Human gastric tissue biopsies were collected from GC or GC-free patients enrolled at Monash Medical Centre (Melbourne, Australia) undergoing surgical resection or upper gastrointestinal endoscopy (16, 17). Patients with a history of taking nonsteroidal anti-inflammatory drugs, proton pump inhibitors, or antibiotics were excluded. Biopsies were either snap-frozen in liquid nitrogen, or fixed in formalin prior to tissue embedding (paraffin) and sectioning on slides. Histopathological assessment and H. pylori status was determined on fresh GC patient tissue biopsies used for NLRP3 immunohistochemical staining (n = 6) as before (31). Full and informed patient consent was obtained, and biopsy collections were approved by the Monash Health Human Research Ethics Committee (13058A).

The gp130 ^F/F mice have been previously reported (30). Mice homozygous null for Nlrp3 (32) were kindly provided by K. Fitzgerald (University of Massachusetts Chan Medical School, USA), and were used to generate gp130 ^F/F:Nlrp3^-/- mice on a mixed 129Sv × C57BL/6 background. Mice were housed under specific pathogen-free conditions on a 12-h light/dark cycle, and all animal experiments were approved by the Monash University Monash Medical Centre “B” Animal Ethics Committee.

RNA sequencing (RNA-seq) gene counts for NLRP3 and clinical data from The Cancer Genome Atlas (TCGA)-STAD cohort were obtained using the TCGAbiolinks R package. Counts were processed using the edgeR package. A DGEList object was created from the counts and gene annotation information was obtained using the Homo.sapiens package. Samples were grouped according to whether they were normal tissue (n = 32) or primary tumor (n = 375). Normalization factors were calculated using the TMM method and robust dispersions were estimated. Expression levels are expressed as log2 counts per million. The “Gastric Cancer Project ‘08 - Singapore Patient Cohort” (GSE15459) dataset was sourced for NLRP3 gene expression profiling in tumor and non-tumor tissue (33). GC patient overall survival in both patient cohorts was assessed by Kaplan–Meier analysis on patients stratified into “low” versus “high” tumoral gene expression levels for NLRP3 as described previously (13).

Total protein lysates were prepared from snap-frozen mouse gastric tissues, and were subjected to immunoblotting and analysis as described before (17, 29). Antibodies used were those against mouse Caspase-1 (AdipoGen, AG-20B-0042-C100), mouse IL-18 (BioVision, 5180R-100), mouse IL-1β (R&D Systems, BAF401), mouse phospho(p)-NF-κB p65 (Ser536) (Cell Signaling Technology, 3031S), total NF-κB p65 (Cell Signaling Technology, 6956S), mouse ASC (AdipoGen, AG-25B-0006-C100), and α-tubulin (Abcam, Ab6160). Protein bands were visualized using the ECL method for Caspase-1 (p45/p20) or, for all other immunoblots, the Odyssey Infrared Imaging System (LI-COR). The bands were quantified using ImageJ software (34).

Following formalin fixation and paraffin embedding (FFPE), histological assessment of mouse stomachs was performed blinded on 4- to 6-mm hematoxylin and eosin (H&E)-stained tissue sections. Immunohistochemistry on mouse gastric tissue sections was performed with primary antibodies against proliferating cell nuclear antigen (PCNA; Abcam, Ab18197), CD45 (550539) and B220 (550286) (BD BioSciences), CD3 (Abcam, Ab11089), cleaved Caspase-3 (9661S) and cleaved Caspase-8 (8592S) (Cell Signaling Technologies), and NLRP3 (R&D Systems, MAB7578), as described previously (17, 29). Immunohistochemistry on human gastric tissue sections was performed with a primary antibody against NLRP3 (R&D Systems, MAB7578). Positive-staining cells were counted manually on 50 counted cells per high-power field (HPF; n = 20) of 350 μm × 250 μm as described previously (17).

Snap-frozen mouse stomach tissues were subjected to mechanical homogenization on ice using a T 10 ULTRA-TURRAX^® instrument (IKA), following which total RNA was isolated using TRI Reagent^® Solution (Sigma), and subsequent on-column RNeasy^® Mini Kit RNA clean-up and DNase treatment (Qiagen). Total RNA was transcribed with the Transcriptor High Fidelity cDNA Synthesis Kit (Sigma-Aldrich), and quantitative real-time PCR (qPCR) was performed on cDNA with SYBR Green chemistry (Life Technologies) using the 7900HT Fast Real-Time System (Applied Biosystems), and the ViiA 7 Real-Time PCR System (ThermoFisher Scientific). Data acquisition and analyses were undertaken using the Applied Biosystems Sequence Detection System Version 2.4 software and ViiA 7 software. Sequences for mouse and human primers have been previously published (17, 29).

