KATO III cells (gastric carcinoma, derived from a metastatic site and poorly differentiated) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and maintained in ISCOVE (Euroclone, ECM0192, Italy), supplemented with 20% (v/v) heat inactivated fetal bovine serum (FBS, Euroclone, ECS0180D, South America, origin EU approved) and Penicillin-Streptomycin solution (100 U/mL penicillin and 100 μg/mL streptomycin) (Euroclone, ECB3001D, Italy).

Blood samples were collected from healthy donors (aged 25–40 years) at Centro Trasfusionale, Ospedale S. Maria della Speranza, Battipaglia (Italy) in accordance with the agreement for the use of residual biological materials for research purposes (Prot, n. 0329152, Rep. 783/2023). Human peripheral blood mononuclear cells (PBMCs) were isolated using a previously published protocol (8). Cells were seeded into a petri dish overnight before stimulation.

Gastric mucosoids were generated from organoids using a previously established protocol (22). The organoids themselves were originally established from gastric cells isolated from resected tissue of obese patients at Helios Klinikum (Berlin-Buch, Germany) (Supplementary Table S1). Tissue collection was performed with the approval of the Ethics Committee of Charité University Medicine, Berlin (EA1/129/12).

For mucosoid cultures, 200,000 cells derived from antrum organoids were resuspended in 200 μL of culture medium (prepared as described in 22) and seeded onto collagen-coated (Gibco A10644-01, 15 μg/cm²) transwell inserts (Millipore, PIHP01250) placed in 24-well plates.

The space between the filter and the well was filled with 400 μL of culture medium (Supplementary Table S2). On day 3 post-seeding, the medium overlying the cells was removed from the well insert to start the Air-Liquid Interface (ALI) culture. Subsequently, the medium below the filter was replaced with 500 μL of medium in which the ROCKi concentration was reduced from 9 µM to 1.8 µM (Supplementary Table S2). At this step, only half of the medium volume was replaced regularly to maintain cell-secreted factors in the medium. Foveolar phenotype differentiation was induced by culturing them in a medium lacking WNT3A and R-spondin 1 for one week.

All cells were cultured at 37°C in a humidified incubator with 5% CO₂.

Helicobacter pylori P12 strain from a German patient with a duodenal ulcer was cultured on selective Columbia agar (Oxoid, CM0331, Basingstoke, Hampshire, UK) containing 7% (v/v) defibrinated horse blood (ThermoFisher Scientific, SR0050D, The Netherlands) supplemented with an antibiotic mix (DENT, Oxoid, SR0147E, Basingstoke, Hampshire, UK). Bacterial plates were incubated for 3–4 days in a capnophilic atmosphere with 10% CO₂. Once confluent on the plate, bacteria were scraped using Brain Heart Infusion (BHI, Oxoid, CM1135, Basingstoke, Hampshire, UK) with 10% (v/v) FBS (Euroclone, ECS0180D, South America, origin EU approved) and measured at an optical density at 600 nm (OD₆₀₀), considering 1 OD₆₀₀ = 1 × 10^⁸ bacteria/mL.

To obtain conditioned medium, PBMCs (2 × 10^⁶ cells/well), cultured in 12-well plates, were infected with H. pylori at a multiplicity of infection (MOI) of 1:1. The culture supernatant was collected 24 hours post-infection, and IFNγ concentration was measured to assess PBMCs activation, as previously described (8).

Total RNA was extracted using EuroGOLD RNA Pure reagent (Euroclone, S.p.A; EMR506200, Pero, Italy) according to the manufacturer’s instructions and quantified using a NanoDrop 1000 (ThermoFisher Scientific). Its integrity was assessed by electrophoresis on a 1% agarose gel. For cDNA synthesis, 0.2-1 µg of total RNA was reverse-transcribed using M-MLV Reverse Transcriptase (GeneSpin S.r.l., Milan, Italy). Quantitative PCR (qPCR) was performed on a QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific, MA, USA). Appropriate cDNA dilutions were used for each target gene in a 12 µL reaction volume, using Luna Universal qPCR Master Mix (New England Biolabs, USA). Data from three independent experiments, each performed in technical duplicates, were analyzed using the ΔΔCt method. HPRT1 was used as the endogenous reference gene. Primer sequences are listed in Supplementary Table S3.

