and their target genes, while taking into account the directional change of the expression. Thus, up-or down-regulation of particular genes would indicate activation or inhibition of the corresponding regulators upstream of these genes. After 2 h of infection, the transcriptional upstream regulators EGFR, EGR1, FOXL2, FOXO3, IL1A, IL1B and RELA were activated in MKN-28 cells infected with either H. pylori wt or the ΔhtrA mutant, while WWTR1 was inhibited (Fig. 7, Tables S7 and S8). The EFNA5 and FAS genes were downregulated in MKN-28 cells only after 2-h infection with the H. pylori wt strain. In contrast, infection with the ΔhtrA mutant resulted in activation of FOXO1 and SYVN1, but inhibited IL1RN and S100A6. After 6-h infection, both H. pylori wt and ΔhtrA mutant led to activation of ECSIT, FOXM1, MITF, MYBL2, MYC and TRAF2 genes (Fig. 7, Tables S9 and S10). On the other hand, both bacteria inhibited The node size reflects the gene number involved in the process. The red and green numbers associated with genes indicate the up-and down-regulation levels in log 2 fold change. (C) TNF signaling pathway (hsa04668) affected by H. pylori ΔhtrA infection for 2-h was enriched using the KEGG enrichment assay. Red and green colors reflect up-and down-regulation levels in log 2 fold change, respectively. the upstream regulators CDKN1A, DUSP1, KDM5B, MAVS, MRTFB, NR3C1, NUPR1, TAZ, TEAD1 and TEAD2. Six hours of infection with the H. pylori wt strain activated upstream regulators ATF6 and CIP2A while TEAD3 and TEAD4 were inhibited. During infection with the ΔhtrA mutant, CBX5 was activated and CLU, KLF5, SP1, TCF7L2 and UBE2I were inhibited. Finally, the TNF upstream regulator was activated in all experimental groups with the exception of 6-h infection with the H. pylori wt strain.

