Scholarlink

Background <p>Liver cancer stem cells (CSCs) play a pivotal role in the initiation, progression, and recurrence of hepatocellular carcinoma (HCC), but their molecular and metabolic heterogeneity remains largely undefined. Characterizing subpopulations with enriched stemness features is crucial for understanding liver cancer biology and developing targeted therapies.</p> Methods <p>We integrated two single-cell RNA sequencing datasets (GSE149614 and GSE156625) of HCC to identify malignant cell subpopulations through clustering and annotation. Mitochondrial scoring, CytoTRACE, pseudotime, and pathway enrichment analyses were employed to systematically assess stemness-related features and metabolic features of each cluster. Survival analysis and prognostic modeling based on key cluster signature genes led to the development and validation of the Cluster 3 and Cluster 4 Stemness-based Prognostic Signature (C3C4-SPS).</p> Results <p>A total of 4,155 malignant cells were re-clustered into eight subpopulations, as revealed by tSNE visualization, highlighting pronounced intratumoral heterogeneity. Mitochondrial scoring demonstrated that Cluster 3 and Cluster 4 had significantly higher metabolic activity than other clusters. CytoTRACE analysis identified higher enrichment of stemness-associated features in these two clusters, and pseudotime trajectory analysis located them at the origins of distinct differentiation pathways, suggesting that they may represent stemness-enriched subpopulations, though this interpretation requires experimental confirmation. Functional enrichment revealed that Cluster 3 was predominantly involved in oxidative phosphorylation (respiratory electron transport chain and ATP synthesis), whereas Cluster 4 was more active in mitochondrial fatty acid β-oxidation and tRNA aminoacylation pathways. Both clusters showed high expression of core stem cell markers (POU5F1, SOX2, MYC, CD44). Multi-cohort survival analyses demonstrated that high expression of signature gene sets from Cluster 3 and Cluster 4 was closely associated with unfavorable prognosis in HCC patients. The C3C4-SPS model effectively stratified patients by risk, with high-risk groups exhibiting significantly worse outcomes, and performed robustly in ROC and PCA validation.</p> Conclusion <p>This study systematically characterized the molecular and metabolic heterogeneity of liver cancer stem cell subpopulations and elucidated their decisive roles in tumor progression and patient prognosis. These findings are primarily based on computational analyses and warrant further experimental validation. The C3C4-SPS model provides a novel tool for prognostic assessment and potentially individualized therapy targeting stemness-associated features in hepatocellular carcinoma.</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

Dissecting liver cancer stem cell heterogeneity via single-cell RNA sequencing and developing a stemness-based prognostic model

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月16日

查看全文
  • Jinhan Zhao,
  • Jiaqi Feng,
  • Pengpeng Zhang,
  • Xinyi Wu,
  • Bin Zhang,
  • Lichao Sun

摘要

Background

Liver cancer stem cells (CSCs) play a pivotal role in the initiation, progression, and recurrence of hepatocellular carcinoma (HCC), but their molecular and metabolic heterogeneity remains largely undefined. Characterizing subpopulations with enriched stemness features is crucial for understanding liver cancer biology and developing targeted therapies.

Methods

We integrated two single-cell RNA sequencing datasets (GSE149614 and GSE156625) of HCC to identify malignant cell subpopulations through clustering and annotation. Mitochondrial scoring, CytoTRACE, pseudotime, and pathway enrichment analyses were employed to systematically assess stemness-related features and metabolic features of each cluster. Survival analysis and prognostic modeling based on key cluster signature genes led to the development and validation of the Cluster 3 and Cluster 4 Stemness-based Prognostic Signature (C3C4-SPS).

Results

A total of 4,155 malignant cells were re-clustered into eight subpopulations, as revealed by tSNE visualization, highlighting pronounced intratumoral heterogeneity. Mitochondrial scoring demonstrated that Cluster 3 and Cluster 4 had significantly higher metabolic activity than other clusters. CytoTRACE analysis identified higher enrichment of stemness-associated features in these two clusters, and pseudotime trajectory analysis located them at the origins of distinct differentiation pathways, suggesting that they may represent stemness-enriched subpopulations, though this interpretation requires experimental confirmation. Functional enrichment revealed that Cluster 3 was predominantly involved in oxidative phosphorylation (respiratory electron transport chain and ATP synthesis), whereas Cluster 4 was more active in mitochondrial fatty acid β-oxidation and tRNA aminoacylation pathways. Both clusters showed high expression of core stem cell markers (POU5F1, SOX2, MYC, CD44). Multi-cohort survival analyses demonstrated that high expression of signature gene sets from Cluster 3 and Cluster 4 was closely associated with unfavorable prognosis in HCC patients. The C3C4-SPS model effectively stratified patients by risk, with high-risk groups exhibiting significantly worse outcomes, and performed robustly in ROC and PCA validation.

Conclusion

This study systematically characterized the molecular and metabolic heterogeneity of liver cancer stem cell subpopulations and elucidated their decisive roles in tumor progression and patient prognosis. These findings are primarily based on computational analyses and warrant further experimental validation. The C3C4-SPS model provides a novel tool for prognostic assessment and potentially individualized therapy targeting stemness-associated features in hepatocellular carcinoma.