To obtain total protein extracts, cells were resuspended in 1X Laemmli buffer (50 mM Tris-Cl pH 6.8; 2% w/v SDS; 0.1% bromophenol blue; 10% v/v glycerol and 100 mM β-Mercaptoethanol). Mucus from mucosoids was collected and diluted in Laemmli buffer to a final concentration of 2X. Samples were sonicated (1 minute, 10 sec pulse on, 10 sec pulse off, amplitude 28%, Vibra-Cell Sonics). Protein samples were then incubated at 100°C for 5 minutes, centrifuged at 10,000g for 10 minutes to remove cellular debris, loaded into a 10% or 15% polyacrylamide gel according to the proteins molecular weight for electrophoretic separation, and then transferred to an Amersham™ Protran™ Premium 0.45 μm NC (GE Healthcare Life Sciences, GE10600003, Germany) nitrocellulose membrane. Ponceau red staining (0.1% solution in 1% v/v acetic acid) for 5 min was used to evaluate the transfer efficiency. After blocking with 5% w/v of non-fat dry milk (BioRad) for 1 hour or 3% w/v of Bovine Serum Albumine (BSA) (Sigma-Aldrich) for 1 hour the membranes were incubated overnight with primary antibodies at 4°C. After washing three times with TBS-T (Tris-Buffered Saline with 0.05% Tween 20) for 10 min each, the membranes were incubated with secondary antibodies for 1 hour at room temperature, washed three times, and then visualized using a LAS 4000 digital imaging system (GE Healthcare, Waukesha, WI, USA). Band densities were quantified by ImageJ (National Institutes of Health, USA).

A full list of antibodies is provided in Supplementary Table S4.

For reporter gene assays, KATO III cells (1.5 × 10⁵ cells/well) were seeded in 24-well plates. After 24 hours, cells were transfected using Lipofectamine 2000 reagent (Invitrogen, #11668-027, USA), according to the manufacturer’s protocol. Cells were co-transfected with 1 μg of a pGL3-based plasmid (Promega, #E1751, USA) containing a 1 kb fragment of the TFF1 promoter (from –1036 bp upstream) driving luciferase expression, and 0.1 μg of a plasmid encoding β-galactosidase, used to normalize transfection efficiency. The day after transfection, cells were treated with the DNA methylation inhibitor 5-aza-2’-deoxycytidine (5-AZA, 10 μM) for a total of 24 hours, together with a 2-hour IFNγ (10 ng/ml) pulse.

Luciferase activity was measured using the Dual-Light Luciferase/β-Galactosidase Assay Kit (Applied Biosystems, Foster City, CA, USA). Relative transcriptional activity was expressed as the ratio of firefly luciferase to β-galactosidase activity. All measurements were performed in quadruplicate, and light emission was detected using an EnSpire Alpha Multimode Plate Reader (PerkinElmer, Waltham, MA, USA).

ChIP assays were performed using a SimpleChIP^® Enzymatic Chromatin IP Kit (Magnetic Beads) (Cell Signaling, 9003, USA) following the manufacturer’s protocol. Briefly, KATO III cells were treated with 1% paraformaldehyde for protein-DNA crosslinking and then with glycine (Cell Signaling, 7005S, USA) for neutralization. Samples were treated with Micrococcal Nuclease (Cell Signaling, 10011, USA) for 20 min at 37°C to obtain DNA fragments of approximately 150–900 bp. Nuclei were sonicated to break the nuclear membrane and chromatin was immunoprecipitated with anti-C/EBPβ antibody (Cell Signaling, #90081S, USA), anti-H3S10ph antibody (Cell Signaling, 53348, USA), anti-H3K27me3 antibody (Abcam, ab6002) or anti-H3K9ac antibody (Abcam, ab10812) respectively or control IgG overnight at 4°C and then incubated for 2 hours with ChIP-Grade Protein G Magnetic Beads (Cell Signaling, 9006, USA). DNA was eluted, de-crosslinked for 2 hours at 65°C, and purified. The enrichment of C/EBPβ binding or specific histone modifications at the TFF1 gene locus was quantitatively assessed by real-time PCR using the primer sequences listed in Supplementary Table S5.

