Methylation Subgroups Research Articles

Spinal cord atypical teratoid/rhabdoid tumors in children: Clinical, genetic, and outcome characteristics in a representative European cohort.

Case reports have portrayed spinal cord atypical teratoid/rhabdoid tumor (spATRT) as an aggressive form of ATRT. We conducted a retrospective European survey to collect data on clinical characteristics, molecular biology, treatment, and outcome of children with intramedullary spATRT. Scrutinizing a French national series and the European Rhabdoid Registry database, we identified 13 patients (median age 32 months; metastatic disease at diagnosis, n=6). Systemic postoperative chemotherapy was administered to all patients; three received intrathecal therapy and six were irradiated (craniospinal, n=3; local, n=3). Median observation time was 8 (range, 1-93) months. Progression-free and overall survival rates at 1 and (2 years) were 35.2% ± 13.9% (26.4% ± 12.9%) and 38.5% ± 13.5% (23.1% ± 11.7%). Four patients (ATRT-SHH, n=2; ATRT-MYC, n=1; DNA methylation subgroup not available, n=1) achieved complete remission (CR); two of them are alive in CR 69 and 72 months from diagnosis. One patient relapsed after CR and is alive with progressive disease (PD) and one died of the disease. Three patients (ATRT-MYC, n=2; subgroup not available, n=1) died after 7 to 22 months due to PD after having achieved a partial remission (n=1) or stabilization (n=2). Five patients (ATRT-MYC, n=2; subgroup not available, n=3) developed early PD and died. One patient (ATRT-MYC) died of intracerebral hemorrhage prior to response evaluation. Long-term survival is achievable in selected patients with spATRT using aggressive multimodality treatment. Larger case series and detailed molecular analyses are needed to understand differences between spATRT and their inracranial counterparts and the group of extradural malignant rhabdoid tumors.

DNA Methylation as a Robust Classifier of Epithelial Ovarian Cancer.

Epithelial ovarian cancers (EOC) can be defined across different histologic subtypes by specific DNA methylation profiles which correlate with outcome. Identification of distinct EOC methylation subgroups emphasizes the role of epigenetic remodeling to establish unique EOC etiology during transformation, and harbors the potential to direct personalized medicine and patient care.See related article by Bodelon et al., p. 5937.

P14.48 Identification of subsets of IDH-mutant glioblastomas with distinct epigenetic and copynumber alterations and stratified clinical risks

Abstract BACKGROUND IDH-mutant glioblastoma is classified by the 2016 CNS WHO as a group with good prognosis. However, the actual number of cases examined in the literature is relatively small. We hypothesize that IDH-mutant glioblastoma is not a uniform group and should be further stratified. MATERIAL AND METHODS We conducted methylation profiles and estimated copy number variations in 64 IDH-mutant glioblastomas. RESULTS Our results showed that 10.9%, 53.1%, and 35.9% of tumors belonged to Codel, G-CIMPhigh, and G-CIMP-low methylation subgroups, respectively. G-CIMP-low subgroup was associated with significantly worse OS as compared to G-CIMP-high (P=0.005) and Codel groups (P=0.009). CDKN2A deletion (37.5%) was the most common gene copy number variation, and was associated with G-CIMP-low subgroup (P=0.001). Other frequent copy number changes included MET (4.7%), CCND2 (17.2%), PDGFRA (14.1%), CDK4 (12.5%), and EGFR (12.5%) amplification. Both CDKN2A deletion (P=0.008) and MET amplification (P<0.001) were associated with poor OS in IDH-mutant glioblastomas. Combined epigenetic signatures and gene copy number variations separated IDH-mutant glioblastomas into Group 1 (Codel), Group 2 (GCIMP- high), Group 3 (G-CIMP-low without CDKN2A and MET alterations), and Group 4 (G-CIMP-low with CDKN2A/MET alteration). Survival analysis revealed Group 1 had a favorable OS (median survival: 41.3 months), while Groups 2 and 3 exhibited an intermediate OS (median survival: 20.6 and 21.8 months, respectively). Group 4 exhibited the worst OS (median survival: 8.4 months). Multivariable analysis confirmed the independent prognostic significance of our Groups. CONCLUSION IDH-mutant glioblastomas should be stratified for risk with combined epigenetic signature and CDKN2A/MET status and some cases have poor outcome.

Integrated Multi-Omics Characterization of Neurofibromatosis Type 1 Related Peripheral Nerve Sheath Tumors Identifies Two Distinct Malignant Peripheral Nerve Sheath Tumors Subtypes with Sonic Hedgehog or WNT Pathway Activation

Abstract INTRODUCTION Malignant peripheral nerve sheath tumors (MPNSTs) are highly aggressive Schwann-cell derived sarcomas and the leading cause of mortality in neurofibromatosis Type 1. In many cases, MPNSTs develop through malignant transformation of an underlying neurofibroma (NF). The genomic alterations and molecular pathways driving malignant transformation is not well understood. In this study, we leverage multi-platform genomic and epigenomic profiling of human NFs and MPNSTs to identify targetable molecular pathways that lead to malignant transformation. METHODS Fresh-frozen tumor (N = 108; 19 MPNSTs, 18 premalignant NFs, 37 benign peripheral nerve NFs, and 34 cutaneous NFs) and matched-controls were allocated for NGS platforms including methylation-profiling, RNA, whole exome and ChIP sequencing. In addition, functional validation of identified pathways was performed using CRISPR/cas9 knockout of relevant sonic hedgehog (SHH) and WNT pathway genes in immortalized neurofibroma cell lines. RESULTS Unsupervised consensus clustering of peripheral nerve tumors identified 4 distinct methylation subgroups, with MPNSTs clustering in either Methyl_Grp-2B or Methyl_Grp-3. Transcriptomic analysis of the MPNSTs, identified SHH pathway activation in Methyl_Grp-3 MPNSTs (NES = 1.9952, P < .0001, FDR < 0.001) and WNT pathway activation in Methyl_Grp-2B MPNSTs (NES = 1.7630, P < .0001, FDR = 0.0058). Single sample gene set enrichment analysis demonstrates that there is a strong negative correlation between SHH and WNT pathway activation in MPNSTs (r = −0.78, P = .007). Methyl_Gr-3 MPNSTs harbor more PTCH1 copy number loss (62.5% vs 0%, P < .05) and PTCH1 promoter hypermethylation (mean beta value: 0.2899 vs. 0.06119, P < .05). PTCH1 knockout in immortalized neurofibroma cell lines increased expression of SHH pathway target genes (GLI1, GLI2, and SMO), and increased cellular proliferation and invasiveness. CONCLUSION We found that MPNSTs can be classified into two molecular subgroups driven by SHH or WNT pathway activation. Furthermore, PTCH1 loss is one mechanism that mediates SHH pathway activation. These results suggest a targeted approach should be taken for treating MPNSTs, and SHH and WNT pathway inhibition may be promising avenue for therapeutic development.

