Protein Programs in H. pylori Injury

Evidence

Two independent primary studies examined protein changes during Helicobacter pylori-associated gastric injury at different scales. A July 31 tissue survey mapped proteins across histopathological stages in people. A February full-text study combined human biopsy staining with epithelial gene perturbation and proteomics in mice. The studies converge on immune remodeling after infection, but only the mouse experiment tested a candidate regulator.

The fresh study analyzed gastric tissue from 22 patients divided among four groups: H. pylori-negative chronic non-atrophic gastritis, H. pylori-positive chronic non-atrophic gastritis, H. pylori-positive chronic atrophic gastritis, and intestinal-type gastric cancer. Using label-free liquid chromatography-tandem mass spectrometry, the researchers quantified 6,019 proteins. They then compared adjacent stages with functional enrichment, temporal clustering, and protein-interaction network analysis.

The number of differentially expressed proteins varied sharply by comparison. The study reported 771 between infected and uninfected chronic gastritis, 101 between infected atrophic and non-atrophic gastritis, and 535 between gastric cancer and infected atrophic gastritis. Immune-response pathways rose with infection, whereas oxidative-phosphorylation signals declined progressively across the staged groups. A temporal cluster containing ribosome-biogenesis proteins increased from precancerous stages to cancer; NIP7 and PDCD11 emerged as network hubs. Seven hub genes were also higher in an external analysis of TCGA and GTEx RNA data, although that transcript comparison was not protein-level replication.

The independent study focused on hypusination, a post-translational modification of EIF5A made from spermidine through the rate-limiting enzyme deoxyhypusine synthase, or DHPS. In representative gastric-biopsy images from three to four people per group, DHPS and hypusinated EIF5A staining was greater in H. pylori gastritis than in uninfected tissue, appearing in epithelial cells and immune infiltrates. This small human component established localization, not an estimate of effect size across lesion stages.

To test epithelial function, the researchers reduced Dhps specifically in the gastric epithelium of mice. Animals received an H. pylori strain twice and were studied eight weeks later. Among infected animals, the analysis included 20 controls and 30 epithelial-Dhps-deficient mice. Bacterial burden did not differ detectably between genotypes, but blinded histology showed less gastric inflammation in the Dhps-deficient group. That separation matters because it links epithelial hypusination to tissue inflammation without attributing the result to lower colonization.

Tandem-mass-tag proteomics of isolated gastric epithelial cells provided a molecular readout. In uninfected Dhps-deficient cells, 79 proteins increased and 183 decreased; the latter included several ribosomal proteins. After infection, 110 proteins rose in control cells and 174 in Dhps-deficient cells, with 37 shared. Although more proteins met the induction threshold in the deficient group, shared immune mediators were generally expressed at lower levels than in infected controls. Infection reduced 170 proteins in controls but 61 in deficient cells. Pathway analysis of the genotype contrast found infection-related processes increased while inflammation-related processes were mainly reduced.

Analysis — Distinguishing Progression From Inflammatory Control

The cross-study inference is that the gastric proteome contains at least two analytically distinct layers: stage-associated patterns observed across human lesions and an epithelial inflammatory program that can be perturbed without changing bacterial burden. This is an interpretation across studies, not a demonstrated causal chain. The human survey's infection-linked immune enrichment is directionally consistent with the mouse study's reduced inflammation and altered immune-protein response after epithelial Dhps reduction. The connection makes hypusination a specific candidate controller of part of the infection response. However, the human survey did not measure DHPS activity or hypusinated EIF5A, and the mouse study did not reproduce the progressive oxidative-phosphorylation decline or establish that its ribosomal changes match the NIP7-PDCD11 cluster. A decisive bridge would measure hypusination across human lesion stages, resolve epithelial and immune-cell signals separately, and test whether Dhps perturbation changes the stage-linked protein modules in independent organoid or longitudinal models. Until then, convergence supports a research hypothesis, not a progression mechanism or therapeutic conclusion.

Limitations

The July 31 paper was available to ingestion only as a PubMed abstract. It supports the stated design, protein counts, pathway directions, and hub-gene summary, but not assessment of group sizes, normalization, multiple-testing choices, effect sizes, or sensitivity analyses. Its 22-person cross-sectional sample is small for four histological groups. Adjacent-stage comparisons do not follow individuals over time, and external RNA-expression agreement cannot validate protein abundance or biomarker performance.

The hypusination study's human evidence consisted of representative staining from only three to four people per group. Its causal experiment used mice, partial epithelial Dhps reduction, an eight-week infection endpoint, and pooled proteomic samples reported as two per group. It measured inflammation rather than precancerous progression or cancer. The two studies used different species, tissues, proteomic workflows, and endpoints, with no shared cohort or direct replication. Neither establishes a diagnostic marker, a safe intervention, or clinical benefit.