中文摘要

目的：肿瘤微环境（Tumor immune microenvironment，TME）对肿瘤的发生发展起到至关重要的作用，与肿瘤患者临床预后密切相关。肿瘤免疫治疗主要利用肿瘤免疫微环境中的免疫细胞对肿瘤细胞进行特异性识别和攻击，其免疫效应机制与免疫微环境中CD8阳性T淋巴细胞和肿瘤相关巨噬细胞有关。CD8阳性T淋巴细胞和肿瘤相关巨噬细胞作为免疫微环境中主要的效应细胞成分，影响着肺腺癌(lung adenocarcinoma，LUAD）的发生发展。本研究旨在探索LUAD中与CD8阳性T淋巴细胞和肿瘤相关巨噬细胞浸润相关的关键基因。

方法：本研究中我们使用肿瘤基因组图谱(The Cancer Genome Atlas，TCGA)数据库中肺腺癌患者相关数据进行了分析，筛选与临床预后相关的、与肿瘤微环境相关的基因。Kaplan-Meier生存曲线、多因素Cox回归用于确定与临床预后相关的基因。使用CIBERSORT算法，分析目的基因与CD8阳性T淋巴细胞和CD163阳性肿瘤相关巨噬细胞浸润显著相关的基因的相关性。我们使用肺腺癌组织芯片、PrognoScan数据库、TIMER数据库进行外部验证。进一步地，我们对目的基因的差异表达进行了富集分析。最后，结合目的基因构建预测模型，通过ROC曲线和校准曲线对模型效能进行评估。

结果：根据TCGA数据库中515例LUAD患者数据，我们提取了20530个基因表达数据，分析了各基因表达与预后的关系，与T分期、N分期、M分期、CD8阳性T淋巴细胞和CD163阳性肿瘤相关巨噬细胞的相关性。我们发现了Ras激活蛋白样3 (RAS protein activator like 3，RASAL3)在LUAD组织中表达低于正常组织。RASAL3高表达患者预后更好，RASAL3的低表达与患者更晚的T分期，M分期，TNM分期和更差的病理分级相关。TCGA-LUAD患者的CD8 mRNA表达水平与RASAL3 mRNA表达呈正相关。运用CIBERSORT算法分析显示RASAL3表达与CD8阳性T淋巴细胞浸润正相关。在TCGA数据库队列中，CD8阳性T淋巴细胞的标志物CD8的 mRNA表达水平与RASAL3 mRNA表达呈正相关。进一步地，我们使用组织芯片、PrognoScan数据库中两个LUAD队列数据集(GSE13213和GSE31210)、TIMER数据库进行验证，结果趋势与TCGA数据库、CIBERSORT算法一致。对TCGA中肺腺癌患者RASAL3的表达进行GSEA分析，发现RASAL3异常表达有可能影响脂代谢重编程。

此外，在TCGA数据库中，我们还发现LUAD组织中1,4-α-葡聚糖分支酶 1(1,4-alpha-glucan branching enzyme 1 ，GBE1)的表达水平显著高于癌旁正常组织。GBE1的高表达与LUAD中较差的总体生存（OS）相关。而且，GBE1的高表达与患者更晚的T分期，N分期，M分期，TNM分期和更差的病理分级相关。GBE1与LUAD中CD163阳性肿瘤相关巨噬细胞的浸润水平呈正相关。运用CIBERSORT算法分析显示GBE1表达与M2型巨噬细胞，CD4阳性记忆激活T细胞和静息树突状细胞的浸润呈正相关。我们使用PrognoScan数据库、TIMER数据库、组织芯片进行验证。分析PrognoScan数据库中三个LUAD队列（HARVARD-LC，GSE31210，jacob-00182-UM）的结果表明， LUAD患者中GBE1的高表达者预后更差。通过TIMER数据库分析了GBE1表达与各肿瘤免疫细胞浸润水平之间的相关性，结果显示，GBE1表达与巨噬细胞浸润正相关。使用TIMER数据库探究了GBE1与各肿瘤免疫细胞的免疫标记基因的关系，发现GBE1与LUAD中的M2型巨噬细胞免疫标记基因CD163，VSIG4和MS4A4A呈正相关。TCGA队列中的CD163 mRNA表达也与GBE1 mRNA表达正相关。我们通过使用基于组织芯片TMA的IHC进一步证实了GBE1表达与CD163阳性肿瘤相关巨噬细胞浸润之间的关系。进一步地，对TCGA中肺腺癌患者GBE1的差异表达进行GSEA分析，发现GBE1异常表达可能是通过mTORC1信号通路，参与脂肪酸代谢。

我们进一步联合RASAL3与GBE1的表达与传统的TNM分期构建“改良TNM分期”预测模型，校准曲线的结果表明，改良TNM分期的校准曲线和对角线拟合良好，对于五年生存率的预测较为准确。改良TNM分期的AUC值高于传统TNM分期，但是并未达到显著差异。

结论：本研究发现RASAL3与肺腺癌中CD8阳性T细胞浸润成正相关，GBE1表达与CD163阳性肿瘤相关巨噬细胞浸润正相关。RASAL3表达与GBE1表达可用以预测肺腺癌的预后情况，可为进一步探索肺腺癌中与CD8阳性T淋巴细胞和CD163阳性肿瘤相关巨噬细胞浸润相关的肿瘤免疫微环境机制提供新的方向，并为肺腺癌的预后预测提供了新的手段。RASAL3与GBE1有可能成为LUAD的预后和免疫生物标志物。

Abstract

Objective: The tumor immune microenvironment (TME) plays a crucial role in the development and progression of cancer, and is closely associated with clinical outcomes in cancer patients. Tumor immunotherapy mainly relies on the recognition and attack of tumor cells by immune cells in the TME, and its immunological mechanisms are related to CD8⁺ T lymphocytes and tumor-associated macrophages in the immune microenvironment. CD8⁺ T lymphocytes and tumor-associated macrophages (TAMs) are the main effector cell components in the TME, and they affect the occurrence and development of lung adenocarcinoma (LUAD). The aim of this study is to explore key genes associated with the infiltration of CD8⁺ T lymphocytes and TAMs in LUAD.

