PO.MD01.02 · 分子诊断与数据
早发性头颈部鳞状细胞癌中独特的临床、分子及多组学生物标志物
Distinct clinical, molecular and multi-omics biomarkers in early-onset head and neck squamous cell carcinoma
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
背景:早发性头颈部鳞状细胞癌(EOHNSCC)的发病率在全球范围内上升,且往往独立于烟草和酒精等传统危险因素。EOHNSCC 是否携带与典型发病头颈部鳞状细胞癌(TOHNSCC)不同的分子和多组学特征仍知之甚少。
方法:我们分析了 The Cancer Genome Atlas(TCGA)Firehose Legacy 和 AACR GENIE 数据集,以比较 EOHNSCC(18-49 岁)和 TOHNSCC(>49 岁)HNSCC。特征包括体细胞突变、拷贝数改变(CNA)、结构变异、基因融合、DNA 甲基化、mRNA 表达和蛋白表达。使用 Fisher 精确检验和 Wilcoxon 秩和检验评估 HNSCC 发病组间差异的显著性(p < 0.05)。通路富集利用 GO、KEGG、Reactome 和 CORUM 数据库。多变量 logistic 回归评估 EOHNSCC 与临床协变量之间的关联。
结果:在两个数据集中,我们的研究队列由 2,472 例 HNSCC 个体组成(N AACR GENIE = 1,961;N TCGA = 511)。在 AACR GENIE 中,EOHNSCC 患者更可能为女性(aOR = 1.49,95% CI:1.09-2.02)、非白人(aOR = 1.99,95% CI:1.38-2.83),并具有较低的突变负荷(aOR = 0.44,95% CI:0.29-0.64)。然而,在 TCGA 中,EO-HNSCC 患者具有较低组织学分级的肿瘤(G3 对比 G1 aOR = 0.38,95% CI:0.14-1.00),更可能接受新辅助治疗(aOR = 17.02,95% CI:3.32-101.65),且几乎没有或没有既往吸烟史(戒烟者对比非吸烟者:aOR = 0.28,95% CI:0.13-0.57)。TCGA 中 EOHNSCC 的分子标志物特征为 GDNF、FRYL、KIAA0586、CADPS、MYF5 和 DCTN1 的并发突变和 mRNA 过表达,富集于缺氧相关通路。CNA 与表达的整合突显了 ASH2L 和 SERPIND1,它们与细胞周期调控信号相关。甲基化驱动的表达涉及运动相关基因(SPIPM6、SPAG6、ODAD3、RSPH4A),富集于纤毛运动和 DAP12 信号。HSPA2 低甲基化与 HSPA2 CNA 和 mRNA 表达上调相关,与癌症侵袭和转移相关。具有 mRNA-蛋白一致性的基因包括 NOTCH1、SYK、SERPINE1、IRS1、CHEK2、EIF4E;与翻译起始相关。在 AACR GENIE 中,EOHNSCC 显示出较高的 KCNIP1 突变,而 TOHNSCC 则富集于 NOTCH1、FAT1、KMT2D 和 PIK3CA。值得注意的是,FAT1 是两个数据集中唯一均发生突变的基因;且一致地存在于 TOHNSCC 中。
结论:EOHNSCC 表现出涉及缺氧、细胞运动、细胞周期调控、癌症转移和翻译起始的独特多组学特征。候选生物标志物包括 GDNF、ASH2L、SPAG6、HSPA2、EIF4E 和 KCNIP1。数据集之间的重叠极少,突显了对预后和治疗意义进行验证和探索的必要性。
查看英文原文 English abstract
Background Early-onset head and neck squamous cell carcinoma (EOHNSCC) incidence is rising globally, often independent of traditional risk factors such as tobacco and alcohol. Whether EOHNSCC harbors distinct molecular and multi-omics signatures compared to typical-onset HNSCC (TOHNSCC) remains poorly understood.
Methods We analyzed The Cancer Genome Atlas (TCGA) Firehose Legacy and AACR GENIE datasets to compare EOHNSCC (18-49 years) and TOHNSCC (>49 years) HNSCC. Features included somatic mutations, copy number alterations (CNAs), structural variants, gene fusions, DNA methylation, mRNA expression, and protein expression. HNSCC onset group differences were assessed for significance using Fisher's exact and Wilcoxon rank sum tests (p < 0.05). Pathway enrichment leveraged GO, KEGG, Reactome, and CORUM databases. Multivariable logistic regression evaluated associations between EOHNSCC and clinical covariates.
Results Across both datasets, our study cohort was composed of 2,472 individuals with HNSCC (N AACR GENIE = 1,961; N TCGA =511). In AACR GENIE, EOHNSCC patients were more likely female (aOR = 1.49, 95% CI: 1.09-2.02), non-white (aOR = 1.99, 95% CI: 1.38-2.83), and had lower mutation burden (aOR = 0.44, 95% CI: 0.29-0.64). However, in TCGA, EO-HNSCC patients had lower histological grade tumors (G3 vs G1 aOR = 0.38, 95% CI: 0.14-1.00), more likely to receive neoadjuvant therapy (aOR = 17.02, 95% CI: 3.32-101.65), and had little to no prior smoking history (reformed vs non-smoker: aOR = 0.28, 95% CI: 0.13-0.57). Molecular markers in TCGA for EOHNSCC was characterized by concurrent mutation and mRNA overexpression in GDNF, FRYL, KIAA0586, CADPS, MYF5 , and DCTN1 , enriched in hypoxia-related pathways. Integration of CNAs and expression highlighted ASH2L and SERPIND1 , which are associated with cell cycle regulation signaling. Methylation-driven expression involved motility genes ( SPIPM6, SPAG6, ODAD3, RSPH4A ), enriched in cilium movement and DAP12 signaling. HSPA2 hypomethylation was correlated with HSPA2 CNA and upregulated mRNA expression, linked to cancer invasion and metastasis. Genes with mRNA-protein concordance included NOTCH1, SYK, SERPINE1, IRS1, CHEK2, EIF4E ; associated with translation initiation. In AACR GENIE, EOHNSCC showed higher KCNIP1 mutation, while TOHNSCC was enriched for NOTCH1, FAT1, KMT2D , and PIK3CA . Of note, FAT1 was the only gene mutated in both datasets; consistently in TOHNSCC.
Conclusions EOHNSCC exhibits distinct multi-omics signatures involving hypoxia, cell motility, cell cycle regulation, cancer metastasis, and translation initiation. Candidate biomarkers include GDNF, ASH2L, SPAG6, HSPA2, EIF4E and KCNIP1 . Minimal overlap across datasets underscores the need for validation and exploration of prognostic and therapeutic implications.
利益披露 Disclosure
J. O. Odera, None..
R. Jiang, None..
M. C. Byrd, None..
O. L. Osazuwa-Peters, None.