Discussion
Early work on the interference of H. pylori infection with the host at the transcriptional level was mainly based on microarrays (Guillemin et al., 2002;Israel et al., 2001;Mills et al., 2001;Pachathundikandi et al., 2013). For example, Mueller and co-workers analyzed gene expression profiles to predict histopathological stages in a mouse model of MALT lymphoma (Mueller et al., 2003); DNA microarrays of gastric biopsies obtained from H. pylori-infected rhesus macaques provided new insights in host factors involved in pathogenesis (Huff et al., 2004); and a set of host genes was identified to be transcribed independently of the phosphorylation state of the CagA protein, a major H. pylori virulence factor (El-Etr et al., 2004). Later, the genome-wide map of H. pylori transcriptional start sites and operons was described by using differential RNA-seq (Sharma et al., 2010), which subsequently led to the discovery of small RNAs controlling expression of the major H. pylori virulence factors CagA and VacA (Eisenbart et al., 2020). Previous work using transcriptomic and proteomic approaches mainly focused on studying the disruption of host molecular pathways by H. pylori CagA and/or its associated T4SS (Chichirau et al., 2020;El-Etr et al., 2004;Glowinski et al., 2014;Guillemin et al., 2002;Wen et al., 2007). Furthermore, RNA-seq has been used to study the global immune responses to H. pylori in B cells (Chichirau et al., 2020) and macrophages (Tubau-Juni et al., 2020). In the present study, we aimed to investigate changes in molecular signaling in human gastric epithelial cells in response to H. pylori infection, with special focus on serine protease HtrA. Since the most significant changes in transcriptional regulation upon H. pylori infection appear at first hours, we aimed to analyze gastric MKN-28 cells after either 2 or 6 h of infection.
HtrA proteases are known to play a vital role in bacterial survival, in particular under stress conditions (Hansen and Hilgenfeld, 2013), but in some microbes they are also known as important virulence factors. HtrAs mediate secretion of other virulence factors, for example in Streptococcus pyogenes, or contribute to biofilm formation as was shown for S. mutans Biswas and Biswas, 2005;Cole et al., 2007). Another function of HtrA involves direct damage of cellular tight junctions of the host epithelium, for instance by Salmonella enterica, Shigella flexneri, C. jejuni or H. pylori (Harrer et al., 2017(Harrer et al., , 2019Hoy et al., 2012).
The most prominent changes in the MKN-28 transcriptome in response to H. pylori infection occurred between the 2-h and 6-h infection time points, independent of the htrA gene (Fig. 1). However, significant changes were observed in the number of DEGs at the 6-h time point, which was higher in MKN-28 cells infected with H. pylori ΔhtrA (n ¼ 7880; adjusted n ¼ 2329) compared to infection with the wt bacteria (n ¼ 7681; adjusted n ¼ 2228) (Fig. 2A). Interestingly, after adjusting for p 0.05 and log 2 fold change of -0.58 and !0.58, the significant DEGs (n ¼ 2722) sub-clustered the infections with H. pylori wt and ΔhtrA mutant into two distinct groups, even though the difference was minor (Fig. 2B). The separation of these two groups may possibly be explained by the effect of accumulating HtrA after secretion by the bacteria. Each individual H. pylori cell constantly secretes the HtrA protein, which results in a steady increase of the secreted HtrA amount over time (Neddermann and Backert, 2019). The bacteria replicate, and more bacterial cells secrete more HtrA molecules, which results in an increasing effect of HtrA on the MKN-28 expression profile, possibly both directly and indirectly. HtrA damages the cell junctions by cleavage of the adherens junction protein E-cadherin and tight junction proteins claudin-8 and occludin, which triggers a cellular response. In addition and possibly even more important, damage of the cell junctions allows paracellular transmigration of H. pylori through the opened junctions to the basolateral side of the epithelium where H. pylori employs its T4SS to deliver the oncogenic effector protein CagA into the host cell (Tegtmeyer et al., 2017b).
A more detailed analysis of DEGs indicated on a strong immune response in MKN-28 cells after 2-h infection based on increased Fig. 7. The most significant upstream regulators in MKN-28 cells affected by H. pylori infection as determined by the Ingenuity Pathway Analysis (IPA). The threshold for including upstream regulators in the analysis was designed for the activation Z-score (activation state of an upstream regulator based on the regulation direction associated with the relationship between the regulator and the corresponding DEGs): <2 (inhibited) or >2 (activated). The P value of overlap (significance of overlap between the dataset genes and the genes that are regulated by an upstream regulator) < 0.05 was considered as significant. The red or blue colored genes correspond to the activated or inhibited transcription regulators, respectively. The GREP1 gene (in black) was activated in MKN-28 infected by ΔhtrA mutant for 2 h, but downregulated after 6 h infection with H. pylori wt. transcription of IL1A, IL8, IL24, CXCL2, CXCL3 and CCL20 (Fig. 3A and  B). These data are in line with previous reports that indicated a strong IL-8 expression in response to cagPAI-positive H. pylori (Crabtree et al., 1995;Eftang et al., 2012). In our analysis, IL8 expression was also among the most affected at both 2-and 6-h infection time points, confirming its major role in H. pylori-induced inflammation. Interestingly, expression of the chemokine (C-X-C motif) receptor 2 (CXCR2) ligands CXCL2 and CXCL3 has been recently shown to contribute to the malignant progression of gastric cancer (Yamamoto et al., 2019). In addition, CC chemokine ligand 20 (CCL20) expression was shown to be upregulated by H. pylori in an cagPAI-dependent but CagA-independent manner (Yoshida et al., 2009). In our experiments, the pro-inflammatory response of MKN-28 gastric cells after 2-h infection was comparable between H. pylori wt and ΔhtrA mutant infections. Both wt and ΔhtrA strains activated transcription of the EREG and EPGN genes both of which contribute to stimulation of the epidermal growth factor receptor gene EGFR. Elevated epiregulin (EREG) expression predicts poor prognosis in gastric cancer (Xia et al., 2019), while epigen (EPGN) is suspected to play a role in the development of lung carcinomas (Fujimoto et al., 2005).
