Background

Hepatocellular carcinoma (HCC) is one of the most common malignancies and inflicts high mortality worldwide. The effect of tumor microenvironment components on HCC oncogenesis remains unknown. In particular, the nonleukocyte portion of the stromal fraction (SF) is poorly understood.

Methods

We comprehensively evaluated the proportional cell counts and gene expression data from The Cancer Genome Atlas (TCGA) to examine the contributions of cell components to the tumor microenvironment. Single-cell sequencing data from the Gene Expression Omnibus (GEO) were also analyzed to verify the association between the nonleukocyte SF and genes. We classified HCC using a hierarchical clustering method based on diversity of nonleukocyte SF–related gene expression among different components, and we used an appropriate GEO dataset to verify the clusters with a support vector machine (SVM) model. The prognosis of subtypes and their relationship with tumor microenvironmental cell proportions, clinicopathogenesis factors, and other indicators were evaluated.

Results

Based on linear regression, 711 genes related to nonleukocyte SF were selected from the TCGA dataset. We classified HCC into three subtypes using genes related to the nonleukocyte SF. Additionally, the GEO single-cell sequencing data confirmed the relationship between genes and the nonleukocyte SF. The tumor microenvironment of Type 2 contained the most significant mutually reinforcing interaction between the nonleukocyte SF and tumor cells. Meanwhile, Type 2 patients had the poorest prognosis and the most severe tumor-node-metastasis (TNM) stages, histological grades, etc. The analysis based on the GEO dataset verified the classification results with an SVM model. Type 2 was associated with worse clinicopathological characteristics, including tumor grading and staging, than the other types. In addition, the pathway analysis revealed that signals related to the SF and cell proliferation were significantly enhanced in Type 2 compared to the other group, which consisted of Types 1 and 3.

Conclusion

The nonleukocyte SF in the tumor microenvironment contributed greatly to HCC oncogenesis. We can use these HCC classification criteria to stratify patients into subtypes for personalized treatment.

中文翻译：

---

具有不同比例的非白细胞基质细胞的肝细胞癌亚型的临床和基因组学格局。

背景

肝细胞癌（HCC）是最常见的恶性肿瘤之一，并在全球范围内造成很高的死亡率。肿瘤微环境成分对肝癌发生的影响尚不清楚。特别是，对基质部分（SF）的非白细胞部分了解甚少。

方法

我们全面评估了癌症基因组图谱（TCGA）的比例细胞计数和基因表达数据，以检查细胞成分对肿瘤微环境的贡献。还分析了来自Gene Expression Omnibus（GEO）的单细胞测序数据，以验证非白细胞SF与基因之间的关联。我们基于不同成分之间非白细胞SF相关基因表达的多样性，使用分级聚类方法对HCC进行分类，并使用适当的GEO数据集通过支持向量机（SVM）模型验证了聚类。评估亚型的预后及其与肿瘤微环境细胞比例，临床病理因素和其他指标的关系。

结果

基于线性回归，从TCGA数据集中选择了与非白细胞SF相关的711个基因。我们使用与非白细胞SF相关的基因将HCC分为三种亚型。此外，GEO单细胞测序数据证实了基因与非白细胞SF之间的关系。2型肿瘤微环境在非白细胞SF和肿瘤细胞之间具有最显着的相互促进相互作用。同时，2型患者的预后最差，肿瘤淋巴结转移（TNM）分期，组织学等级等最为严重。基于GEO数据集的分析使用SVM模型验证了分类结果。2型与其他类型相比，其临床病理特征（包括肿瘤分级和分期）更差。此外，

结论

肿瘤微环境中的非白细胞SF极大地促进了HCC的发生。我们可以使用这些HCC分类标准将患者分为亚型进行个性化治疗。