Identification of a novel autophagy signature for predicting survival in patients with lung adenocarcinoma

Jin Duan; Youming Lei; Guoli Lv; Yinqiang Liu; Wei Zhao; Qingmei Yang; Xiaona Su; Zhijian Song; Leilei Lu; Yunfei Shi

doi:10.7717/peerj.11074

Identification of a novel autophagy signature for predicting survival in patients with lung adenocarcinoma

PeerJ. 2021 Apr 21:9:e11074. doi: 10.7717/peerj.11074. eCollection 2021.

Authors

Jin Duan^#¹, Youming Lei^#¹, Guoli Lv¹, Yinqiang Liu¹, Wei Zhao¹, Qingmei Yang¹, Xiaona Su², Zhijian Song³, Leilei Lu³, Yunfei Shi¹

Affiliations

¹ Department of Geriatric Thoracic Surgery, The First Hospital of Kunming Medical University, Kunming City, Yunnan Province, P.R. China.
² Department of Cancer Center, Daping Hospital, Army Medical University, Chongqing, China.
³ Origimed Co. Ltd., Shanghai, P.R. China.

^# Contributed equally.

Abstract

Background: Lung adenocarcinoma (LUAD) is the most commonhistological lung cancer subtype, with an overall five-year survivalrate of only 17%. In this study, we aimed to identify autophagy-related genes (ARGs) and develop an LUAD prognostic signature.

Methods: In this study, we obtained ARGs from three databases and downloaded gene expression profiles from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) database. We used TCGA-LUAD (n = 490) for a training and testing dataset, and GSE50081 (n = 127) as the external validation dataset.The least absolute shrinkage and selection operator (LASSO) Cox and multivariate Cox regression models were used to generate an autophagy-related signature. We performed gene set enrichment analysis (GSEA) and immune cell analysis between the high- and low-risk groups. A nomogram was built to guide the individual treatment for LUAD patients.

Results: We identified a total of 83 differentially expressed ARGs (DEARGs) from the TCGA-LUAD dataset, including 33 upregulated DEARGs and 50 downregulated DEARGs, both with thresholds of adjusted P < 0.05 and |Fold change| > 1.5. Using LASSO and multivariate Cox regression analyses, we identified 10 ARGs that we used to build a prognostic signature with areas under the curve (AUCs) of 0.705, 0.715, and 0.778 at 1, 3, and 5 years, respectively. Using the risk score formula, the LUAD patients were divided into low- or high-risk groups. Our GSEA results suggested that the low-risk group were enriched in metabolism and immune-related pathways, while the high-risk group was involved in tumorigenesis and tumor progression pathways. Immune cell analysis revealed that, when compared to the high-risk group, the low-risk group had a lower cell fraction of M0- and M1- macrophages, and higher CD4 and PD-L1 expression levels.

Conclusion: Our identified robust signature may provide novel insight into underlying autophagy mechanisms as well as therapeutic strategies for LUAD treatment.

Keywords: Autophagy; Gene expression omnibus database; Gene set enrichment analysis; Immune cell analysis; LASSO Cox regression; Lung adenocarcinoma; Molecular biomarkers; Multivariate cox regression analyses; Prognosis; The Cancer Genome Atlas.

Grants and funding

This work was supported by the Scientific Research Projects of Institutions of Medical and Health Institutions in Yunnan Province-The Role of NOD-like Receptors and Inflammatory Bodies in the Development of Xuanwei Lung Cancer (2016NS017), the Study of genetic risk of inflammatory body associated genes (D-2017013), the Kunming Medical Association Special Project for Applied Basic Research in Yunnan Province (2017FE467), the Yunnan Province Health and Family Planning Commission Medical Reserve Talents Plan (H-201703), the 2018 CSCO-Qilu Cancer Research Fund Project (Y-Q201802-011), and the research about PDGFRB functions on lung squamous cell carcinoma progression and its potential usage as a clinical lung squamous cell carcinoma marker (2017BS029). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.