All statistical analyses were performed using GraphPad Prism V8.0.2 software. Statistical significance (p < 0.05) between the means of two groups was determined using Student’s t-tests (normal distribution) or Mann–Whitney U tests (abnormal distribution), and matched datasets involved Wilcoxon signed-rank tests. Statistical significance between the means of multiple groups was determined using ordinary one-way ANOVA (normal distribution) or Kruskal–Wallis (abnormal distribution) tests. All data are presented as the mean ± standard error of the mean (SEM) from at least 3 technical replicates. The log-rank test was used to calculate the statistical significance of the difference in survival between two groups. All data are presented as standard error of the mean (SEM).

Analysis of intestinal-type GC patient data within The Cancer Genome Atlas (TCGA) revealed that NLRP3 mRNA levels were comparable between total gastric tumor tissues of patients (n = 375) compared to non-tumor gastric tissues (n = 32) ( Figure 1A ). In addition, NLRP3 expression was similar in gastric tumor tissues compared to their paired adjacent non-tumor tissues (n = 27) ( Figure 1B ). Conversely, in a second independent intestinal-type GC patient cohort, the “Gastric Cancer Project ‘08 - Singapore Patient Cohort” (GSE15459) (33), comprising gastric tumor tissues (n = 177) and non-tumor gastric tissues (n = 92), NLRP3 mRNA levels were significantly upregulated in total gastric tumor tissues ( Figure 1C ). Similarly, analysis of NLRP3 expression in 83 patients for which there were paired tumor and adjacent non-tumor tissues indicated that NLRP3 was significantly upregulated in tumors from 82% (68/83) of patients ( Figure 1D ). The altered tumoral expression of NLRP3 did not align with a specific disease stage, since NLRP3 mRNA expression levels were comparable in early (stage I and II) and advanced (stage III and IV) GC patients ( Figure 1E ). We also observed that in these independent GC patient cohorts, the segregation of GC patient primary tumors into either low or high NLRP3 gene expression indicated that NLRP3 mRNA levels are not prognostic for overall patient survival outcomes ( Figure 1F ). Furthermore, immunohistochemical staining on a third cohort (Australian) of GC patients (16, 17) indicated that NLRP3 protein expression levels were elevated in patient tumors compared to matched non-tumor gastric tissues, with increased numbers of diffusely stained NLRP3-positive cells observed throughout the transformed glandular epithelium and in immune cells in the stroma of patient tumor tissues ( Figures 1G–I ). Collectively, these divergent data suggest that NLRP3 may not have clinical significance in human GC.

Since we have previously shown that elevated expression of the ASC inflammasome adaptor promotes gastric tumorigenesis in the spontaneous gp130 ^F/F GC mouse model (29), we next assessed the expression of Nlrp3 in these mice by qPCR. In this model, antral gastric intestinal-type tumors spontaneously develop from 1.5 months of age onwards with 100% penetrance, and progressively increase in size through to 6 months of age (13, 35). At 1, 3, and 6 months of age, gastric Nlrp3 mRNA levels were significantly elevated (6- to 8-fold) in gp130 ^F/F tumor-free gastric antrum (1 month old) or gp130 ^F/F gastric antral tumors (3 and 6 months old) compared to either adjacent non-tumor antral tissue, or normal antral tissues of wild-type (WT) mice ( Figure 2A ). In support of these gene expression data, immunohistochemistry indicated that the total number of NLRP3 positively stained cells was significantly greater in gp130 ^F/F antral tumors compared to normal antral tissue from WT mice, with predominant staining observed in the gastric tumor epithelium of gp130 ^F/F mice ( Figures 2B, C ).

To determine whether increased gastric tumoral expression of NLRP3 contributed to the pathogenesis of GC, we generated NLRP3-deficient gp130 ^F/F mice (gp130 ^F/F:Nlrp3 ^-/-). At 3 months of age (early-stage tumorigenesis), the stomachs and gastric tumors of gp130 ^F/F:Nlrp3 ^-/- mice were comparable in size and weight compared to age-matched gp130 ^F/F littermate mice, as was the incidence of tumors ( Figures 2D–I ). Similarly, gastric tumorigenesis was also comparable in gp130 ^F/F:Nlrp3 ^-/- and gp130 ^F/F mice at 6 months of age (late-stage tumorigenesis), as evidenced by comparable stomach size and gastric tumor burden and incidence among age-matched mice of both genotypes ( Figures 3A–F ). Since tumorigenesis in gp130 ^F/F mice is driven by IL-11-mediated hyper-activation of the STAT3 signaling pathway (11, 30), we also investigated whether the loss of NLRP3 affected this pro-tumorigenic signaling axis in the stomach. However, the expression of STAT3-target genes Il11 and Socs3 was similar among gp130 ^F/F:Nlrp3 ^-/- and gp130 ^F/F tumor and matched non-tumor gastric tissues at 3 and 6 months of age ( Figures 3G, H ). Therefore, these data indicate that NLRP3 does not contribute the onset and growth of gastric tumors in the gp130 ^F/F GC model.

