A gene linked to ferroptosis is reduced in gastric cancer, yet patients with more of it tend to have better outcomes. This study asks whether AKR1C2 could connect tumor prognosis with the immune response. Gastric cancer is one of the most common malignant tumors in the world and ranks as the fifth most diagnosed cancer worldwide.

Surgical treatment, chemotherapy, and immunotherapy have been reported to be effective treatment strategies, but the survival rate of gastric cancer is still low. That is why the study focuses on exploring a new biomarker for predicting the prognosis of gastric cancer.

Ferroptosis is a regulated cell death form strongly related to the accumulation of ferrous iron and the peroxidation of lipid, ending with mitochondrial dysfunction. AKR1C2 is an aldo-keto reductases family 1 member C2 that has been reported to cause the mutual translation between sex hormones and associated inactive metabolites.

Its substrate specificity is quite different from other AKR1C members. Studies in thyroid carcinoma, acute myeloid leukemia, and esophageal squamous cell carcinoma have linked AKR1C2 with prognosis, but few studies have reported its correlation with gastric cancer patients. This gap in gastric cancer is the specific problem the study addresses.

The study used bioinformatics tools to explore the underlying mechanism of AKR1C2 in gastric cancer and found that downregulated AKR1C2 could be a prognostic prediction tool. It also investigated clinical characteristics, used gene-set enrichment analysis to reveal biological functions and signaling pathways, and evaluated tumor-infiltrating immune cells.

The stated goal was to assess AKR1C2 as a potential predictive prognostic biomarker and its role in immunotherapy of gastric cancer patients. The analysis searched three gastric cancer datasets from the Gene Expression Omnibus: GSE26942, GSE112369, and GSE33651.

Differentially expressed genes between normal gastric tissues and gastric cancer tissues were analyzed using a p-value below zero point zero five and an absolute logFC threshold of zero point eight. Venn analysis then investigated co-differentially expressed genes between the ferroptosis-related gene dataset and the three GEO datasets.

In addition to other datasets, the researchers considered TCGA, a public database that provides clinical information concerning thirty-three different cancer types. Table one summarizes the three Gene Expression Omnibus datasets used to study gene expression in gastric cancer.

GSE26942, on platform GPL6947, includes 202 cancer and 12 normal samples, with 105 up-regulated and 588 down-regulated genes; GSE112369 and GSE33651 provide the other sample sets and their corresponding differentially expressed genes. This matters because these datasets form the basis for comparing cancer with normal tissue before identifying shared genes and investigating ferroptosis-related mechanisms.

The three GEO datasets were downloaded, and the stated p-value and absolute logFC thresholds were used to identify differentially expressed genes between the gastric cancer group and the normal tissue group. GSE26942 contained one hundred five up-regulated genes and five hundred eighty-eight down-regulated genes; GSE112369 contained seven hundred ten up-regulated and one thousand one hundred forty down-regulated genes.

GSE33651 contained nine hundred fifty-nine up-regulated genes and one hundred forty-three down-regulated genes. Because ferroptosis may participate in the development and progression of certain tumors, the study used Venn analysis to identify co-differentially expressed genes between the ferroptosis-related dataset and the three GEO datasets.

The Venn plot showed two down-regulated genes with potential significance in gastric cancer: AKR1C2 and MUC1. No up-regulated co-differentially expressed genes existed between the three GEO datasets and the ferroptosis-related gene dataset. Figure one uses a Venn diagram to compare downregulated co-expressed genes from a ferroptosis-related gene dataset with three gastric cancer GEO datasets: GSE112369, GSE33651, and GSE26942.

The shared intersection contains two genes, AKR1C2 and MUC1, highlighted by the arrow. This matters because the overlap identifies candidate genes consistently connected with both ferroptosis-related biology and gastric cancer expression changes, motivating the subsequent prognosis and tissue-expression analyses.

Kaplan-Meier analysis examined overall survival, first progression, and post-progression survival for AKR1C2 and MUC1 in gastric cancer patients. Higher AKR1C2 expression was correlated with better overall survival, first progression, and post-progression survival in all gastric cancer patients.

For MUC1, lower expression was correlated with better overall survival, first progression, and post-progression survival. Together, these results indicated that AKR1C2 expression may predict a good prognosis in gastric cancer patients. Figure two uses six Kaplan–Meier plots to relate gene expression to outcomes in gastric cancer: overall survival, first progression, and post-progression survival.

In panels A through C, higher AKR1C2 expression is associated with hazard ratios of zero point seven, zero point six seven, and zero point six five, respectively. Panels D through F show MUC1 with hazard ratios of one point six one, one point six one, and two point zero eight, indicating that the two markers carry distinct prognostic associations across these clinical endpoints.

Across GSE26942, GSE112369, and GSE33651, AKR1C2 was expressed much more highly in normal tissues than in gastric cancer groups, with p below zero point zero five. Figure three compares AKR1C2 expression between normal gastric tissue and gastric cancer across three GEO datasets, TNMplot gene-chip and RNA-seq data, GEPIA two point oh, and TCGA.

