单细胞转录组分析揭示胃肠道间质瘤肿瘤相关巨噬细胞异质性与伊马替尼耐药关系

Single-Cell Transcriptomic Analysis Reveals the Heterogeneity of Tumor-Associated Macrophages and Their Association with Imatinib Resistance in Gastrointestinal Stromal Tumors

摘要: 目的：胃肠道间质瘤(GIST)是胃肠道中最常见的间质肿瘤。尽管选择性酪氨酸激酶抑制剂伊马替尼显著延长了大多数患者的生存期，但原发性耐药和获得性耐药性仍然是一个主要的治疗挑战。越来越多的证据表明，肿瘤免疫微环境参与影响治疗反应及肿瘤耐药，肿瘤相关巨噬细胞(tumor-associated macrophages, TAMs)在其中起重要作用。然而，GIST中TAM亚群的功能多样性和动态演变仍有待进一步研究。方法：我们从基因表达综合数据库(GEO)中获得了GIST样本的单细胞RNA测序(scRNA-seq)数据，包括伊马替尼敏感患者和耐药患者的测序数据。经过严格的质量控制，我们进行了降维、无监督聚类和细胞类型注释。提取TAM，并根据其转录组特征将其进一步分类为不同的亚群。使用基因集变异分析(GSVA)对TAM亚群进行功能表征，以鉴定各个亚群中富集的信号通路。使用Monocle2进行拟时序分析，以推断巨噬细胞亚群的动态分化路径。同时，我们描述了GIST患者中肿瘤细胞亚型的转录和功能异质性。基于配体–受体相互作用模型，使用CellChat推断TAM和肿瘤细胞之间的细胞间通讯模式。结果：我们鉴定出5个转录组特征不同的TAM亚群，它们具有不同的功能特征。炎症细胞因子富集型TAMs (inflammatory cytokine-enriched tumor-associated macrophages, Inflam-TAMs)在伊马替尼敏感患者肿瘤组织中比例较高，而干扰素预激型TAMs (interferon-primed tumor-associated macrophages, IFN-TAMs)在伊马替尼耐药患者肿瘤组织中占主导地位。拟时序分析提示不同TAM状态之间可能存在连续性变化，IFN-TAMs相关状态可能出现在较后阶段。细胞间通讯分析提示，Inflam-TAMs与肿瘤细胞之间的预测互作相对较弱，并提示TAMs与特定肿瘤亚群之间可能存在潜在信号轴。结论：本研究表明，GIST中TAMs具有明显异质性，不同TAM亚群在伊马替尼敏感与耐药样本中的分布模式不同。进一步分析提示，TAM状态变化及其与肿瘤细胞的相互作用可能与GIST耐药相关微环境有关。

Abstract: Objective: Gastrointestinal stromal tumor (GIST) is the most common mesenchymal tumor of the gastrointestinal tract. Although imatinib, a selective tyrosine kinase inhibitor, has significantly improved the survival of patients with GIST, primary and acquired resistance remain major therapeutic challenges. Increasing evidence suggests that the tumor immune microenvironment is involved in treatment response and drug resistance, among which tumor-associated macrophages (TAMs) play an important role. However, the functional diversity and dynamic evolution of TAM subsets in GIST remain insufficiently understood. Methods: Single-cell RNA sequencing (scRNA-seq) data of GIST samples were obtained from the Gene Expression Omnibus (GEO), including samples from imatinib-sensitive and imatinib-resistant patients. After quality control, dimensionality reduction, unsupervised clustering, and cell-type annotation were performed. TAMs were extracted and further classified into distinct subsets according to their transcriptomic characteristics. Gene set variation analysis (GSVA) was used to characterize the functional features of TAM subsets and identify enriched signaling pathways. Monocle2 was applied for pseudotime analysis to infer the potential state-transition trend of macrophage subsets. In addition, the transcriptional and functional heterogeneity of tumor cell subpopulations was analyzed. CellChat was used to infer potential ligand-receptor-mediated communication patterns between TAMs and tumor cells. Results: Five TAM subsets with distinct transcriptomic features were identified. Inflammatory cytokine-enriched TAMs (Inflam-TAMs) were more abundant in imatinib-sensitive tumor samples, whereas interferon-primed TAMs (IFN-TAMs) predominated in imatinib-resistant tumor samples. Pseudotime analysis suggested potential continuity among different TAM states, with the IFN-TAM-associated state tending to appear at a later stage. Cell-cell communication analysis indicated relatively weak predicted interactions between Inflam-TAMs and tumor cells, and suggested potential signaling axes between TAMs and specific tumor subpopulations. Conclusion: This study demonstrated marked heterogeneity of TAMs in GIST, with distinct distribution patterns of TAM subpopulations between imatinib-sensitive and imatinib-resistant samples. Further analyses suggested that TAM state transitions and their interactions with tumor cells may be associated with the imatinib resistance-related tumor microenvironment in GIST.