Methylation Datasets Research Articles

Improve consensus partitioning via a hierarchical procedure.

Consensus partitioning is an unsupervised method widely used in high-throughput data analysis for revealing subgroups and assigning stability for the classification. However, standard consensus partitioning procedures are weak for identifying large numbers of stable subgroups. There are two major issues. First, subgroups with small differences are difficult to be separated if they are simultaneously detected with subgroups with large differences. Second, stability of classification generally decreases as the number of subgroups increases. In this work, we proposed a new strategy to solve these two issues by applying consensus partitioning in a hierarchical procedure. We demonstrated hierarchical consensus partitioning can be efficient to reveal more meaningful subgroups. We also tested the performance of hierarchical consensus partitioning on revealing a great number of subgroups with a large deoxyribonucleic acid methylation dataset. The hierarchical consensus partitioning is implemented in the R package cola with comprehensive functionalities for analysis and visualization. It can also automate the analysis only with a minimum of two lines of code, which generates a detailed HTML report containing the complete analysis. The cola package is available at https://bioconductor.org/packages/cola/.

Polyunsaturated Fatty Acid-Enriched Lipid Fingerprint of Glioblastoma Proliferative Regions Is Differentially Regulated According to Glioblastoma Molecular Subtype.

Glioblastoma (GBM) represents one of the deadliest tumors owing to a lack of effective treatments. The adverse outcomes are worsened by high rates of treatment discontinuation, caused by the severe side effects of temozolomide (TMZ), the reference treatment. Therefore, understanding TMZ’s effects on GBM and healthy brain tissue could reveal new approaches to address chemotherapy side effects. In this context, we have previously demonstrated the membrane lipidome is highly cell type-specific and very sensitive to pathophysiological states. However, little remains known as to how membrane lipids participate in GBM onset and progression. Hence, we employed an ex vivo model to assess the impact of TMZ treatment on healthy and GBM lipidome, which was established through imaging mass spectrometry techniques. This approach revealed that bioactive lipid metabolic hubs (phosphatidylinositol and phosphatidylethanolamine plasmalogen species) were altered in healthy brain tissue treated with TMZ. To better understand these changes, we interrogated RNA expression and DNA methylation datasets of the Cancer Genome Atlas database. The results enabled GBM subtypes and patient survival to be linked with the expression of enzymes accounting for the observed lipidome, thus proving that exploring the lipid changes could reveal promising therapeutic approaches for GBM, and ways to ameliorate TMZ side effects.

A pan-tissue DNA methylation atlas enables in silico decomposition of human tissue methylomes at cell-type resolution

Bulk-tissue DNA methylomes represent an average over many different cell types, hampering our understanding of cell-type-specific contributions to disease development. As single-cell methylomics is not scalable to large cohorts of individuals, cost-effective computational solutions are needed, yet current methods are limited to tissues such as blood. Here we leverage the high-resolution nature of tissue-specific single-cell RNA-sequencing datasets to construct a DNA methylation atlas defined for 13 solid tissue types and 40 cell types. We comprehensively validate this atlas in independent bulk and single-nucleus DNA methylation datasets. We demonstrate that it correctly predicts the cell of origin of diverse cancer types and discovers new prognostic associations in olfactory neuroblastoma and stage 2 melanoma. In brain, the atlas predicts a neuronal origin for schizophrenia, with neuron-specific differential DNA methylation enriched for corresponding genome-wide association study risk loci. In summary, the DNA methylation atlas enables the decomposition of 13 different human tissue types at a high cellular resolution, paving the way for an improved interpretation of epigenetic data.

