---
title: "Glycolysis and T cell Gene Signature for Prognosis and Therapy Prediction in Pancreatic Cancer"
id: "frontiers-in-immunology-12-glycolysis-and-t-cell-associated-gene-signature-predicts-prognosis-and"
canonical_url: "https://medichelpline.com/clinical-feed/frontiers-in-immunology-12-glycolysis-and-t-cell-associated-gene-signature-predicts-prognosis-and"
content_type: "clinical_feed_article"
specialty: "Oncology"
source_name: "Frontiers in Immunology"
source_url: "https://www.frontiersin.org/articles/10.3389/fimmu.2026.1918556"
published_at: "2026-08-28T00:00:00.000Z"
evidence_level: "Journal Feed"
license: "CC-BY-NC-4.0 / Informational Use"
---
# Glycolysis and T cell Gene Signature for Prognosis and Therapy Prediction in Pancreatic Cancer
## Provenance & Clinical Metadata
- **Canonical URL:** https://medichelpline.com/clinical-feed/frontiers-in-immunology-12-glycolysis-and-t-cell-associated-gene-signature-predicts-prognosis-and
- **Specialty:** [Oncology](https://medichelpline.com/clinical-feed/oncology.md)
- **Primary Source:** Frontiers in Immunology
- **Source URL:** [Original Journal Publication](https://www.frontiersin.org/articles/10.3389/fimmu.2026.1918556)
- **Published At:** 2026-08-28T00:00:00.000Z
- **Evidence Rating:** Journal Feed
## Executive GIST (TL;DR)
- The provided source content did not include the manuscript body, results, methods, data, or conclusions for the article titled "Glycolysis and T cell-associated gene signature predicts prognosis and therapeutic responses in pancreatic cancer." - The only available items from the source were site navigation and journal metadata (Frontiers in Immunology) and the article URL; no study details were present in the supplied content. - Essential elements normally required to summarize or interpret the study—sample size, cohorts, gene lists, computational methods, prognostic model performance, validation approach, statistical outcomes, and therapeutic-response associations—were not reported in the provided source text. - Because the manuscript text and data were not included, no factual statements about associations between **glycolysis**, **T cell** gene signatures, and outcomes in **pancreatic cancer** can be made beyond the article title. - To access verifiable findings, one must retrieve the full article via the publisher link or the journal site; the provided URL points to Frontiers in Immunology but the content provided here lacked the article body. - Without the full article, details such as gene signature composition, prognostic accuracy, therapeutic-prediction performance, methods for signature derivation, and clinical or translational recommendations cannot be summarized or paraphrased from this source.
## Clinical Analysis & Structured Key Points
Frontiers | Glycolysis and T cell-associated gene signature predicts prognosis and therapeutic responses in pancreatic cancer ORIGINAL RESEARCH article Front. Immunol. , 28 August 2026 Sec. Cancer Immunity and Immunotherapy Volume 17 - 2026 | https://doi.org/10.3389/fimmu.2026.1918556 Published in Frontiers in Immunology Cancer Immunity and Immunotherapy 7 impact factor 11.3 citescore Part of a Research Topic Metabolic crosstalk between tumors and regulatory immune cells Submission open 20k views 12 articles Editor & Reviewers Edited by S Z Shaoquan Zheng Reviewed by S D Shirong Ding X W Xiumei Wang H Z Huijuan Zhou Outline Figures and Tables Figure 1 View in article Figure 2 View in article Figure 3 View in article Figure 4 View in article Figure 5 View in article Figure 6 View in article Figure 7 View in article Figure 8 View in article ORIGINAL RESEARCH article Front. Immunol. , 28 August 2026 Sec. Cancer Immunity and Immunotherapy Volume 17 - 2026 | https://doi.org/10.3389/fimmu.2026.1918556 Glycolysis and T cell-associated gene signature predicts prognosis and therapeutic responses in pancreatic cancer W L Wancheng Li 1 † D F Dongao Fan 1 † Y D Yan Du 2 L L Lin Li 3 W L Wenjia Li 4 * W Z Wence Zhou 1,2 * 1. The Second School of Clinical Medicine, Lanzhou University, Lanzhou, China 2. Department of General Surgery, Lanzhou University Second Hospital, Lanzhou, China 3. Department of Medical Insurance Office, The Gansu Provincial Maternity and Child-Care Hospital, Lanzhou, China 4. Department of Anesthesiology, Lanzhou University Second Hospital, Lanzhou, China See more