Background

There exists considerable evidence conforming that autophagy may play an important role in the biological process of breast cancer. This study aimed to construct and evaluate a novel autophagy-related gene signature as a potential prognostic factor and therapeutic target in breast cancer patients based on high-throughput sequencing datasets.

Materials & methods

Autophagy-related genes obtained from the Human Autophagy Database and high-sequencing data obtained from The Cancer Genome Atlas (TCGA) were analyzed to identify differential expressed genes (DEGs) between tumor and normal tissues. Then GO and KEGG analysis were performed to explore potential biological and pathological functions of DEGs. Autophagy-related prognostic genes were identified by univariate COX regression analysis. Subsequently stepwise model selection using the Alkaike information criterion (AIC) and multivariate COX regression model was performed to construct autophagy-related gene signature. Then patients were divided into high- and low-risk groups based on the risk score identified by the autophagy-related gene signature. Multivariate COX regression model and stratification analysis were used to specify the prognostic value of this gene signature in whole cohort and various subgroups. T-test and ANOVA analysis were used to compare the expression differences of continuous variables (5 prognostic genes and risk score) in binary and multiple category groups respectively. Kaplan-Meier analysis, log-rank tests and the area under receiver operating characteristic (ROC) curve (AUC) were conducted to validate the accuracy and precise of the autophagy-related gene signature based on GSE20685 and GSE21653 datasets.

Results

We profiled autophagy-related DEGs in normal and breast tumor tissues. GO and KEGG analysis indicated that autophagy-related DEGs might participate in breast cancer occurrence, development and drug resistance. Then we identified five autophagy-related genes (EIF4EBP1, ATG4A, BAG1, MAP1LC3A and SERPINA1) that had significantly prognostic values for breast cancer. Autophagy-related gene signature was constructed and patients were divided into high- and low- risk groups based on their risk score. Patients in the high-risk group tended to have shorter overall survival (OS) and relapse-free survival (RFS) times than those in the low-risk group (OS: HR = 1.620, 95%CIs: 1.345–1.950; P < 0.001; RFS: HR = 1.487, 95%CIs: 1.248–1.771, P < 0.001). Autophagy-related gene signature had significant prognostic value in stratified subgroups especially in advanced breast cancer subgroups (T3–4; N2–3; stage III–IV). Its prognostic value was further confirmed in two GEO validation datasets (GSE20685: P = 6.795e-03; GSE21653: P = 1.383e-03). Finally, association analysis between clinicopathological factors and gene signature showed the risk score was higher in patients with ER/PR negative, higher clinical stage or T stage (P < 0.01).

Conclusion

We established and confirmed a novel autophagy-related gene signature for patients with breast cancer that had independent survival prognostic value especially in advanced breast cancer subgroups. Our research might promote the molecular mechanism study of autophagy-related genes in breast cancer.

中文翻译：

---

乳腺癌患者新型自噬相关基因签名的建立和验证。

背景

有大量证据表明自噬可能在乳腺癌的生物学过程中发挥重要作用。这项研究旨在基于高通量测序数据集，构建和评估新型自噬相关基因标记，作为乳腺癌患者的潜在预后因素和治疗靶标。

材料与方法

分析了从人类自噬数据库获得的自噬相关基因和从癌症基因组图谱（TCGA）获得的高测序数据，以鉴定肿瘤与正常组织之间的差异表达基因（DEG）。然后进行GO和KEGG分析，以探索DEGs的潜在生物学和病理功能。通过单变量COX回归分析确定自噬相关的预后基因。随后使用Alkaike信息标准（AIC）和多元COX回归模型进行逐步模型选择，以构建自噬相关基因签名。然后根据通过自噬相关基因签名确定的风险评分将患者分为高风险和低风险组。使用多元COX回归模型和分层分析来确定该基因标记在整个队列和各个亚组中的预后价值。用T检验和ANOVA分析比较连续变量（5个预后基因和风险评分）在二元组和多组中的表达差异。进行了Kaplan-Meier分析，对数秩检验和接收器工作特征（ROC）曲线下面积（AUC），以验证基于GSE20685和GSE21653数据集的自噬相关基因签名的准确性和精确性。

结果

我们分析了正常和乳腺肿瘤组织中的自噬相关的DEG。GO和KEGG分析表明，自噬相关的DEG可能参与了乳腺癌的发生，发展和耐药性。然后，我们鉴定了五个自噬相关基因（EIF4EBP1，ATG4A，BAG1，MAP1LC3A和SERPINA1），它们对乳腺癌的预后具有重要意义。构建了自噬相关的基因签名，并根据患者的风险评分将其分为高风险和低风险组。与低风险组相比，高风险组患者的总生存期（OS）和无复发生存期（RFS）时间往往较短（OS：HR = 1.620，95％CI：1.345–1.950； P < 0.001； RFS：HR = 1.487，95％CI：1.248–1.771，P <0.001）。自噬相关基因签名在分层亚组中具有重要的预后价值，尤其是在晚期乳腺癌亚组中（T3-4，N2-3，III-IV期）。在两个GEO验证数据集中进一步证实了其预后价值（GSE20685：P = 6.795e-03； GSE21653：P = 1.383e-03）。最后，临床病理因素与基因特征之间的相关性分析表明，ER / PR阴性，临床分期或T分期较高的患者，其风险评分较高（P <0.01）。

结论

我们为乳腺癌患者建立并证实了一种新的自噬相关基因签名，该签名具有独立的生存预后价值，尤其是在晚期乳腺癌亚组中。我们的研究可能会促进乳腺癌中自噬相关基因的分子机制研究。