Characterization of DNA methylation as well as mico-RNA expression and screening of epigenetic markers in adipogenesis

This study aimed to use bioinformatics methods to characterize epigenetic changes in terms of micro-RNA(miRNA) expression and DNA methylation during adipogenesis. The mRNA and miRNA expression microarray and DNA methylation dataset were obtained from the GEO database. Differentially expressed genes (DEGs), differentially expressed miRNAs (DEMs) and differentially methylated probes (DMPs) were filtered using the limma package. The R language cluster profile package was used for functional and enrichment analysis. A protein–protein interaction (PPI) network was constructed using STRING and visualized in Cytoscape. The Connection map (CMap) website tool was used to screen potential therapeutic drugs for adipogenesis. When comparing the early and late stages of adipogenesis, 111 low miRNA targeted upregulated genes and 64 high miRNA targeted downregulated genes were obtained, as well as 663 low-methylated high-expressed genes and 237 high-methylated low-expressed genes. In addition, 41 genes (24 upregulated and 17 downregulated) were simultaneously regulated by abnormal miRNA changes and DNA methylation. Ten chemicals were identified as putative therapeutics for adipogenesis. In addition, among the dual-regulated genes identified, CANX, HNRNPA1, MCL1, and PPIF may play key roles in the epigenetic regulation of adipogenesis and may serve as aberrant methylation or miRNA targeting biomarkers.

Ferroptosis-associated DNA methylation signature predicts overall survival in patients with head and neck squamous cell carcinoma

BackgroundHead and neck squamous cell carcinoma (HNSCC) is a common cancer characterized by late diagnosis and poor prognosis. The aim of this study was to identify a novel ferroptosis-related DNA methylation signature as an alternative diagnosis index for patients with HNSCC.MethodsMethylome and transcriptome data of 499 HNSCC patients, including 275 oral squamous cell carcinoma (OSCC) samples, were obtained from The Cancer Genome Atlas (TCGA). An additional independent methylation dataset of 50 OSCC patients from the NCBI Gene Expression Omnibus (GEO) database was used for validation. As an index of ferroptosis activity, the ferroptosis score (FS) of each patient was inferred from the transcriptome data using single-sample gene set enrichment analysis. Univariate, multivariate, and LASSO Cox regression analyses were used to select CpG sites for the construction of a ferroptosis-related DNA methylation signature for diagnosis of patients.ResultsWe initially inferred the FS of each TCGA HNSCC patient and divided the samples into high- and low-FS subgroups. Results showed that the high-FS subgroup displayed poor overall survival. Moreover, 378 differentially methylated CpG sites (DMCs) were identified between the two HNSCC subgroups, with 16 selected to construct a 16-DNA methylation signature for risk prediction in HNSCC patients using the LASSO and multivariate Cox regression models. Relative operating characteristic (ROC) curve analysis showed great predictive efficiency for 1-, 3-, and 5-year HNSCC survival using the 16-DNA methylation signature. Its predictive efficiency was also observed in OSCC patients from the TCGA and GEO databases. In addition, we found that the signature was associated with the fractions of immune types in the tumor immune microenvironment (TIME), suggesting potential interactions between ferroptosis and TIME in HNSCC progression.ConclusionsWe established a novel ferroptosis-related 16-DNA methylation signature that could be applied as an alternative tool to predict prognosis outcome in patients with HNSCC, including OSCC.

Trans-omics analyses revealed key epigenetic genes associated with overall survival in secondary progressive multiple sclerosis

BackgroundSecondary progressive multiple sclerosis (SPMS) is the second most common presentation of multiple sclerosis (MS) and is characterized by a gradually deteriorating disease with or without relapses. Approximately 80% of patients with relapsing-remitting MS (RRMS) develop SPMS within 20 years. Epidemiological investigations have revealed an average 7-year life expectancy decrease (more severe in progressive subtypes) in patients with MS. Studies have focused on the neurodegenerative pathogenesis of SPMS; and epigenetic changes have been associated with disease progression in neurodegenerative disorders. However, the evidence for the association between epigenetic changes and SPMS is scarce. Thus, in this study we aimed to identify the key epigenetic genes in SPMS. MethodsWe downloaded DNA methylation and gene expression matrices from the Gene Expression Omnibus (GEO) database. We used bioinformatic analyses to identify key epigenetic genes associated with overall survival (OS) in patients with SPMS. ResultsWe found 49 differentially methylated positions (DMPs) between the SPMS and control GSE40360 datasets. We used the wANNOVAR server to obtain 64 methylated genes. We merged the gene expression datasets (GSE131282 and GSE135511) in the NetworkAnalyst platform and found 12,442 differentially-expressed genes (DEGs) between SPMS and controls using the Fisher's method, fixed effect model, Vote counting, and direct merging methods. Moreover, we identified 21 epigenetic genes (all hyper-methylated) after an integrating analysis of DMPs and DEGs of patients with SPMS. We established an epigenetic gene signature associated with the OS of patients with SPMS including six hyper-methylated genes (ITGA6, PPP1R16B, RNF126, ABHD8, FOXK1, and SLC6A19) based on the LASSO-Cox method. The calculated individual risk scores were associated with Oss, and we divided patients into high- and low-risk groups on the basis of the mean cut-off value. The six key epigenetic genes were significantly associated with gender, disease duration, and age at death via Spearman correlation analyses. In addition, survival analyses revealed a significant OS difference between high- and low-risk groups. The ROC curves indicated good performance for this predictive model. ConclusionWe identified 21 hyper-methylated genes in patients with SPMS via an integrated analysis of DNA methylation and gene expression datasets. We identified a six-epigenetic gene signature that predicts the individual OS with good accuracy. These results indicated that epigenetic modifications play a vital role in the disease progression of SPMS.

