CD73/adenosine pathway-related gene signature for intratumor heterogeneity and immune infiltration in prognosis of lung adenocarcinoma.

Abstract

e20524 Background: Adenosine is an immunosuppressive molecule that can inhibit the activity of immune cells, particularly T cells, and promote tumor growth. CD73, also known as ecto-5'-nucleotidase, is a cell-surface enzyme that plays a role in the adenosine pathway can catalyzes the conversion of AMP to adenosine. Establish and evaluate a prognosis model based on CD73/ Adenosine pathway will help to understanding of the biological mechanisms in lung adenocarcinoma (LUAD) and providing more accurate predictions for patients' treatment. Methods: The RNA sequencing (RNA-seq), whole exome sequencing (WES) and clinical data were obtained from LUAD cohorts of the Cancer Genome Atlas (TCGA). CD73/ Adenosine pathway-related genes were identified by screening for genes that were in both the CD73 STRING network (confidence = 0.9, 1st shell ≤50 interactors) and the KEGG purine metabolism pathway gene list(map00230). Least absolute shrinkage and selection operator (LASSO) regression was conducted to develop a CD73/ Adenosine pathway-related risk score (CD73RS). Time-dependent ROC curve and survival analysis were conducted in LUAD-TCGA cohort to verify the predictive performance of CD73RS. The MATH (Mutant-Allele Tumor Heterogeneity) ITH scores were calculated to evaluate intratumor heterogeneity (ITH) in LUAD-TCGA WES "maf" files using the R package "maftools" and the "math.score" function. TIMER 2.0 database(http://timer.cistrome.org/) was conducted to estimation of tumor microenvironment to further understand the molecular features of CD73RS. An independent cohort (GSE72094) was used to verify the prognostic effect of CD73RS. Results: A prognosis prediction model based on 21 CD73/ Adenosine pathway-related genes was established using LASSO regression. Kaplan-Meier survival analysis of the risk model from LUAD-TCGA cohort showed that the high-risk group had a significantly worse prognosis (HR = 3.298, P = 1.26e-08). The area under the curve (AUC) in time-dependent ROC curve at 1, 3 and 5 years were turn out to be 0.794, 0.722 and 0.686, respectively. The validation set also showed worse prognosis in high-risk group(HR = 1.473,p = 0.0403). In the high-risk group, there was a higher ITH score compared to the low-risk group (p = 0.0094), along with lower infiltration levels of B cells(p = 7.23e-15), CD4+ T cells(p = 0.003), macrophage cells(p = 0.016), and myeloid dendritic cells(p = 0.037). Conclusions: We established and validated a CD73/ Adenosine pathway risk model that exhibits a robust predictive performance for the outcome of patients with LUAD. The CD73/Adenosine pathway has been shown to play a role in both self-driven genomic changes, such as intra-tumor heterogeneity (ITH), as well as non-autonomous processes, such as immune microenvironment adaptation, which contribute to tumor plasticity and development.