The characteristics of metabolic related pathways under different anoikis activities revealed by pan-cancer analyses.

Abstract

3150 Background: Anoikis is usually activated to eliminate extracellular matrix-detached cells. Cancers develop several mechanisms to escape from this cellular death signaling pathway which was associated with metabolic reprogramming in tumor development and progression. Comprehensively understanding the characteristics of anoikis activities in pan cancers facilitating better clinical treatment strategies. Methods: RNA-Seq data and related patient information were collected from 33 cancer types in TCGA database. We also obtained anoikis related genes (ARGs) from Harmonizome database. The related pathways of epithelial-mesenchymal transition (EMT) and reactive oxygen species (ROS) genes were collected from MSigDB. Autophagy genes were from HADb and metabolic intermediates were from published literatures. Firstly, pathway scores of anoikis and the related pathways were computed through single sample gene set enrichment analysis (ssGSEA). Then, survival analyses based on the anoikis scores were performed to achieve the threshold between the high anoikis (HA) or low anoikis (LA) sample groups. To study the drug efficacy, cell line transcriptome and IC50 data were obtained from GDSC database. Pearson correlations were computed between anoikis and other related pathways or immune cell infiltrations in each sample group. Results: Most ARGs were up-regulated in tumors, the same as the HA sample group. Differentially expressed genes between HA and LA were mostly involved in metabolic super-pathways, including metabolic intermediates (TCA cycle, carbohydrate, lipid, nucleotide, vitamin cofactor, and amino acid, etc.) and promoting anoikis-resistance pathways including EMT, inflammation and autophagy. Higher scores of anoikis related pathways were identified in HA across various cancers. Kaplan-Meier survival analysis revealed that HA in KIRC, KIRP, LAML, LGG, LUAD, PAAD, STAD, THYM, and UVM cancers had significantly worse clinical outcomes (p < 0.05). Functional enrichment analysis of ARGs through GSEA suggested that anoikis resistance was promoted, whereas anoikis was retarded in those cancers. However, patients in HA of several cancers had longer survives than LA, which may owe to the higher infiltrating levels of anti-tumor immune cells including CD8+ T, CD4+ memory activated T cells, particularly in BRCA, MESO, SARC, and UCEC cancers. Drug efficacy analysis revealed that BCL2 gene was significantly associated with sensitive drug response on Venetoclax, ABT737, and Tozasertib. And TLE1 and CAV1 were associated with Tozasertib resistance and AZD5991 resistance, respectively. Conclusions: The characteristics of anoikis activities and metabolic related pathways during tumor progression were comprehensively studied and the prognosis value of anoikis was further assessed by drug efficacy.