Efficient gradient boosting for prognostic biomarker discovery.

A gradient boosting decision tree (GBDT) is a powerful ensemble machine-learning method that has the potential to accelerate biomarker discovery from high-dimensional molecular data. Recent algorithmic advances, such as extreme gradient boosting (XGB) and light gradient boosting (LGB), have rendered the GBDT training more efficient, scalable and accurate. However, these modern techniques have not yet been widely adopted in discovering biomarkers for censored survival outcomes, which are key clinical outcomes or endpoints in cancer studies. In this paper, we present a new R package 'Xsurv' as an integrated solution that applies two modern GBDT training frameworks namely, XGB and LGB, for the modeling of right-censored survival outcomes. Based on our simulations, we benchmark the new approaches against traditional methods including the stepwise Cox regression model and the original gradient boosting function implemented in the package 'gbm'. We also demonstrate the application of Xsurv in analyzing a melanoma methylation dataset. Together, these results suggest that Xsurv is a useful and computationally viable tool for screening a large number of prognostic candidate biomarkers, which may facilitate future translational and clinical research. 'Xsurv' is freely available as an R package at: https://github.com/topycyao/Xsurv. Supplementary data are available at Bioinformatics online.

Identification of key genes associated with survival of glioblastoma multiforme using integrated analysis of TCGA datasets

Background and ObjectiveGlioblastoma (GBM) is the most aggressive type of brain tumor. In spite of having various treatment options, GBM patients usually have a poor prognosis. Genetic markers play a vital role in the progression of the disease. Identification of these novel molecular biomarkers is essential to explain the mechanisms or improve the prognosis of GBM. Advances in high throughput genomic technologies enable the analysis of the varied types of omics data to find biomarkers in GBM. Although data repositories like The Cancer Genome Atlas (TCGA) are rich sources of such multi-omics data, integrating these different genomic datasets of varying quality and patient heterogeneity is challenging. MethodsMulti-omics gene expression datasets from TCGA consisting of DNA methylation, RNA sequencing, and copy number variation (CNV) of GBM patient is obtained to carry out the analysis. The Cox proportional hazards regression model is developed in R to identify significant genes from diverse datasets associated with the patient's survival. (Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) is used as an estimator for the model. Validation is performed to determine the accuracy and corresponding prediction error. ResultsFive key genes are identified from DNA Methylation and RNA sequencing datasets those are ANK1, HOXA9, TOX2, CXCR6, PIGZ, and L3MBTL, KDM5B, CCDC138, NUS1P1, and ARHGAP42, respectively. Higher expression values of these genes determine better survival of the GBM patients. Kaplan-Meier estimate curves show the exact correlation. Lower values of AIC and BIC determine the suitability of the model. The prediction model is validated on the test set and signifies a low error rate. Copy number variation data is also analysed to find the significant chromosomal location of GBM patients associated with chromosome 2,5,6,7,12,13, respectively. Among all nine CNV locations are found to be influencing the progression of GBM. ConclusionIntegrated analysis of multiple omics dataset is carried out to identify significant genes from DNA Methylation and RNA sequencing profiles of 76 common individuals. Copy number variation dataset for the same patients is analyzed to recognize notable locations associated with 22 chromosomes. The survival analysis determines the correlation of these biomarkers with the progression of the disease.