Considering that NLRP3 activity is mediated via inflammasome complexes, and the gastric tumor phenotype in gp130 ^F/F mice is associated with inflammasome activation (29), we investigated whether the ablation of NLRP3 affected inflammasome activation during gastric tumorigenesis. For this purpose, we compared the expression and/or activation status of key inflammasome components, ASC and Caspase-1, along with other inflammasome-associated PRRs, in gastric tissues from 6-month-old (mo) gp130 ^F/F:Nlrp3 ^-/- and gp130 ^F/F mice. Immunoblotting demonstrated that Caspase-1 activation levels, measured by detection of the cleavage of pro-Caspase-1 (p45) to its mature form (p20 subunit), were comparable between either gastric tumor or non-tumor tissues from gp130 ^F/F:Nlrp3 ^-/- and gp130 ^F/F mice ( Figures 4A, B ). Also, immunoblotting indicated comparable protein levels of ASC among gp130 ^F/F versus gp130 ^F/F:Nlrp3 ^-/- gastric tumor lysates ( Figures 4A, C ). Gene expression analyses by qPCR also indicated that NLRP3 deficiency did not significantly alter the mRNA levels of other inflammasome-associated PRRs in tumor samples from gp130 ^F/F mice ( Supplementary Figure 1 ). Furthermore, phosphorylation levels of NF-κB, which is activated downstream of inflammasomes to promote gastric tumorigenesis in the gp130 ^F/F model (29), were also similar between gp130 ^F/F and gp130 ^F/F:Nlrp3 ^-/- gastric tumor or non-tumor samples ( Figures 4A, D ). Immunohistochemistry for cleaved Caspase-1 also confirmed comparable numbers and staining intensity for cleaved Caspase-1-positive cells throughout the tumor epithelium of gp130 ^F/F:Nlrp3 ^-/- and gp130 ^F/F mice ( Figures 4E, F ). Taken together, these findings further suggest that NLRP3 is not the primary inflammasome-associated PRR that contributes to gastric inflammasome activity and associated tumorigenesis in the gp130 ^F/F model.

Despite NLRP3 ablation having no impact on gastric tumorigenesis in gp130F/F mice, it remained possible that NLRP3 may contribute to a subset of tumor-associated cellular processes in the gastric compartment. We first investigated whether NLRP3 deficiency in gp130 ^F/F mice impacted on tumor-associated inflammation. However, immunohistochemistry analyses indicated comparable numbers of infiltrating CD45+ leukocytes, B220+ B cells, and CD3+ T cells in the gastric mucosa of 3mo and 6mo gp130 ^F/F versus gp130 ^F/F:Nlrp3 ^-/- mice ( Figures 5A–F and Supplementary Figures 2A–D ). The comparable gastric inflammation in gp130 ^F/F and gp130 ^F/F:Nlrp3 ^-/- mice was also confirmed at the molecular level, with qPCR analyses indicating similar mRNA levels of inflammatory genes (Il6, Ifng, Il10, Il17a, and Tnfa) among both genotypes at 3 and 6 months of age ( Figure 5G and Supplementary Figure 2E ).

Since elevated gastric epithelial cell proliferation and survival are prominent features of inflammasome-associated gastric tumorigenesis in gp130 ^F/F mice (29), we also investigated whether NLRP3 promoted these oncogenic cellular processes in the stomach. However, immunohistochemistry indicated similar PCNA⁺ proliferating cell numbers in the epithelium of gastric tumors of gp130 ^F/F and gp130 ^F/F:Nlrp3 ^-/- 3mo and 6mo mice ( Figure 6A, B and Supplementary Figures 3A, B ). Also, similar numbers of apoptotic cleaved Caspase-3⁺ and cleaved Caspase-8⁺ cells were observed in 3mo and 6mo gp130 ^F/F and gp130 ^F/F:Nlrp3 ^-/- mouse gastric tumors ( Figures 6C–F and Supplementary Figures 3C, D ). We also note that the expression of several angiogenesis-related genes (Cxcl1, Cxcl2, Mmp2, Mmp9, and Vegf) was comparable among gastric tumor and non-tumor tissues from 3mo and 6mo gp130 ^F/F and gp130 ^F/F:Nlrp3 ^-/- mice ( Figure 6G and Supplementary Figure 3E ). Collectively, these data suggest that NLRP3 does not contribute to key inflammasome-driven cellular processes that promote gastric tumorigenesis.