Lung cancer is the leading cause of cancer morbidity and mortality in men, whereas, in women, it ranks third for incidence, and second for mortality. Due to the global increase in the number of smokers, the burden of lung cancer will likely continue to increase in the coming years primarily in developing countries1. Adenocarcinoma is the most common pathological type and is the most common type of peripheral lung cancer. Due to the lack of specific symptoms in the early stage of lung adenocarcinoma, about 70% of patients are advanced at diagnosis, losing the opportunity of radical surgery, while, radio-chemotherapy either or alone are the main treatment methods for these patients. The main mechanism of radio-chemotherapy is to break DNA double-strand, then induces DNA damages, leading to the irreversible death in the tumor cells14. DNA damage repair machinery, as an important mechanism to maintain genome stability, could repair damaged DNA, and plays an important role in tumor genesis, development, metastasis and prognosis15. However, due to the abnormally activated DNA damage repair, tumor cells are tolerance and resistant toward radiotherapy and chemotherapy, reduced efficacy of radio-chemotherapy, leading to poor prognosis. Studies have indicated that targeted therapy based on the DNA damage response (DDR) in cancers offers the potential for a greater therapeutic window16, 17. Thus it is crucial for looking for and identifying DNA damage repair genes in specific tumor.
Qi Liu et al analyzed the expression of DNA damage repair genes in glioma whose data were downloaded from TCGA and CGGA databases and constructed a risk model based on DNA damage repair related genes, which could predict the prognosis of astrocytoma patients with IDH mutation more accurately18. Similarly, Xin Wang constructed a risk prediction model composed of 11 DNA damage repair genes, which could be used as a prognostic indicator for colon cancer patients19. Therefore, an in-depth understanding of the gene expression profile of DNA damage repair related genes in LUAD patients is helpful to accurately evaluate the prognosis and provide a new idea for clinical treatment.
In the current study, we downloaded 513 DNA damage repair related genes from the MSigDB database, and found that 74 of them were significantly associated with the prognosis of LUAD patients. Then, we constructed a risk prediction model of DNA damage repair genes, which composed of three genes (PLK1, NEIL3 and EXO1) by Lasso - Cox regression analysis. PLK1 encodes a serine/threonine protein kinase, which is involved in the cell mitosis process and could promote the proliferation of tumor cells and inhibit their apoptosis. Elodie Montaudon et al found that the PLK1 inhibitor volasertib significantly inhibited tumor growth in breast cancer patients with CCND1-amplified20. Moataz Reda found that targeting PLK1 enhances radiosensitivity in non-small cell lung cancer21. NEIL3 encodes a DNA glycosylase, which initiates the first step of base excision repair by cleavage of bases damaged by reactive oxygen species and introduction of DNA strand breaks through associated lyase reactions. In pancreatic cancer, lung adenocarcinoma, low-grade glioma, renal clear cell carcinoma, and renal papillary cell carcinoma, patients with high expression of NEIL3 have poor overall survival and it may be related to homologous recombination and mismatch repair genes22. Exo1 is a potential regulator of telomerase activity and cell survival, which is expected to be a new target for tumor therapy. Exo1 expression was significantly associated with lymph node MET and disease-free survival. Fei Luo et al 23 found that Exo1 overexpression was closely related to the prognosis of prostate cancer patients and may be a potential biomarker of prostate cancer. In this paper, the risk score of each patient was calculated based on the expression of PKL1, NEIL3 and EXO1 and their respective COEF values. All of PKL1, NEIL3 and EXO1 are associated with prognosis of cancers. Thus, the results suggest that the model could accurately predict the prognosis of LUAD patients. The ROC curves of 1, 3 and 5 years confirmed that the model had good predictive performance and was verified in both internal and external validation sets. It is expected to guide clinical treatment strategies for LUAD patients. In the mechanism analysis, we found that the high-risk group was mainly enriched in mTOR, Myc, G2M and E2F signaling pathways. Abnormal activation of these four pathways in tumor cells regulates the metabolism and cell cycle, and promotes the proliferation of tumor cells24–27.
In conclusion, a risk model of PLK1-Neil3-Exo1 DNA damage repair related genes was constructed in this study, which could accurately divide LUAD patients into two groups of high and low risk and predict the prognosis of LUAD patients. The results were based on the large database of TCGA and GEO biological information. Although this model has been verified internally and externally, and has good accuracy, it still needs to be supported by a large number of clinical data and multi-center data. We will conduct further clinical verification of this model in the future.