Workflow to Mine Frequent DNA Co-methylation Clusters in DNA Methylome Data.

The advances in high-throughput nucleotide sequencing technology revolutionized biomedical research. Vast amount of genomic data rapidly accumulates in a daily basis, which in turn calls for the development of powerful bioinformatics tools and efficient workflows to analyze them. One of the approaches to address the "big data" issue is to mine highly correlated clusters/networks of biological molecules, which may provide rich yet hidden information about the underlying functional, regulatory, or structural relationships among genes, proteins, genomic loci or various types of biological molecules or events. A network mining algorithm lmQCM has recently been developed, which can be applied to mine tightly connected correlation clusters (networks) in large biological data with big sample size, and it guarantees a lower bound of the cluster density. This algorithm has been used in a variety of cancer transcriptomic data to mine gene co-expression networks (GCNs), but it can be applied to any correlational matrix. lmQCM is available through R package lmQCM as well as the online tool package TSUNAMI ( https://biolearns.medicine.iu.edu ).In this study, the purpose is to establish an analytical workflow to apply lmQCM for frequent (consensus)cluster mining in multiple DNA methylation datasets in different cancers and extract the underlying common co-methylation networks for genes.Specifically, the workflow is applied to analyze DNA methylome data across different cancer types using lmQCM. It mines frequent DNA methylation clusters based on individual clustering mining results, identifying common as well as distinctive DNA methylation patterns among different cancer types. This workflow has successfully identified frequent GCNs in 33 types of cancers, thus proven to be a powerful tool to analyze large biological data. It helps to identify common features as well as distinctions among different diseases, disease subtypes, or among different biological processes. The resulted frequent clusters may provide new insights on the pathway/function networks. In the case of disease study, the results lead to new directions for biomarker and drug target discovery. The advantages of this workflow include the highly efficient processing of large biological data generated from high-throughput experiments, quick identification of highly correlated interaction networks, substantial reduction of the data dimensionality to a manageable number of variables for downstream comparative analysis, and consequently increased statistical power for detecting differences between conditions.

Screening of Crucial Differentially-Methylated/Expressed Genes for Alzheimer's Disease.

Background: We aimed to make an integrated analysis of published transcriptome and DNA methylation dataset to ascertain the key differentially methylated and differentially expressed genes for Alzherimer's disease (AD). Methods: Two gene expression microarrays and 1 gene methylation microarray were downloaded for identification of differentially expressed genes and differentially methylated genes. Then, we used various biological information databases to annotate the functions of the differentially-methylated/expressed genes, and screen out key genes and important signaling pathways. Finally, we validate the differentially-methylated/expressed genes in the additional online datasets and in blood from AD patients.Results: A total of 8 hub hypomethylated-high expression genes were obtained, including Rac family small GTPase 2, FGR proto-oncogene, Src family tyrosine kinase, LYN proto-oncogene, Src family tyrosine kinase, protein kinase C delta, myosin IF, integrin subunit alpha 5, semaphorin 4D, and growth arrest specific protein 7. Some enriched signaling pathways of hypomethylated-high expression genes were identified, including regulation of actin cytoskeleton, chemokine signaling pathway, Fc gamma R-mediated phagocytosis, and axon guidance. Conclusion: Differentially-methylated/expressed genes are likely to be associated with AD.