Article metrics View details Abstract Background: Pancreatic cancer (PC) is a fatal malignancy, with glycolysis and T cells playing crucial roles in its pathogenesis. This study explored prognosis-related genes in glycolysis and T cells in PC using bioinformatics methods. Methods: We first quantified distinct T-cell subsets via immune infiltration analysis of the TCGA-PAAD cohort and extracted T cell-related genes (T-RGs). Differentially expressed genes (DEGs) were obtained from GSE28735 and TCGA-PAAD. Intersecting these DEGs with T-RGs and glycolysis-related genes (G-RGs) yielded candidate genes. We screened prognostic genes using univariate Cox and LASSO regression, built a prognostic model in TCGA-PAAD and validated it in GSE57495. A nomogram was constructed for survival prediction, its reliability verified by calibration and ROC curves. We compared immune microenvironment, pathway enrichment, mutation landscape and drug sensitivity between high- and low-risk subgroups, and built lncRNA-miRNA-mRNA regulatory networks. In vitro and in vivo mouse tumorigenesis assays explored GPR87’s potential involvement in glycolytic activity and CD8 + T cell infiltration. Seahorse analysis, glucose uptake detection and immunohistochemistry further explored roles in glucose metabolism and anti-tumor immunity. Results: MET , KDELR3 , AK4 , and GPR87 were determined as prognostic genes. In the training set, this exploratory prognostic model showed moderate predictive performance, with area under the curve values of 0.72, 0.71, and 0.71 at 1, 2, and 3 years, respectively. The risk score and N-stage were identified as independent prognosis predictors, and the developed nomogram demonstrated moderate predictive performance. Functional pathways revealed enrichment in 44 pathways. There were 17 differential immune cells. In addition, risk scores were correlated with 28 immune checkpoints. The regulatory network comprised 26 miRNAs and 50 lncRNAs. Furthermore, computationally predicted correlations between prognostic genes and 60 drugs were identified in different risk groups. In vitro and in vivo experiments demonstrated that GPR87 knockdown inhibits the proliferation, migration, and clonogenic ability of PC cells while promoting their apoptosis. Furthermore, GPR87 expression was negatively correlated with CD8 + T cell, and GPR87 knockdown inhibited glycolysis in pancreatic cancer cells. Conclusion: MET , KDELR3 , AK4 , and GPR87 were identified as candidate prognostic genes in PC, providing preliminary hypothesis-generating insights that require validation in independent institutional or prospective clinical cohorts before any clinical application. 1 Introduction Pancreatic cancer (PC) is a highly aggressive malignancy of the digestive system. Due to its subtle early clinical manifestations, as well as the absence of effective screening methods, most patients are only diagnosed at intermediate or advanced stages, resulting in the loss of opportunities for radical resection ( 1 , 2 ). Only approximately 20% of patients are eligible for surgical resection at diagnosis, and even after surgery, recurrence is common, contributing to a 5-year survival rate below 13%. Therefore, there is an urgent need to identify novel prognostic biomarkers and therapeutic strategies to enhance the clinical management of PC. In recent years, tumor immunotherapy has achieved remarkable success in multiple malignancies; however, its clinical efficacy in PC remains limited ( 3 ). This poor therapeutic efficacy is largely attributed to the unique immunosuppressive tumor microenvironment (TME) of PC ( 4 ). In the PC TME, cytotoxic CD8 + T cells, which represent the core effector cells of antitumor immunity, are frequently infiltrated but predominantly exist in a state of functional exhaustion or dysfunction, with severely impaired proliferation, cytokine production, and cytotoxic capacity ( 5 , 6 ). Thus, investigating the mechanisms underlying T cell dysfunction is critical to overcoming immunotherapy resistance and improving the prognosis of patients with PC. Metabolic reprogramming is a hallmark of PC ( 7 ). Notably, aberrantly elevated aerobic glycolysis (the Warburg effect), driven by oncogenic mutations such as KRAS, provides energy and biosynthetic precursors for rapid tumor proliferation and profoundly shapes the immunosuppressive TME ( 