ADP-heptose attenuates Helicobacter pylori-induced dendritic cell activation

ABSTRACT Sophisticated immune evasion strategies enable Helicobacter pylori (H. pylori) to colonize the gastric mucosa of approximately half of the world’s population. Persistent infection and the resulting chronic inflammation are a major cause of gastric cancer. To understand the intricate interplay between H. pylori and host immunity, spatial profiling was used to monitor immune cells in H. pylori infected gastric tissue. Dendritic cell (DC) and T cell phenotypes were further investigated in gastric organoid/immune cell co-cultures and mechanistic insights were acquired by proteomics of human DCs. Here, we show that ADP-heptose, a bacterial metabolite originally reported to act as a bona fide PAMP, reduces H. pylori-induced DC maturation and subsequent T cell responses. Mechanistically, we report that H. pylori uptake and subsequent DC activation by an ADP-heptose deficient H. pylori strain depends on TLR2. Moreover, ADP-heptose attenuates full-fledged activation of primary human DCs in the context of H. pylori infection by impairing type I IFN signaling. This study reveals that ADP-heptose mitigates host immunity during H. pylori infection.


Introduction
Helicobacter pylori (H.pylori) is one of the most successful human pathogens, infecting approximately 50% of the world's population.This unique pathogen ensures its own survival within the stomach by perfectly adapting to and shaping its niche. 1 Chronic infection with H. pylori is asymptomatic in most cases, but 10-15% of infected individuals develop gastric ulcers and 1-3% stomach cancer, 2 which is the third leading cause of cancer-related deaths worldwide. 3,4Accordingly, the International Agency for Research on Cancer has declared H. pylori to be a class 1 (definite) carcinogen. 5Although, H. pylori is masterful at evading immune recognition and manipulating host immune responses, 6 infection still results in potent inflammatory responses in cells of the innate and adaptive immune system. 7,8 key feature of the innate immune system's rapid response to infection is its ability to recognize conserved pathogen-associated molecular patterns (PAMPs).9,10 In this context, metabolic intermediates of lipopolysaccharide (LPS) biosynthesis are attracting intense interest as they have been identified as novel PAMPs of various Gram-negative bacteria, including H. pylori.The immunostimulatory capacities of these metabolites were initially described in studies involving infections with Neisseria gonorrhoeae, Neisseria meningitidis, Shigella flexneri, Salmonella enterica serovar Typhimurium or H. pylori, showing that NF-κB activation as well as secretion of proinflammatory cytokines is induced by HMP (D-glycero-β-D-manno-heptose 1-monophosphate) or HBP (D-glycero-β-D-manno-heptose 1-biphosphate).[11][12][13][14][15][16] However, soon afterward, in 2018, it was shown that only one heptose metabolite, namely ADP-heptose (ADP-glycero-β-D-manno-heptose), could effectively enter the host cell cytosol to robustly activate inflammatory responses.HBP, in contrast, was found to require conversion to ADP-heptose-7-phosphate by host adenylyltransferase to elicit a pro-inflammatory response.Although HBP indeed triggers inflammatory signals, the effects of HBP are modest in comparison to those induced by ADP-heptose. Idenification of a physical interaction of ADPheptose with the N-terminal domain of the serine/threonine-specific kinase alpha kinase 1 (ALPK1) led to assignment of the ADP-heptose /ALPK1 complex to the PAMP -pattern recognition receptor (PRR) family.[17][18][19] Recognition of ADP-heptose by its cognate receptor results in phosphorylation of TRAF-interacting protein with forkhead-associated domain (TIFA) to trigger its oligomerization, leading to the recruitment of various other downstream interaction partners to form a multiprotein-complex termed the TIFAsome.16 Based on detailed mechanistic studies it is now well-established that TIFA forms a complex with TNF receptor-associated factor (TRAF) 6 or TRAF2, which in turn mediates the induction of classical or alternative NF-κB signaling, respectively, and promotes the secretion of CXCL8 in epithelial cells.11,14,16,20 Moreover, in H. pylori-infected epithelial cells the ALPK1/ TIFA/NF-κB axis has been shown to induce DNA damage via the induction of R-loop-associated replication stress.21 To date, most studies conferring PAMP status to ADP-heptose have relied on epithelial cell models.Interestingly, at the time that ALPK1 was identified as the cognate receptor of ADP-heptose, Ryzhakov and colleagues reported that Alpk1 −/− mice are more susceptible to Helicobacter hepaticus-induced colitis and that this condition is mainly driven by Alpk1 deficiency in hematopoietic cells.In particular, they identified that Alpk1 −/− macrophages produce increased amounts of inflammatory cytokines, including Interleukin 12 (IL-12), upon H. hepaticus infection and thus concluded that ALPK1 suppresses IL-12mediated, Th1-driven intestinal inflammation.22 Their study suggests that in hematopoietic cells, ALPK1 signaling attenuates rather than promotes inflammatory responses.
To further our understanding of the role of ADP-heptose beyond the gastric epithelium, we aimed to investigate its capacity to activate human dendritic cells (DCs), the most adept antigen-presenting cells, by using H. pylori infection as a model system.As sentinels of the immune system, DCs are not only well-equipped with a plethora of PRRs, but they are also essential for maintaining the delicate balance between activating potent inflammatory responses and inducing tolerance in different pathological settings.However, activation of DCs in the context of H. pylori infection is controversial, since several studies have concluded that H. pylori induces tolerogenic DCs, [23][24][25] while others clearly show that H. pylori infection results in the secretion of a variety of pro-inflammatory mediators and the expression of co-stimulatory surface molecules. 7,8,26,27In this study, we show that DCs are recruited in high numbers into H. pylori-infected gastric mucosa, and that they are potently activated upon H. pylori infection in an in vitro model of the gastric epithelial environment.Moreover, we report that ADPheptose barely activates human DCs compared to the well-studied bacterial PAMP LPS from E. coli.Instead, and even more intriguingly, ADP-heptose potently suppresses the Th-1-associated inflammatory response in DCs.Our mechanistic investigations revealed that TLR2 promotes H. pylori uptake and the enhanced CD40 expression upon infection with the ADPheptose deficient H. pylori mutant.Moreover, ADP-heptose dampens H. pylori-induced type I interferon (IFN) signaling, resulting in decreased IL-12p70 secretion upon H. pylori infection.Thus, rather than initiating potent inflammatory responses in professional antigenpresenting cells similar to other bona fide PAMPs, ADP-heptose appears to exert an immune-dampening effect on human DCs and subsequent T cell responses.