DNA methylation-based classification and identification of renal cell carcinoma prognosis-subgroups

BackgroundRenal cell carcinoma (RCC) is the most common kidney cancer and includes several molecular and histological subtypes with different clinical characteristics. The combination of DNA methylation and gene expression data can improve the classification of tumor heterogeneity, by incorporating differences at the epigenetic level and clinical features.MethodsIn this study, we identified the prognostic methylation and constructed specific prognosis-subgroups based on the DNA methylation spectrum of RCC from the TCGA database.ResultsSignificant differences in DNA methylation profiles among the seven subgroups were revealed by consistent clustering using 3389 CpGs that indicated that were significant differences in prognosis. The specific DNA methylation patterns reflected differentially in the clinical index, including TNM classification, pathological grade, clinical stage, and age. In addition, 437 CpGs corresponding to 477 genes of 151 samples were identified as specific hyper/hypomethylation sites for each specific subgroup. A total of 277 and 212 genes corresponding to DNA methylation at promoter sites were enriched in transcription factor of GKLF and RREB-1, respectively. Finally, Bayesian network classifier with specific methylation sites was constructed and was used to verify the test set of prognoses into DNA methylation subgroups, which was found to be consistent with the classification results of the train set. DNA methylation-based classification can be used to identify the distinct subtypes of renal cell carcinoma.ConclusionsThis study shows that DNA methylation-based classification is highly relevant for future diagnosis and treatment of renal cell carcinoma as it identifies the prognostic value of each epigenetic subtype.

Impact of predictive impact of MGMT promoter methylation in malignant astrocytomas depends on the methylation subgroup.

2013 Background: O6-methylguanine DNA-methyl transferase (MGMT) status is predictive for alkylating chemotherapy in most series, but there are non-benefitting subgroups. Despite multiple attempts, MGMT has not been unambiguously established as a predictive biomarker for patients with malignant gliomas. Further, these tumors are to be better classified according to global methylation profiles. Methods: Long-term efficacy data of the NOA-08 trial (NCT01502241) that compared efficacy and safety of radiotherapy (RT, n= 176) to temozolomide (TMZ, n= 193) in patients > 65 years with anaplastic astrocytoma (AA) or GB as well as genome-wide DNA methylation patterns and copy number variations assessed by methylation arrays in a biomarker subset ( n= 104) and an independent cohort ( n= 380) have been used to assess the interaction between MGMT status and methylation subgroups. Results: In the long-term update of NOA-08 patients with MGMT methylated tumors had longer OS and EFS when treated with TMZ (18.4 [13.9-24.4] months and 8.5 [6.9-13.3] months) versus RT (9.6 [6.4-13.7] months and 4.8 [4.3-6.2] months, HR 0.44 [0.27-0.70], p < 0.001 for OS and 0.46 [0.29-0.73], p = 0.001 for EFS). These data compared favorably with recently published data from patients treated with chemoradiation (Perry et al. NEJM 2017). Importantly, only patients with glioblastomas of the methylation class receptor tyrosine kinase II (RTKII) and mesenchymal but not RTK I demonstrated the predictive impact of MGMT in the NOA and the independent validation cohort. Conclusions: MGMT promoter methylation as a strong but methylation subclass-dependent predictive biomarker for the use of alkylating chemotherapy in malignant gliomas. The data call for embedding of MGMT tests into global methylation analyses for all patients with malignant gliomas potentially treated with alkylating chemotherapy.

Identification of subsets of IDH-mutant glioblastomas with distinct epigenetic and copy number alterations and stratified clinical risks.

Background IDH-mutant glioblastoma is classified by the 2016 CNS WHO as a group with good prognosis. However, the actual number of cases examined in the literature is relatively small. We hypothesize that IDH-mutant glioblastoma is not a uniform group and should be further stratified.MethodsWe conducted methylation profiles and estimated copy number variations of 57 IDH-mutant glioblastomas.ResultsOur results showed that 59.6% and 40.4% of tumors belonged to glioma-CpG island methylator phenotype (G-CIMP)-high and G-CIMP-low methylation subgroups, respectively. G-CIMP-low subgroup was associated with significantly worse overall survival (OS) as compared to G-CIMP-high (P = .005). CDKN2A deletion (42.1%) was the most common gene copy number variation, and was significantly associated with G-CIMP-low subgroup (P = .004). Other frequent copy number changes included mesenchymal–epithelial transition (MET) (5.3%), CCND2 (19.3%), PDGFRA (14.0%), CDK4 (12.3%), and EGFR (12.3%) amplification. Both CDKN2A deletion (P = .036) and MET amplification (P < .001) were associated with poor OS in IDH-mutant glioblastomas. Combined epigenetic signature and gene copy number variations separated IDH-mutant glioblastomas into Group 1 (G-CIMP-high), Group 2 (G-CIMP-low without CDKN2A nor MET alteration), and Group 3 (G-CIMP-low with CDKN2A and/or MET alteration). Survival analysis revealed Groups 1 and 2 exhibited a favorable OS (median survival: 619 d [20.6 mo] and 655 d [21.8 mo], respectively). Group 3 exhibited a significant shorter OS (median survival: 252 d [8.4 mo]). Multivariable analysis confirmed the independent prognostic significance of our Groups.Conclusions IDH-mutant glioblastomas should be stratified for risk with combined epigenetic signature and CDKN2A/MET status and some cases have poor outcome.

