Background/Objectives:We aimed to discover genes with bimodal expression linked to patient outcomes, to reveal underlying oncogenotypes and identify new therapeutic insights in lung adenocarcinoma (LUAD).Methods:We performed meta-analysis to screen LUAD datasets for prognostic genes with bimodal expression patterns. Kynureninase (KYNU), a key enzyme in tryptophan catabolism, emerged as a top candidate. We then examined its relationship with LUAD mutations, metabolic alterations, immune microenvironment states, and expression patterns in human and mouse models using bulk and single-cell transcriptomics, metabolomics, and preclinical model datasets. Pan-cancer prognostic associations were also assessed.Results:Model-based clustering ofKYNUexpression outperformed median-based dichotomization in prognostic accuracy.KYNUwas elevated in tumors withKEAP1andSTK11co-mutations but remained a strong independent prognostic marker. Metabolomic analysis showed thatKYNU-high tumors had increased anthranilic acid, a catalytic product, while maintaining stable kynurenine levels, suggesting a compensatory mechanism sustaining immunosuppressive signaling. Single-cell and bulk data showedKYNUexpression was cancer cell-intrinsic in immune-cold tumors and myeloid-derived in immune-infiltrated tumors. In murine LUAD models,Kynuexpression was predominantly immune-derived and uncoupled from Nrf2/Lkb1 signaling, indicating poor model fidelity.KYNU’s prognostic associations extended across cancer types, with poor outcomes in pancreatic and kidney cancers but favorable outcomes in melanoma, underscoring the need for lineage-specific considerations in therapy development.Conclusions:KYNUis a robust prognostic biomarker and potential immunometabolic target in LUAD, especially inSTK11andKEAP1co-mutated tumors. Its cancer cell-intrinsic expression and immunosuppressive metabolic phenotype offer translational potential, though species-specific expression patterns pose challenges for preclinical modeling.
背景/目的：本研究旨在发现与患者预后相关的双峰表达基因，以揭示潜在的致癌基因型，并为肺腺癌（LUAD）提供新的治疗思路。 方法：我们通过荟萃分析筛选LUAD数据集，寻找具有双峰表达模式的预后相关基因。犬尿氨酸酶（KYNU）作为色氨酸分解代谢的关键酶，成为首要候选基因。随后，我们利用批量及单细胞转录组学、代谢组学数据以及临床前模型数据集，探究了KYNU与LUAD基因突变、代谢改变、免疫微环境状态的关系，并分析了其在人类和小鼠模型中的表达模式。同时，评估了KYNU在泛癌种中的预后关联性。 结果：基于模型的KYNU表达聚类分析在预后预测准确性上优于基于中位数的二分法。在携带KEAP1和STK11共突变的肿瘤中，KYNU表达升高，且仍是一个强力的独立预后标志物。代谢组学分析显示，KYNU高表达肿瘤中其催化产物邻氨基苯甲酸增加，而犬尿氨酸水平保持稳定，提示存在维持免疫抑制信号的代偿机制。单细胞及批量数据显示，在免疫"冷"肿瘤中，KYNU表达具有癌细胞内在性；而在免疫浸润肿瘤中，则主要来源于髓系细胞。在小鼠LUAD模型中，Kynu表达主要来源于免疫细胞，且与Nrf2/Lkb1信号通路无关，表明模型保真度较低。KYNU的预后关联性在不同癌种中具有差异：在胰腺癌和肾癌中与不良预后相关，而在黑色素瘤中则与良好预后相关，这凸显了在治疗开发中需考虑谱系特异性。 结论：KYNU是LUAD中一个稳健的预后生物标志物和潜在的免疫代谢治疗靶点，尤其在STK11和KEAP1共突变肿瘤中。其癌细胞内在性表达和免疫抑制代谢表型具有转化潜力，但物种特异性表达模式对临床前建模提出了挑战。
肿瘤(癌症)患者之家