Methods: In this study, we analyzed lung adenocarcinoma patients' data from The Cancer Genome Atlas (TCGA) database to screen genes related to clinical prognosis and tumor microenvironment. Kaplan-Meier survival curves and multivariate Cox regression were used to identify genes associated with clinical prognosis. The CIBERSORT algorithm was used to analyze the correlation between the target genes and CD8-positive T lymphocytes and CD163-positive tumor-associated macrophages infiltration. External validation was performed using lung adenocarcinoma tissue microarrays, the PrognoScan database, and the TIMER database. Furthermore, enrichment analysis was conducted on differentially expressed target genes. Finally, a prediction model was constructed based on the target genes, and the model's performance was evaluated using ROC and calibration curves.

Results: According to the data of 515 LUAD patients from the TCGA database, we extracted expression data of 20,530 genes and analyzed their relationship with prognosis, as well as their correlation with T-stage, N-stage, M-stage, CD8-positive T lymphocytes, and CD163-positive tumor-associated macrophages. We found that RAS protein activator like 3 (RASAL3 ) was expressed at lower levels in LUAD tissues compared to normal tissues. Patients with high RASAL3 expression had better prognosis, while low RASAL3 expression was associated with later T-stage, M-stage, TNM stage, and worse pathological grade. The mRNA expression levels of CD8 in TCGA-LUAD patients were positively correlated with RASAL3 mRNA expression. CIBERSORT analysis showed that RASAL3 expression was positively correlated with CD8-positive T lymphocyte infiltration. In the TCGA database cohort, the mRNA expression levels of CD8, a marker for CD8-positive T lymphocytes, were positively correlated with RASAL3 mRNA expression. Furthermore, we validated our findings using tissue microarrays, two LUAD datasets (GSE13213 and GSE31210) from the PrognoScan database, and the TIMER database, and the results were consistent with those from the TCGA database and the CIBERSORT algorithm. GSEA analysis of RASAL3 expression in LUAD patients from the TCGA database suggested that abnormal RASAL3 expression may affect lipid metabolism reprogramming.

       In addition, we found that the expression level of 1,4-alpha-glucan branching enzyme 1 (GBE1) in LUAD tissue was significantly higher than that in adjacent normal tissue in the TCGA database. The high expression of GBE1 was associated with poor overall survival (OS) in LUAD. Furthermore, high expression of GBE1 was correlated with later T stage, N stage, M stage, TNM stage, and worse pathological grade in LUAD patients. GBE1 was positively correlated with the infiltration level of CD163-positive tumor-associated macrophages in LUAD. CIBERSORT analysis showed that the expression of GBE1 was positively correlated with the infiltration of M2 macrophages, CD4-positive memory activated T cells, and resting dendritic cells in LUAD. We validated our findings using the PrognoScan database, TIMER database, and tissue microarray. Analysis of three LUAD cohorts (HARVARD-LC, GSE31210, jacob-00182-UM) in the PrognoScan da

tabase showed that patients with high expression of GBE1 had worse prognosis. Analysis of the correlation between GBE1 expression and tumor immune cell infiltration levels in TIMER database showed that GBE1 expression was positively correlated with macrophage infiltration in LUAD. GBE1 was also found to be positively correlated with immune marker genes CD163, VSIG4, and MS4A4A of M2 macrophages in LUAD through TIMER database analysis. The expression of CD163 mRNA in TCGA cohort was also positively correlated with GBE1 mRNA expression. Furthermore, we used tissue microarray-based IHC to confirm the relationship between GBE1 expression and the infiltration of CD163-positive tumor-associated macrophages. Finally, GSEA analysis of differential expression of GBE1 in lung adenocarcinoma patients in TCGA revealed that abnormal expression of GBE1 may be involved in fatty acid metabolism through the mTORC1 signaling pathway.

       We further constructed an "improved TNM staging" prediction model by combining the expression of RASAL3 and GBE1 with the traditional TNM staging. The calibration curve showed a good fit between the improved TNM staging and the diagonal line, and the model provided a more accurate prediction of the five-year survival rate. Although the AUC value of the improved TNM staging was higher than that of the traditional TNM staging, the difference was not statistically significant.

Conclusion: This study reveals a positive correlation between RASAL3 expression and CD8⁺ T cell infiltration in lung adenocarcinoma, as well as a positive correlation between GBE1 expression and CD163⁺ tumor-associated macrophage infiltration. The expression levels of RASAL3 and GBE1 can be used to predict the prognosis of lung adenocarcinoma, providing new avenues for exploring the mechanisms of tumor immune microenvironment related to CD8⁺ T lymphocytes and CD163⁺ tumor-associated macrophages in lung adenocarcinoma. Moreover, RASAL3 and GBE1 may serve as prognostic and immune biomarkers for LUAD.