The 6-h infection time point was marked by downregulation of CTGF, H1F0, CYR61, CTH, RGS2, DKK1, DNAJB9, ARG2 and CDH5. Of interest, Dickkopf-related protein 1 (DKK1), an important inhibitor of carcinogenic Wnt signaling, was recently shown to be inhibited in intestinal metaplasia via promoter methylation (Lu et al., 2020). This study reported DKK1 downregulation in response to bile acid stimulation, while we observed DKK1 downregulation after exposure to H. pylori. In addition, H1F0, which encodes histone H1.0, was shown to be silenced in various cancers (Torres et al., 2016) and could be a potential target molecule during H. pylori pathogenesis. Numerous other genes such as ESPL1, DOCK8, ARID3A, PFKFB4, ANKZF1, BANP and GRB7 showed elevated expression rates, some of which are known or suspected to be involved in cancer development and/or progression. Of those, dedicator of cytokinesis 8 (DOCK8) is a guanine nucleotide exchange factor (GEF) that activates a number of small G proteins, including Rac1 and Cdc42, which are known to be manipulated by H. pylori during rearrangement of the actin cytoskeleton of the host cell (Churin et al., 2001), suggesting that increased DOCK8 transcription might exacerbate the effect. The Warburg pathway enzyme 6-phosphofructo-2-kinase/fructose-2, 6-bisphosphatase 4 (PFKFB4) is a kinase of the sugar metabolism that affects gene regulation by activating transcription of the oncogenic steroid receptor coactivator-3 (SRC-3), which promotes the development of tumour metastases in breast cancer (Dasgupta et al., 2018). Elevated PFKFB4 transcription levels strongly correlated with shorter disease-free survival and overall breast cancer survival (Yao et al., 2019). Finally, growth factor receptor bound 7 (GRB7), a multidomain adaptor protein that was previously shown to interact with the H. pylori CagA effector protein (Selbach et al., 2009), is a critical mediator of EGFR/ErbB signaling involved in cancer development (Chu et al., 2019).
Interestingly, MKN-28 cells infected with ΔhtrA at both 2-h and 6-h time points exhibited increased TNF expression implying that HtrA may be involved in TNF signaling, a view that was supported by the observed deregulation of ZFP36, BCL3 and TNF during infection with the ΔhtrA mutant. In addition, a GO enrichment assay of H. pylori-induced biological pathways revealed a distinct, ΔhtrA-specific group of enriched GO terms that represented "Cell death" or "Apoptotic process", suggesting that H. pylori's HtrA might interfere with host cell apoptosis. Considering that TNF plays a crucial role in the regulation of both NF-κBinduced inflammation and caspase-mediated apoptosis (Lee et al., 2016), disturbance of the TNF pathway by HtrA explains the observed enrichment of apoptosis GO terms. In agreement, analysis of the upstream transcriptional regulators using IPA indicated downregulation of the EFNA5 and FAS upstream regulators by H. pylori wt, but not ΔhtrA bacteria. Fas, a member of the tumor necrosis factor receptor superfamily, is a cell surface "death receptor" which upon activation induces caspase-mediated apoptosis (Meynier and Rieux-Laucat, 2019). H. pylori appears to inhibit Fas receptor expression in an HtrA-dependent manner, which leads to inhibition of Fas-mediated apoptosis, underpinning the above observations. In addition, infection with the H. pylori ΔhtrA mutant led to activation of the FOXO1 upstream regulator, another transcription factor involved in apoptosis regulation. Interestingly, FOXO1 was previously shown to be deregulated upon H. pylori infection (Tabassam et al., 2012). H. pylori wt infection activated the Cancerous Inhibitor of PP2A (CIP2A) upstream regulator that is overexpressed in gastric cancer. Intriguingly, CIP2A expression was shown to depend on H. pylori CagA (Zhao et al., 2010), however, HtrA might represent another bacterial factor disturbing the CIP2A pathway. Other upstream regulators such as CLU, KLF5, SP1 and TCF7L2 that were inhibited by ΔhtrA mutant infection have also been previously shown to be involved in apoptosis regulation (Deniaud et al., 2006;Li et al., 2014Li et al., , 2018Mustafi et al., 2017). H. pylori ΔhtrA-inhibited SP1 likely decreased expression of the downstream cystathionine γ-lyase (CTH) as observed above. In this regard, H. pylori was recently shown to evade the host immune response by inducing CTH in macrophages (Gobert et al., 2019), and therefore, HtrA could play a role in SP-mediated CTH expression. Finally, infection with H. pylori ΔhtrA activated TNF throughout the infection, which substantiates the above-discussed considerations on the potential role of HtrA in the TNF-mediated inflammatory response.
We suggest that HtrA may affect host cellular signaling by either direct interactions with host target molecules or, more likely, by facilitating enhanced delivery of CagA and/or other virulence factors into the host cytoplasm. Indeed, HtrA-dependent cleavage of cellular junctions was shown to promote H. pylori paracellular transmigration, which facilitates basolateral injection of CagA into the epithelial cells (Tegtmeyer et al., 2017b). Alternatively, HtrA might affect cellular signaling directly via cleavage of E-cadherin, a tumour suppressor protein, the disruption of which appears to play a significant role in carcinogenesis signaling. In addition, we propose that the proteolytic activity of HtrA might also be involved in interfering with the TNF pathway. For instance, trypsin-like proteases were previously shown to cleave the TNF molecule from the N-terminus that subsequently resulted in its inactivation (Nakamura and Komiya, 1996). Proteolysis of TNF was also shown by a cysteine protease from the pathogen Porphyromonas gingivalis, confirming the biological relevance of such a process (Park et al., 2016). Thus, functional inactivation of TNF through cleavage by HtrA might also contribute to the observed host cell apoptosis. The mechanism(s), by which HtrA affects host signaling, whether directly or indirectly, should be unraveled in further studies. Summarizing the above discussed observations, we propose that upon H. pylori infection, secreted HtrA in addition to the tight junctions cleavage, significantly contributes to disturbance of host cellular signaling, which overall impacts disease development.