Isolation of primary human immune cells from peripheral blood
This study was conducted in accordance with the approved guidelines of the World Medical Association's Declaration of Helsinki and the guidelines of the Ethics Committee of the Province of Salzburg.No additional form of consent is required from anonymized blood donors for scientific use of leukapheresis products according to Austrian national regulations.
Human primary CD1c + (BDCA1 + ) DCs were isolated from fresh buffy coats obtained from the Blood Bank Salzburg.Briefly, PBMCs were isolated from human leukapheresis products via gradient density centrifugation using HistoPaque-1077 (Sigma-Aldrich 10,771).Thereafter, DCs were isolated via magnetic cell sorting according to the manufacturer's instructions using the Human CD1c (BDCA1) + Dendritic Cell Isolation Kit (Miltenyi Biotec, 130-119-475).DCs were then cultured in RPMI-1640 (Sigma-Aldrich, R0083) supplemented with 10% heat-inactivated fetal calf serum (Biowest, BS-2020-500) and 1% L-glutamine (Sigma-Aldrich, G7513).After isolation, DCs were incubated between 1 and 4 h at 37°C until further use.For qPCR and co-culture experiments, DCs were seeded at a density of 10 5 cells/mL, and for flow cytometry and Western Blot analysis at 2 × 10 5 cells/mL.Primary human naïve CD4 + T cells or total T cells were isolated from PBMCs as described above via magnetic cell sorting according to the manufacturer's instructions using a Human Naïve CD4 + T Cell Isolation Kit II (Miltenyi Biotec, 130-094-131) or Human Pan T Cell Isolation Kit (Miltenyi Biotec, 130-096-535), respectively.T cells were subsequently incubated with DCs as described below.

DC/T cell co-culture
DCs were infected with H. pylori for 18 h before Penicillin and Streptomycin (1×, Sigma-Aldrich, P4333) was added to the medium, to avoid direct infection of naïve T cells.After another 6 h of incubation, freshly isolated allogenic naïve CD4 + T cells or total T cells were added to the DC culture at a ratio of 1:10 (10 5 DCs: 10 6 T cells).Co-culture was performed for 6 days, thereafter cytokine and surface marker expression of T cells was analyzed by flow cytometry.

Bacterial culture and infection
Helicobacter pylori P12 and its isogenic mutant (ΔrfaE) 16 were cultivated under microaerophilic conditions on GC agar plates containing 10% horse serum (Biowest, S0910) and selective antibiotics (Kanamycin).Microaerophilic conditions were obtained using Thermo Scientific Oxoid AnaeroJars (Thermo Fisher Scientific) and Oxoid CampyGen 3.5 L sachets (Thermo Fisher Scientifc, CN0025A) and bacteria were cultured over a period of 3 days.One day prior to DC infection, bacteria were re-plated and allowed to grow for another 24 h.Acinetobacter lwoffii (DSM 2403) was plated the day before infection of eukaryotic cells on nutrient agar plates and incubated at 37°C and 5°C CO 2 .On the day of infection, cotton buds (Paul Boettger, EH12.1) were used to collect bacteria from plates.Bacteria were resuspended in 1 mL phosphate-buffered saline (PBS, Sigma-Aldrich, D8537) and cell density was determined via spectrophotometric measurement at OD600 (BioPhotometer Plus, Eppendorf).Using an inhouse calibration curve bacterial cell numbers were calculated.For all experiments including DCs, an MOI of 5, for infection of mucosoids an MOI of 100 and for PBMCs an MOI of 0.1 was applied.

Mucosoid culture and infection
Human gastric tissue specimens were provided by the Department of General, Visceral and Thoracic Surgery, Paracelsus Medical University/Salzburger Landeskliniken (SALK) from individuals undergoing gastrectomy or sleeve resection.The corresponding ethics application was approved by the local ethics committee (EK Nr: 1003/2021, Ethikkommission für das Bundesland Salzburg).

Luminex assay
Measurement of chemokine and cytokine secretion was performed via bead-based multiplex assay using both an Inflammation 20-Plex Human ProcartaPlex Panel (Thermo Fisher Scientific, EPX200 -12,185-901) or a Cytokine/Chemokine/ Growth Factor 45-Plex Human ProcartaPlex Panel (Thermo Fisher Scientific, EPX450 -12,171-901).All following incubation steps are performed on an orbital shaker (450-600 rpm) at indicated temperatures and for specified time periods.For both assays, appropriate amounts of beads were collected in 1.5 ml tube and washed once by adding PBS containing 0.05% Tween-20 (Sigma-Aldrich, P7949) followed by centrifugation.Thereafter, beads were resuspended in Assay Buffer and transferred to 96-well V bottom plates (Greiner BioOne 651,101).Upon addition of standards and samples, the plate was incubated overnight at 4°C.The next day, three washing steps were performed, by adding 150 µl wash buffer/well followed by centrifugation (1650 g, 5 min, 4°C).Supernatants were removed by flipping the plate.Incubation with the biotinylated Detection Antibody for 30 min at room temperature was followed by another three washing steps.Thereafter, streptavidin-labeled PE was added, and the plate was incubated for another 30 min at room temperature.The plate was washed thrice before beads were resuspended in Reading Buffer and wells were measured on a Luminex MagPix instrument (Luminex).The browserbased ProcartaPlex Analyst software was used to analyze the raw data.

Granzyme B enzyme linked immunosorbent assay (ELISA)
Granzyme B secretion was analyzed using the Granzyme B DuoSet ELISA (R&D Systems, DY2906-05) according to manufacturer's instructions.

Flow cytometry analysis of intracellular cytokine and granzyme B production
Three hours prior to isolation, co-cultured and polarized T cells were stimulated with phorbol 12-myristate 13-acetate (PMA, final concentration 50 ng/mL, Sigma-Aldrich, P1585), ionomycin (1 µg/mL, Sigma-Aldrich, I0634) and brefeldin A (10 µg/mL, Sigma-Aldrich, B7651).Thereafter, cells were isolated and washed with PBS and transferred to a 96-well V-bottom plate (Greiner BioOne 651,101).Cells were then stained with viability dye eFluor 780 (Thermo Fisher Scientific, 65-0865-14) and incubated for 15 min at 4°C.After another wash, F c receptor block (Biolegend 422,302) was added according to the manufacturer's instructions and incubated with the cells for 15 min at 4°C.Then, the cells were washed again and surface antigen-binding antibody-fluorophore conjugates were appropriately diluted in PBS to prepare a surface staining mix, of which 30 µL were added to each well and the cells were resuspended.Cells were incubated with the surface staining mix for 30 min at 4°C in the dark.Subsequently, cells were washed with PBS, after which 100 µL Fixation/ Permeabilization solution (BD Biosciences 554,714) was applied and the cells were incubated for 15 min at 4°C.Cells were then washed twice with Perm/Wash buffer (BD Biosciences 554,714) and afterward resuspended in 50 µL intracellular staining mix, which was prepared beforehand by diluting intracellular antigenbinding antibody-fluorophore conjugates in Perm/Wash buffer.Cells were incubated with the intracellular staining mix for 30 min at 4°C and then washed again with Perm/Wash buffer, after which they were resuspended in 120 µL PBS-EDTA (3 mM) and examined using a Cytoflex S cytometer (Beckman-Coulter) or a Canto II flow cytometer (BD Biosciences).The following antibody-fluorophore conjugates were utilized in the analysis: CD3 PerCP/Cy5.5 (Biolegend 300,429, clone: UCHT1), CD4 BV510 (Biolegend 344,633, clone: SK3), CD4 FITC (Thermo Fisher Scientific, 11-0048-42, clone: OKT4), CD8 BV510 (BD Biosciences 563,919, clone SK1 (RUO)), IFNγ PE (Biolegend 506,507, clone: B27), Granzyme B AF700 (BD Biosciences 560,213, clone: GB11).Data analysis was performed on FlowJo 10 software.