Experiments were performed in three independent biological replicates; each measured in technical triplicate or quadruplicate. Results are presented as mean ± standard deviation (SD). Data analysis and graph generation were conducted using Microsoft Excel and GraphPad Prism 4 (GraphPad Software, La Jolla, CA, USA). Figures were assembled using Adobe Illustrator. Statistical significance was assessed using either an unpaired Student’s t-test or one-way ANOVA, as appropriate, with a p-value ≤ 0.05 considered statistically significant.

Our recent findings suggest that IFNγ, within the inflammatory microenvironment, drives the silencing of the tumor suppressor gene TFF1 via the transcription factor C/EBPβ (8). Here, we investigated whether a brief exposure to IFNγ could lead to long-lasting phenotypic changes, with a focus on the sustained repression of TFF1 expression.

To this end, KATO III gastric cancer cells were exposed to 10 ng/mL IFNγ for different pulses (30 minutes or 2 hours), followed by thorough washing and subsequent culture in cytokine-free medium for 24, 48, and 72 hours. As shown in Supplementary Figure S1A, both 30 min and 2-hour pulses were sufficient to induce a marked and sustained reduction in TFF1 protein levels, comparable to those observed under continuous IFNγ exposure. Notably, a 2-hour IFNγ pulse reduced TFF1 expression for up to 120 hours after cytokine washout.

Based on these results, a 2-hour IFNγ pulse with a 72 hours washout was selected for subsequent experiments (Supplementary Figures S1B, C). As shown in Figure 1, both TFF1 protein (Figures 1A, B) and mRNA levels (Figure 1C) remained markedly suppressed for up to 72 hours following treatment. In contrast, C/EBPβ transcript levels exhibited a strong and progressive increase, reaching up to a nine-fold induction 72 hours after IFNγ withdrawal (Figure 1D). Concomitantly, levels of STAT1 and phosphorylated STAT1 remain elevated even after IFNγ withdrawal (Supplementary Figure S2).

These findings indicate that even a brief pulse of IFNγ is sufficient to trigger a transcriptional program resulting in the persistent repression of TFF1. This long-lasting effect is likely mediated by sustained activation of transcriptional regulators such as C/EBPβ, despite the absence of ongoing cytokine stimulation.

The prolonged repression of TFF1 following IFNγ treatment prompted us to explore the potential involvement of epigenetic mechanisms. While only a limited number of studies have examined histone modification dynamics in epithelial cells in response to inflammatory cues, available evidence suggests that chromatin remodeling plays a central role in regulating gene expression under these conditions (23).

Therefore, we examined whether IFNγ suppresses TFF1 gene expression through chromatin-mediated mechanisms. Specifically, we focused on three histone H3 modifications implicated in TFF1 transcriptional regulation in various cellular contexts: phosphorylation at serine 10 (H3S10ph), trimethylation at lysine 27 (H3K27me3), and acetylation at lysine 9 (H3K9ac) (24–27).

Notably, IFNγ has been shown to promote H3K27 trimethylation, contributing to stable gene silencing in macrophages (28). Moreover, Li and colleagues (27) demonstrated that increased levels of H3K9ac and H3S10ph on the TFF1 promoter are associated with transcriptional activation.

To assess whether IFNγ influences these histone modifications at the global level, we performed Western blot analyses following a 2-hour cytokine pulse and subsequent incubation in fresh medium for 24 or 72 hours. As shown in Figure 2B, we observed a significant reduction in H3S10ph and a marked increase in H3K27me3 levels after 72 hours, suggesting a shift toward a more repressive chromatin state. In contrast, no notable changes were observed after 24 hours (Figure 2A), indicating that these epigenetic alterations require an extended post-treatment incubation period to become evident. This result suggests that the downregulation of TFF1 observed during the first 24 hours after cytokine withdrawal is not yet maintained through histone modifications, but is more likely driven by alternative mechanisms, such as the cell-surface catch and release of cytokines mediated by phosphatidylserine (3).

Spatial and temporal changes in histone modifications at the TFF1 gene were assessed by chromatin immunoprecipitation (ChIP) in KATO III cells collected 72 hours after a 2-hour IFNγ pulsed treatment. Two sets of primers were designed to target the upstream regulatory region, and one set targeted the first exon of TFF1. Additionally, three primer sets previously described by Li et al. (27), targeting the second and third exons (primers d, e, and f), were included (Figure 3A).

