Background: Esophageal squamous cell carcinoma (ESCC) is an aggressive and fatal malignancy that leads to epithelial cancer. The association between epithelial cell heterogeneity, prognosis, and immune response in this cancer remains uncertain. This study aimed to investigate epithelial cell heterogeneity in ESCC and develop a predictive risk model using the identified cell types.
Methods: Single-cell RNA sequencing (scRNA-seq) and differential ESCC gene data were accessed from the Gene Expression Omnibus. Functional enrichment analysis, inferCNV, cell development trajectories, and intercellular communication were analyzed following epithelial cell characterization. Differentially expressed ESCC (n = 773) and epithelial cell marker genes (n = 3407) were intersected to obtain core genes, and epithelial cell-related prognostic genes were identified. LASSO regression analysis was used to construct a prognostic model. The external dataset GSE53624 was used to further validate the stability of the model. Drug sensitivity predictions, and immune cell infiltration were analyzed. Molecular docking clarified the possible therapeutic role of β-sitosterol in ESCC. Finally, wound healing assay, cell colony, and transwell assay were constructed to detect the effects of the core gene PDLIM2 on ESCC cell proliferation, invasion, and migration.
Results: Eight cell clusters were identified, and epithelial cells were categorized into tumor and paratumor groups. The tumor group possessed more chromosomal variants than the paratumor group. Epithelial cells were associated with multiple cell types and significantly correlated with the Wnt, transforming growth factor, and epidermal growth factor signaling pathways. From 231 intersected genes, five core genes were screened for use in the risk model: CTSL, LAPTM4B, MYO10, NCF2, and PDLIM2. These genes may contribute to the cancerous transformation of normal esophageal epithelial cells and thereby act as biomarkers and potential therapeutic targets in patients with ESCC. β-Sitosterol furthermore displayed excellent docking potential with these genes. Meanwhile, further experiments demonstrated that the gene PDLIM2 plays a major role in the progression of oesophageal squamous carcinoma.
Conclusion: We successfully developed a risk model for the prognosis of ESCC based on epithelial cells that addresses the response of ESCC to immunotherapy and offers novel cancer treatment options.



中文翻译：

---

使用上皮细胞的预后风险模型将 β-谷甾醇识别为食管鳞状细胞癌的潜在治疗靶点

背景：食管鳞状细胞癌（ESCC）是一种侵袭性、致命性恶性肿瘤，可导致上皮癌。这种癌症的上皮细胞异质性、预后和免疫反应之间的关联仍不确定。本研究旨在研究食管鳞癌中的上皮细胞异质性，并使用已识别的细胞类型开发预测风险模型。
方法：从基因表达综合库获取单细胞 RNA 测序 (scRNA-seq) 和差异 ESCC 基因数据。在上皮细胞表征后，对功能富集分析、inferCNV、细胞发育轨迹和细胞间通讯进行了分析。将差异表达的 ESCC (n = 773) 和上皮细胞标志基因 (n = 3407) 相交以获得核心基因，并鉴定上皮细胞相关的预后基因。 LASSO回归分析用于构建预后模型。使用外部数据集GSE53624进一步验证模型的稳定性。分析药物敏感性预测和免疫细胞浸润。分子对接阐明了β-谷甾醇在食管鳞癌中可能的治疗作用。最后构建伤口愈合实验、细胞集落实验、Transwell实验，检测核心基因PDLIM2对ESCC细胞增殖、侵袭和迁移的影响。
结果：鉴定出八个细胞簇，并将上皮细胞分为肿瘤组和肿瘤旁组。肿瘤组比肿瘤旁组拥有更多的染色体变异。上皮细胞与多种细胞类型相关，并且与Wnt、转化生长因子和表皮生长因子信号通路显着相关。从 231 个交叉基因中，筛选出 5 个核心基因用于风险模型：CTSL、LAPTM4B、MYO10、NCF2 和 PDLIM2。这些基因可能有助于正常食管上皮细胞的癌变，从而充当食管鳞癌患者的生物标志物和潜在的治疗靶点。 β-谷甾醇还表现出与这些基因优异的对接潜力。同时，进一步的实验表明，PDLIM2基因在食管鳞癌的进展中发挥着重要作用。
结论：我们成功开发了一种基于上皮细胞的 ESCC 预后风险模型，该模型解决了 ESCC 对免疫治疗的反应，并提供了新的癌症治疗选择。