GeoMx digital spatial profiling
All patients gave informed consent for gastroscopy and endoscopic biopsy.As it was neither a clinical drug trial nor an epidemiological study, no ethical approval is required (see: https://www.salzburg.gv.at/themen/gesundheit/einrichtungen/ethikkommis sion.) and no further approval of the study by the local ethics committee was necessary.
Cells were analyzed using a Zeiss Observer Z1 fluorescence microscope equipped with an Abberior Instruments STEDYCON unit for confocal and super-resolution STED microscopy.Representative confocal images were taken with a 100× oil objective from single focal z-planes and/or by generating confocal z-stacks.All images were post-processed with Fiji (ImageJ1.54f)and Inkscape.

Metabolite analysis
Bacteria were cultivated on GC agar plates containing 10% horse serum and selective antibiotics, if appropriate.Bacteria were cultured over a period of 72 h and then re-plated and allowed to grow overnight prior harvest for metabolite measurement.10 9 bacterial cells were harvested from the plates and used for metabolite analysis.To confirm the presence or absence of ADP-heptose in the different bacterial samples, a targeted HPLC-MS assay was employed: the bacterial samples were washed once by addition of 1 mL of 185 mmol/L ammonium bicarbonate (Sigma-Aldrich) and subsequent centrifugation at 13,680 g for 20 seconds at 4°C.After removal of the supernatant, the bacteria were resuspended in 500 µL ice-cold methanol (Sigma-Aldrich) containing 10 µmol/L 3-nitro-L-tyrosine (Sigma-Aldrich) as an internal standard.Subsequently the bacteria were lysed by sonication in a Bioruptor device for 5 min.Afterwards the samples were centrifugated at 15,000 g for 10 min at RT.The supernatant was transferred into a fresh tube, evaporated to dryness at 40°C and subsequently resuspended in 50 µL of 20 mmol/L triethylamine (Sigma-Aldrich) in water adjusted to pH 6.0 with acetic acid (Sigma Aldrich), which was also used as solvent A for the chromatography.As a positive control, an ADP-heptose standard (InvivoGen) was utilized.The separation was carried out following Rautengarten et al. 29 : 10 µL of sample were injected into an eFluor™ 3000 Rapid Separation system (Thermo Fisher Scientific) operated at a flow rate of 200 μL/min and a temperature of 28°C employing a Fluor™ Fluor™ Hydro-RP column (150 × 2.0 mm i.d., 4 µm particle size, 80 Å pore size) with a security guard cartridge C18 (4.0 × 3.0 mm i.d., Phenomenex).Acetonitrile (VWR) was used as solvent B. The separation started isocratically at 100.0%A for 5 min followed by a linear gradient of 0-5.0%B in 10 min, a second linear gradient from 5.0-30.0%B in 15 min, a second isocratic phase at 90.0% B for 3 min, and re-equilibration of the column at 100% A for 7 min.The total runtime for one sample was 40 min.UV detection was carried out at 262 nm.Mass spectrometry was conducted as described by Pfannkuch et al. 19 by targeted selected ion monitoring on a Thermo UltiMate™ Q Phenomenex® Hybrid Quadrupole-Synergi™ mass spectrometer equipped with a Thermo Scientific™ Ion Exactive™ ion source with a heated electrospray ionization (HESI) probe.The source heater temperature was set to 250°C, spray voltage to −4.0 kV, sheath gas flow to 35 arbitrary units, auxiliary gas flow of 5 arbitrary units, capillary temperature to 350°C and S-lens RF level to 60.0.The m/z values set on the inclusion list were 618.08500 (ADP-heptose) and 225.05110 (3-nitro-L-tyrosine) with an isolation window of 2.0 m/z.The resolution was set to 17,500 at m/z 200, the automatic gain control target was set to 1e6 charges with a maximum injection time of 200 ms.Data acquisition was conducted using Thermo Orbitrap™ Scientific™ 7.2 CDS), data analysis was done with Max™ 3.0.63software (Thermo Fisher Scientific, Waltham, MA, USA) with a mass tolerance of 20 ppm.

Whole cell proteomics
Primary DCs were infected with H. pylori wt or the ΔrfaE mutant for 16 h.To obtain sufficient cell numbers (~1.4-1.8 × 10 6 DCs), DCs of two to three donors were pooled and subsequently subjected to proteomics analyses.Three individual experiments including two to three donors per experiment were performed.Sample preparation has been performed as previously described. 30Cs were lysed in 100 mmol/L triethylammonium bicarbonate buffer (TEAB pH 7.55; Sigma-Aldrich) containing 5% (w/w) sodium dodecyl sulfate (Sigma-Aldrich) and 1× cOmplete Mini EDTA-free protease inhibitor cocktail (Roche).Additionally, the samples were heated for 5 min at 95°C, followed by sonication in a Bioruptor device (Diagenode) for 5 min.Thereafter, the samples were centrifuged at 14,000 g and the protein content in the supernatant was quantified using a Pierce BCA Protein assay kit (Thermo Fisher Scientific).Next, the lysates were treated with 40 mmol/L dithiothreitol (Sigma-Aldrich) for 10 min at 95°C in order to reduce disulfide bonds, followed by alkylation of the cysteine residues by adding 80 mmol/L iodoacetamide and incubation at 22°C in the dark for 10 min.By acidification to pH ≤ 1 with 12% (v/v) ortho-phosphoric acid (Merck) and by a 1:7 (v/v) dilution in 100 mmol/L TEAB (pH 7.55) in 90% (v/v) methanol (Sigma-Aldrich), the proteins were precipitated.In order to purify proteins, suspension trapping was employed utilizing S-Trap mini columns (Protifi) according to the manufacturer's guidelines.Proteolysis was done employing the protease trypsin (sequencing grade modified, porcine, Promega) at a protease/ protein ratio of 1:20 (w/w) at 37°C for 15 h.A vacuum centrifuge was utilized to dry the peptides at 45°C.Thereafter samples were resuspended in 100 mmol/L TEAB (pH 8.5) to obtain a concentration of 1.0 µg/µL.Each sample was analyzed in three technical replicates by highperformance liquid chromatography coupled to mass spectrometry (HPLC-MS).In detail, 1.0 µg of sample was injected into an UltiMate™ 3000 RSLCnano System (Thermo Fisher Scientific) and chromatographically separated by employing reversed phase HPLC using an Acclaim™ PepMap™ 100 C18 HPLC column (500 × 0.075 mm i.d., Thermo Fisher Scientific).For separation, solvent A [0.10% aqueous formic acid (Sigma-Aldrich)] and solvent B [0.10% formic acid in acetonitrile (VWR)] were pumped in the following order at a flow rate of 300 nL/ min: a linear gradient from 1.0-22.0%B in 150 min, followed by a second linear gradient from 22.0-40.0%B in 70.0 min, and a third linear gradient from 40.0-90.0%B in 10.0 min.Thereafter, the column was flushed with 90.0%B for 20 min and re-equilibrated with 1.0% B for 50 min.The temperature of the column was maintained at 50°C.The HPLC system was hyphenated to a Q Exactive™ Plus Hybrid Quadrupole-Orbitrap™ mass spectrometer with a Nanospray Flex™ ion source (both from Thermo Fisher Scientific).The ion source was equipped with a SilicaTip emitter (360 µm o.d., 20 µm i.d. and a tip i.d. of 10 µm, CoAnn Technologies Inc.).The spray settings were as follows: voltage of 1.5 kV, S-lens RF level of 55.0 and capillary temperature of 320°C.The mass spectrometry data were acquired in datadependent mode.Scan cycles consisted of: full scan (range of m/z 400-2,000 and a resolution setting of 70,000 at m/z 200), followed by 15 data-dependent higher-energy collisional dissociation (HCD) scans (2.0 m/z isolation window at 32% normalized collision energy, resolution setting of 17,500 at m/z 200).The automatic gain control (AGC) target was set to 3e6 charges for the full scan, for the HCD scans to 1e5 charges and a maximum injection time of 100 ms for both.Additionally, we excluded already fragmented precursor ions for 30 seconds.Thermo Scientific™ Chromeleon™ 7.2 CDS (Thermo Fisher Scientific) was used for data acquisition.

