Glioblastoma (GBM) is a highly malignant and aggressive form of brain cancer, presenting significant clinical challenges due to its resistance to conventional therapies and poor prognosis. GBM is known for its heterogeneity at the molecular and cellular levels, making it crucial to explore its diverse characteristics comprehensively.
Traditional approaches to studying GBM have included genomic and transcriptomic analyses, but these have provided limited insights into the complex interplay of factors contributing to its aggressiveness and therapeutic resistance. As a result, this study aimed to extend the scope of investigation by integrating multiple "omics" approaches, including genomics, transcriptomics, proteomics, phosphoproteomics, epigenomics, metabolomics, and lipidomics, to gain a more holistic understanding of GBM.
By delving deep into the molecular and cellular features of GBM, including its immune microenvironment, molecular pathways, epigenetic modifications, and potential therapeutic targets, this study sought to uncover novel insights that could inform the development of more effective treatments for this devastating disease. The multifaceted analysis of GBM in this study aimed to shed light on its complexities and provide a foundation for future research and personalized approaches to GBM therapy.
Case. Proteogenomic and metabolomic characterization ofhuman glioblastoma
Research Materials
The study analyzed a cohort of GBM samples, incorporating multi-omics data from a diverse set of patients. This included bulk RNA and protein expression data, single-cell RNA sequencing (snRNA-seq) on 18 GBM samples, as well as analyses of histone acetylation, lipid composition, and metabolomic features. The samples were categorized into distinct immune-based GBM subtypes.
Technical Methods
Immune Subtypes Analysis: The study employed single-sample gene set enrichment analysis (GSEA) by xCell to categorize GBM samples into four immune-based subtypes (im1, im2, im3, and im4) based on immune cell enrichment scores. These subtypes exhibited variations in immune cell composition, with im1 having higher levels of microglia, macrophages, and lymphocytes.
Single-Cell RNA Sequencing (snRNA-seq): SnRNA-seq was conducted on 18 GBM samples to gain insights into the tumor microenvironment. TAMs were identified as the predominant non-neoplastic cell population in the GBM TME, and differences in immune cell infiltration were observed across immune subtypes.
Morphologic Differences Analysis: The study employed a deep learning model to analyze morphological differences between immune subtypes. Notably, im4 tumors displayed large cells, including giant cells, while inflammatory cell fractions varied across subtypes.
Mesenchymal Tumor and Microenvironment Characteristics: The mesenchymal subtype of GBM was characterized by upregulated hypoxia pathways, increased angiogenesis, and complex macrophage activation and polarization patterns.
Differential Histone Acetylation: Histone acetylation patterns were explored across GBM tumors, with some tumors exhibiting elevated acetylation of specific histones. The analysis suggested potential connections between histone acetyltransferases (HATs), bromodomain-containing proteins (BRDs), and H2B acetylation sites.
Lipid Composition and Metabolomic Features: The study identified variations in lipid and metabolite profiles among GBM subtypes, with mesenchymal tumors showing distinct lipid characteristics.
Oncogenic Pathways Analysis: The research examined key oncogenic pathways, including RTK/RAS, PI3K/AKT, and p53/cell cycle, and identified genetic alterations and expression patterns associated with specific GBM subtypes.
Therapeutic Opportunities Identification: The study integrated genetic alterations, RNA, protein, and phosphosite data to identify potential therapeutic targets and candidate compounds for GBM treatment.
LINCS Analysis: The Library of Integrated Network-Based Cellular Signatures (LINCS) was utilized to predict compounds that could reverse tumor signatures in specific genetic or expression contexts.
Results
Protein Genomics and Metabolomics Characterize Molecular Subtypes of Glioblastoma
A multi-omics study was conducted on 99 cases of malignant glioblastoma (GBM) patients and 10 healthy control brain tissue samples. This study encompassed genomics, transcriptomics, proteomics (using TMT), modified proteomics (phosphorylation, acetylation), metabolomics, and lipidomics. Three subtypes, nmf1 (neuronal-like; n=29), nmf2 (mesenchymal-like; n=37), and nmf3 (classical-like; n=26), were observed in IDH-WT tumors, similar to genetic subtypes. Pathway enrichment analysis based on RNA, protein, and phosphorylation site abundance revealed that nmf1 was enriched in synaptic vesicle cycling and neuronal transmission, nmf2 was enriched in innate immunity, including neutrophil degranulation, phagocytosis, extracellular matrix organization, while nmf3 showed enrichment in mRNA splicing and RNA metabolism.