Post-Traumatic Stress Disorder (PTSD) biomarker identification using a deep learning model

Post-traumatic stress disorder (PTSD) is a neuropsychological disorder in which individuals struggle to recover from a traumatizing event. It affects a significant population, including COVID-19 patients, frontline health workers, and war veterans. Given biases associated with self-assessment and diagnosis of PTSD, researchers are actively searching for unbiased biological markers (biomarkers) for predicting PTSD status. The Systems Biology of PTSD Consortium has collected candidate biomarkers for PTSD using molecular and clinical measurements of male war veterans between the ages of 20 and 60. PTSD-positive and negative subjects were separated based on the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) Clinician Administered PTSD Scale (CAPS) scores, derived from structured interviews to measure an individual’s abundance of symptoms, such as re-experiencing, flashbacks, and hyperarousal. CAPS scores higher than 40 were considered PTSD-positive and below 20 were considered negative. We created artificial neural network models to classify PTSD-positive and negative individuals based on metabolomics, micro-RNA (miRNA), protein expression, endocrine markers, and DNA methylation datasets. Model training involved 64 iterations of a Bayesian Hyperparameter Optimization algorithm with 5-fold cross-validation. Each model was calibrated based on cross-validation performance and variance across iterations and then fit to the entire respective dataset (76 PTSD-positive, 76 PTSD-negative). We applied the trained models to an independent validation cohort to assess accuracy on unseen datasets. The top performing datasets from the validation cohort based on classification accuracy were metabolomics (65.2%) and protein expression (61.8%). We anticipate the candidate biomarkers identified in this and future studies will assist with the diagnosis of PTSD.

A bespoke target selection tool to guide biomarker discovery in tubo-ovarian cancer

IntroductionCancers presenting at advanced stages inherently have poor prognosis. High grade serous carcinoma (HGSC) is the most common and aggressive form of tubo-ovarian cancer. Clinical tests to accurately diagnose and monitor this condition are lacking. Hence, development of disease-specific tests are urgently required. MethodsThe molecular profile of HGSC during disease progression was investigated in a unique patient cohort. A bespoke data browser was developed to analyse gene expression and DNA methylation datasets for biomarker discovery. The Ovarian Cancer Data Browser (OCDB) is built in C# with a.NET framework using an integrated development environment of Microsoft Visual Studio and fast access files (.faf). The graphical user interface is easy to navigate between four analytical modes (gene expression; methylation; combined gene expression and methylation data; methylation clusters), with a rapid query response time. A user should first define a disease progression trend for prioritising results. Single or multiomics data are then mined to identify probes, genes and methylation clusters that exhibit the desired trend. A unique scoring system based on the percentage change in expression/methylation between disease stages is used. Results are filtered and ranked using weighting and penalties. ResultsThe OCDB’s utility for biomarker discovery is demonstrated with the identified target OSR2. Trends in OSR2 repression and hypermethylation with HGSC disease progression were confirmed in the browser samples and an independent cohort using bioassays. The OSR2 methylation biomarker could discriminate HGSC with high specificity (95%) and sensitivity (93.18%). ConclusionsThe OCDB has been refined and validated to be an integral part of a unique biomarker discovery pipeline. It may also be used independently to aid identification of novel targets. It carries the potential to identify further biomarker assays that can reduce type I and II errors within clinical diagnostics.

Profiling DNA Methylation in Human Naïve Pluripotent Stem Cells.

DNA methylation represents one of the best characterized epigenetic modifications. Inparticular, global demethylation is a common feature of epigenetic reprogramming to naïve pluripotency in human and mouse pluripotent stem cells. In parallel to the global changes, several locus-specific changes to the DNA methylation landscape occur and also loss of imprinting has been observed in naïve human pluripotent stem cells. The current gold standard to assess and quantitively map DNA methylation is bisulfite sequencing. Various protocols are available for genome-wide bisulfite sequencing and here I describe an optimized method based on Post Bisulfite Adapter Tagging (PBAT) for low amounts of DNA or cells, with as little as 50 cells as minimum requirement, and with the possibility to process a large number of samples in parallel. I outline the basic bioinformatic steps needed to process rawIllumina sequencing data and then describe the inital steps ofthe analysis of DNA methylation datasets, including an assessment of imprint control regions.

