Supplementary Material for: Identification of A Novel Activated NK-Associated Gene Score Associated with Diagnosis and Biological Therapy Response in Ulcerative Colitis

Introduction: Natural killer (NK) cells are associated with the pathogenesis of ulcerative colitis (UC); however, their precise contributions remain unclear. The present study aimed to investigate the diagnostic value of the activated NK-associated gene (ANAG) score in UC and evaluate its predictive value in response to biological therapy.
Methods: Bulk RNA-seq and scRNA-seq datasets were obtained from the Gene Expression Omnibus (GEO) and Single Cell Portal (SCP) databases. In the bulk RNA-seq, differentially expressed genes (DEGs) were screened by the “Batch correction” and “Robust rank aggregation” (RRA) methods. The immune infiltration landscape was estimated using single-sample gene set enrichment analysis (ssGSEA) and CIBERSORT. DEGs that correlated with activated NK cells were identified as activated NK-associated genes (ANAGs). Protein-protein interaction (PPI) analysis and least absolute shrinkage and selection operator (LASSO) regression were used to screen key ANAGs and establish an ANAG score. The expression levels of the four key ANAGs were validated in human samples by real-time quantitative polymerase chain reaction (RT-qPCR) and immunofluorescence. The potential therapeutic drugs for UC were identified using the DSigDB database. Through scRNA-seq data analysis, the cell scores based on the ANAGs were calculated by “AddModuleScore” and “AUCell”.
Results: Immune infiltration analysis revealed a higher abundance of activated NK cells in non-inflamed UC tissues (ssGSEA, P<0.001; CIBERSORT, P<0.01). Fifty-four DEGs correlated with activated NK cells were identified as ANAGs. The ANAG score was established using four key ANAGs (SELP, TIMP1, MMP7, and ABCG2). The ANAG scores were significantly higher in inflamed tissues (P<0.001) and in biological therapy non-responders (NR) tissues before treatment (golimumab, P<0.05; ustekinumab, P<0.001). The ANAG score demonstrated an excellent diagnostic value (AUC = 0.979). Patients with higher ANAG scores before treatment were more likely to experience a lack of response to golimumab or ustekinumab (golimumab, P<0.05; ustekinumab, P<0.001).
Conclusion: This study established a novel ANAG score with the ability to precisely diagnose UC and distinguish the efficacy of biological treatment.

引言：自然杀伤（NK）细胞与溃疡性结肠炎（UC）的发病机制相关，但其确切作用仍不明确。本研究旨在探讨活化NK细胞相关基因（ANAG）评分对UC的诊断价值，并评估其对生物治疗应答的预测价值。
方法：从基因表达综合（GEO）数据库与单细胞门户（SCP）数据库获取批量RNA测序（bulk RNA-seq）及单细胞RNA测序（scRNA-seq）数据集。在bulk RNA-seq数据中，通过批次校正与稳健秩聚合（RRA）方法筛选差异表达基因（DEGs）。采用单样本基因集富集分析（ssGSEA）与CIBERSORT评估免疫浸润特征。将与活化NK细胞相关的DEGs鉴定为活化NK细胞相关基因（ANAGs）。通过蛋白质相互作用（PPI）分析与最小绝对收缩和选择算子（LASSO）回归筛选关键ANAGs，并构建ANAG评分模型。采用实时定量聚合酶链反应（RT-qPCR）与免疫荧光在人体样本中验证4个关键ANAGs的表达水平。利用DSigDB数据库筛选UC的潜在治疗药物。通过scRNA-seq数据分析，采用AddModuleScore与AUCell方法计算基于ANAGs的细胞评分。
结果：免疫浸润分析显示，非炎症性UC组织中活化NK细胞的丰度更高（ssGSEA：P<0.001；CIBERSORT：P<0.01）。共鉴定出54个与活化NK细胞相关的DEGs作为ANAGs。通过4个关键ANAGs（SELP、TIMP1、MMP7及ABCG2）构建ANAG评分模型。炎症组织与治疗前生物治疗无应答者（NR）组织的ANAG评分显著更高（戈利木单抗组：P<0.05；乌司奴单抗组：P<0.001）。ANAG评分展现出优异的诊断价值（曲线下面积AUC=0.979）。治疗前ANAG评分较高的患者更可能对戈利木单抗或乌司奴单抗治疗无应答（戈利木单抗组：P<0.05；乌司奴单抗组：P<0.001）。
结论：本研究构建了一种新型ANAG评分，可精准诊断UC并区分生物治疗的疗效。