8 ). By competing for glucose and secreting lactate, tumor cells create a nutrient-depleted, acidic milieu that suppresses T cell activation, impairs effector function, and promotes the expansion of immunosuppressive subsets such as regulatory T cells ( 9 ). These findings indicate a close and dynamic crosstalk between tumor glycolytic metabolism and T cell immune function. Therefore, identifying prognostic biomarkers associated with this immunometabolic regulation would aid in elucidating the mechanisms of immune cell dysfunction and provide novel targets for personalized therapy ( 10 ). Currently, bioinformatics approaches are widely used to identify novel biomarkers and potentially effective small-molecule drugs in PC ( 10 , 11 ). In this study, we aimed to identify candidate genes associated with glycolysis-related genes (G-RGs) and T cell-related genes (T-RGs) in patients with PC via transcriptomic analysis. Univariate Cox regression and least absolute shrinkage and selection operator (LASSO) regression were applied to screen prognostic genes and establish a prognostic model for PC. We further analyzed the enriched pathways, immune profiles, and TME characteristics of patients with PC stratified by the risk score. Finally, in vitro and in vivo experiments preliminarily verified the effects of GPR87, a gene included in the risk score, on the malignant biological behaviors of PC. 2 Materials and methods 2.1 Data source Transcriptomic data from patients with PC and normal pancreatic tissue samples were obtained from the Gene Expression Omnibus (GEO) ( https://www.ncbi.nlm.nih.gov/geo/ ) and The Cancer Genome Atlas (TCGA) ( https://gdc.cancer.gov/ ). The GSE28735 dataset (45 PC samples and 45 normal samples) served as the primary dataset for differential expression analysis ( 12 , 13 ). TCGA-PAAD (177 PC samples with survival information and 4 normal samples) was used as the training cohort ( 14 ). Three independent GEO datasets were used for validation: GSE62452 (65 PC cases with complete survival data) ( 15 ), GSE57495 (63 PC samples with survival data) ( 16 ), and GSE79668 (49 PC samples with survival information ( 17 ). Samples with missing survival time, survival status, or key clinical variables were excluded prior to analysis. For TCGA-PAAD, differential expression analysis was performed on count data using DESeq2, whereas all downstream analyses (including GSEA, GSVA, LASSO modeling, and survival analysis) used FPKM-normalized expression data. A total of 296 glycolysis-related genes (G-RGs) were retrieved from the Molecular Signatures Database (MSigDB v7.5.1) ( https://www.gsea-msigdb.org/gsea/msigdb/ ) after removing duplicates ( 18 ). Each dataset was independently analyzed without cross-platform integration or batch correction. Detailed data download and preprocessing procedures are provided in the Supplementary Methods . 2.2 Identification of T-RGs The single-sample gene set enrichment analysis (ssGSEA) algorithm was applied to TCGA-PAAD to compute enrichment scores for 28 immune cell types across PC and normal samples using the R package GSVA ( 19 ). Immune cells showing significant abundance differences between the two groups were identified as differential immune cells (Wilcoxon test, p 0.3 and p 0.5 and p 0.05) were retained as prognosis-related genes. LASSO regression was subsequently performed using “glmnet” (v 4.1-4) ( 29 ) with the optimal lambda value (lambda.min) determined via 10-fold cross-validation to identify prognostic genes and control for overfitting. The risk score was calculated as: where “coef” represents the risk coefficient for each prognostic gene, and “X” represents the expression level of that gene in the respective sample. Expression differences in prognostic genes were further evaluated in the GSE28735 and GSE62452 datasets (Wilcoxon test, p 0.6) using “survivalROC” (v 1.0.3.1) ( 32 ). Heatmaps were plotted using “pheatmap.” In addition, the median risk score threshold from TCGA-PAAD was applied to the GSE79668 dataset for risk stratification and validation analysis. 