Risk-focused differences in molecular processes implicated in SARS-CoV-2 infection: corollaries in DNA methylation and gene expression

BackgroundUnderstanding the molecular basis of susceptibility factors to the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection is a global health imperative. It is well-established that males are more likely to acquire SARS-CoV-2 infection and exhibit more severe outcomes. Similarly, exposure to air pollutants and pre-existing respiratory chronic conditions, such as asthma and chronic obstructive respiratory disease (COPD) confer an increased risk to coronavirus disease 2019 (COVID-19).MethodsWe investigated molecular patterns associated with risk factors in 398 candidate genes relevant to COVID-19 biology. To accomplish this, we downloaded DNA methylation and gene expression data sets from publicly available repositories (GEO and GTEx Portal) and utilized data from an empirical controlled human exposure study conducted by our team.ResultsFirst, we observed sex-biased DNA methylation patterns in autosomal immune genes, such as NLRP2, TLE1, GPX1, and ARRB2 (FDR < 0.05, magnitude of DNA methylation difference Δβ > 0.05). Second, our analysis on the X-linked genes identified sex associated DNA methylation profiles in genes, such as ACE2, CA5B, and HS6ST2 (FDR < 0.05, Δβ > 0.05). These associations were observed across multiple respiratory tissues (lung, nasal epithelia, airway epithelia, and bronchoalveolar lavage) and in whole blood. Some of these genes, such as NLRP2 and CA5B, also exhibited sex-biased gene expression patterns. In addition, we found differential DNA methylation patterns by COVID-19 status for genes, such as NLRP2 and ACE2 in an exploratory analysis of an empirical data set reporting on human COVID-9 infections. Third, we identified modest DNA methylation changes in CpGs associated with PRIM2 and TATDN1 (FDR < 0.1, Δβ > 0.05) in response to particle-depleted diesel exhaust in bronchoalveolar lavage. Finally, we captured a DNA methylation signature associated with COPD diagnosis in a gene involved in nicotine dependence (COMT) (FDR < 0.1, Δβ > 0.05).ConclusionOur findings on sex differences might be of clinical relevance given that they revealed molecular associations of sex-biased differences in COVID-19. Specifically, our results hinted at a potentially exaggerated immune response in males linked to autosomal genes, such as NLRP2. In contrast, our findings at X-linked loci such as ACE2 suggested a potentially distinct DNA methylation pattern in females that may interact with its mRNA expression and inactivation status. We also found tissue-specific DNA methylation differences in response to particulate exposure potentially capturing a nitrogen dioxide (NO2) effect—a contributor to COVID-19 susceptibility. While we identified a molecular signature associated with COPD, all COPD-affected individuals were smokers, which may either reflect an association with the disease, smoking, or may highlight a compounded effect of these two risk factors in COVID-19. Overall, our findings point towards a molecular basis of variation in susceptibility factors that may partly explain disparities in the risk for SARS-CoV-2 infection.

A deep learning framework with an embedded-based feature selection approach for the early detection of the Alzheimer's disease

Ageing is associated with various ailments including Alzheimer ’s disease (AD), which is a progressive form of dementia. AD symptoms develop over a period of years and, unfortunately, there is no cure. Existing AD treatments can only slow down the progression of symptoms and thus it is critical to diagnose the disease at an early stage. To help improve the early diagnosis of AD, a deep learning-based classification model with an embedded feature selection approach was used to classify AD patients. An AD DNA methylation data set (64 records with 34 cases and 34 controls) from the GEO omnibus database was used for the analysis. Before selecting the relevant features, the data were preprocessed by performing quality control, normalization and downstream analysis. As the number of associated CpG sites was huge, four embedded-based feature selection models were compared and the best method was used for the proposed classification model. An Enhanced Deep Recurrent Neural Network (EDRNN) was implemented and compared to other existing classification models, including a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), and a Deep Recurrent Neural Network (DRNN). The results showed a significant improvement in the classification accuracy of the proposed model as compared to the other methods.

