摘要
      胰腺腺癌（PAAD）是一种快速恶化且预后不良的癌性生长。研究人员正在研究PAAD中的自噬，以确定一种新的生物标志物和治疗靶点。使用多变量Cox回归分析构建了自噬相关基因（ARG）总生存率（OS）模型。癌症基因组图谱（TCGA）-PAAD队列被用作训练组，作为模型构建的基础。该预测模型通过几个外部数据集进行了验证。为了评估模型性能，对受试者工作特性曲线（ROC）进行了分析。研究人类蛋白图谱（HPA）和癌症细胞系百科全书（CCLE）以验证ARG表达对癌症细胞的影响。通过使用CIBERSORT比较高危组和低危组之间的免疫浸润水平。通过基因本体论生物过程（GO-BP）和京都基因和基因组百科全书（KEGG）分析进一步分析低/高危组之间的差异表达基因（DEG），这些分析用于在连接性图（CMap）中识别潜在的小分子化合物，然后用PANC-1细胞进行半最大抑制浓度（IC50）检查。风险评分最终计算如下：BAK1 × 0.34 + ITGA3型 × 0.38 + 袋3 × 0.35 + APOL1基因 × 0.26–槽口24 × 在多变量Cox回归中，ITGA3和RAB24均为独立的预后因素。高危组的每个PAAD队列的OS明显短于低风险组。高危组表现出几种免疫细胞类型的浸润，包括幼稚B细胞（p = 0.003）、浆细胞（p = 0.044）和CD8 T细胞（几乎显著，p = 0.080）。较高的NK细胞浸润水平（p = 0.025），静息巨噬细胞（p = 0.020）和肥大细胞（p = 0.007）在高危组中的表达高于低危组。在CCLE和HPA数据库中，标志性ARGs的体外和体内表达是一致的。前3个富集的基因本体论生物过程（GO BP）是信号释放、跨突触信号调节和化学突触传递的调节，前3个富含的京都基因和基因组百科全书（KEGG）途径是MAPK、cAMP和细胞粘附分子。还鉴定了四种潜在的靶向ARGs的小分子化合物（哌拉西嗪、长春花碱、紫藤素A和hecogenin）。综合这些结果，我们的研究表明，ARG信号可能是一个有用的预后指标，并揭示PAAD患者的潜在治疗靶点。
Abstract
Adenocarcinoma of the pancreas (PAAD) is a cancerous growth that deteriorates rapidly and has a poor prognosis. Researchers are investigating autophagy in PAAD to identify a new biomarker and treatment target. An autophagy-related gene (ARG) model for overall survival (OS) was constructed using multivariate Cox regression analyses. A cohort of the Cancer Genome Atlas (TCGA)-PAAD was used as the training group as a basis for model construction. This prediction model was validated with several external datasets. To evaluate model performance, the analysis with receiver operating characteristic curves (ROC) was performed. The Human Protein Atlas (HPA) and Cancer Cell Line Encyclopedia (CCLE) were investigated to validate the effects of ARGs expression on cancer cells. Comparing the levels of immune infiltration between high-risk and low-risk groups was finished through the use of CIBERSORT. The differentially expressed genes (DEGs) between the low-/high-risk groups were analyzed further via Gene Ontology biological process (GO-BP) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses, which were used to identify potential small-molecule compounds in Connectivity Map (CMap), followed by half-maximal inhibitory concentration (IC50) examination with PANC-1 cells. The risk score was finally calculated as follows: BAK1 × 0.34 + ITGA3 × 0.38 + BAG3 × 0.35 + APOL1 × 0.26–RAB24 × 0.67519. ITGA3 and RAB24 both emerged as independent prognostic factors in multivariate Cox regression. Each PAAD cohort had a significantly shorter OS in the high-risk group than in the low-risk group. The high-risk group exhibited infiltration of several immune cell types, including naive B cells (p = 0.003), plasma cells (p = 0.044), and CD8 T cells (nearly significant, p = 0.080). Higher infiltration levels of NK cells (p = 0.025), resting macrophages (p = 0.020), and mast cells (p = 0.007) were found in the high-risk group than the low-risk group. The in vitro and in vivo expression of signature ARGs was consistent in the CCLE and HPA databases. The top 3 enriched Gene Ontology biological processes (GO-BPs) were signal release, regulation of transsynaptic signaling, and modulation of chemical synaptic transmission, and the top 3 enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were MAPK, cAMP, and cell adhesion molecules. Four potential small-molecule compounds (piperacetazine, vinburnine, withaferin A and hecogenin) that target ARGs were also identified. Taking the results together, our research shows that the ARG signature may serve as a useful prognostic indicator and reveal potential therapeutic targets in patients with PAAD.

https://www.nature.com/articles/s41598-022-11050-w