Data analysis
For data evaluation, MaxQuant 2.0.1.0 31was used in default settings for label-free quantification (LFQ), employing a protein database from the Uniprot consortium 32 Homo sapiens from 19.12.2022 applying a 1% false discovery rate.The obtained protein groups were processed using the Perseus software platform. 33First, the protein groups were filtered by removing proteins that were only identified by site and reverse sequence matches.Next, the intensities were log2 transformed and normalized by subtraction of the median.
Analysis for differential protein expression was performed using linear models as part of the limma package 34 in R. Initial data exploratory analysis revealed clustering of data by the experimental run, indicating that a batch-effect was present.The batcheffect was corrected for by the removeBatchEffect() function within limma.Uniprot protein IDs were matched with their respective gene IDs downloaded from the Uniprot database 32 for Homo sapiens and Helicobacter pylori reviewed (SwissProt) entries.Gene set enrichment analysis was performed using the fgsea 35 package.The following databases were downloaded from the EnrichR database 36,37 for the fgsea analysis: WikiPathways_2019_Human, NCI-Nature _2016, TRRUST_Transcription_Factors_2019, MSigDB_Hallmark_2020, GO_Cellular_Com-ponent_2018, CORUM, KEGG_2019_Human, TRANSFAC_and_JASPAR_PWMs, ENCODE_and_ ChEA_Consensus_TFs_from_ChIP-X, GO_ Biological_Process_2018, GO_Molecular_Function_ 2018, Human_Gene_Atlas.

Code and data availability
The raw mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium (http://proteomecentral.proteomex change.org.)via the PRIDE partner repository 38 with the dataset identifier PXD046603.All analysis scripts for the differential expression analysis and gene set enrichment analysis can be freely accessed in the following GitHub repository: https://github.com/VSchaepertoens/proteomics_limma.

Statistics
Statistical details can be found in the Figure Legends.Data are presented as dots representing individual donors with bars indicating the mean ± standard deviation (SD).GraphPad Prism 10 Software was used for statistical.Student's T-test was performed to compare two groups while differences between multiple stimulation groups were analyzed via repeated measures one-way ANOVA including appropriate post-hoc tests.The term "repeated-measures" strictly applies only when treatments are given repeatedly to each subject, which was not done in this study.The term randomized block is used when you randomly assign treatments within each group (block) of matched subjects.As we are using GraphPad Prism for statistical analysis and the analyses are identical for repeated-measures and randomized block experiments, and Prism always uses the term repeatedmeasures, we also chose this term.p values < 0.05 were considered significant (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.001).

Dendritic cells are present in Helicobacter pylori-infected gastric epithelium
Using the GeoMx digital spatial profiler, we analyzed gastric biopsies from gastritis patients, categorized as H. pylori positive (Hp + ) or H. pylori negative (Hp − ).Immunohistochemistry and immunofluorescence staining reveals infiltration by immune cells (CD45 + ) into the gastric tissue of both Hp − and Hp + gastritis patients (Figure 1(a,b), Supplementary Figure S1).To classify the infiltrating immune cells, CD45 + cells of selected regions of interest (ROI) were harvested.Three representative ROIs for one Hp − and one Hp + patient are outlined with white lines in Figure 1(a).Immune cell profiling was performed on selected ROIs of Hp − and Hp + gastritis patients (Figure 1(c)).Spatial profiling of gastric biopsies revealed that several types of immune cells, including CD11c + cells, are recruited at a higher rate to the mucosa of H. pylori positive compared to H. pylori negative individuals (Figure 1(d)).The significant increase of the DC lineage marker CD11c in H. pylori positive compared to H. pylori negative tissue indicates that more DCs are migrating into the tissue upon H. pylori infection (Figure 1(d)).Additionally, the expression of CD11c (DCs) correlates with CD3 and CD20 (T and B lymphocytes respectively) within the regions analyzed (Figure 1(e)).Interestingly, we also observed significant correlations between CD11c and HLA-DR, the T cell marker CD4, and the proliferation marker Ki-67 (Figure 1(e,f)), suggesting crosstalk between DCs and adaptive immune cells in the selected regions.
Most studies investigating the interaction of H. pylori with human DCs make use of a widely accepted DC-model, monocyte-derived DCs (moDCs), which are in vitro differentiated for 7 days in the presence of IL-4 and GM-CSF and characterized by the expression of CD1a.Thorough investigations of DCs in the gastric mucosa revealed, however, that gastric DCs are positive not only for CD11c but also CD1c, 39,40 in contrast to CD1a + moDCs.Given that gastric DCs and primary CD1c + DCs directly isolated from the peripheral blood share characteristic markers, blood-derived CD1c + DCs are a suitable model to study the effects of H. pylori on human DCs.In previous studies, we have already shown that CD1c + DCs purified from peripheral blood are activated in vitro upon H. pylori infection. 8To compare infection of primary DCs in single culture with DC activation in the gastric context, we developed a mucosoid/DC co-culture platform based on the mucosoid culture system developed by Boccellato and colleagues, 28 which provides a highly physiological representation of the gastric lining (Figure 1 (g)).We added human primary DCs in a drop of Matrigel on the basal side of the mucosoid, just below the filter on which the mucosoid is cultured (Figure 1(g)).This co-culture revealed that infection with H. pylori at the apical side of the mucosoid robustly activates DCs cultured at the basal side, as indicated by increased CD80, CD86, and PD-L1 expression (Figure 1(h)).This activation signature broadly recapitulates the DC phenotype generated when DCs are infected in single culture (Figure 1(i)).This set of data reveals that DCs are highly abundant in the gastric mucosa of H. pylori + gastritis patients and that H. pylori infection of primary CD1c + DCs resembles the DC activation in the context of an in vitro representation of the H. pylori infected gastric lining.

ADP-heptose dampens Th1-associated responses during bacterial infection
Upon infection of the gastric mucosa, a critical step in host pathogen recognition is the detection of common microbial structures, known as PAMPs, by both the gastric epithelium and innate immune cells.While there are many well-characterized PAMPs that have been studied for decades, it is only in recent years that the bacteria-derived sugar ADP-heptose has been proposed as a new PAMP that activates and modifies epithelial cells. 17Accordingly, several publications have described the consequences of ADP-heptose stimulation in the epithelial compartment; however, information on its effects on hematopoietic, especially primary human innate immune cells, is scarce.Given the high abundance of DCs in the gastric mucosa of H. pylori infected patients, this study aims to thoroughly investigate the capacity of ADP-heptose to trigger the activation of human primary DCs.Since ADP-heptose is an intermediate in LPS biosynthesis, the recognition of ADP-heptose by DCs will mostly occur in the context of bacterial infection.Thus, we first monitored DC activation upon infection with H. pylori wild type (wt) or an H. pylori mutant that fails to synthesize ADP-heptose (ΔrfaE).Contrary to our expectation, infection with the ΔrfaE mutant (i.e., absence of ADP-heptose) resulted in more pronounced cytokine secretion (Figure 2(a)), altered chemokine release (Figure 2(b)) as well as increased surface marker expression (Figure 2(c)) compared to infection with H. pylori wt.In particular, the release of IL-1β, IL-12p70, TNFα, IL-10 and CCL3 as well as expression of CD40 and CD80 was enhanced, while CCL2 and CXCL1 levels were decreased and CXCL8 remained unchanged (Figure 2(a-c)).To further investigate these unexpected findings, we infected DCs with A. lwoffii, a Gram-negative bacterial species that is naturally devoid of ADP-heptose, 41,42 as verified by mass spectrometry (Figure 2(d)).Similar to infection with the H. pylori ΔrfaE mutant, infection of DCs with A. lwoffii resulted in enhanced secretion of IL-12p70 as well as CD40 expression compared to infection with H. pylori wt (Figure 2(e)).As we monitored a correlation of DCs and T cells in gastric biopsies of H. pylori infected patients (Figure 1(e,f)) we next investigated whether the increase in DC activation affects the capacity of DCs to induce T cell polarization.We therefore performed co-culture experiments involving DCs and allogenic naïve CD4 + T cells.We observed increased IFNγ production upon infection with ADP-heptose-deficient bacteria (H.pylori ΔrfaE mutant or A. lwoffii) compared to H. pylori wt (Figure 2(f,g)).Finally, we investigated whether the presence of ADP-heptose during infection of the mucosoid/DC culture affects the activation of T cells.Therefore, we infected the mucosoid/DC culture with H. pylori wt or ΔrfaE mutant for 40 h and then isolated the DCs for subsequent co-culture with naïve CD4 + T cells (Figure 2(h)).Subsequently, we analyzed CXCL8 expression in the epithelial compartment of the mucosoid/DC co-cultures upon infection with H. pylori wt or the ΔrfaE mutant as well as the T cell phenotype.In line with previous studies on the effects of ADP-heptose in epithelial cells, 17 this experiment revealed that infection with the H. pylori ΔrfaE mutant results in reduced CXCL8 expression in the mucosoid compared to H. pylori wt infection (Figure 2(i)); however, we found that DCs isolated from the mutant-infected cultures are more potent inducers of IFNγ production in T cells (Figure 2(j)).These data suggest that the presence of ADP-heptose during bacterial infection promotes the release of CXCL8 from epithelial cells, but simultaneously attenuates DC activation and subsequent T cell stimulation, rather than acting as a classical immunostimulatory PAMP in DCs.

ADP-heptose reduces pathogen-induced dendritic cell activation
In epithelial cells, others showed that ADP-heptose can effectively enter a host cell to trigger inflammation via its cognate receptor ALPK1, activating the classical or alternative NF-κB signaling pathway. 15,17,19,20Since we observed increased DC activation upon infection with ADP-heptose-deficient bacteria, we next asked whether ADPheptose alone has any effect on human primary DCs.As sentinels of the immune system, DCs are well-equipped with PRRs and very sensitive to minute amounts of PAMPs. 43Accordingly, when we exposed DCs to a low concentration of E. coli LPS, it induced gene transcription and secretion of a wide variety of pro-inflammatory mediators as well as surface expression of co-stimulatory molecules (Figure 3(a), Supplementary Figure S2).In contrast, ADP-heptose induced none of the tested pro-inflammatory mediators after 4 to 48 h of infection (Figure 3(a), Supplementary Figure S2) except for low levels of CXCL8, as reported in epithelial cells. 17,19This indicates a highly limited capacity of ADP-heptose to activate human primary DCs.However, in line with previous reports in epithelial cells, 16 we found that the minimal amounts of CXCL8 released by DCs upon ADPheptose treatment depend on the ALPK1/TIFA/ TAK1 pathway, as pharmacological inhibition of TAK1 by Takinib diminishes CXCL8 secretion (Figure 3(b)).Still, although ADP-heptose stimulation induced the production of minimal levels of CXCL8 (Figure 3(a)), this appears to be negligible in the context of infection, as we found no difference in CXCL8 secretion between wt-and ΔrfaEinfected DCs (Figure 2(b)).
As ADP-heptose clearly does not act as a potent immunostimulatory PAMP in DCs, we investigated whether the addition of exogenous ADPheptose could attenuate DC activation in response to the H. pylori ΔrfaE mutant or A. lwoffii.Therefore, we infected human DCs with H. pylori wt or the ΔrfaE mutant in the presence of exogenous ADP-heptose.Intriguingly, we found that the up-regulation of IL-12p70 and CD40 attributed to ΔrfaE-infected DCs was reversed by the addition of ADP-heptose, to levels observed following H. pylori wt infection (Figure 3(c,d)).Similarly, the addition of exogenous ADP-heptose also reduced the cytokine secretion and surface marker expression in response to the ADP-heptosedeficient bacterium A. lwoffii (Figure 3(e,f)).To investigate whether exogenous ADP-heptose dampens cytokine and chemokine secretion via the conventional ALPK1/TIFA/NF-κB pathway, we cultured DCs in the presence of exogenous ADPheptose as well as the TAK1 inhibitor Takinib.Co-treatment of DCs with ADP-heptose and Takinib did not affect the ADP-heptose-induced inhibition of IL-12p70 secretion (Figure 3(g)), indicating that the inhibitory functions of ADP-heptose on DC activation are independent of TAK1 signaling.

TLR2 mediates Helicobacter pylori uptake in human dendritic cells and enhances CD40 expression
Prior to the identification of ALPK1 as the cognate receptor for ADP-heptose, Ryzhakov and colleagues described that murine Alpk1 -/-macrophages infected with Helicobacter hepaticus express significantly higher levels of pro-inflammatory mediators than wt counterparts. 22Thus, the ALPK1 knockout in macrophages mirrors the phenotypic observations in DCs infected with the ADPheptose deficient H. pylori mutant.Interestingly, the latter study also showed that TLR2 inhibition reduced the activation of H. hepaticus infected Alpk1 -/-macrophages by a previously unknown mechanism, 22 suggesting that TLR2 signaling promotes hyperactivation of Alpk1 -/-macrophages.To study the potential role of TLR2 in DC activation upon infection with the ADP-heptose deficient H. pylori mutant, we first monitored TLR2 expression levels.These analyses revealed that while TLR2 mRNA expression was increased upon infection with both H. pylori strains (Figure 4(a)), levels on the cell surface were significantly diminished compared to untreated DCs (Figure 4(b)).To investigate whether TLR2 is internalized upon H. pylori infection, TLR2 localization as well as bacterial uptake by DCs was analyzed by confocal fluorescence microscopy.These analyses revealed that after 1 h of infection, TLR2 is found on the cell surface but also inside the cell upon H. pylori wt and mutant infection (Figure 4(c)).Moreover, both H. pylori strains were localized within DCs 1 h post-infection (Figure 4(d)).To investigate whether TLR2 mediates H. pylori uptake, a TLR2 neutralizing antibody was added to the culture 20 min prior to infection and bacterial uptake was visualized 1 h post-infection.Antagonizing TLR2 during H. pylori infection of human DCs resulted in reduced intracellular levels of both H. pylori wt and the ADP-heptose deficient mutant (Figure 4 (e)), indicating that TLR2 is involved in bacterial uptake upon H. pylori infection of primary human DCs independent of ADP-heptose.Finally, to assess the impact of TLR2 activation and subsequent bacterial uptake on DC activation, we monitored the DC phenotype upon infection in presence of the TLR2 neutralizing antibody.TLR2 neutralization significantly decreased the expression of H. pylori ΔrfaE-induced CD40 after 16 h of infection, while IL-12p70 secretion was not altered (Figure 4(f,g)).In contrast, in H. pylori wtinfected DCs, neither CD40 expression nor IL-12p70 release was affected by TLR2 neutralization.This set of data clearly shows that TLR2 is involved in H. pylori uptake and that TLR2 signaling is particularly important for enhanced CD40 expression in DCs infected with the ADP-heptose deficient mutant.