Deciphering a TB-related DNA methylation biomarker and constructing a TB diagnostic classifier

We systemically identified tuberculosis (TB)-related DNA methylation biomarkers and further constructed classifiers for TB diagnosis. TB-related DNA methylation datasets were searched through October 3, 2020. Limma and DMRcate were employed to identify differentially methylated probes (DMPs) and regions (DMRs). Machine learning methods were used to construct classifiers. The performance of the classifiers was evaluated in discovery datasets and a prospective independent cohort. Eighty-nine DMPs and 24 DMRs were identified based on 67 TB patients and 45 healthy controls from 4 datasets. Nine and three DMRs were selected by elastic net regression and logistic regression, respectively. Among the selected DMRs, two regions (chr3: 195635643–195636243 and chr6: 29691631–29692475) were differentially methylated in the independent cohort (p = 4.19 × 10−5 and 0.024, respectively). Among the ten classifiers, the 3-DMR logistic regression classifier exhibited the strongest performance. The sensitivity, specificity, and area under the curve were, respectively, 79.1%, 84.4%, and 0.888 in the discovery datasets and 64.5%, 90.3%, and 0.838 in the independent cohort. The differential diagnostic ability of this classifier was also assessed. Collectively, these data showed that DNA methylation might be a promising TB diagnostic biomarker. The 3-DMR logistic regression classifier is a potential clinical tool for TB diagnosis, and further validation is needed.

Abstract 14166: Epigenetic Biomarkers Correlate With Post-stroke Long-term Memory Recovery

Introduction: Epigenetic markers including 5-metylcytosine (5mC) can reflect disease severity and treatment effect. Stroke is a multifactorial disease causing aberrant DNA methylation (DNAm) and memory damage. Hypothesis: Models built for 5mC biomarker panel signatures may be useful for quantitative prediction of post-stroke long-term memory recovery in response to treatment. Methods: Stroke was induced with a middle cerebral artery occlusion (MCAO)(18 rats: sham(6), stroke(6), and stroke with treatment (6)). Xe-liposomes were intravenously administrated repetitively for 3 days after MCAO. At day 35, memory was assessed and scaled using a novel objective recognition test. Brain hippocampus and cortex tissue samples were prepared for whole-genome bisulfite sequencing (WGBS) to generate quantitative DNA methylation datasets in response to repetitive intermittent Xe administration. The resulting WGBS data were evaluated and quantitatively associated with memory recovery. Results: A post-stroke long-term memory outcome model was generated with scaled data using the following inputs: (1) good recovery, (2) moderate recovery, (3) edge, (4) light damage, and (5) severe damage. Differential DNAm 5mC analysis was performed comparing MCAO affected hemisphere with sham and Xe-treated samples. Initial 5mC differential analysis identified 6,175 (Cortex), and 9,037 (Hippocampus) CpG sites that were different between groups. Subsequently the LASSO algorithm was applied and selected 22 CpG sites that correlated with the long-term memory recovery scale. PCA analysis using the 22 CpGs demonstrated separation of memory outcomes. We have also built a predictive linear regression model of the 22 CpGs to assess memory loss or recovery in response to treatment. Analysis showed 7 out of 22 CpGs are conserved in humans (orthologs) for 30 bp regions. Conclusions: Epigenetic 5mC marker panels have the potential for quantitating post-stroke memory damages/recovery.

Novel Mechanisms of Acalabrutinib Resistance in Patients with Chronic Lymphocytic Leukemia By Whole Genome Methylome Sequencing

Novel Mechanisms of Acalabrutinib Resistance in Patients with Chronic Lymphocytic Leukemia By Whole Genome Methylome Sequencing

CD19-CAR T Cells Develop Exhaustion Epigenetic Programs during a Clinical Response

