Table_3_The Landscape of Immune Cells Infiltrating in Prostate Cancer.docx

BackgroundThis study was to explore the infiltration pattern of immune cells in the prostate cancer (PCa) microenvironment and evaluate the possibility of specific infiltrating immune cells as potential prognostic biomarkers in PCa.
MethodsInfiltrating percentage of 22 immune cells were extracted from 27 normalized datasets by CIBERSORT algorithm. Samples with CIBERSORT p-value < 0.05 were subsequently merged and divided into normal or tumor groups. The differences of 22 immune cells between normal and tumor tissues were analyzed along with potential infiltrating correlations among 22 immune cells and Gleason grades. SNV data from TCGA was used to calculate the TMB score. A univariate and multivariate regression were used to evaluate the prognostic effects of immune cells in PCa.
ResultsTen immune cells with significant differences were identified, including seven increased and three decreased infiltrating immune cells from 190 normal prostate tissues and 537 PCa tissues. Among them, the percentage of infiltration of resting NK cells increased the most, whereas the percentage of infiltration of resting mast cells decreased the most. In normal tissues, CD8+ T cells had the strongest infiltrating correlation with monocytes, while activated NK cells and naive B cells were the highest in PCa tissues. Moreover, the infiltration of five immune cells was significantly associated with TMB score and mutations of immune gene change the infiltration of immune cells. The Area Under Curve (AUC) of the multivariate regression model for the five- and 10-year survival prediction of PCa reached 0.796 and 0.862. The validation cohort proved that the model was reproducible.
ConclusionsThis study demonstrated that different infiltrating immune cells in prostate cancer, especially higher infiltrating M1 macrophages and neutrophils in PCa tissue, are associated with patients’ prognosis, suggesting that these two immune cells might be potential targets for PCa diagnosis and prognosis of treatment.

背景 本研究旨在探究前列腺癌（prostate cancer, PCa）微环境的免疫细胞浸润模式，并评估特定浸润免疫细胞作为前列腺癌潜在预后生物标志物的可行性。方法 本研究通过CIBERSORT算法从27个标准化数据集中提取22种免疫细胞的浸润占比。将CIBERSORT检验P值<0.05的样本合并，随后分为正常组织组与肿瘤组织组。分析正常与肿瘤组织间22种免疫细胞的浸润差异，并探究22种免疫细胞之间以及免疫细胞与格里森分级的潜在浸润相关性。从癌症基因组图谱（The Cancer Genome Atlas, TCGA）获取单核苷酸变异（Single Nucleotide Variant, SNV）数据以计算肿瘤突变负荷（Tumor Mutational Burden, TMB）评分。采用单因素与多因素回归分析评估免疫细胞在前列腺癌中的预后价值。结果 本研究从190例正常前列腺组织与537例前列腺癌组织中，筛选出10种存在显著浸润差异的免疫细胞，其中7种免疫细胞的浸润水平上调，3种免疫细胞的浸润水平下调。其中，静息自然杀伤细胞的浸润占比升高幅度最大，静息肥大细胞的浸润占比降低幅度最为显著。在正常前列腺组织中，CD8阳性T细胞与单核细胞的浸润相关性最强；而在前列腺癌组织中，活化自然杀伤细胞与初始B细胞的浸润占比最高。此外，5种免疫细胞的浸润水平与肿瘤突变负荷评分显著相关，免疫基因的突变可改变免疫细胞的浸润状态。本研究构建的前列腺癌5年与10年生存率预测多因素回归模型的受试者工作特征曲线下面积（Area Under Curve, AUC）分别达0.796与0.862。验证队列结果证实该模型具有良好的可重复性。结论 本研究证实，前列腺癌中不同的浸润免疫细胞，尤其是前列腺癌组织中浸润水平升高的M1型巨噬细胞与中性粒细胞，与患者预后密切相关，提示这两种免疫细胞可作为前列腺癌诊断与治疗预后评估的潜在靶点。