OR02-1 DNA Methylation Profiling in Pediatric Adrenocortical Tumors Reveals Distinct Methylation Signatures with Prognostic Significance: A Report from the International Pediatric Adrenocortical Tumor Registry

Pediatric adrenocortical tumors (ACTs) are heterogeneous and have a poor prognosis. Overall survival is around 60%. In pediatric ACT that is managed only with surgery, features associated with adverse outcome include age greater than 3 years, non-functioning tumors, and metastatic (Stage IV) or residual tumor after surgery (Stage III). These factors individually do not account for most treatment failures, however. As current histopathologic classification systems are hindered by poor reproducibility and lack of predictive value, new prognostic markers are critically needed. In adults, methylation profiling can distinguish benign from malignant ACT, but similar data on pediatric ACT have not been reported. We performed DNA methylation profiling (MethylationEPIC BeadChip Array) on 52 pediatric ACTs and 13 normal pediatric adrenocortical tissue samples. The median age of 38 girls and 14 boys was 2.8 years (range, 0.3-17). ACTs were histologically classified as carcinoma (n=29), adenoma (n=10), and indeterminate (n=13). Overall survival was used as a primary endpoint. Unsupervised analysis of the 20,000 most variably methylated probes segregated cases into two distinct subgroups (designated A1 and A2). Differentially methylated genes (q-value<0.05) between A1 and A2 groups included those associated with IGF-, mTOR, WNT, and HGF/cMET signaling pathways. Other genes commonly mutated in pediatric malignancies (DICER1, EZH2,EGFR, PTEN, RB1, TERT, SF1, and MEN1) were also differentially methylated. Normal adrenal cortex samples segregated with the A2 group. The A1 subgroup had unique copy number changes, and recurrent global hypomethylation of chromosomes 4, 13, and 18. Moreover, the A1 group was significantly associated with higher frequency of CTNNB1 mutations (OR = 4.2, p = 0.03), lower frequency of TP53 mutations (p = 0.04), and use of chemotherapy (OR = 9.7, p = 0.0007). Methylation subgroups were not significantly associated with sex, histopathologic classification, disease stage, or ATRX status. In a univariable analysis, vascular invasion (HR=6.2, p=0.02), tumor weight (HR=10.9, p=0.001), tumor size (HR=11.3, p=0.002), tumor volume (cm3, HR=8.53, p=0.008), diffuse necrosis (HR=9.21, p=0.03), proliferative index by Ki-67 immunohistochemical staining (HR=7.1, p=0.02), disease stage III/IV (HR = 9.8, p = 0.03), age > 4 years (HR = 16.8, p = 0.008) and A1 methylation group (HR 7.7, p=0.0011) were each significantly correlated with overall survival. A1 subgroup remained significantly associated with overall survival in a series of models that adjusted for features including: clinical diagnosis, histopathologic characteristics, Ki67 index, and advanced stage. Methylation profiling provides prognostic significance that is independent of most current predictive variables, and in the future, may serve as a powerful biomarker in the prognostication of pediatric ACT.

An integrated transcriptome analysis in T-cell acute lymphoblastic leukemia links DNA methylation subgroups to dysregulated TAL1 and ANTP homeobox gene expression.

Classification of pediatric T‐cell acute lymphoblastic leukemia (T‐ALL) patients into CIMP (CpG Island Methylator Phenotype) subgroups has the potential to improve current risk stratification. To investigate the biology behind these CIMP subgroups, diagnostic samples from Nordic pediatric T‐ALL patients were characterized by genome‐wide methylation arrays, followed by targeted exome sequencing, telomere length measurement, and RNA sequencing. The CIMP subgroups did not correlate significantly with variations in epigenetic regulators. However, the CIMP+ subgroup, associated with better prognosis, showed indicators of longer replicative history, including shorter telomere length (P = 0.015) and older epigenetic (P < 0.001) and mitotic age (P < 0.001). Moreover, the CIMP+ subgroup had significantly higher expression of ANTP homeobox oncogenes, namely TLX3, HOXA9, HOXA10, and NKX2‐1, and novel genes in T‐ALL biology including PLCB4, PLXND1, and MYO18B. The CIMP− subgroup, with worse prognosis, was associated with higher expression of TAL1 along with frequent STIL‐TAL1 fusions (2/40 in CIMP+ vs 11/24 in CIMP−), as well as stronger expression of BEX1. Altogether, our findings suggest different routes for leukemogenic transformation in the T‐ALL CIMP subgroups, indicated by different replicative histories and distinct methylomic and transcriptomic profiles. These novel findings can lead to new therapeutic strategies.

DNA Methylation Identifies Genetically and Prognostically Distinct Subtypes of MDS

Background: Recurrent somatic mutations in patients with myelodysplastic syndromes (MDS) implicate several epigenetic regulators in the molecular pathobiology of these disorders. We hypothesized that identification of MDS subtypes defined by DNA methylation patterns could improve prognostication and enhance our understanding of MDS disease biology. Methylation assays that investigate regions of the genome beyond promoters may better inform which regulatory regions possess biologic importance in the development and progression of MDS. Additionally, computational methods that agnostically approach large methylation datasets are needed when attempting to classify clinically and genetically heterogeneous disease populations such as MDS.Methods: To measure DNA methylation, we used bisulfite padlock probes (BSPP), a targeted enrichment technique that captures approximately 500,000 unique CpGs in regions known to contain differentially methylated regulatory regions. BSPP was performed on bone marrow mononuclear DNA from 141 treatment-naïve patients with MDS that were genetically and clinically annotated in prior studies (Bejar et al. N Engl J Med 2011 and J Clin Oncol 2012). We used Onco-GPS, a computational method that identifies shared underlying cellular states among heterogeneous sample populations in an unsupervised manner, to identify DNA methylation subtypes of MDS (Kim et al. Cell Systems 2017). This method uses non-negative matrix factorization (NMF) followed by consensus clustering of the resulting NMF components. The optimal number of NMF components and consensus clusters were chosen to maximize the cophenetic correlation coefficient. A Fisher's Exact test was used to test for enrichment of genetic lesions in specific clusters. Univariate and multivariable models of overall survival (OS) were constructed using cox proportional hazards regression. Variables with univariate p-values <0.2 were evaluated for inclusion in the final multivariable model and the final model was constructed using a forward and backward stepwise procedure based on the Aikake Information Criterion.Results: We found that 5 NMF components and 5 consensus clusters maximized the cophenetic correlation coefficient, indicating the most stable partitioning of our dataset (Figures 1A-1B). The samples making up each cluster were enriched largely in a single NMF component, indicating that the underlying DNA methylation loci contributing to each cluster are distinct. Specific genetic and cytogenetic abnormalities were significantly enriched in different clusters, including enrichment of mutations in TP53 and U2AF1 in cluster 2; EZH2, ASXL1, RUNX1, SRSF2 in cluster 3; TET2 in cluster 4; SF3B1 in cluster 5; and abnormal and complex karyotypes enriched in clusters 1 and 2 (Figure 1C-1D). Kaplan-Meier curves revealed two major patterns of survival with clusters 2 and 3 displaying inferior OS when compared with clusters 1, 4, and 5. We combined clusters with similar median OS into “High Risk” and “Low Risk” cluster groups. In univariate Cox proportional hazards regression, High Risk and Low Risk cluster groups had significant differences in OS (HR, 1.95; 95% CI,1.24-3.08; p=0.003). In a multivariable model including elements of the IPSS-R, cluster risk group remained a significant predictor of OS (HR, 2.02; 95% CI, 1.25-3.27, p=0.004). We determined cluster-specific differentially methylated regions (DMRs) and found divergent patterns of methylation among clusters with the majority of DMRs located at non-promoter regulatory regions with high enrichment at both active and bivalent enhancers (as defined in a reference epigenome from mobilized CD34+ cells).Conclusion: Using BSPP and novel computational methods, we found that DNA methylation of MDS patients identifies clinically relevant subtypes not otherwise distinguished by current prognostic models. Our analysis also shows that while genetic abnormalities can be associated with specific DNA methylation changes, patients with diverse genetic lesions can converge on common DNA methylation states, reflecting shared pathogenic mechanisms and clinical outcomes. Ongoing work to annotate the biological differences between the DNA methylation subgroups identified will enhance our understanding of the role of DNAm in prognosis and the underlying biology of MDS. [Display omitted] DisclosuresZhang:Singlera Genomics: Consultancy, Equity Ownership, Membership on an entity's Board of Directors or advisory committees. Bejar:AbbVie/Genentech: Consultancy, Honoraria; Modus Outcomes: Consultancy; Takeda: Research Funding; Celgene: Consultancy, Honoraria; Astex/Otsuka: Consultancy, Honoraria; Genoptix: Consultancy; Foundation Medicine: Consultancy.