Background:CD19-CAR T-cell therapy has emerged as a curative approach for patients with relapsed/refractory B-cell malignancies, including lymphoma and acute lymphoblastic leukemia (ALL). However, many patients develop recurrent disease after initial responses. Therapeutic failure is most likely due to T cell extrinsic and intrinsic mechanisms. While CD19-negative relapse has emerged as a major extrinsic mechanism, T cell intrinsic mechanisms remain elusive. T-cell exhaustion is driven by epigenetic programs, however epigenetic changes in CAR T cells pre and post infusion have not been determined longitudinally. Thus, the goal of this study was to determine the epigenetic landscape of CD19-CAR T cells pre and post infusion as an initial step to elucidate intrinsic mechanisms that limit CAR T-cell effector functions in humans.Methods/Results:A longitudinal analysis of CD8+ CD19-CAR T cell epigenetic changes was performed by whole-genome DNA methylation profiling of CAR T cells during manufacturing and from peripheral blood mononuclear cells (PBMCs) of infused patients. DNA methylation profiling was performed on samples from 15 patients enrolled on our institutional, autologous CD19-CAR T cell therapy study (NCT03573700). CAR T cell expansion and persistence was determined by measuring vector copy numbers in the PBMCs of treated patients. B cell aplasia was tracked by monitoring B cell reconstitution as an additional measure of CD19-CAR T cell function. We had previously established novel exhaustion DNA methylation datasets that delineate between progenitor and fully exhausted T cells. These datasets served as a guide for stratifying our post-infusion CAR T cells along the exhaustion developmental trajectory. Lastly, publicly available single-cell transcriptional profiles of CD19-CAR T cells from patients were mined to validate expression of the transitory exhaustion signature.Our data show that CD19-CAR T cells lose repressive DNA methylation at effector loci (e.g. PRF1, TBET) while gaining methylation at genes (e.g. LEF1, TCF7) associated with memory potential. We confirmed that these epigenetic changes are coupled to endogenous human T cell effector and memory differentiation by cross-referencing our epigenetic data with publicly available transcriptional profiles for antigen-specific effector and long-lived memory CD8 T cells from individuals vaccinated for yellow fever. Furthermore, we show that CAR T cells were unable to mount an in vivo recall response after recovery of antigen-positive B cells or disease relapse. This observation, coupled to the fact that these patients' CAR T cells developed exhaustion-associated DNA methylation programs, further supports the broader conclusion that CAR T cells acquire stable epigenetic exhaustion programs that limit their protective capacity.Conclusion:Our data demonstrate for the first time that CAR T cell expansion during a clinical response is coupled to progressive restriction of gene regulatory programs that control T cell fate potential. Furthermore, we show that CD19-CAR T cells undergo exhaustion epigenetic programs that are coupled to an inability to mount a recall response in the presence of antigen-positive normal and malignant B cells. This work lays the foundation for future efforts aimed at improving the efficacy of cellular therapy by reversing epigenetic programs that are coupled to CAR T-cell exhaustion. DisclosuresTriplett: Miltenyi: Other: Travel, meeting registration. Gottschalk: Other: Other: patents and patent applications in the field of cancer cell and gene therapy ; Catamaran Bio: Consultancy; Novartis: Consultancy; Tidal: Consultancy; Immatics: Membership on an entity's Board of Directors or advisory committees; Tessa Therapeutics: Consultancy. Youngblood: ElevateBio: Consultancy; ElevateBio: Honoraria; AstraZeneca: Honoraria; Cell Signaling: Honoraria; Merck: Research Funding; other: Other: patents and patent applications in the field of cancer cell and gene therapy .

Blood Immune Cell Composition Associated with Obesity and Drug Repositioning Revealed by Epigenetic and Transcriptomic Conjoint Analysis.

This research was designed to analyze the composition of immune cells in obesity and identify novel and potent drugs for obesity management by epigenetic and transcriptomic conjoint analysis. DNA methylation data set (GSE166611) and mRNA expression microarray (GSE18897) were obtained from the Gene Expression Omnibus database. A total of 72 objects (35 obese samples and 37 controls) were included in the study. Immune cell composition analysis, drug repositioning, and gene set enrichment analysis (GSEA) were performed using CIBERSORT, connectivity map (CMap), and GSEA tools. Besides, we performed a single-cell RNA-seq of the immune cells from whole blood samples obtained from one obese patient and one healthy control. mRNA levels of drug target genes were analyzed by qPCR assay in blood samples from six patients and six healthy controls. Immune cell composition analysis found that CD8 + T cells and NK cells were significantly lower in the obese group. 11 drugs/compounds are considered to possess obesity-control potential, such as atorvastatin. Moreover, the expression of drug targets (STAT3, MCL1, PMAIP1, SOD2, FOX O 3, FOS, FKBP5) in obese patients were higher than those in controls. In conclusion, immune cells are potential therapeutic targets for obesity. Our results also contribute to accelerate research on drug development of obesity.