Phosphorylation of histone H3 at serine 10 (H3S10ph) is widely associated with transcriptional activation (29). Notably, following the transient cytokine stimulation, a significant reduction in H3S10ph levels was observed both at the TFF1 promoter and within the coding region (Figure 3B). These findings are consistent with previous evidence implicating H3S10ph in both transcription initiation and elongation (27). Given that H3S10ph often functions in cooperation with H3K9 acetylation (30), we hypothesized a similar trend for this modification in the context of the TFF1 promoter. Indeed, H3K9ac levels were significantly decreased in the coding region of TFF1, whereas at the promoter region, the reduction did not reach statistical significance (Figure 3C). This observation aligns with studies suggesting that H3K9ac, along with other acetylation marks, contributes to chromatin accessibility and may recruit transcriptional co-regulators, particularly during the elongation phase (31).

In contrast, trimethylation of H3K27 (H3K27me3), a mark typically associated with transcriptional repression and stable gene silencing, did not show significant alterations at any region of the TFF1 gene following IFNγ exposure (Figure 3D). The apparent discrepancy between the global increase H3K27me3 (Figure 2) and the absence of enrichment at the TFF1 locus (Figure 3D) suggests that IFNγ promotes a widespread chromatin repressive state, whereas the sustained silencing of TFF1 is driven primarily by specific histone modifications and DNA methylation.

Given the established role of IFNγ in promoting TFF1 silencing (8), we sought to determine whether cytokine-induced repression of TFF1 in our model also involves alterations in DNA methylation. This hypothesis is supported by evidence linking chronic Helicobacter pylori infection to CpG island methylation of tumor suppressor genes (20, 21), as well as by prior findings showing that TFF1 expression is critically regulated by promoter methylation (17).

In our study, we observed that TFF1 expression was downregulated in gastric cancer-derived KATO III cells upon exposure to IFNγ alone, whereas in primary human gastric mucosoids, its repression required a combination of TNF-α, IL-1β, and IFNγ (8). These findings highlight how cellular context (transformed versus primary epithelial cells) can influence cytokine responsiveness. They also emphasize the importance of modelling complex inflammatory environments, as exposure to a single cytokine may not be sufficient to replicate the multifaceted signaling that occurs during chronic infection. Our results suggest a central role for the host inflammatory response in silencing TFF1 during the later stages of infection.

Following this hypothesis, we investigated TFF1 expression in mucosoid cultures exposed to conditioned media derived from H. pylori-infected immune cells. Specifically, human peripheral blood mononuclear cells (PBMCs) isolated from healthy donors were infected with H. pylori at a multiplicity of infection (MOI) of 1:1. The conditioned medium was collected 24 hours post-infection and used to treat both KATO III and human gastric mucosoids.

KATO III cells showed reduced TFF1 expression after a 2-hour pulse of conditioned medium from H. pylori-infected PBMCs followed by a 72-hour washout, similar to the effect observed with continuous 72-hour exposure to the conditioned medium. This reduction was also comparable to that seen with IFNγ treatment (Supplementary Figure S3).

As shown in Figure 7, exposure of mucosoid cells to conditioned medium from H. pylori-infected PBMCs (Figure 7A) resulted in a marked downregulation of TFF1 expression at 3- and 5-days post-incubation (Figure 7B). Consistently, Western blot analysis of the mucus fraction (Figure 7C) revealed a significant decrease in secreted TFF1 levels in treated cultures compared to controls.

To determine whether the effects of inflammatory mediators persist after their removal, differentiated mucosoids were exposed to conditioned medium from H. pylori-infected PBMCs for three days. Following this treatment, the medium was replaced with cytokine-free medium, and the mucosoids were cultured for an additional three days (Figure 8A). As shown in Figure 8B, IL-8 expression was transiently induced during treatment but returned to baseline after the recovery period. In contrast, TFF1 expression continued to decline even after cytokine withdrawal (Figures 8D, E), whereas C/EBPβ remained persistently upregulated (Figure 8C). These findings reinforce the notion that, in primary epithelial cells, durable repression of TFF1 requires the complex cytokine milieu produced during infection rather than IFNγ alone.

The results indicate that this unique in vitro model (primary human gastric mucosoids incubated with conditioned media from H. pylori-infected human immune cells, enriched in pro-inflammatory mediators) provides a more physiologically relevant representation of chronic inflammation. This highlights the importance of considering the combined effects and interactions of multiple factors when examining gene expression changes under inflammatory conditions.