摘要

    背景：前列腺癌（PCa）是全世界男性中最普遍的癌症之一。自噬相关基因（ARG）可能在前列腺癌的多种生物学过程中发挥重要作用。但是，ARG
表达签名很少用于研究自噬与PCa预后之间的关系。
这项研究旨在鉴定和评估预后性ARGs的特征，以预测总体生存率（OS）和无病
PCa患者的生存率（DFS）。
方法：首先，从人类自噬数据库中获得了总共234个自噬相关基因。然后，
根据癌症基因组图谱（TCGA）在前列腺癌患者中鉴定了差异表达的ARGs
数据库。进行单因素和多因素Cox回归分析筛查中心性预后ARGs。
总生存期和无病生存期，并建立了预后模型。最后，两者之间的相关性
进一步分析了预后模型和临床病理参数，包括年龄，T状况，N状况和
格里森得分。
结果：基于五个ARG（FAM215A，FDD，MYC，RHEB，
和ATG16L1），并从OS角度将前列腺癌患者显着分为高危和低危人群
（HR = 6.391，95％CI = 1.581–25.840，P <0.001）。接收器工作特性曲线（AUC）下方的区域
预测模型为0.84。在T3-4中，与OS相关的预测模型值高于T1-2（P = 0.008），
并且格里森分数> 7高于≤7（P = 0.015）。此外，基于22个ARG（ULK2，NLRC4，MAPK1，ATG4D，MAPK3，ATG2A，ATG9B，FOXO1，PTEN，HDAC6，PRKN，
HSPB8，P4HB，MAP2K7，MTOR，RHEB，TSC1，BIRC5，RGS19，RAB24，PTK6和NRG2），AUC为0.85（HR = 7.407，95％
CI = 4.850–11.320，P <0.001），这与T状态（P <0.001），N状态（P = 0.001）和格里森评分密切相关
（P ＜0.001）。
结论：我们基于ARGs的预测模型是整体生存和生存的可靠的预后和预测工具。
前列腺癌患者的无病生存期。
关键字：TCGA，GEO，前列腺癌，生存，自噬

    Background: Prostate cancer (PCa) is one of the most prevalent cancers that occur in men worldwide. Autophagyrelated genes (ARGs) may play an essential role in multiple biological processes of prostate cancer. However, ARGs
expression signature has rarely been used to investigate the association between autophagy and prognosis in PCa.
This study aimed to identify and assess prognostic ARGs signature to predict overall survival (OS) and disease-free
survival (DFS) in PCa patients.
Methods: First, a total of 234 autophagy-related genes were obtained from The Human Autophagy Database. Then,
diferentially expressed ARGs were identifed in prostate cancer patients based on The Cancer Genome Atlas (TCGA)
database. The univariate and multivariate Cox regression analysis was performed to screen hub prognostic ARGs for
overall survival and disease-free survival, and the prognostic model was constructed. Finally, the correlation between
the prognostic model and clinicopathological parameters was further analyzed, including age, T status, N status, and
Gleason score.
Results: The OS-related prognostic model was constructed based on the fve ARGs (FAM215A, FDD, MYC, RHEB,
and ATG16L1) and signifcantly stratifed prostate cancer patients into high- and low-risk groups in terms of OS
(HR=6.391, 95% CI=1.581– 25.840, P<0.001). The area under the receiver operating characteristic curve (AUC) of
the prediction model was 0.84. The OS-related prediction model values were higher in T3-4 than in T1-2 (P=0.008),
and higher in Gleason score > 7 than ≤ 7 (P=0.015). In addition, the DFS-related prognostic model was constructed based on the 22 ARGs (ULK2, NLRC4, MAPK1, ATG4D, MAPK3, ATG2A, ATG9B, FOXO1, PTEN, HDAC6, PRKN,
HSPB8, P4HB, MAP2K7, MTOR, RHEB, TSC1, BIRC5, RGS19, RAB24, PTK6, and NRG2), with AUC of 0.85 (HR=7.407, 95%
CI=4.850–11.320, P<0.001), which were frmly related to T status (P<0.001), N status (P=0.001), and Gleason score
(P<0.001).
Conclusions: Our ARGs based prediction models are a reliable prognostic and predictive tool for overall survival and
disease-free survival in prostate cancer patients.
Keywords: TCGA, GEO, Prostate cancer, Survival, Autophagy

    https://link.springer.com/content/pdf/10.1186/s12967-020-02323-x.pdf