Li Chenyang, Hong Wei, Reuben Alexandre, Wang Linghua, Maitra Anirban, Zhang Jianjun, Cheng Chao
Department of Genomic Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA.
Graduate School of Biomedical Sciences, The University of Texas MD Anderson Cancer Center UTHealth Houston, Houston, TX 77030, USA.
bioRxiv. 2024 Jun 27:2024.06.21.600089. doi: 10.1101/2024.06.21.600089.
Accumulating evidence suggests that the tumor immune microenvironment (TIME) significantly influences the response to immunotherapy, yet this complex relationship remains elusive. To address this issue, we developed TimiGP-Response (TIME Illustration based on Gene Pairing designed for immunotherapy Response), a computational framework leveraging single-cell and bulk transcriptomic data, along with response information, to construct cell-cell interaction networks associated with responders and estimate the role of immune cells in treatment response. This framework was showcased in triple-negative breast cancer treated with immune checkpoint inhibitors targeting the PD-1:PD-L1 interaction, and orthogonally validated with imaging mass cytometry. As a result, we identified CD8+ GZMB+ T cells associated with responders and its interaction with regulatory T cells emerged as a potential feature for selecting patients who may benefit from these therapies. Subsequently, we analyzed 3,410 patients with seven cancer types (melanoma, non-small cell lung cancer, renal cell carcinoma, metastatic urothelial carcinoma, hepatocellular carcinoma, breast cancer, and esophageal cancer) treated with various immunotherapies and combination therapies, as well as several chemo- and targeted therapies as controls. Using TimiGP-Response, we depicted the pan-cancer immune landscape associated with immunotherapy response at different resolutions. At the TIME level, CD8 T cells and CD4 memory T cells were associated with responders, while anti-inflammatory (M2) macrophages and mast cells were linked to non-responders across most cancer types and datasets. Given that T cells are the primary targets of these immunotherapies and our TIME analysis highlights their importance in response to treatment, we portrayed the pan-caner landscape on 40 T cell subtypes. Notably, CD8+ and CD4+ GZMK+ effector memory T cells emerged as crucial across all cancer types and treatments, while IL-17-producing CD8+ T cells were top candidates associated with immunotherapy non-responders. In summary, this study provides a computational method to study the association between TIME and response across the pan-cancer immune landscape, offering resources and insights into immune cell interactions and their impact on treatment efficacy.
越来越多的证据表明，肿瘤免疫微环境（TIME）对免疫治疗反应有显著影响，但这种复杂关系仍难以捉摸。为解决这一问题，我们开发了TimiGP-Response（基于基因配对设计用于免疫治疗反应的TIME图谱），这是一个计算框架，利用单细胞和批量转录组数据以及反应信息，构建与反应者相关的细胞-细胞相互作用网络，并估计免疫细胞在治疗反应中的作用。该框架在使用靶向PD-1:PD-L1相互作用的免疫检查点抑制剂治疗的三阴性乳腺癌中得到展示，并用成像质谱流式细胞术进行了正交验证。结果，我们确定了与反应者相关的CD8+ GZMB+ T细胞，其与调节性T细胞的相互作用成为选择可能从这些治疗中获益的患者的一个潜在特征。随后，我们分析了3410例接受各种免疫治疗和联合治疗以及几种化疗和靶向治疗作为对照的七种癌症类型（黑色素瘤、非小细胞肺癌、肾细胞癌、转移性尿路上皮癌、肝细胞癌、乳腺癌和食管癌）患者。使用TimiGP-Response，我们在不同分辨率下描绘了与免疫治疗反应相关的泛癌免疫格局。在TIME水平上，CD8 T细胞和CD4记忆T细胞与反应者相关，而抗炎（M2）巨噬细胞和肥大细胞在大多数癌症类型和数据集中与无反应者相关。鉴于T细胞是这些免疫治疗的主要靶点，且我们的TIME分析突出了它们在治疗反应中的重要性，我们描绘了40种T细胞亚型的泛癌格局。值得注意的是，CD8+和CD4+ GZMK+效应记忆T细胞在所有癌症类型和治疗中都至关重要，而产生IL-17的CD8+ T细胞是与免疫治疗无反应者相关的顶级候选细胞。总之，本研究提供了一种计算方法来研究TIME与泛癌免疫格局中反应之间的关联，为免疫细胞相互作用及其对治疗效果的影响提供了资源和见解。