Mutations in driver genes alter the abundance and phosphorylation of oncogenic proteins
Connecting gene alterations (mutations, CNVs, fusions, and SVs) to RNA, protein expression, and phosphorylation levels revealed 95 cis-phosphorylation events. Strong cis-effects were observed between EGFR and PDGFRA, with significant increases in RNA and protein expression, corresponding to elevated phosphorylation at S1166 and S1067/S1070. In trans-effects, EGFR gene alterations led to decreased CTNNB1 mRNA levels but increased protein levels, along with elevated PTPN11 and PLCG1 phosphorylation levels. TP53 mutant tumors displayed more phosphorylation sites in DNA repair genes, but protein expression levels remained unchanged. Tumor suppressors RB1, NF1, PTEN, and ATRX showed good consistency between gene alterations and decreases in RNA, protein, and phosphorylation levels.
Activation of RKT cascade signaling in GBM
Genomic loci associated with RTKs such as EGFR, PDGFRA, and MET are frequently amplified in GBM. All 45 EGFR SV tumors had copy number amplifications, indicating high concordance between SVs and CNVs. Tumors with EGFR mutations and SV variants exhibited high RNA, protein, and Y1172 phosphorylation levels, indicating activation of the EGFR pathway. Furthermore, in GBM, PTPN11, GAB1, and GRB2 formed complexes regulated by RTKs, activating the RAS pathway.
GBM Immune Subtyping
Different immune markers and epigenetic events characterize GBM immune subtypes. Using single-sample gene set enrichment analysis (GSEA), four distinct immune subtypes were identified. Immune subtype 1 (im1) had an overall higher score, with elevated levels of small glial cells, macrophages, and lymphocytes. Immune subtypes 2 and 3 (im2 and im3) displayed varying proportions of macrophages and lymphocytes, with im2 having higher macrophage levels and im2 lower lymphocyte levels. Im3 had a higher neuronal score. Immune subtype 4 (im4) exhibited lower enrichment in all immune cell types compared to the other subtypes.
Histone Acetylation Cell Typing
Differential histone acetylation is associated with specific subtypes and pathways, as histone acetylation regulates gene expression but is often dysregulated in cancer. Over 30 acetylation sites were identified on histones H1, H2A, H2B, H3.3, and H4, categorizing them into four tumor subtypes. Histone acetylation levels were generally upregulated in tumors. Pathways related to immune infiltration such as ferroptosis, mast cell, and ROS pathways were negatively correlated with H2B acetylation, while splicing, nuclear receptor, and SUMOylation modifications showed positive correlations. Two key proteins in the SUMOylation pathway, SUMO1 and UBE2I, were upregulated in samples with high H2B acetylation. These observations suggest that H2B acetylation contributes to distinguishing immune cells from other cell types.
Metabolomic Analysis of GBM Subtypes
The authors identified 582 lipids in 75 tumor and 7 normal control tissue samples, with over 500 lipids showing significant differences across the four subtypes in multi-omics analysis. The mesenchymal subtype displayed high levels of triglycerides (TGs) and low levels of phosphatidylcholines (PCs) among phospholipids. The pre-neuronal subtype enriched in very-long-chain fatty acid lipids and glycerophospholipids with long-chain polyunsaturated fatty acids. Regarding metabolites, the pre-neuronal-like cluster showed significantly elevated levels of creatinine and homocysteine, and there were differences in metabolite expression between IDH mutant and IDH-WT tumors.
Key Oncogenic Pathways and Therapeutic Targets
Three crucial oncogenic signaling pathways in GBM are RTK/RAS, PI3K/AKT, and p53/cell cycle. In the RTK/RAS pathway, typical tumors primarily exhibited EGFR amplification, while pre-neuronal and IDH mutant tumors showed PDGFRA amplification, both with high RNA, protein, and phosphorylation site abundance for EGFR and PDGFRA. In the PI3K pathway, PTEN was downregulated in pre-neuronal, mesenchymal, and classical tumors, potentially leading to AKT1 and AKT2 activation through PIP3. In the p53/cell cycle pathway, subtype-specific amplifications were observed, with increased MDM2 protein expression in mesenchymal, pre-neuronal, and classical subtypes, and increased MDM4 protein expression in pre-neuronal and classical neurotypes.
Reference
- Wang, Liang-Bo, et al. "Proteogenomic and metabolomic characterization of human glioblastoma." Cancer cell 39.4 (2021): 509-528.
You may be interested in:
