Integrative analysis of DNA methylation suggests down-regulation of oncogenic pathways and reduced somatic mutation rates in survival outliers of glioblastoma

Comprehensive molecular characterization of long-term glioblastoma survivors

Fewer than 5 % glioblastoma (GBM) patients survive over five years and are termed long-term survivors (LTS), yet their molecular background is unclear. The present cohort included 72 isocitrate dehydrogenase (IDH)-wildtype GBM patients, consisting of 35 LTS and 37 short-term survivors (STS), and we employed whole exome sequencing, RNA-seq and DNA methylation array to delineate this largest LTS cohort to date. Although LTS and STS demonstrated analogous clinical characters and classical GBM biomarkers, CASC5 (P = 0.002) and SPEN (P = 0.013) mutations were enriched in LTS, whereas gene-to-gene fusions were concentrated in STS (P = 0.007). Importantly, LTS exhibited higher tumor mutation burden (P < 0.001) and copy number (CN) increase (P = 0.013), but lower mutant-allele tumor heterogeneity score (P < 0.001) and CN decrease (P = 0.026). Additionally, LTS demonstrated hypermethylated genome (P < 0.001) relative to STS. Differentially expressed and methylated genes both enriched in olfactory transduction. Further, analysis of the tumor microenvironment revealed higher infiltration of M1 macrophages (P = 0.043), B cells (P = 0.016), class-switched memory B cells (P = 0.002), central memory CD4+ T cells (P = 0.031) and CD4+ Th1 cells (P = 0.005) in LTS. We also separately analyzed a subset of patients who were methylation class-defined GBM, contributing 70.8 % of the entire cohort, and obtained similar results relative to prior analyses. Finally, we demonstrated that LTS and STS could be distinguished using a subset of molecular features. Taken together, the present study delineated unique molecular attributes of LTS GBM.

Clinicopathological and molecular landscape of 5-year IDH-wild-type glioblastoma survivors: A multicentric retrospective study

Five-year glioblastoma (GBM) survivors (LTS) are the minority of the isocitrate dehydrogenase (IDH)-wild-type GBM patients, and their molecular fingerprint is still largely unexplored. This multicenter retrospective study analyzed a large LTS-GBM cohort from nine Italian institutions and molecularly characterized a subgroup of patients by mutation, DNA methylation (DNAm) and copy number variation (CNV) profiling, comparing it to standard survival GBM. Mutation scan allowed the identification of pathogenic variants in most cases, showing a similar mutational spectrum in both groups, and highlighted TP53 as the most commonly mutated gene in the LTS group. We confirmed DNAm as a valuable tool for GBM classification with a diagnostic refinement by using brain tumor classifier v12.5. LTS were more heterogeneous with more cases classified as diffuse pediatric high-grade glioma subtypes and having peculiar CNVs. We observed a global higher methylation in CpG islands and in gene promoters of LTS with methylation levels of distinct gene promoters correlating with prognosis.

Effect of different iodine levels on the DNA methylation of PRKAA2, ITGA6, THEM4 and PRL genes in PI3K-AKT signaling pathway and population-based validation from autoimmune thyroiditis patients

PURPOSE

Autoimmune thyroiditis (AIT) is one of the most common autoimmune endocrine diseases. The currently recognized causes are genetic susceptibility, environmental factors and immune disorders. It is important to clarify the pathogenesis for the prevention, diagnosis, treatment of AIT and scientific iodine supplementation. This study analyzed the DNA methylation levels of PRKAA2, ITGA6, PRL and THEM4 genes related to PI3K-AKT signaling pathway, compared the DNA methylation levels between cases and controls from different water iodine levels in Shandong Province of China, and evaluated the contribution of PI3K-AKT signaling pathway-related genes in AIT.

METHODS

A total of 176 adult AIT patients were included from three different water iodine areas, and 176 healthy controls were included according to gender, age and BMI. According to the results of the Illumina Methylation 850 K BeadChip in our previous research, the significant methylation differences of genes on the PI3K-AKT signaling pathway related to AIT were determined. The MethylTarget™ assay was used to detect the methylation levels of the target genes, and real-time PCR experiments were used to verify the mRNA expression levels.

RESULTS

Compared with the control group, PRKAA2_3 and 15 CpG sites were hyper-methylated. ITGA6 gene and 2 CpG sites were hypo-methylated in AIT cases. The mRNA expression of ITGA6 gene was negatively correlated with the DNA methylation levels of ITGA6 gene and 2 CpG sites. Compared with cases and controls in areas with different water iodine levels, methylation differences were mainly in PRKAA2 and ITGA6 genes. The methylation levels of PRKAA2_1 and PRKAA2_3 were positively correlated with age. The methylation levels of PRL and THEM4 genes were negatively correlated with age. The methylation level of PRKAA2_3 was positively correlated with FT₄.

CONCLUSION

In summary, we identified aberrant DNA methylation levels of PRKAA2 and ITGA6 genes related to PI3K-AKT signaling pathway in the blood of AIT patients. Both iodine supplementation after long-term iodine deficiency and iodine excess can affect the DNA methylation levels of PRKAA2 and ITGA6 genes, and the former affects more obviously. In ITGA6 gene, this aberrant epigenetic modification is associated with the increased mRNA expression.

Comprehensive omics analyses profile genesets related with tumor heterogeneity of multifocal glioblastomas and reveal LIF/CCL2 as biomarkers for mesenchymal subtype

Rationale: Around 10%-20% patients with glioblastoma (GBM) are diagnosed with more than one tumor lesions or multifocal GBM (mGBM). However, the understanding on genetic, DNA methylomic, and transcriptomic characteristics of mGBM is still limited.

Methods: In this study, we collected nine tumor foci from three mGBM patients followed by whole genome sequencing, whole genome bisulfite sequencing, RNA sequencing, and immunohistochemistry. The data were further examined using public GBM databases and GBM cell line.

Results: Analysis on genetic data confirmed common features of GBM, including gain of chr.7 and loss of chr.10, loss of critical tumor suppressors, high frequency of PDGFA and EGFR amplification. Through profiling DNA methylome of individual tumor foci, we found that promoter methylation status of genes involved in detection of chemical stimulus, immune response, and Hippo/YAP1 pathway was significantly changed in mGBM. Although both CNV and promoter methylation alteration were involved in heterogeneity of different tumor foci from same patients, more CNV events than promoter hypomethylation events were shared by different tumor foci, implying CNV were relatively earlier than promoter methylation alteration during evolution of different tumor foci from same mGBM. Moreover, different tumor foci from same mGBM assumed different molecular subtypes and mesenchymal subtype was prevalent in mGBM, which might explain the worse prognosis of mGBM than single GBM. Interestingly, we noticed that LIF and CCL2 was tightly correlated with mesenchymal subtype tumor focus in mGBM and predicted poor survival of GBM patients. Treatment with LIF and CCL2 produced mesenchymal-like transcriptome in GBM cells.

Conclusions: Together, our work herein comprehensively profiled multi-omics features of mGBM and emphasized that components of extracellular microenvironment, such as LIF and CCL2, contributed to the evolution and prognosis of tumor foci in mGBM patients.

Overcoming the Odds: Toward a Molecular Profile of Long-Term Survival in Glioblastoma

For over a century, gliomas were characterized solely by histologic features. With the publication of the WHO Classification of Tumours of the Central Nervous System, Revised 4th Edition in 2016, integrated histologic and molecular diagnosis became the norm, providing improved tumor grading and prognosis with IDH1/2 (isocitrate dehydrogenase 1 and 2) mutation being the most significant prognostic feature in all grades of adult diffuse glioma. Since then, much work has been done to identify additional molecular prognostic features, but the bulk of the progress has been made in defining aggressive features in lower grade astrocytoma. Although there have been several large case series of glioblastomas with long-term survival (LTS; overall survival ≥36 months), less is known about the clinical and molecular features of these cases. Herein, we review 19 studies examining LTS glioblastoma patients from 2009 to 2020 that include variable molecular analysis, including 465 cases with survival of 36 months or more (total n = 2328). These studies suggest that while there is no definitive molecular signature of long survival, younger age, IDH mutation, and MGMT (methyl guanine methyl transferase) promoter hypermethylation are associated with longer overall survival, and in IDH-wildtype tumors, chromosome 19/20 co-gain and lack of EGFR amplification, chromosome 7 gain/10 loss, and TERT promoter mutation are associated with LTS.

Molecular Correlates of Long Survival in IDH-Wildtype Glioblastoma Cohorts

IDH-wildtype glioblastoma is a relatively common malignant brain tumor in adults. These patients generally have dismal prognoses, although outliers with long survival have been noted in the literature. Recently, it has been reported that many histologically lower-grade IDH-wildtype astrocytomas have a similar clinical outcome to grade IV tumors, suggesting they may represent early or undersampled glioblastomas. cIMPACT-NOW 3 guidelines now recommend upgrading IDH-wildtype astrocytomas with certain molecular criteria (EGFR amplifications, chromosome 7 gain/10 loss, and/or TERT promoter mutations), establishing the concept of a "molecular grade IV" astrocytoma. In this report, we apply these cIMPACT-NOW 3 criteria to 2 independent glioblastoma cohorts, totaling 393 public database and institutional glioblastoma cases: 89 cases without any of the cIMPACT-NOW 3 criteria (GBM-C0) and 304 cases with one or more criteria (GBM-C1-3). In the GBM-C0 groups, there was a trend toward longer recurrence-free survival (median 12-17 vs 6-10 months), significantly longer overall survival (median 32-41 vs 15-18 months), younger age at initial diagnosis, and lower overall mutation burden compared to the GBM-C1-3 cohorts. These data suggest that while histologic features may not be ideal indicators of patient survival in IDH-wildtype astrocytomas, these 3 molecular features may also be important prognostic factors in IDH-wildtype glioblastoma.

modelBuildR: an R package for model building and feature selection with erroneous classifications

BACKGROUND

Model building is a crucial part of omics based biomedical research to transfer classifications and obtain insights into underlying mechanisms. Feature selection is often based on minimizing error between model predictions and given classification (maximizing accuracy). Human ratings/classifications, however, might be error prone, with discordance rates between experts of 5-15%. We therefore evaluate if a feature pre-filtering step might improve identification of features associated with true underlying groups.

METHODS

Data was simulated for up to 100 samples and up to 10,000 features, 10% of which were associated with the ground truth comprising 2-10 normally distributed populations. Binary and semi-quantitative ratings with varying error probabilities were used as classification. For feature preselection standard cross-validation (V2) was compared to a novel heuristic (V1) applying univariate testing, multiplicity adjustment and cross-validation on switched dependent (classification) and independent (features) variables. Preselected features were used to train logistic regression/linear models (backward selection, AIC). Predictions were compared against the ground truth (ROC, multiclass-ROC). As use case, multiple feature selection/classification methods were benchmarked against the novel heuristic to identify prognostically different G-CIMP negative glioblastoma tumors from the TCGA-GBM 450 k methylation array data cohort, starting from a fuzzy umap based rough and erroneous separation.

RESULTS

V1 yielded higher median AUC ranks for two true groups (ground truth), with smaller differences for true graduated differences (3-10 groups). Lower fractions of models were successfully fit with V1. Median AUCs for binary classification and two true groups were 0.91 (range: 0.54-1.00) for V1 (Benjamini-Hochberg) and 0.70 (0.28-1.00) for V2, 13% (n = 616) of V2 models showed AUCs < = 50% for 25 samples and 100 features. For larger numbers of features and samples, median AUCs were 0.75 (range 0.59-1.00) for V1 and 0.54 (range 0.32-0.75) for V2. In the TCGA-GBM data, modelBuildR allowed best prognostic separation of patients with highest median overall survival difference (7.51 months) followed a difference of 6.04 months for a random forest based method.

CONCLUSIONS

The proposed heuristic is beneficial for the retrieval of features associated with two true groups classified with errors. We provide the R package modelBuildR to simplify (comparative) evaluation/application of the proposed heuristic (http://github.com/mknoll/modelBuildR).

Predictive value of MGMT promoter methylation on the survival of TMZ treated IDH-mutant glioblastoma

Objective

O6methylguanine-DNA methyltransferase (MGMT) promoter methylation is a biomarker widely used to predict the sensitivity of IDH-wildtype glioblastoma to temozolomide therapy. Given that the IDH status has critical effects on the survival and epigenetic features of glioblastoma, we aimed to assess the role of MGMT promoter methylation in IDH-mutant glioblastoma.

Methods

This study included 187 IDH-mutant glioblastomas and used 173 IDH-wildtype glioblastomas for comparison. Kaplan-Meier curves and multivariate Cox regression were used to study the predictive effects.

Results

Compared with IDH-wildtype glioblastomas, IDH-mutant glioblastomas showed significantly higher (P < 0.0001) MGMT promoter methylation. We demonstrated that MGMT promoter methylation status, as determined by a high cutoff value (≥30%) in pyrosequencing, could be used to significantly stratify the survival of 50 IDH-mutant glioblastomas receiving temozolomide therapy (cohort A); this result was validated in another cohort of 25 IDH-mutant glioblastomas (cohort B). The median progression-free survival and median overall survival in cohort A were 9.33 and 13.76 months for unmethylated cases, and 18.37 and 41.61 months for methylated cases, and in cohort B were 6.97 and 9.10 months for unmethylated cases, and 23.40 and 26.40 months for methylated cases. In addition, we confirmed that the MGMT promoter methylation was significantly (P = 0.0001) correlated with longer OS in IDH-mutant patients with GBM, independently of age, gender distribution, tumor type (primary or recurrent/secondary), and the extent of resection.

Conclusions

MGMT promoter methylation has predictive value in IDH-mutant glioblastoma, but its cutoff value should be higher than that for IDH-wildtype glioblastoma.