ADP-heptose attenuates type I IFN signaling in dendritic cells
To further elucidate how ADP-heptose exerts its unexpected attenuating effect on H. pylori-induced DC activation, we subjected untreated, H. pylori wt-and ΔrfaE-infected DCs to whole cell proteomics.We found that most proteins were regulated similarly by infection with wt and the ΔrfaE mutant, which was to be expected as they differ only in their capacity to synthesize ADP-heptose (Supplementary Figure S3).Analysis of the proteins that are differentially expressed upon infection with the two genotypes revealed that most of these proteins are more abundant in wt-infected cells (Figure 5(a)).However, as we observed a more pronounced DC activation in ΔrfaE-infected cells, we reasoned that this would likely be caused by the proteins that were more abundant after infection with the H. pylori mutant, which included ISG15, MX1and MX2 (Figure 5(b)).Accordingly, we performed a gene enrichment analysis to identify signaling pathways contributing to the pronounced DC activation (Figure 5(c)).Interestingly, we found that the significantly regulated pathways are almost exclusively associated with type I IFN signaling (Figure 5(c)).Type I IFN signaling is known to promote the formation of IFNstimulated gene factor 3 (ISGF3), a transcriptionally active multi-protein-complex consisting of STAT1, STAT2 and IRF9.Translocation of ISGF3 into the nucleus and subsequent binding to IFN-stimulated response elements (ISRE) drives the transcription of type I IFN and other IFN-stimulated genes (ISGs). 44he fact that type I IFN signaling is activated by both H. pylori wt and the ADP-heptose deficient ΔrfaE mutant, but more pronounced in the latter infection, is evident not only from the proteomics data (Figure 5(b)) but can also be shown by enhanced phosphorylation as well as protein expression of STAT1, STAT2 and IRF9 (Figure 5 (d)).Moreover, the expression of type I IFN target genes, IFNA2 and ISG15, as well as IFNα and CXCL10 secretion is significantly increased by infection with the ΔrfaE mutant compared to H. pylori wt infection (Figure 5(e,f)).Furthermore, it is well established that pathogen-induced type I IFN results in ISG expression but also in IFNα and IFNβ secretion resulting in an autocrine amplification loop. 45This amplification loop seems to be activated upon infection with the ADP-heptose -deficient mutant, as blocking of the type I IFN receptor decreases IFN signaling as well as IFNdependent target gene expression (Supplementary Figure S4).To prove that ADP-heptose is indeed capable of dampening bacteria-induced type I IFN signaling, we treated ΔrfaE-infected DCs with ADPheptose and analyzed type I IFN responses.As expected, the addition of ADP-heptose to ΔrfaEinfected DCs reduced the secretion of IFNα and CXCL10 (Figure 5(g)) as well as type I IFN signaling to the levels observed in H. pylori wt infections (Figure 5(h)).This set of data suggests a role for type I IFN in H. pylori-induced immune responses and identifies ADP-heptose as a negative regulator of type I IFN signaling.
We next characterized the contribution of type I IFNs to the H. pylori-induced DC phenotype.We found that addition of IFNα and -β increases wtinduced IL-12p70 secretion to levels obtained upon infection with the ADP-heptose-deficient mutant (Figure 5(i)).These data indicate that activation of the type I IFN signaling pathway is a pre-requisite for full-fledged DC activation in the context of H. pylori infection.Moreover, our flow cytometry data (Figure 5(j)) as well as the proteomics data revealed that HLA-class I family members are increased upon H. pylori infection (Figure 5(k)).Although type II conventional DCs used in this study are predominantly reported to be involved in activation of CD4 + T cells, 46 the fact that HLAclass I expression (Figure 5(b,k)) as well as the secretion of type I IFNs is significantly enhanced in mutant-infected DCs (Figure 5(f)), prompted us to investigate the impact of ADP-heptose on CD8 + T cell activation.In line with our expectations, PBMCs infected with the H. pylori mutant lacking ADP-heptose secreted significantly higher levels of Granzyme B (GZMB) (Figure 5(l)).In addition, coculture of mutant-infected DCs with total CD3 + T cells revealed a significant increase in Granzyme B production in CD8 + T cells in comparison to coculture with wt-infected DCs (Figure 5(m)), suggesting that ADP-heptose dampens the H. pyloriinduced CD8 + T cell activity.Accordingly, while we already showed that several immune cell subsets are recruited at a higher rate to the gastric epithelium of individuals suffering from H. pylori positive gastritis, we did not observe a significant increase in CD8 + T cells compared to H. pylori negative individuals (Figure 1(d)).In line with this, we do not observe a clear correlation between CD8 and Granzyme B in H. pylori positive gastritis, while CD8 and Granzyme B do correlate in H. pylori negative gastritis samples (Figure 5(n)), indicating that H. pylori infection might impair the effector functions of CD8 + T cells in the gastric mucosa.Taken together, these findings suggest that ADP-heptose prevents fully functional type I IFN signaling in H. pylori-infected DCs, which impairs DC activation and the subsequent CD8 + T cell responses.

