Spermine oxidase promotes Helicobacter pylori-mediated gastric carcinogenesis through acrolein production

Abstract

Helicobacter pylori is the primary cause of gastric cancer, and there is a need to discover new molecular targets for therapeutic intervention in H. pylori disease progression. We have previously shown that spermine oxidase (SMOX), the enzyme that catabolizes the back-conversion of the polyamine spermine to spermidine, is upregulated during infection and is associated with increased cancer risk in humans. We sought to determine the direct role of SMOX in gastric carcinogenesis during H. pylori infection. In this study, we demonstrate that transgenic FVB/N insulin-gastrin (INS-GAS) mice that develop gastric carcinoma with H. pylori infection were protected from cancer development with Smox deletion. RNA sequencing revealed that genes associated with the immune system and cancer were downregulated in the infected Smox^-/- mice. Furthermore, there was a decrease in cell proliferation and DNA damage in infected Smox^-/- animals. There was significant generation of adducts of the highly reactive electrophile acrolein, a byproduct of SMOX activity, in gastric tissues from H. pylori-infected humans and wild-type, but not Smox^-/- mice. Genetic deletion of Smox in murine organoids or chemical inhibition of SMOX in human gastric epithelial cells significantly reduced generation of acrolein induced by H. pylori. Additionally, acrolein-induced DNA damage in gastric epithelial cells was ablated with the electrophile scavenger 2-hydroxybenzylamine (2-HOBA). Gastric acrolein adduct levels were attenuated in infected INS-GAS mice treated with 2-HOBA, which exhibit reduced gastric carcinoma. These findings implicate SMOX and acrolein in H. pylori-induced carcinogenesis, thus indicating their potential as therapeutic targets.

Fig. 1. Effect of Smox deletion on H. pylori pathogenesis in a gastric cancer model.

Smox mRNA expression (A; n = 3 per genotype) and polyamine concentrations (B; n = 5 per genotype) were determined in the gastric tissues of naïve WT and Smox^–/– animals by RT-real-time PCR and mass spectrometry, respectively. Then, animals were infected or not with H. pylori PMSS1 for 8 weeks. H. pylori colonization was determined by serial dilution and culture (C). Tissues stained by H&E (D) were used to analyze the frequency of dysplasia and cancer (E) and their extent (F); ND no dysplasia; in D, scale bars represent 100 μm (left and center) and 50 μm (right) and white arrows indicate cancerous cells invading into the lamina propria. In figures with bars and dots, all values reported as mean ± SEM. Statistical analyses, where shown, ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 by Student’s t test (A–C) and one-way ANOVA and Tukey test (F); each symbol represents a different mouse. In E, ^**P < 0.01 by Chi-square comparing ND, LGD, and IMC between the two infected genotypes; ^§§§P < 0.001 by Fisher’s exact test comparing cancer versus no cancer.

Fig. 2. Transcriptional regulation orchestrated by SMOX in infected mice.

A Transcriptomic analysis using RNA sequencing of infected and uninfected gastric tissues; n = 4–5 mice per genotype. The volcano plots show differential gene signatures in gastric tissues from infected mice compared with uninfected and infected mice (fold change >1.3; FDR < 0.05). Genes of interest bolded. B, C IPA was performed on the differential transcriptome dataset obtained from the gastric tissues from the infected WT and Smox^–/– mice and the pathways related to diseases and functions are broken down into immune (B) and cancer (C) related pathways. Blue bars, z-score < −2; gray bars, z-score < 0. Gene expression by RT-real-time PCR of genes encoding for chemokines (D), carcinogenic markers (E), proinflammatory cytokines (F), T cell cytokines (G), and anti-inflammatory markers (H). All values reported as mean ± SEM. Statistical analyses, where shown, ^*P < 0.05, ^**P < 0.01, ^***P < 0.001 determined by one-way ANOVA and Tukey test or Kruskal–Wallis test followed by a Mann–Whitney U test.

Fig. 3. Cell proliferation and DNA damage in GECs.

Representative images of gastric tissues immunostained for Ki-67 (A) and pH2AX (B); n = 3 uninfected mice and n = 7–8 infected mice per genotype; border and symbol color indicate stage of disease: ND, green; LGD, yellow; IMC, red. Scale bars, 50 μm. These stainings were used to quantify the number of Ki-67 (C) and pH2AX- (D) positive GECs from five high power fields per mouse. Each dot represents a mouse. Values are reported as mean ± SEM. Statistical analyses, where shown, ^*P < 0.05, ^**P < 0.01 determined by one-way ANOVA and a Tukey test. E Correlation between pH2AX and Ki-67 levels. Correlation and significance were determined by a simple linear regression.

Fig. 4. H. pylori induces SMOX-mediated acrolein production.

A Acrolein quantification by mass spectrometry in the gastric tissues of experimental animals; n = 3 per genotype. B Representative immunofluorescence images of acrolein adducts from WT and Smox^–/– mice infected or not with H. pylori; data are representative of n = 2 uninfected mice and n = 3 infected mice per genotype. AcroleinRED assay was performed on gastric organoids generated from naïve mice and infected or not ex vivo for 24 h with H. pylori (C) and the staining was quantified (D); color of symbol indicates organoids derived from same mouse but different passages/experiments, with n = 2 animals per genotype and n = 3 replicates per mouse. AGS cells pretreated or not with the SMOX inhibitor MDL 72527 (25 μM) for 24 h followed by infection or not with H. pylori for 24 h were then stained with AcroleinRED (E) and the staining was quantified (F). Human gastric organoids isolated from normal patient biopsies infected or not with H. pylori were stained with AcroleinRED (G) and the staining was quantified (H). I The antral gastric tissues from 3 normal and 5 patients with H. pylori (Hp) gastritis were immunostained for acrolein adducts. In all representative fluorescence images, AcroleinRED or acrolein adducts are depicted in red and nuclei are stained in blue with DAPI. Scale bars, 50 μm (B and I, bottom row) or 100 μm (all other images). In all bar graphs, values are reported as mean ± SEM. Statistical analyses, where shown, ^*P < 0.05, ^****P < 0.0001 determined by one-way ANOVA and Tukey test (A, D, F) or Student’s t test (H).

Fig. 5. Effect of 2-HOBA on DNA damage.

AGS cells pretreated or not with 2-HOBA (100 μM) for 2 h and then treated or not with acrolein (10 μM) for 24 h; cells were stained by the AcroleinRED assay (A), and the staining was quantified (B). C, D AGS cells were pretreated or not with 2-HOBA (100 μM) for 2 h and then exposed to acrolein (10 μM) and/or infected with H. pylori. After 24 h, cells were immunostained for pH2AX (green; C) and the fluorescence was quantified (D) from 3 independent experiments. E Representative immunofluorescence images of acrolein adducts in the gastric tissues of INS-GAS mice infected or not with H. pylori ± 2-HOBA; data are representative of n = 2 uninfected mice and n = 3 infected animals per group. Scale bars, 50 μm (C and E, bottom row) and 100 μm (all other images). Data represent the mean ± SEM. Statistical analyses, where shown, ^**P < 0.01, ^****P < 0.0001 by one-way ANOVA and Tukey test (B), and two-way ANOVA and Tukey test (D).