Scholarlink

Purpose <p>Low-grade gliomas(LGGs) show significant clinical and molecular heterogeneity, complicating progression prediction with conventional indicators. Metabolic reprogramming, a cancer hallmark, is linked to immune microenvironment remodeling, yet its role in LGG prognostic modeling remains underexplored. This study aims to develop a robust metabolism-related prognostic signature and elucidate its interaction with the immune microenvironment.</p> Materials and methods <p>Multi-omics data from 1322 LGG patients were obtained from public databases, including the Cancer Genome Atlas (TCGA), the Chinese Glioma Genome Atlas (CGGA), and others. Metabolism-related genes were identified using three strategies: (1) differential expression analysis; (2) univariate Cox regression; and (3) weighted gene co-expression network analysis (WGCNA). Overlapping genes were further refined using protein-protein interaction (PPI) network analysis and four algorithms. We systematically compared 101 machine learning algorithms and selected the Cox model with likelihood-based boosting (CoxBoost) and Ridge regression (Ridge) to construct the hub metabolism-related gene risk score (HMRG-RS).</p> Result <p>A total of 7 hub metabolic genes were identified <i>(TYMS</i>,<i> PLA2G5</i>,<i> GPX7</i>,<i> GLRX</i>,<i> CYP17A1</i>,<i> ALOX15B</i>,<i> ACACB</i>). HMRG-RS demonstrated good prognostic predictive performance across multiple external validation cohorts, with an average concordance index (C-index) of 0.723 and 1/3/5-year area under the receiver operating characteristic curve (AUC) of 0.778/0.797/0.745. Patients in the high-risk group exhibited significantly shorter survival and an immunosuppressive microenvironment characterized by M2 macrophage enrichment and increased tumor mutational burden(TMB). Notably, the prognostic value of HMRG-RS and the metabolic subtypes it characterizes were significantly dependent on Isocitrate dehydrogenase 1 (<i>IDH1</i>) mutation status. Drug sensitivity analysis revealed differential responsiveness to specific chemotherapeutic/targeted agents (e.g., AZD6482, fluvastatin) across risk groups. Molecular docking further predicted multiple therapeutic compounds (e.g., prunellin, mometasone, isoliquiritigenin) with high affinity for pivotal metabolic genes. Single-cell analysis confirmed high expression of hub metabolism-related genes (HMRGs) in myeloid cells (particularly metabolically active protumor M2 macrophages), implicating them in lipid metabolism reprogramming and immune evasion.</p> Conclusion <p>This study constructed and validated a metabolism-driven prognostic model. The model enables prognostic stratification of LGG patients and links high-risk scores to metabolic dysregulation and an immunosuppressive microenvironment characterized by M2 macrophage enrichment, based on multi-omics data. Mechanistic exploration indicates this association is particularly pronounced in myeloid cells, predominantly within metabolism-related M2 macrophage subpopulations. Furthermore, computational analysis suggests differences in drug sensitivity between risk groups and identifies potential therapeutic compounds, providing clues for future exploration of therapeutic strategies targeting metabolic-immune interactions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Development and validation of a prognosis model for low-grade gliomas based on metabolic gene risk scoring and immune microenvironment interaction

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月10日

查看全文
  • Haobin Liu,
  • Yuxiao Wu,
  • Haoyu Sun,
  • Xiao Han,
  • Qian Liu,
  • Yuening Zhang,
  • Jinling Zhang

摘要

Purpose

Low-grade gliomas(LGGs) show significant clinical and molecular heterogeneity, complicating progression prediction with conventional indicators. Metabolic reprogramming, a cancer hallmark, is linked to immune microenvironment remodeling, yet its role in LGG prognostic modeling remains underexplored. This study aims to develop a robust metabolism-related prognostic signature and elucidate its interaction with the immune microenvironment.

Materials and methods

Multi-omics data from 1322 LGG patients were obtained from public databases, including the Cancer Genome Atlas (TCGA), the Chinese Glioma Genome Atlas (CGGA), and others. Metabolism-related genes were identified using three strategies: (1) differential expression analysis; (2) univariate Cox regression; and (3) weighted gene co-expression network analysis (WGCNA). Overlapping genes were further refined using protein-protein interaction (PPI) network analysis and four algorithms. We systematically compared 101 machine learning algorithms and selected the Cox model with likelihood-based boosting (CoxBoost) and Ridge regression (Ridge) to construct the hub metabolism-related gene risk score (HMRG-RS).

Result

A total of 7 hub metabolic genes were identified (TYMS, PLA2G5, GPX7, GLRX, CYP17A1, ALOX15B, ACACB). HMRG-RS demonstrated good prognostic predictive performance across multiple external validation cohorts, with an average concordance index (C-index) of 0.723 and 1/3/5-year area under the receiver operating characteristic curve (AUC) of 0.778/0.797/0.745. Patients in the high-risk group exhibited significantly shorter survival and an immunosuppressive microenvironment characterized by M2 macrophage enrichment and increased tumor mutational burden(TMB). Notably, the prognostic value of HMRG-RS and the metabolic subtypes it characterizes were significantly dependent on Isocitrate dehydrogenase 1 (IDH1) mutation status. Drug sensitivity analysis revealed differential responsiveness to specific chemotherapeutic/targeted agents (e.g., AZD6482, fluvastatin) across risk groups. Molecular docking further predicted multiple therapeutic compounds (e.g., prunellin, mometasone, isoliquiritigenin) with high affinity for pivotal metabolic genes. Single-cell analysis confirmed high expression of hub metabolism-related genes (HMRGs) in myeloid cells (particularly metabolically active protumor M2 macrophages), implicating them in lipid metabolism reprogramming and immune evasion.

Conclusion

This study constructed and validated a metabolism-driven prognostic model. The model enables prognostic stratification of LGG patients and links high-risk scores to metabolic dysregulation and an immunosuppressive microenvironment characterized by M2 macrophage enrichment, based on multi-omics data. Mechanistic exploration indicates this association is particularly pronounced in myeloid cells, predominantly within metabolism-related M2 macrophage subpopulations. Furthermore, computational analysis suggests differences in drug sensitivity between risk groups and identifies potential therapeutic compounds, providing clues for future exploration of therapeutic strategies targeting metabolic-immune interactions.