Multi-Omics Analysis Based on Genomic Instability for Prognostic Prediction in Lower-Grade Glioma

Abstract

Background: Lower-grade gliomas (LGGs) are a heterogeneous set of gliomas. One of the primary sources of glioma heterogeneity is genomic instability, a novel characteristic of cancer. It has been reported that long noncoding RNAs (lncRNAs) play an essential role in regulating genomic stability. However, the potential relationship between genomic instability and lncRNA in LGGs and its prognostic value is unclear.Methods: In this study, the LGG samples from The Cancer Genome Atlas (TCGA) were divided into two clusters by integrating the lncRNA expression profile and somatic mutation data using hierarchical clustering. Then, with the differentially expressed lncRNAs between these two clusters, we identified genomic instability-related lncRNAs (GInLncRNAs) in the LGG samples and analyzed their potential function and pathway by co-expression network. Cox and least absolute shrinkage and selection operator (LASSO) regression analyses were conducted to establish a GInLncRNA prognostic signature (GInLncSig), which was assessed by internal and external verification, correlation analysis with somatic mutation, independent prognostic analysis, clinical stratification analysis, and model comparisons. We also established a nomogram to predict the prognosis more accurately. Finally, we performed multi-omics-based analyses to explore the relationship between risk scores and multi-omics data, including immune characteristics, N6-methyladenosine (m6A), stemness index, drug sensitivity, and gene set enrichment analysis (GSEA).Results: We identified 52 GInLncRNAs and screened five from them to construct the GInLncSig model (CRNDE, AC025171.5, AL390755.1, AL049749.1, and TGFB2-AS1), which could independently and accurately predict the outcome of patients with LGG. The GInLncSig model was significantly associated with somatic mutation and outperformed other published signatures. GSEA revealed that metabolic pathways, immune pathways, and cancer pathways were enriched in the high-risk group. Multi-omics-based analyses revealed that T-cell functions, m6A statuses, and stemness characteristics were significantly disparate between two risk subgroups, and immune checkpoints such as PD-L1, PDCD1LG2, and HAVCR2 were significantly highly expressed in the high-risk group. The expression of GInLncSig prognostic genes dramatically correlated with the sensitivity of tumor cells to chemotherapy drugs.Conclusion: A novel signature composed of five GInLncRNAs can be utilized to predict prognosis and impact the immune status, m6A status, and stemness characteristics in LGG. Furthermore, these lncRNAs may be potential and alternative therapeutic targets.

TGFB2-AS1 binding to MED1 promotes doxorubicin-induced cardiomyocyte apoptosis via BMP7 pathway

“…Recent studies found the role of long non-coding RNA in doxorubicin-induced cardiomyopathy, lncRNA NONMMUT015745 inhibits doxorubicin-mediated cardiomyocytes apoptosis by regulating the Rab2A-p53 axis [27], METTL14 silencing suppresses the ferroptosis of cardiomyocyte in doxorubicin-induced mice[28].TGFB2-AS1 has been reported in breast cancer, glioma, lung adenocarcinoma, gastric cancer, and other tumor diseases, and more importantly, the limited molecular biological role of TGFB2-AS1 in some diseases has been illustrated [29]. Studies have shown that mutations in the TGFBR1 and TGFB2-AS1 genes were associated with myopia in preschoolers, mediated by the TGF-β signaling pathway [30]; loss of TGFB2-AS1 signi cantly promoted Apoptosis of HepG2 liver cancer cells and inhibited migration and invasion [31], and TGFB2-AS1 interacted with the Core subunit of SWI/SNF complex, SMARCA4 to inhibit the expression of target genes TGFB2 and SOX2, ultimately inhibiting the breast Cancer progression [32].…”

Immune checkpoint therapy for solid tumours: clinical dilemmas and future trends

“… 242 A novel signature of five GlnLncRNAs has been screened to predict prognosis in glioma. 243 , 244 Mutations of seven genes have been detected with better OS in gastric cancer patients receiving anti-PD-1/PD-L1 treatment. 245 In addition, the deep learning method has also been used to predict the clinical benefits of anti-PD-L1/PD-1 therapy in advanced NSCLC patients.…”

A machine learning and deep learning-based integrated multi-omics technique for leukemia prediction

“…ML has made considerable strides in the previous several decades, developing complicated learning algorithms and effective pre-processing methods [ 22 , 23 ]. The development of Artificial Neural Networks (ANNs) into Deep Neural Network (DNN) topologies with gradually boosted DL learning capabilities was one of these breakthroughs [ 24 , 25 ].…”