Cell line, bacteria and culture conditions
Human MKN-28 cells (JCRB, #0253), originally isolated from gastric adenocarcinoma, were used in this work for the infection experiments. Cells were cultured in RPMI-1640 medium, containing 4 mM glutamine (Invitrogen, Karlsruhe/Germany), and 10% FCS (Invitrogen, Karlsruhe/ Germany) at 37 C. MKN-28 cells were grown until the formation of proper monolayers as described (Tegtmeyer et al., 2017b). Briefly, the cells were cultured on cell culture inserts with 3 μm pore size (Millipore, Burlington, Massachusetts, USA) to confluent monolayers and were subsequently incubated for another 14 days to allow for cell polarization. A transepithelial electrical resistance (TER) of !150 Ω/cm 2 indicated the formation of polarized cell layers . H. pylori strain N6 and H. pylori N6ΔhtrA (Zawilak-Pawlik et al., 2019) with protease knockout were grown on horse serum GC agar plates supplemented with nystatin (1 μg/mL), vancomycin (10 μg/mL) and trimethoprim (5 μg/mL), and if necessary with 4 μg/mL chloramphenicol.

MKN-28 infection with H. pylori
For infection H. pylori wt and H. pylori ΔhtrA were grown for 2 days at 37 C in anaerobic chambers containing a CampyGen gas mix (Oxoid, Wesel/Germany) (Wiedemann et al., 2012). H. pylori was harvested and resuspended in phosphate buffered saline (PBS, pH 7.4) using sterile cotton swabs (Carl Roth, Karlsruhe/Germany). The bacterial concentration was measured in a spectrophotometer as optical density (OD) at 600 nm (Eppendorf, Hamburg/Germany). Apical marker expression such as microvilli and tight junction formation were routinely checked as described (Tegtmeyer et al., 2017b). Infections were performed from the apical side at a multiplicity of infection (MOI) of 100 as described (Kim et al., 2013), unless indicated otherwise. All infection assays were repeated four times.