Discussion
In this study, we showed that the newly described PAMP, ADP-heptose, does not have the activating capacity of a bona fide PAMP in primary human DCs.Instead, ADP-heptose attenuates bacteriainduced activation of DCs and the subsequent Th1associated T cell response.The effects of H. pyloriderived PAMPs such as ADP-heptose on DCs are particularly interesting in light of the finding that DCs as well as several types of lymphocytes are significantly more abundant in H. pylori-infected gastritis samples compared to H. pylori negative biopsies, which underscores the pivotal role of the immune system in the H. pylori-induced chronic inflammatory response.
The recognition of PAMPs by their cognate PRRs is a key feature of the innate immune system, enabling rapid anti-microbial responses to a wide variety of pathogens.While it is well established that PAMPs are potent activators of DCs, our data suggest that ADP-heptose exerts unexpected and yet unexplored effects in primary DCs.In epithelial cells, ADP-heptose is recognized by ALPK1 and induces pro-inflammatory responses.Moreover, H. pylori mutants devoid of ADP-heptose are less potent in triggering inflammatory responses in epithelial cells. 15,17The ADP-heptose receptor ALPK1 has also been described to drive inflammatory responses, as infection with the Gram-negative bacterium Burkholderia cenocepacia increased cytokine and chemokine expression in the lungs of wild-type mice, while these responses were compromised in Alpk1 −/− mice. 17Additionally, initial studies using THP-1 cells, as a surrogate for human innate immune cells, as well as primary monocytes and monocyte-derived macrophages suggested that ADP-heptose exerts similar pro-inflammatory effects on leukocytes. 47Taken together, these studies all point toward an immune-stimulatory role for ADP-heptose and ALPK1 signaling.On the other hand, recent reports have significantly increased our understanding of the multiple functions of ALPK1.Ryzhakov and colleagues showed that ALPK1 attenuates Th1 responses to Helicobacter hepaticus, by impairing the activation of innate immune cells, suggesting a suppressive role of ALPK1 in hematopoietic cells. 22Moreover, a recent report identifies ALPK1 signaling to be crucial for maintaining intestinal homeostasis after infection with the Gram-negative commensal Akkersmania muciniphila. 48Further elaborating on potential immunomodulatory effects of ADPheptose/ALPK1, we show here that ADP-heptose alone is a poor activator of primary DCs.Since we could not detect ALPK1 or TIFA by whole-cell proteomics of H. pylori infected DCs, the poor activating capacity of ADP-heptose alone may be due to limited availability of the pathway in human primary DCs.Furthermore, in the context of a bacterial infection, it could be speculated that H. pylori is rapidly engulfed by DCs due to the strong phagocytic activity of DCs.This rapid uptake by the cell may locate ADP-heptose to phagosomes rather than the cytosol, making it accessible to receptors other than cytosolic ALPK1.Strikingly, bacteria devoid of ADP-heptose are even more potent in activating DCs and subsequent Th1-associated immune responses, while supplementation of ADP-heptose seems to reverse this enhanced effect on DC activation.Thus, existing data allow classification of ADP-heptose as a PAMP in the sense of a conserved microbial substance shared by different bacterial species, which is activating to epithelial cells.Regarding the activation of DCs, however, the ADP-heptose /ALPK1 axis does not seem to possess proinflammatory characteristics of a classical PAMP/ PRR family member.
The importance of TLR2 for the activation of DCs in the context of H. pylori infection has been previously demonstrated by using TLR2 knock-out mice and by blocking TLR2 in human DCs.While significant decreases in IL-1β, IL-6, IL-12, IL-23, and TNFα release were observed in H. pyloriinfected murine BM-DCs from TLR2ko mice, 49,50 in particular the release of TNFα and GM-CSF was significantly impaired upon TLR2 blocking in human CD1c + DCs. 8 In addition, it has been shown that increased macrophage activation observed in Alpk1 −/− mice upon infection with Helicobacter hepaticus was blocked in the presence of anti-TLR2 antibodies. 22Along this line TLR2blocking also impairs the activation of DCs infected with the ADP-heptose deficient mutant.In an attempt to explain the contribution of TLR2 during H. pylori infection, we show here that TLR2 is involved in bacterial uptake, accordingly TLR2 has already been described to mediate uptake of Staphylococcus aureus and Pseudomonas aeruginosa. 51,52nterestingly, a recent study reports that H. pylori has the capacity to inhibit signaling induced by the intracellular PRRs STING and RIG-I, both of which are known to induce type I IFNs. 53While it is known that H. pylori induces type I IFNs, 54,55 Dooyema and colleagues showed that STING agonist-induced expression of ISGs is reduced upon co-infection with H. pylori in human organoids. 53In line with this finding, we report here that addition of ADP-heptose dampens the increased activation of type I IFN signaling upon infection with the ADP-heptose deficient mutant ΔrfaE to levels observed upon wt infection.This suggests that during infection with H. pylori wt, ADP-heptose might cause low levels of type I IFN signaling.Type I IFNs are well-established in antiviral defense, but when it comes to host susceptibility to bacterial infections, type I IFNs can have protective or deleterious effects.Recent advances reporting harmful consequences describe that type I IFNs suppress antibacterial activity upon infection with both Gram-positive and Gram-negative bacteria 56 and exacerbate tuberculosis in a coinfection setting. 57Accordingly, diminished susceptibility to tuberculosis has been reported in humans harboring genetic variation in the IFNAR1 gene. 58In contrast, type I IFNdependent signaling was shown to be triggered by microbiota in the skin and contributes to wound healing. 59Moreover, STING-mediated type I IFN responses correlate with protection against Streptococcus pyogenes infection. 60In the context of H. pylori, a protective role was suggested as well, in that IFNAR-deficient mice exhibit increased susceptibility to H. pylori. 614][65] Yet, upon chronic infection with H. pylori, CD8 + T cells lose their resident memory phenotype and seem to be replaced by CD4 + T cells. 62,66lthough conventional DCs used in this study, are not known as the main activators of CD8 + T cells, recently a novel DC phenotype has been described, termed ISG + DC, which are cDC2s characterized by increased expression of IFN stimulated genes. 67In a tumor context, MHC class I dressed ISG + DCs exert enhanced capacity to stimulate CD8 + T cells, resulting in effective anti-tumor T cell immunity. 67Moreover, DCs from H. pyloriinfected mice are less potent in promoting tumorspecific CD8 + T cell proliferation, resulting in impaired efficacy of cancer immunotherapies. 68In light of these studies, the finding that ADP-heptose attenuates type I IFN signaling upon H. pylori infection is of great interest, as it suggests a potential mode of action by which H. pylori mitigates CD8 + T cell responses.Further studies are necessary to identify whether the ADP-heptose driven attenuation of type I IFN responses mediates the effects on CD8 + T cell responses and to clarify whether ADP-heptose affects host susceptibility to H. pylori.
Taken together, our study findings emphasize the role of DCs in the gastric mucosa and show for the first time that ADP-heptose, despite its potent proinflammatory functions in epithelial cells, does not act as a bona fide PAMP in human primary DCs.Rather, ADP-heptose attenuates bacteria-induced immune responses by disrupting the positive feedback loop enforced by type I IFN signaling, thus dampening DC activation and the subsequent T cell response.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Dendritic cells are recruited and activated upon Helicobacter pylori infection.(a) Immunohistochemical and immunofluorescence staining of FFPE sections of gastric biopsies from three H. pylori negative (Hp − ) and three H. pylori positive (Hp + ) gastritis patients.One representative out of 3 is shown.Five regions of interest (ROIs) per patient were harvested, selected ROIs are outlined with white lines in the lower panel.(b) Fluorescence staining of gastric biopsies using PanCK for epithelial cells (green), CD45 to identify immune cells (red) and Syto13 nuclear staining (blue).(c) Graphical depiction of GeoMx technology.CD45-positive areas were exposed to UV light to cleave DNA tags and oligos were quantified using the nCounter system.(d) CD45 + areas of 15 ROIs from Hp − and Hp + biopsies (3 patients, 5 ROIs per patient) were harvested and analyzed using the nCounter system.Fold change vs. significance of all tested markers in CD45 + regions comparing Hp + and Hp − gastritis patients are shown.(e) Pearson correlation of all tested markers in the CD45 + regions of Hp + positive gastritis patients.Only significant correlations (p ≤ 0.05) are depicted).Red/blue annotations indicate positive/negative correlations, respectively.(f) Correlation of normalized counts of CD11c and CD4 or HLA-DR, respectively.Dots represent values of 15 individual ROIs.(g) Schematic depiction of the mucosoid/DC co-culture model.The mucosoid is cultured on a filter and infected with H. pylori, and DCs are cultivated in a drop of matrigel on the basolateral side.Created in BioRender.Neuper, T. (2022) BioRender.com/u31t534(h,i) Surface marker expression of DCs cultured in the mucosoid/DC model (h) or in DC single culture (i) upon H. pylori infection.One representative donor of 2 (h) and 4 (i) individual donors is shown.

