TLS相关特征基因：胃癌预后与精准治疗的新型生物标志物

TLS-related signature genes: novel biomarkers for prognosis and precision therapy of gastric cancer

摘要

目的：基于三级淋巴结构（TLS）相关基因构建胃癌（GC）预后风险预测模型，并探讨该模型预测GC预后的价值。方法：从TCGA数据库和GEO数据库中下载GC组的基因表达数据和临床信息，通过多因素Cox回归分析，设计一个以TLS相关基因为中心的预后风险预测模型。利用该模型预测GC患者的预后风险和化疗反应，并用ROC曲线对风险预测模型进行效能评估。结果：在443例GC基因表达数据中找到366个TLS相关基因，利用单因素Cox回归筛选出COL5A1、HGF、AR、MPO、AXIN2、CTLA4、BRIP1、TLR7、F5、THP0、SERPINE1、MAPK10、NOTCH3、ENG、CRYAB、COL4A1、COL4A5、COL1A1、TYK2等44个与预后相关的基因，通过多因素Cox和SHAP筛选出排名靠前的前19个基因，并构建预后风险预测模型，风险评分=-0.765 × COL5A1-0.512 × HGF-0.509 ×AR+0.264× MPO-0.111 × AXIN2 - 0.390 × CTLA4 - 0.422 × BRIP1 + 0.703 × TLR7 + 0.133 × F5 + 0.221 × THPO + 0.389 × SERPINE1 + 1.110 × MAPK10 + 0.386 × NOTCH3 - 0.459 × ENG - 0.259 × CRYAB + 0.401 × COL4A1 + 0.212 × COL4A5 + 0.458 × COL1A1 - 0.496 × TYK2。风险评分升高与患者更差的预后、更抑制的免疫微环境以及降低的药物敏感性相关，同时伴随较低的肿瘤突变负荷水平。与传统的TNM分期相比，该预后风险预测模型表现出更高的一致性指数和更优的受试者工作特征曲线（ROC）下面积（AUC=0.72）。结论：基于TLS相关基因构建的预后风险模型对GC的预后评估具有良好预测性能，并能为免疫治疗和化疗决策提供参考。

Abstract

Objective: To construct a prognostic risk prediction model for gastric cancer(GC) based on tertiary lymphoid structure(TLS)-related genes, and explore the predictive value of this model in GC prognosis. Methods: Gene expression profiles and clinical information for GC were downloaded from The Cancer Genome Atlas(TCGA) and the Gene Expression Omnibus(GEO). A TLS-related gene-centered prognostic risk model was developed using multivariable Cox regression. The model was applied to predict prognosis risk and chemotherapy response, and its performance was evaluated using receiver operating characteristic(ROC) curves. Results: Among 443 GC gene expression datasets, 366 TLS-related genes were identified. Univariate Cox regression screened 44 prognostic genes, including COL5A1, HGF, AR, MPO, AXIN2, CTLA4, BRIP1, TLR7, F5, THP0, SERPINE1, MAPK10, NOTCH3, ENG, CRYAB, COL4A1, COL4A5, COL1A1 and TYK2, etc. Nineteen top-ranked genes were further selected using multivariable Cox regression and SHAP, and a prognostic risk model was established with the following formula: Risk score=−0.765×COL5A1−0.512×HGF−0.509×AR+0.264×MPO−0.111×AXIN2−0.390×CTLA4−0.422×BRIP1+0.703×TLR7+0.133×F5+0.221×THPO+0.389×SERPINE1+1.110×MAPK10+0.386×NOTCH3−0.459×ENG−0.259×CRYAB+0.401×COL4A1+0.212×COL4A5+0.458×COL1A1−0.496×TYK2. A higher risk score was associated with poorer prognosis, a more immunosuppressive tumor microenvironment, reduced drug sensitivity, and lower tumor mutational burden(TMB). Compared with conventional TNM staging, the model showed a higher concordance index and superior ROC performance(AUC=0.72). Conclusion: The prognostic risk model constructed based on TLS-related genes exhibits favorable predictive performance for prognostic evaluation of GC, and can provide a reference for decision-making regarding immunotherapy and chemotherapy.

关键词

胃癌 / 三级淋巴结构 / 预后风险模型 / 肿瘤免疫微环境 / 肿瘤突变负荷

Key words

Gastric cancer / Tertiary lymphoid structures / Prognostic risk model / Tumor immune microenvironment / Tumor mutational burden

参考文献

[1] Sundar R , Nakayama I , Markar S R , et al. Gastric cancer[J]. Lancet, 2025, 405(10494): 2087-2102. DOI: 10.1016/S0140—6736(25)00052—2.

[2] Al—Batran S E , Homann N , Pauligk C , et al. Perioperative chemotherapy with fluorouracil plus leucovorin, oxaliplatin, and docetaxel versus fluorouracil or capecitabine plus cisplatin and epirubicin for locally advanced, resectable gastric or gastro—oesophageal junction adenocarcinoma (FLOT4): a randomised, phase 2/3 trial[J]. Lancet, 2019, 393(10184): 1948-1957. DOI: 10.1016/S0140—6736(18)32557—1.