RNA-seq and differential gene expression
After 2 or 6 h of infection, MKN-28 cells were harvested. Total RNA was extracted using the RNeasy Mini kit (Qiagen) as described (Heimesaat et al., 2014). Illumina sequencing libraries were constructed according to the manufacturer's instructions, and were subjected to single-end sequencing (101 bp) on a HiSeq-2500 platform (Illumina, San Diego, CA). Quality filtering was performed using cutadapt v. 1.9.1 (Martin, 2011). The reads were mapped against the human reference genome (Ensembl GRCh37, release 87) using STAR aligner v. 2.5.2 b (Dobin et al., 2013), and a STAR genome directory created by supplying the Ensembl gtf annotation file (release 87) for GRCh37. Read counts per gene were obtained using featureCounts v. 1.5.1 (Liao et al., 2014) and the Ensembl gtf annotation file.
The subsequent analyses were performed using R version 4.1.1 (Team, 2020) in the RStudio platform (Team, 2021). In particular, differential expression analysis was performed with the DESeq2 package v.1.32.0 (Love et al., 2014). DESeq2 performs by default shrinkage of fold changes and "independent filtering", i.e. it finds a threshold on the mean normalized counts that optimizes the number of differentially expressed genes (DEGs). The list of the most significant DEGs with a log 2 Fold-Change threshold of <1 or >1 are presented in Table S2. A set of representative genes showing the expression levels of all 20 samples along with their computed log 2 fold changes are presented in Fig. S5 to illustrate the data quality and signal differences between different groups. Volcano plots representing the most significant DEGs were constructed using EnhancedVolcano, an R package version 1.10.0 (Blighe et al., 2021). Venn diagrams were drawn using ggvenn, an R package version 0.1.9, available from https://github.com/yanlinlin82/ggvenn.

Pathway analysis
The functional enrichment analysis was performed in RStudio by the Gene Ontology (GO) overrepresentation analysis using ClusterProfiler, an R package version 4.0.5 (Yu et al., 2012). The grch37 table with human annotations based on genome assembly GRCH37 from Ensembl was loaded from the annotables library (an R package version 0.1.91) and further used for conversion of gene IDs. The p values in GO enrichment analysis were adjusted using the Benjamini-Hochberg (BH) false discovery rate. To analyze the pathways enriched with DEGs we used KEGG (Kyoto Encyclopedia of Genes and Genomes) annotation data supported in ClusterProfiler. To visualize selected pathways, the Pathview, an R package (version 1.32.0) was used (Luo and Brouwer, 2013). Alternatively, the DEGs were analyzed using Ingenuity Pathway Analysis (IPA) Tool (Kramer et al., 2014), in particular, to determine the upstream regulators which were significantly affected by H. pylori infection. Log 2 fold change values and corresponding identifiers of DEGs were used as an input data for IPA. Upstream regulator analysis was performed to predict the top significantly activated (Z-score ! 2) and inhibited (Z-score À2) upstream regulators corresponding to the input data, considering the direction of change, i.e. up-or downregulation. The overlap p values reflect statistically significant overlap between the dataset genes and the genes that are regulated by an upstream regulator, and were calculated using Fisher's Exact Test. The upstream regulators with the false discovery rate (p value) < 0.05 and the overlap p value < 0.05 were considered significant.