DNA Methylation Subgroups in Juvenile Myelomonocytic Leukemia: An International Collaborative Analysis and Development of a Common Diagnostic Platform

Juvenile myelomonocytic leukemia (JMML) is a rare myeloproliferative neoplasm of early childhood characterized by Ras pathway mutations and aberrant DNA methylation. Despite allogeneic hematopoietic stem cell transplantation (HSCT) approx. 30% of patients (pts) succumb to resistant leukemia. Slowly regressing disease in the absence of HSCT is, however, noted in some pts. Mutational patterns and known clinical risk factors are insufficient to fully explain this heterogeneity. Three independent recent studies demonstrated the potential of DNA methylation profiling to identify clinically meaningful risk groups (Lipka et al., Nat Commun 2017; Stieglitz et al., Nat Commun 2017; Murakami et al., Blood 2018).The aim of this work was to establish an international consensus for the classification of JMML based on DNA methylation, to systematically describe the biological and clinical features of the subgroups identified and to develop a molecular classifier that enables the prospective assignment of DNA methylation categories.In a collaborative effort, we re-analyzed three well-annotated Illumina 450k methylation array data sets. In total, data from 292 pts with JMML or Noonan syndrome-associated myeloproliferative disorder (NS/MPD; EWOG-MDS, Europe: 147 pts, Nagoya, Japan: 106 pts, UCSF, USA: 39 pts) were available. Methylation data from 285 pts passed our quality control and were included in the present analysis. Of these, 256 (89.8%) pts had been diagnosed with JMML and 29 (10.2%) with NS/MPD. Mean age at diagnosis was 1.9 yrs. Levels of fetal hemoglobin (HbF) were available in 228/285 pts and elevated for age in 61%. Driver mutations affecting the Ras pathway were identified in 92% of JMML pts (PTPN11: 36%, NF1: 11%, NRAS: 15%, KRAS: 15%, CBL: 14%), and monosomy 7 in 14%. Initial exploratory analysis revealed batch effects that could, at least in part, be explained by the different ethnical backgrounds of the pts but also included a technical component. Therefore, we stringently filtered probes containing SNPs and performed batch correction using an empirical Bayes framework (Combat). CpG dinucleotides (CpGs) showing variable methylation across normal hematopoietic cell types were excluded from the analysis to avoid confounding with differential cell type composition between the samples.Unsupervised consensus clustering using the 5000 most variable CpGs identified three stable methylation clusters in this inter-group patient cohort. According to their DNA methylation levels, the clusters were designated “high methylation” (HM), “intermediate methylation” (IM) and “low methylation” (LM). Overall, 29.5%, 23.5%, and 47.0% of the cohort of JMML and NS/MPD pts were assigned to the HM, IM, and LM clusters, respectively. The methylation clusters showed a highly significant association with specific Ras-pathway mutations and with known clinical risk factors. All NS/MPD pts clustered with the LM subgroup and showed the lowest methylation levels. The HM cluster was enriched for somatic PTPN11 mutations, elevated HbF and older age, while the LM cluster was enriched for NRAS and CBL mutations, normal HbF and age <2yrs. KRAS mutations as well as monosomy 7 were significantly enriched in the IM cluster. Comparison of clusterassignments made in the present study to those from the original publications found only 3.9% (11/285) discordant cases.This encouraged us to establish a low-dimensional DNA methylation classifier for JMML that could be applied in a prospective setting. Therefore, we split the intergroup cohort randomly into a training (N=202) and a validation cohort (N=83) and trained an extreme gradient boosting machine (xgboost). When using as few as 10 CpGs, the resulting model predicted methylation clusters with 94% accuracy in the validation cohort.This inter-group meta-analysis is the single largest dataset of JMML patients to date and demonstrates the robustness of the JMML methylation signature across continents of origin and its power to identify biologically and clinically distinct JMML subgroups. Our ongoing efforts include developing a sequencing-based assay to predict DNA methylation clusters and to validate this assay using an independent cohort of 60 pts from the three participating research groups. This assay would allow for inclusion of DNA methylation as a biomarker into prospective clinical trials to assist in molecularly driven risk-stratification of JMML. DisclosuresNiemeyer:Celgene: Consultancy, Membership on an entity's Board of Directors or advisory committees.

DNA Methylation and Genetic Profiles in 320 Patients with Myelodysplastic Syndromes

MDS is a heterogeneous group of myeloid neoplasms caused by genetic and epigenetic alterations. During the past decade, the major driver mutations in MDS have been fully investigated. However, the role of epigenetic alterations, particularly those of DNA methylation, has less intensively been studied, even though abnormal DNA methylation has long been implicated in the pathogenesis of MDS. In this study, we analyzed DNA methylation status of bone marrow mononuclear cells from 320 cases with MDS-SLD (n = 7), MDS-RS (n = 63), MDS-MLD (n = 51), MDS-EB (n = 186), MDS-U (n = 1), and MDS with isolated del(5q) (n = 12), using Illumina 450K methylation array. Mutations in major driver genes (51 genes) and abnormal genomic copy numbers were also interrogated using targeted-capture sequencing.Using unsupervised consensus clustering, we identified 3 subgroups showing unique DNA methylation profiles. Subsequently, we assessed differentially methylated positions (DMPs) associated with each subgroup. Differentially hypermethylated positions (hyper-DMPs) were significantly more enriched in Group 3 (n = 82) (P < 0.001), while differentially hypomethylated positions (hypo-DMPs) were more prominent in Group 1 (n = 125). Group 1 was significantly enriched for SF3B1 (46%) mutations (q < 0.01), while Group 2 (n = 131) was characterized by the enrichment of ASXL1 (38%), RUNX1 (30%), TP53 (26%), STAG2 (15%), and SETBP1 (6.7%) mutations (q < 0.01). In contrast, Group 3 (n = 64) was significantly enriched for TET2 (67%) and IDH1/2 (12% and 15%, respectively) mutations (q < 0.01), suggesting a strong association between DNA methylation and gene mutations. To further elucidate mutation-specific DNA methylation patterns, supervised analysis was performed for each mutation. As expected from their enrichment in Group 3 (q < 0.01), TET2 and IDH1/2 mutations were significantly associated with hyper-DMPs (P < 0.001) involving 1891 and 8330 promotor sites, respectively. Conspicuously, among these hypermethylated promoter sites, >1616 were commonly hypermethylated, strongly supporting the common impact of TET2 and IDH1/2 mutations on deregulated DNA methylation. To clarify prognostic impact of abnormal DNA methylation, we first interrogated the correlation between unique methylation subgroups and revised IPSS. Patients with very low or low risk were significantly dominant (74%) in Group 1 (q < 0.01), and very high or high risk cases were significantly enriched (68%) in Group 2 (q < 0.01). In accordance with this finding, patients in Group 3 showed significantly shorter overall survival (OS) compared to Group 1 (HR: 1.94, 95%CI: 1.11-3.4, P < 0.05) and OS was even worse in Group 2 patients (vs. Group 1: HR: 5.18, 95%CI: 3.21-8.36, P < 0.001).Strong correlations between epigenetic and genetic profiles were further interrogated using a Bayesian statistical model; on the basis of DNA methylation and gene mutations, the original 3 clusters were re-classified into 5 discrete clusters, clusters A, B, C, D, and E (n = 124, 17, 74, 46, and 59, respectively); patients in Group 1 and 3 largely clustered into Cluster A and E, respectively, while Group 2 was further subclassified into clusters B, C, and D. Clusters B and D were characterized by a conspicuos enrichment of DNMT3A (88%) and TP53 (69%) mutations (q < 0.001), while Cluster C was characterized by higher frequency of ASXL1 (71%), RUNX1 (54%), STAG2 (27%), and EZH2 (21%) mutations (q < 0.001). In contrast to significant associations between epigenetic regulators and unique methylation clusters, splice factor mutations tended to be clustered into multiple clusters, depending on type of co-occurring mutations. For example, combined SF3B1 and TET2 mutations (n = 20) were enriched in Cluster A, where highly associated with MDS-RS, while patients with SF3B1 and RUNX1 mutations (n = 9) were more grouped in Cluster C, mostly showing MDS-EB phenotype (89%). Similarly SRSF2 mutations with RUNX1 and/or ASXL1 mutations (n = 36) were enriched in Cluster C, largely associated with MDS-EB phenotype (80%), while those with TET2 or IDH1/2 (n = 39) were mainly grouped into Cluster C, many of which showed MDS-EB phenotype (74%). These findings highlight differential roles of mutated epigenetic regulators and splicing factors in abnormal DNA methylation.In conclusion, we elucidated the collaborative impact of DNA methylation profiles and mutation status on heterogeneous pathogenesis and prognosis in MDS. [Display omitted] DisclosuresNadarajah:MLL Munich Leukemia Laboratory: Employment. Baer:MLL Munich Leukemia Laboratory: Employment. Nakagawa:Sumitomo Dainippon Pharma Co., Ltd.: Research Funding. Inagaki:Sumitomo Dainippon Pharma Co., Ltd.: Employment. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Kern:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Prognostic effect of whole chromosomal aberration signatures in standard-risk, non-WNT/non-SHH medulloblastoma: a retrospective, molecular analysis of the HIT-SIOP PNET 4 trial.

SummaryBackgroundMost children with medulloblastoma fall within the standard-risk clinical disease group defined by absence of high-risk features (metastatic disease, large-cell/anaplastic histology, and MYC amplification), which includes 50–60% of patients and has a 5-year event-free survival of 75–85%. Within standard-risk medulloblastoma, patients in the WNT subgroup are established as having a favourable prognosis; however, outcome prediction for the remaining majority of patients is imprecise. We sought to identify novel prognostic biomarkers to enable improved risk-adapted therapies.MethodsThe HIT-SIOP PNET 4 trial recruited 338 patients aged 4–21 years with medulloblastoma between Jan 1, 2001, and Dec 31, 2006, in 120 treatment institutions in seven European countries to investigate hyperfractionated radiotherapy versus standard radiotherapy. In this retrospective analysis, we assessed the remaining tumour samples from patients in the HIT-SIOP PNET 4 trial (n=136). We assessed the clinical behaviour of the molecularly defined WNT and SHH subgroups, and identified novel independent prognostic markers and models for standard-risk patients with non-WNT/non-SHH disease. Because of the scarcity and low quality of available genomic material, we used a mass spectrometry-minimal methylation classifier assay (MS-MIMIC) to assess methylation subgroup and a molecular inversion probe array to detect genome-wide copy number aberrations. Prognostic biomarkers and models identified were validated in an independent, demographically matched cohort (n=70) of medulloblastoma patients with non-WNT/non-SHH standard-risk disease treated with conventional therapies (maximal surgical resection followed by adjuvant craniospinal irradiation [all patients] and chemotherapy [65 of 70 patients], at UK Children's Cancer and Leukaemia Group and European Society for Paediatric Oncology (SIOPE) associated treatment centres between 1990 and 2014. These samples were analysed by Illumina 450k DNA methylation microarray. HIT-SIOP PNET 4 is registered with ClinicalTrials.gov, number NCT01351870.FindingsWe analysed methylation subgroup, genome-wide copy number aberrations, and mutational features in 136 assessable tumour samples from the HIT-SIOP PNET 4 cohort, representing 40% of the 338 patients in the trial cohort. This cohort of 136 samples consisted of 28 (21%) classified as WNT, 17 (13%) as SHH, and 91 (67%) as non-WNT/non-SHH (we considered Group3 and Group4 medulloblastoma together in our analysis because of their similar molecular and clinical features). Favourable outcomes for WNT tumours were confirmed in patients younger than 16 years, and all relapse events in SHH (four [24%] of 17) occurred in patients with TP53 mutation (TP53mut) or chromosome 17p loss. A novel whole chromosomal aberration signature associated with increased ploidy and multiple non-random whole chromosomal aberrations was identified in 38 (42%) of the 91 samples from patients with non-WNT/non-SHH medulloblastoma in the HIT-SIOP PNET 4 cohort. Biomarkers associated with this whole chromosomal aberration signature (at least two of chromosome 7 gain, chromosome 8 loss, and chromosome 11 loss) predicted favourable prognosis. Patients with non-WNT/non-SHH medulloblastoma could be reclassified by these markers as having favourable-risk or high-risk disease. In patients in the HIT-SIOP PNET4 cohort with non-WNT/non-SHH medulloblastoma, with a median follow-up of 6·7 years (IQR 5·8–8·2), 5-year event-free survival was 100% in the favourable-risk group and 68% (95% CI 57·5–82·7; p=0·00014) in the high-risk group. In the validation cohort, with a median follow-up of 5·6 years (IQR 3·1–8·1), 5-year event-free survival was 94·7% (95% CI 85·2–100) in the favourable-risk group and 58·6% (95% CI 45·1–76·1) in the high-risk group (hazard ratio 9·41, 95% CI 1·25–70·57; p=0·029). Our comprehensive molecular investigation identified subgroup-specific risk models which allowed 69 (51%) of 134 accessible patients from the standard-risk medulloblastoma HIT-SIOP PNET 4 cohort to be assigned to a favourable-risk group.InterpretationWe define a whole chromosomal signature that allows the assignment of non-WNT/non-SHH medulloblastoma patients normally classified as standard-risk into favourable-risk and high-risk categories. In addition to patients younger than 16 years with WNT tumours, patients with non-WNT/non-SHH tumours with our defined whole chromosomal aberration signature and patients with SHH-TP53wild-type tumours should be considered for therapy de-escalation in future biomarker-driven, risk-adapted clinical trials. The remaining subgroups of patients with high-risk medulloblastoma might benefit from more intensive therapies.FundingCancer Research UK, Swedish Childhood Cancer Foundation, French Ministry of Health/French National Cancer Institute, and the German Children's Cancer Foundation.

DNA methylation-based classification of ependymomas in adulthood: implications for diagnosis and treatment

Ependymal tumors are glial tumors that commonly manifest in children and young adults. Their classification has remained entirely morphological until recently, and surgery and radiotherapy are the main treatment options, especially in adults. Here we sought to correlate DNA methylation profiles with clinical and pathological characteristics in the prospective cohort of the German Glioma Network. Tumors from 122 adult patients with myxopapillary ependymoma, ependymoma, anaplastic ependymoma, subependymoma, or RELA fusion-positive ependymoma classified according to the World Health Organization (WHO) 2016 were subjected to DNA methylation profiling using the Illumina HumanMethylation450 BeadChip platform. Molecular data were correlated with histologic features and clinical characteristics. At a median follow-up of 86.7 months, only 22 patients experienced progression (18.0%) and 13 patients (10.7%) died. Each tumor could be assigned to one of the previously defined molecular ependymoma subgroups. All histologic subependymomas corresponded to subependymoma (SE) DNA methylation subgroups, but the reverse was not true: 19 histologic ependymomas (WHO grade II) were allocated to molecular SE groups. Similarly, all histological myxopapillary ependymomas were assigned to the molecularly defined spinal myxopapillary ependymoma (SP-MPE) class, but this molecular subgroup additionally included 15 WHO grade II ependymomas by histology. Overall, WHO grade II ependymomas distributed into 7 molecular subgroups. Most adult patients with ependymoma show a favorable prognosis. Molecular classification may provide diagnostic and prognostic information beyond histology and facilitate patient stratification in future clinical trials. The prognostic significance of a subependymoma or myxopapillary ependymoma DNA methylation phenotype without corresponding histology requires further study. 1. Ependymoma diagnosed in adult patients most often shows a good prognosis. 2. Molecular classification can support diagnostic and prognostic information beyond histology.

Specific breast cancer prognosis-subtype distinctions based on DNA methylation patterns.

Tumour heterogeneity is an obstacle to effective breast cancer diagnosis and therapy. DNA methylation is an important regulator of gene expression, thus characterizing tumour heterogeneity by epigenetic features can be clinically informative. In this study, we explored specific prognosis‐subtypes based on DNA methylation status using 669 breast cancers from the TCGA database. Nine subgroups were distinguished by consensus clustering using 3869 CpGs that significantly influenced survival. The specific DNA methylation patterns were reflected by different races, ages, tumour stages, receptor status, histological types, metastasis status and prognosis. Compared with the PAM50 subtypes, which use gene expression clustering, DNA methylation subtypes were more elaborate and classified the Basal‐like subtype into two different prognosis‐subgroups. Additionally, 1252 CpGs (corresponding to 888 genes) were identified as specific hyper/hypomethylation sites for each specific subgroup. Finally, a prognosis model based on Bayesian network classification was constructed and used to classify the test set into DNA methylation subgroups, which corresponded to the classification results of the train set. These specific classifications by DNA methylation can explain the heterogeneity of previous molecular subgroups in breast cancer and will help in the development of personalized treatments for the new specific subtypes.

Risk-adapted therapy for young children with medulloblastoma (SJYC07): therapeutic and molecular outcomes from a multicentre, phase 2 trial

Young children with medulloblastoma have a poor overall survival compared with older children, due to use of radiation-sparing therapy in young children. Radiotherapy is omitted or reduced in these young patients to spare them from debilitating long-term side-effects. We aimed to estimate event-free survival and define the molecular characteristics associated with progression-free survival in young patients with medulloblastoma using a risk-stratified treatment strategy designed to defer, reduce, or delay radiation exposure. In this multicentre, phase 2 trial, we enrolled children younger than 3 years with newly diagnosed medulloblastoma at six centres in the USA and Australia. Children aged 3-5 years with newly diagnosed, non-metastatic medulloblastoma without any high-risk features were also eligible. Eligible patients were required to start therapy within 31 days from definitive surgery, had a Lansky performance score of at least 30, and did not receive previous radiotherapy or chemotherapy. Patients were stratified postoperatively by clinical and histological criteria into low-risk, intermediate-risk, and high-risk treatment groups. All patients received identical induction chemotherapy (methotrexate, vincristine, cisplatin, and cyclophosphamide), with high-risk patients also receiving an additional five doses of vinblastine. Induction was followed by risk-adapted consolidation therapy: low-risk patients received cyclophosphamide (1500 mg/m2 on day 1), etoposide (100 mg/m2 on days 1 and 2), and carboplatin (area under the curve 5 mg/mL per min on day 2) for two 4-week cycles; intermediate-risk patients received focal radiation therapy (54 Gy with a clinical target volume of 5 mm over 6 weeks) to the tumour bed; and high-risk patients received chemotherapy with targeted intravenous topotecan (area under the curve 120-160 ng-h/mL intravenously on days 1-5) and cyclophosphamide (600 mg/m2 intravenously on days 1-5). After consolidation, all patients received maintenance chemotherapy with cyclophosphamide, topotecan, and erlotinib. The coprimary endpoints were event-free survival and patterns of methylation profiling associated with progression-free survival. Outcome and safety analyses were per protocol (all patients who received at least one dose of induction chemotherapy); biological analyses included all patients with tissue available for methylation profiling. This trial is registered with ClinicalTrials.gov, number NCT00602667, and was closed to accrual on April 19, 2017. Between Nov 27, 2007, and April 19, 2017, we enrolled 81 patients with histologically confirmed medulloblastoma. Accrual to the low-risk group was suspended after an interim analysis on Dec 2, 2015, when the 1-year event-free survival was estimated to be below the stopping rule boundary. After a median follow-up of 5·5 years (IQR 2·7-7·3), 5-year event-free survival was 31·3% (95% CI 19·3-43·3) for the whole cohort, 55·3% (95% CI 33·3-77·3) in the low-risk cohort (n=23) versus 24·6% (3·6-45·6) in the intermediate-risk cohort (n=32; hazard ratio 2·50, 95% CI 1·19-5·27; p=0·016) and 16·7% (3·4-30·0) in the high-risk cohort (n=26; 3·55, 1·66-7·59; p=0·0011; overall p=0·0021). 5-year progression-free survival by methylation subgroup was 51·1% (95% CI 34·6-67·6) in the sonic hedgehog (SHH) subgroup (n=42), 8·3% (95% CI 0·0-24·0%) in the group 3 subgroup (n=24), and 13·3% (95% CI 0·0-37·6%) in the group 4 subgroup (n=10). Within the SHH subgroup, two distinct methylation subtypes were identified and named iSHH-I and iSHH-II. 5-year progression-free survival was 27·8% (95% CI 9·0-46·6; n=21) for iSHH-I and 75·4% (55·0-95·8; n=21) for iSHH-II. The most common adverse events were grade 3-4 febrile neutropenia (48 patients [59%]), neutropenia (21 [26%]), infection with neutropenia (20 [25%]), leucopenia (15 [19%]), vomiting (15 [19%]), and anorexia (13 [16%]). No treatment-related deaths occurred. The risk-adapted approach did not improve event-free survival in young children with medulloblastoma. However, the methylation subgroup analyses showed that the SHH subgroup had improved progression-free survival compared with the group 3 subgroup. Moreover, within the SHH subgroup, the iSHH-II subtype had improved progression-free survival in the absence of radiation, intraventricular chemotherapy, or high-dose chemotherapy compared with the iSHH-I subtype. These findings support the development of a molecularly driven, risk-adapted, treatment approach in future trials in young children with medulloblastoma. American Lebanese Syrian Associated Charities, St Jude Children's Research Hospital, NCI Cancer Center, Alexander and Margaret Stewart Trust, Sontag Foundation, and American Association for Cancer Research.

PD-1 (PDCD1) Promoter Methylation Is a Prognostic Factor in Patients With Diffuse Lower-Grade Gliomas Harboring Isocitrate Dehydrogenase (IDH) Mutations

Immune checkpoints are important targets for immunotherapies. However, knowledge on the epigenetic modification of immune checkpoint genes is sparse. In the present study, we investigated promoter methylation of CTLA4, PD-L1, PD-L2, and PD-1 in diffuse lower-grade gliomas (LGG) harboring isocitrate dehydrogenase (IDH) mutations with regard to mRNA expression levels, clinicopathological parameters, previously established methylation subtypes, immune cell infiltrates, and survival in a cohort of 419 patients with IDH-mutated LGG provided by The Cancer Genome Atlas.PD-L1, PD-L2, and CTLA-4 mRNA expression levels showed a significant inverse correlation with promoter methylation (PD-L1: p=0.005; PD-L2: p<0.001; CTLA-4: p<0.001). Furthermore, immune checkpoint methylation was significantly associated with age (PD-L2: p=0.003; PD-1: p=0.015), molecular alterations, i.e. MGMT methylation (PD-L1: p<0.001; PD-L2: p<0.001), ATRX mutations (PD-L2: p<0.001, PD-1: p=0.001), and TERT mutations (PD-L1: p=0.035, PD-L2: p<0.001, PD-1: p<0.001, CTLA4: p<0.001) as well as methylation subgroups and immune cell infiltrates. In multivariate Cox proportional hazard analysis, PD-1 methylation qualified as strong prognostic factor (HR=0.51 [0.34–0.76], p=0.001).Our findings suggest an epigenetic regulation of immune checkpoint genes via DNA methylation in LGG. PD-1 methylation may assist the identification of patients that might benefit from an alternative treatment, particularly in the context of emerging immunotherapies.

Abstract B58: The effect of differential miRNA expression in glioblastoma multiforme molecular subtypes

Abstract Glioblastoma multiforme (GBM) is characterized by a distinct profile of specific miRNAs that regulate pathways of therapeutic importance, such as cell proliferation, apoptosis, cell cycle regulation, migration, invasion, and angiogenesis. Recently, the molecular classification of gliomas was updated, and three methylation clusters with distinct prognostic values for GBM were characterized. The aim of this research is to evaluate the expression, the functional roles, and the target genes of the miRNAs that have potential value for diagnosis, prognosis, and treatment in patients within the newest molecular subgroup classification of GBM. First, we performed an in silico analysis 606 GBM samples from the Genome Data Commons (GDC) cohort. Based on the methylation subgroups of gliomas recently published (Cecarelli et al., 2016), we performed a comparison of the miRNA expression between LGm4, LGm5, and LGm6 subgroups through Nexus Expression 3.0 (Biodiscovery). The comparison between LGm4 x LGm6 indicated that eight miRNAs were upregulated, while 17 were downregulated. The LGm5 x LGm4 comparison showed two downregulated miRNAs, and the LGm5 x LGm6 comparison showed one upregulated miRNA. By comparing the differentially expressed miRNAs found in our analysis with two other in silico studies, we found an overlap of 54% of the differentially expressed miRNAs. We will now evaluate the expression levels of the upregulated miRNAs obtained through the in silico analysis of samples of the Hospital of Clinics - Faculty of Medicine of Ribeirão Preto (HCFMRP/USP) through a customized nCounter NanoString panel of 24 miRNA. We will also evaluate the plasma exosomal miRNAs from the same patients using the same 24 miRNA nCounter NanoString panel. By identifying the most significant miRNAs regulating cell proliferation and invasive behavior in this tumor, we will be able to use expression levels of the most informative miRNA to both guide the molecular diagnosis of GBM and facilitate future targeted therapeutics for this aggressive disease. Citation Format: Livia Ferreira Silva Verzola, Thiago Vidotto, Jeremy Andrew Squire. The effect of differential miRNA expression in glioblastoma multiforme molecular subtypes [abstract]. In: Proceedings of the AACR International Conference held in cooperation with the Latin American Cooperative Oncology Group (LACOG) on Translational Cancer Medicine; May 4-6, 2017; São Paulo, Brazil. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(1_Suppl):Abstract nr B58.

Characterizing temporal genomic heterogeneity in pediatric high-grade gliomas

Pediatric high-grade gliomas (pHGGs) are aggressive neoplasms representing approximately 20% of brain tumors in children. Current therapies offer limited disease control, and patients have a poor prognosis. Empiric use of targeted therapy, especially at progression, is increasingly practiced despite a paucity of data regarding temporal and therapy-driven genomic evolution in pHGGs. To study the genetic landscape of pHGGs at recurrence, we performed whole exome and methylation analyses on matched primary and recurrent pHGGs from 16 patients. Tumor mutational profiles identified three distinct subgroups. Group 1 (n = 7) harbored known hotspot mutations in Histone 3 (H3) (K27M or G34V) or IDH1 (H3/IDH1 mutants) and co-occurring TP53 or ACVR1 mutations in tumor pairs across the disease course. Group 2 (n = 7), H3/IDH1 wildtype tumor pairs, harbored novel mutations in chromatin modifiers (ZMYND11, EP300 n = 2), all associated with TP53 alterations, or had BRAF V600E mutations (n = 2) conserved across tumor pairs. Group 3 included 2 tumors with NF1 germline mutations. Pairs from primary and relapsed pHGG samples clustered within the same DNA methylation subgroup. ATRX mutations were clonal and retained in H3G34V and H3/IDH1 wildtype tumors, while different genetic alterations in this gene were observed at diagnosis and recurrence in IDH1 mutant tumors. Mutations in putative drug targets (EGFR, ERBB2, PDGFRA, PI3K) were not always shared between primary and recurrence samples, indicating evolution during progression. Our findings indicate that specific key driver mutations in pHGGs are conserved at recurrence and are prime targets for therapeutic development and clinical trials (e.g. H3 post-translational modifications, IDH1, BRAF V600E). Other actionable mutations are acquired or lost, indicating that re-biopsy at recurrence will provide better guidance for effective targeted therapy of pHGGs.

Radiomic subtyping improves disease stratification beyond key molecular, clinical, and standard imaging characteristics in patients with glioblastoma.

The purpose of this study was to analyze the potential of radiomics for disease stratification beyond key molecular, clinical, and standard imaging features in patients with glioblastoma. Quantitative imaging features (n = 1043) were extracted from the multiparametric MRI of 181 patients with newly diagnosed glioblastoma prior to standard-of-care treatment (allocated to a discovery and a validation set, 2:1 ratio). A subset of 386/1043 features were identified as reproducible (in an independent MRI test-retest cohort) and selected for analysis. A penalized Cox model with 10-fold cross-validation (Coxnet) was fitted on the discovery set to construct a radiomic signature for predicting progression-free and overall survival (PFS and OS). The incremental value of a radiomic signature beyond molecular (O6-methylguanine-DNA methyltransferase [MGMT] promoter methylation, DNA methylation subgroups), clinical (patient's age, KPS, extent of resection, adjuvant treatment), and standard imaging parameters (tumor volumes) for stratifying PFS and OS was assessed with multivariate Cox models (performance quantified with prediction error curves). The radiomic signature (constructed from 8/386 features identified through Coxnet) increased the prediction accuracy for PFS and OS (in both discovery and validation sets) beyond the assessed molecular, clinical, and standard imaging parameters (P ≤ 0.01). Prediction errors decreased by 36% for PFS and 37% for OS when adding the radiomic signature (compared with 29% and 27%, respectively, with molecular + clinical features alone). The radiomic signature was-along with MGMT status-the only parameter with independent significance on multivariate analysis (P ≤ 0.01). Our study stresses the role of integrating radiomics into a multilayer decision framework with key molecular and clinical features to improve disease stratification and to potentially advance personalized treatment of patients with glioblastoma.