Across these sources, the plots report lower AKR1C2 expression in gastric cancer, with statistically significant differences, including TNMplot p-values of one point zero nine times ten to the power of minus thirty-two and two point five seven times ten to the power of minus three.

This cross-database consistency supports AKR1C2 as a potentially relevant gastric-cancer-associated marker. Using the TCGA database, the study examined the correlation between AKR1C2 expression and clinical characteristic parameters in gastric cancer patients.

AKR1C2 messenger RNA expression was correlated with N-stage tumor and residual tumor, with p values of zero point zero one four and zero point zero two five. Table two examines whether AKR1C2 expression is associated with clinical characteristics in gastric cancer patients from TCGA, reporting sample size, odds ratios, confidence intervals, and p values.

The authors identify statistically significant associations for N stage, with an odds ratio of one point seven zero two and a p value of zero point zero one four, and residual tumor, with an odds ratio of four point three zero six and a p value of zero point zero two five. The other listed characteristics have p values above zero point zero five.

LinkedOmics was used to explore co-expression genes and biological functions of AKR1C2 in gastric cancer patients. The heatmap showed twenty-nine genes positively correlated with AKR1C2 and thirty-nine genes negatively correlated with AKR1C2. The GO enrichment analysis showed that AKR1C2 co-expressed genes mainly participated in immune-related pathways, including the chemokine-mediated signaling pathway and somatic diversification.

Figure four maps the genes co-expressed with AKR1C2 in gastric cancer. LinkedOmics separates positively and negatively associated genes, while the heatmaps highlight twenty-nine positive and thirty-nine negative correlations; survival heatmaps then show the top fifty genes in each relationship.

GO enrichment points to immune-receptor functions, oxidative-stress responses, and fatty-acid metabolism, and the STRING network connects AKR1C2 with molecules including DHDH, SRD5A1, and SRD5A2. Single-sample GSEA with Spearman correlation was applied to investigate the relationship between AKR1C2 expression and immune infiltration.

AKR1C2 expression was positively correlated with Mast cells, T helper type 17 cells, Eosinophils, B cells, and T follicular helper cells, and negatively correlated with T helper type 2 cells, NK CD56dim cells, activated DC, Treg, cytotoxic cells, T helper type 1 cells, and T cells.

TISIDB identified positive correlation with T helper type 17 cells and negative correlation with T helper type 2 cells and Treg; TIMER reflected a positive correlation with B cells. AKR1C2 expression also had a positive correlation with VSIR, and immunohistochemical analysis found that AKR1C2 and VSIR were both down-regulated in gastric cancer tissues.

Figure five connects AKR1C2 expression with immune infiltration in gastric cancer. Panel A, using ssGSEA correlations, shows positive associations with mast cells, Th17 cells, eosinophils, B cells, and T follicular helper cells, alongside negative associations for several other immune populations.

Panels B and C examine these relationships in scatterplots and TIMER, while panels D and E extend the analysis to VSIR, showing a positive association with AKR1C2. Together, the figure suggests that AKR1C2 may relate to the immune environment of gastric tumors.

Figure six compares normal gastric tissue with gastric cancer tissue using hematoxylin-eosin staining and immunohistochemistry. The staining patterns in panels B and C show AKR1C2 and VSIR expression across the two tissue types, while panel A provides the underlying tissue morphology.

This visual matters because it supplies tissue-level evidence for the study’s investigation of AKR1C2 in gastric cancer and its relationship to immune infiltration and immunostimulatory or immunoinhibitory markers. AKR1C2 expression was positively correlated with immunostimulators including HHLA2, C10orf54, RAET1E, and TNFRSF13B, while negatively related immunoinhibitors included CTLA4, LAG3, CD274, and IDO1.

The top positively correlated chemokines were CXCL14, CCL28, CXCL17, and CCL14. CCR8, CCR1, CCR5, and CXCR6 were receptors strongly negatively correlated with AKR1C2. The results therefore indicated that AKR1C2 was involved in the immune regulation of gastric cancer.

The study identified AKR1C2 and MUC1 as two down-regulated genes shared by the ferroptosis-related dataset and three GEO datasets, then analyzed their correlation with gastric cancer prognosis. Higher AKR1C2 expression was associated with favorable prognosis, and AKR1C2 was expressed higher in the normal group than in the gastric cancer group.

LinkedOmics demonstrated genes positively and negatively correlated with AKR1C2, while ssGSEA and related immune analyses indicated biological and immune-regulatory involvement. The study concluded that AKR1C2 was strongly correlated with the prognosis of gastric cancer patients.

AKR1C2 expression was also correlated with immunostimulators, immunoinhibitors, chemokines, and receptors. Therefore, the ferroptosis-related gene AKR1C2 has the potential to serve as a promising prognostic predictive biomarker for gastric cancer patients.

Across multiple public datasets, AKR1C2 was lower in gastric cancer than normal tissue, associated with better prognosis when more highly expressed, and correlated with several immune features. The evidence supports it as a potential prognostic biomarker, not yet a proven clinical test.