RT-qPCR
Expression levels of ZFP36, TNF, NFKBIA, IL8, DKK1, and DOCK8 were assessed by quantitative reverse transcription PCR (RT-qPCR) in quadruplicates using primers listed in Table S11. Expression of GAPDH served as an internal control. The expression levels were analyzed using SYBR Green PCR master mix in the iCycler/MyiQ Real Time PCR detection system (Bio-Rad, USA) as described (Pachathundikandi et al., 2018). The obtained cycle threshold (CT) values were used to quantify relative expression levels as ΔCT (CT reference gene -CT gene of interest ) and ΔΔCT (ΔCT infectionmean ΔCT mock ) values. The difference between infection groups was analyzed using two-tailed t-test of the corresponding ΔΔCT values. The expression levels were finally presented as fold changes (2^ΔΔCT).

Immunofluorescence microscopy
To assess the amount and distribution of H. pylori over gastric cells, MKN-28 cells were grown on glass coverslips in 12-well plates until the formation of proper monolayers. After infection with H. pylori wt or H. pylori ΔhtrA at MOI of 25 for 2 or 6 h, cells were fixed with cold methanol for 10 min and immunostained using mouse FITC-conjugated α-occludin (#331511, Invitrogen, Waltham, MA, USA). Primary rabbit α-H. pylori (Dako, Glostrup, Denmark) and secondary α-rabbit-AlexaFluor 633 (#A-21070, Invitrogen, Darmstadt, Germany) were used to detect H. pylori cells. Nuclei were counterstained using 1 μg/mL DAPI (4 0 -6diamidino-2-phenylindole dihydrochloride). The images were acquired using confocal laser scanning microscope Leica Stellaris 8 (Leica Microsystems, Wetzlar, Germany) at the Optical Imaging Centre Erlangen (OICE, Erlangen, Germany). LAS AF computer software (Leica Microsystems, Wetzlar, Germany) was used to visualize the obtained data. To analyze the distribution of bacterial cells adhered to MKN-28 monolayer, the number of bacteria in the cell junction area was divided by the number of bacteria outside of the cell junctions area. Welch's ANOVA combined with Holm-Sidak's post-test was used to define the differences between the groups with p 0.05 (*) considered to be significant, and with p > 0.05 to be non-significant (n.s.).

Apoptosis assay
MKN-28 cells were grown in 12-well plates until the formation of proper monolayers before being infected as described above. After infection, cells were washed with HEPES-buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl 2 ) and additionally incubated for 30 min in HEPESbuffer containing 1 μg/mL annexin V conjugated with fluorescein isothiocyanate (FITC) and 1 μg/mL propidium iodide (PI). Subsequently, the cells were washed with HEPES-buffer and analyzed under a DMI4000B epifluorescence microscope (Leica Microsystems, Wetzlar, Germany). Cells stained positive for either annexin V-FITC or PI were considered as undergoing apoptosis/necrosis and were presented as percentage out of all counted cells as described . Welch's ANOVA combined with Holm-Sidak's post-test was used to define the differences between the groups with p 0.05 (*) considered to be significant.

Statistics
The statistical analysis details are explained after every method description where appropriate. The Welch's ANOVA and two-tailed unpaired t-test statistical analyses were performed using GraphPad Prism statistical software version 8.0 (GraphPad Software, United States). Statistical significance was defined as p 0.05 (*) or otherwise nonsignificant (n.s.). Graphs in Fig. 1B, C, 4B, 5D, and 5E are presented as mean values AE standard deviation (SD). In addition, Fig. 4B shows the individual values as black dots.

Data availability
The data discussed in this publication have been deposited in NCBI's Gene Expression Omnibus (Edgar et al., 2002) and are accessible through GEO Series accession number GSE202165 (https://www.ncbi.nlm.nih .gov/geo/query/acc.cgi?acc¼GSE202165).

Declaration of competing interest
The authors declare no conflict of interest.