2.6 Independent prognostic analysis Clinical characteristics including risk score, age, gender, and T and N stage were assessed for compliance with the PH assumption prior to Cox regression (p > 0.05). Univariate Cox regression was performed in TCGA-PAAD to identify variables significantly associated with prognosis (HR ≠ 1, p 0.6), calibration curves at 1, 2, and 3 years, and decision curve analysis (DCA) to assess clinical net benefit. Risk score distributions across clinical subgroups were compared using the Wilcoxon test (p 1 and p 0.3 and p 0.3, p 0.3 and p < 0.05). The Wilcoxon test was applied to compute the disparities in IC 50 for the top 15 drugs (ordered by p-value) between the two groups. For visualization, “ggplot2” was used (p < 0.05). 2.11 Immune cell infiltration algorithms CD8 + T cell infiltration levels in the ICGC-AU, GSE183795, and GSE717295 cohorts were quantified using five independent immune deconvolution algorithms: CIBERSORT, EPIC, MCP-counter, quanTIseq, and xCell, integrated via the R package immunedeconv. The correlation between CD8 + T cell abundance estimates and GPR87 expression was evaluated across all three cohorts. 2.12 Cell and molecular biology analysis 2.12.1 Cell lines, culture conditions, and cell transfection Human PC cell lines BxPC-3 and PANC-1 (KRAS G12D mutation) (Cell Bank of the Chinese Academy of Sciences, Shanghai, China) and hTERT-HPNE cells (Cell Bank of the Chinese Academy of Sciences, Shanghai, China) were used. The BxPC-3 cells were cultured in RPMI-1640, whereas the PANC-1 and HPNE cells were maintained in DMEM, each supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin. Cells were incubated at 37 °C with 5% CO 2 . Small interfering RNAs (siRNAs) (TsingKe, Beijing, China) were transfected using Lipofectamine 3000 (Thermo Fisher Scientific) as per the manufacturer’s protocol. The siRNA transfection concentration was 50 nM. Cells were collected 24 h after transfection, and the interference efficiency was detected by reverse transcription quantitative PCR (RT-qPCR). Transfected cells in the logarithmic growth phase were digested and seeded for subsequent cell phenotypic experiments. 2.12.2 RNA extraction and RT-qPCR Total RNA was extracted with TRIzol (Vazyme), reverse-transcribed using HiScript Q RT SuperMix (Vazyme), and quantified by RT-qPCR with ChamQ SYBR qPCR Master Mix (Vazyme). Normalization was performed using β-actin as the reference gene. The primer and siRNA sequences are provided in Supplementary Table 1 . 2.12.3 Glucose uptake Glucose uptake was detected using the 2-NBDG Glucose Uptake Assay Kit (Servicebio, G1744) following the manufacturer’s protocol. Briefly, after being seeded and attached, cells were starved in glucose-free and serum-free medium for 1 h. The diluted 2-NBDG working solution was then added to incubate the cells at 37 °C for 50 min before fluorescence microscopic observation. Fluorescence intensity of cells in each grou
## Related Clinical Research

- [Shared decision-making and patient involvement in multiple myeloma: cross-country survey findings](https://medichelpline.com/clinical-feed/plos-one-20-enhancing-patient-involvement-through-shared-decision-making-are-we-there-yet.md)
- [One-month physicochemical stability of idarubicin diluted in polyolefin (Freeflex) infusion bags](https://medichelpline.com/clinical-feed/plos-one-21-one-month-physicochemical-stability-study-of-idarubicin-prepared-in-polyolefin.md)
- [Phase III trial protocol: Vaginal radiofrequency plus moisturiser vs moisturiser alone for vulvova](https://medichelpline.com/clinical-feed/bmj-open-10-french-multicentre-randomised-phase-iii-trial-evaluating-the-efficacy-and.md)
- [PD-L1 Expression and Clinical Outcomes in Oral Squamous Cell Carcinoma: Large Cohort Analysis](https://medichelpline.com/clinical-feed/pubmed-42665707.md) (DOI: 10.1007/s10006-026-01628-3)
- [Belzutifan plus lenvatinib vs cabozantinib in previously treated advanced renal cell carcinoma — L](https://medichelpline.com/clinical-feed/pubmed-42586114.md) (DOI: 10.1016/S0140-6736(26)01089-5)

## Navigation
- [← Back to Oncology Feed](https://medichelpline.com/clinical-feed/oncology.md)
- [← All Clinical Specialties](https://medichelpline.com/clinical-feed.md)
## Medical & Regulatory Disclaimer

> [!CAUTION]
> MedicHelpline content is structured for research, educational, and professional discovery purposes. It does not constitute individual medical advice, clinical diagnosis, or treatment recommendations.
> Always verify dosing, contraindications, and regulatory alerts against official product labeling and primary regulatory sources before clinical decision-making.