Figure 2 .
Figure 2. Lack of ADP-heptose increases bacteria-induced Th1-associated immune responses.(a-c) Human CD1c + DCs were infected with H. pylori (wt) or the ADP-heptose deficient mutant ΔrfaE (Δ) at a MOI of 5 for 16 h and cytokine (a) and chemokine (b) as well as surface marker expression (c) was analyzed by multiplex and flow cytometry, respectively (n = 8-21).(d) Extracted ion current chromatograms of ADP-heptose (m/z 618.0850) in bacterial lysates and in a commercially available standard at a concentration of 100 pg/µl.One out of two experiments is shown.(e) DCs were infected with H. pylori (Hp) or A. lwoffii (Al) at an MOI of 5 and IL-12p70 secretion and CD40 expression were analyzed 16 h post infection (n = 4).(f,g) DCs were infected with H. pylori (Hp wt), the ADPheptose deficient mutant (Hp Δ) or A. lwoffii (Al) (MOI5).After 16 h allogenic, naive CD4 + T cells were added and DCs and T cells were co-cultured for another 6 days, before CD4 + /IFNγ + T cells were quantified by flow cytometry (n = 4).(h) Graphical depiction of a mucosoid/DC co-culture, and a DC/T cell co-culture.Created in BioRender.Neuper, T. (2022) BioRender.com/u31t534(i) CXCL8 mRNA expression was analyzed in the mucosoid 40 h after infection with H. pylori (wt) (MOI 100) or the ADP-heptose deficient mutant (Δ) (n = 2).(j) IFNγ production was analyzed by flow cytometry in CD4 + T cells co-cultured with DCs re-isolated from the mucosoid/DC coculture.One out of two independent donors is shown.Bars indicate mean±SD, dots represent individual donors.For statistical analysis repeated measures one-way ANOVA with a šídák's post-hoc test was performed.

Figure 4 .
Figure 4. TLR2 mediated Helicobacter pylori uptake is essential for potent dendritic cell activation.(a) TLR2 mRNA expression upon stimulation with H. pylori wt or mutant (Δ) (MOI 5) at indicated time points.Mean±SD of three individual donors is shown.(b) TLR2 surface expression was monitored by flow cytometry 16 h post-infection (MOI 5) (n = (c,d) H. pylori strains were stained with eFluor670 proliferation dye prior to infection.One hour post infection with H. pylori wt or the ADP-heptose-deficient mutant (Δ) (MOI 5), DCs were subjected to immunofluorescence and stained for CD45, DAPI and TLR2.Internalization of bacteria as well as TLR2 localization was analyzed by confocal fluorescence microscopy.Orthogonal views of confocal z-stacks of one out of three representative donors are shown.Scale bar: 5 µm.(e) DCs were treated with a TLR2 neutralizing antibody 20 min prior to infection.After 1 h of infection with eFluor670 stained H. pylori (MOI 5), DCs were subjected to immunofluorescence and stained for CD45 and DAPI.Maximum intensity projections of confocal z-stacks of one representative out of three donors are shown.Scale bar: 5 µm.(f,g) DCs were treated with a TLR2 neutralizing antibody 20 min prior to infection with H. pylori (wt) or the ADP-heptose deficient mutant (Δ) at an MOI of 5.After 16 h CD40 expression (f) and IL-12p70 secretion (g) was monitored by flow cytometry and multiplex assay (n = 6).For statistical analysis one-way ANOVA with a šídák's post-hoc test was performed.

Figure 5 .
Figure 5. ADP-heptose attenuates dendritic cell activation by suppressing type I IFN signaling.(a) Heatmap showing z-scores of normalized intensities for differentially expressed proteins in DCs upon infection with H. pylori (wt) or the ADP-heptose deficient mutant (Δ) (MOI 5).Red/blue annotation for the direction of change indicates up/down regulation of differentially expressed proteins present in the comparison of Δ vs. wt, respectively (n = 3).(b) Volcano plot displaying proteins in the comparison of H. pylori mutant infected DCs (Δ) vs. H.pylori wildtype infected DCs (wt).The dashed horizontal line indicates the p adj cutoff < 0.05.Black dots represent single proteins and red large dots indicate proteins of interest (n = 3).(c) Bar graph with the normalized enrichment score, showing pathways enriched in DCs infected with the ADP-heptose deficient H. pylori mutant (Δ) (positive NES score) or enriched in H. pylori wt (negative NES score) (n = 3).(d) DCs were infected with H. pylori (wt) or the ADP-heptose deficient mutant (Δ) (MOI 5) and type I IFN signaling was monitored.Using Western Blot STAT1 and STAT2 phosphorylation was analyzed, as well as total protein levels of the ISGF3 components (STAT1, STAT2, IRF9) after 16 h.One out of five representative donors is shown.(e) IFNA2 and ISG15 mRNA expression was monitored by qPCR upon infection (MOI 5) at the indicated time points.Mean±SD of three individual donors is shown.(f) Secretion of IFNα and CXCL10 was monitored after 16 h of infection with the indicated strains (n = 14).(g) DCs were infected with H. pylori wildtype (wt) or the ADP-heptose deficient mutant ΔrfaE (MOI 5) and stimulated with 2.5 or 25 µg/ml ADP-heptose.Cytokine secretion was analyzed by multiplex assay (n = 3).(h) H. pylori (MOI 5) -induced type I IFN signaling (STAT1, STAT2, IRF9) was monitored upon addition of ADP-heptose (25 µg/ml) by Western Blot analysis.One representative out of three donors is shown.(i) DCs were infected with H. pylori (wt) and supplemented with IFNα and IFNβ (10 µg/ml each) or infected with the ADP-heptose deficient mutant (Δ) and IL-12p70 was analyzed after 16 h (n = 4).(j) Histograms of HLA-ABC surface expression after 16 h of infection with H. pylori wt or ΔrfaE (MOI 5) analyzed by flow cytometry.One representative out of three donors is shown.(k) Heatmap displaying log2FC of wt and ΔrfaE infected DCs (MOI 5) compared to untreated DCs.Red annotation indicates upregulation of the indicated proteins.(n = 3).(l) Peripheral blood mononuclear cells (PBMCs) were infected with H. pylori (wt) or the ADP-heptose deficient mutant (Δ) at a MOI of 0.1 and granzyme B (GZMB) secretion was monitored by ELISA after 6 days (n = 7).(m) DCs were infected with the indicated H. pylori strains (MOI 5) for 16 h, before allogenic pan T cells were added.After 6 days of co-culture granzyme B (GZMB) production was monitored using flow cytometry (n = 8).(n) Correlation of normalized counts of CD8 and granzyme B (GZMB) in H. pylori negative (Hp − , left panel) and H. pylori positive (Hp + , right panel) gastritis samples.Dots represent values of individual ROIs.Bars indicate mean±SD, dots represent individual donors.For statistical analysis Student's T-test or repeated measures one-way ANOVA with a šídák's post-hoc test was performed.