[3] Noh S H , Park S R , Yang H K , et al. Adjuvant capecitabine plus oxaliplatin for gastric cancer after D2 gastrectomy(CLASSIC): 5—year follow—up of an open—label, randomised phase 3 trial[J]. Lancet Oncol, 2014, 15(12): 1389-1396. DOI: 10.1016/S1470—2045(14)70473—5.

[4] Qin J , Gong Q , Zhou C , et al. Differential expression pattern of CC chemokine receptor 7 guides precision treatment of hepatocellular carcinoma[J]. Sign Transduct Tar, 2025, 10(1): 229. DOI: 10.1038/s41392—025—02308—6.

[5] Kemi N , Ylitalo O , Väyrynen J P , et al. Tertiary lymphoid structures and gastric cancer prognosis[J]. APMIS, 2023, 131(1): 19-25. DOI: 10.1111/apm.13277.

[6] Stelzer G , Rosen N , Plaschkes I , et al. The GeneCards Suite: From Gene Data Mining to Disease Genome Sequence Analyses[J]. Curr Protoc Bioinformatics, 2016, 54: 1.30.1-1.30.33. DOI: 10.1002/cpbi.5.

[7] Yu G , Wang L G , Han Y , et al. clusterProfiler: an R package for comparing biological themes among gene clusters[J]. OMICS, 2012, 16(5): 284-287. DOI: 10.1089/omi.2011.0118.

[8] Lundberg S M , Lee S I . A unified approach to interpreting model predictions[C]// Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook, NY, USA: Curran Associates Inc., 2017: 4768-4777.

[9] Pepe M S , Janes H , Longton G , et al. Limitations of the odds ratio in gauging the performance of a diagnostic, prognostic, or screening marker[J]. Am J Epidemiol, 2004, 159(9): 882-890. DOI: 10.1093/aje/kwh101.

[10] Gu Z , Eils R , Schlesner M . Complex heatmaps reveal patterns and correlations in multidimensional genomic data[J]. Bioinformatics, 2016, 32(18): 2847-2849. DOI: 10.1093/bioinformatics/btw313.

[11] Smyth G K . Linear models and empirical bayes methods for assessing differential expression in microarray experiments[J]. Stat Appl Genet Mol, 2004, 3: Article3. DOI: 10.2202/1544—6115.1027.

[12] Luchini C , Bibeau F , Ligtenberg M J L , et al. ESMO recommendations on microsatellite instability testing for immunotherapy in cancer, and its relationship with PD—1/PD—L1 expression and tumour mutational burden: a systematic review—based approach[J]. An Oncol, 2019, 30(8): 1232-1243. DOI: 10.1093/annonc/mdz116.

[13] Garnett M J , Edelman E J , Heidorn S J , et al. Systematic identification of genomic markers of drug sensitivity in cancer cells[J]. Nature, 2012, 483(7391): 570-575. DOI: 10.1038/nature11005.

[14] 李治楷， 王建波 . 成熟三级淋巴结构与胃癌新辅助免疫治疗疗效及预后的关系[J]. 中国现代普通外科进展， 2025， 28（8）： 607-611. DOI: 10.3969/j.issn.1009—9905.2025.08.004.

[15] Hu C , You W , Kong D , et al. Tertiary Lymphoid Structure—Associated B Cells Enhance CXCL13+CD103+CD8+Tissue—Resident Memory T—Cell Response to Programmed Cell Death Protein 1 Blockade in Cancer Immunotherapy[J]. Gastroenterology, 2024, 166(6): 1069-1084. DOI: 10.1053/j.gastro.2023.10.022.

[16] Qi L N , Ma L , Wu F X , et al. Clinical implications and biological features of a novel postoperative recurrent HCC classification: A multi—centre study[J]. Liver Int, 2022, 42(10): 2283-2298. DOI: 10.1111/liv.15363.

[17] Bian S , Wang Y , Zhou Y , et al. Integrative single—cell multiomics analyses dissect molecular signatures of intratumoral heterogeneities and differentiation states of human gastric cancer[J]. Nat Sci Rev, 2023, 10(6): nwad094. DOI: 10.1093/nsr/nwad094.

[18] Budczies J , Kazdal D , Menzel M , et al. Tumour mutational burden: clinical utility, challenges and emerging improvements[J]. Nat Rev Clin Oncol, 2024, 21(10): 725-742. DOI: 10.1038/s41571—024—00932—9.

[19] Yang M , Lu Z , Yu B , et al. COL5A1 Promotes the Progression of Gastric Cancer by Acting as a ceRNA of miR—137—3p to Upregulate FSTL1 Expression[J]. Cancers, 2022, 14(13): 3244. DOI: 10.3390/cancers14133244.

[20] Venable E , Knight D R T , Thoreson E K , et al. COL1A1 and COL1A2 variants in Ehlers—Danlos syndrome phenotypes and COL1—related overlap disorder[J]. Am J Med Genet., 2023, 193(2): 147-159. DOI: 10.1002/ajmg.c.32038.

[21] Yu X , Long Y , Chen B , et al. PD—L1/TLR7 dual—targeting nanobody—drug conjugate mediates potent tumor regression via elevating tumor immunogenicity in a host—expressed PD—L1 bias—dependent way[J]. J Immunother Cancer, 2022, 10(10): e004590. DOI: 10.1136/jitc—2022—004590.

基金资助

北京市自然科学基金项目(7252138)