DNA Methylation Data Research Articles

Measuring cancer driving force of chromosomal aberrations through multi-layer Boolean implication networks.

Multi-layer Complex networks are commonly used for modeling and analysing biological entities. This paper presents the advantage of using COMBO (Combining Multi Bio Omics) to suggest a new role of the chromosomal aberration as a cancer driver factor. Exploiting the heterogeneous multi-layer networks, COMBO integrates gene expression and DNA-methylation data in order to identify complex bilateral relationships between transcriptome and epigenome. We evaluated the multi-layer networks generated by COMBO on different TCGA cancer datasets (COAD, BLCA, BRCA, CESC, STAD) focusing on the effect of a specific chromosomal numerical aberration, broad gain in chromosome 20, on different cancer histotypes. In addition, the effect of chromosome 8q amplification was tested in the same TCGA cancer dataset. The results demonstrate the ability of COMBO to identify the chromosome 20 amplification cancer driver force in the different TCGA Pan Cancer project datasets.

Nonlinear DNA methylation trajectories in aging male mice

Although DNA methylation data yields highly accurate age predictors, little is known about the dynamics of this quintessential epigenomic biomarker during lifespan. To narrow the gap, we investigate the methylation trajectories of male mouse colon at five different time points of aging. Our study indicates the existence of sudden hypermethylation events at specific stages of life. Precisely, we identify two epigenomic switches during early-to-midlife (3-9 months) and mid-to-late-life (15-24 months) transitions, separating the rodents’ life into three stages. These nonlinear methylation dynamics predominantly affect genes associated with the nervous system and enrich in bivalently marked chromatin regions. Based on groups of nonlinearly modified loci, we construct a clock-like classifier STageR (STage of aging estimatoR) that accurately predicts murine epigenetic stage. We demonstrate the universality of our clock in an independent mouse cohort and with publicly available datasets.

Neural-net-based cell deconvolution from DNA methylation reveals tumor microenvironment associated with cancer prognosis.

DNA methylation is a pivotal epigenetic modification that defines cellular identity. While cell deconvolution utilizing this information is considered useful for clinical practice, current methods for deconvolution are limited in their accuracy and resolution. In this study, we collected DNA methylation data from 945 human samples derived from various tissues and tumor-infiltrating immune cells and trained a neural network model with them. The model, termed MEnet, predicted abundance of cell population together with the detailed immune cell status from bulk DNA methylation data, and showed consistency to those of flow cytometry and histochemistry. MEnet was superior to the existing methods in the accuracy, speed, and detectable cell diversity, and could be applicable for peripheral blood, tumors, cell-free DNA, and formalin-fixed paraffin-embedded sections. Furthermore, by applying MEnet to 72 intrahepatic cholangiocarcinoma samples, we identified immune cell profiles associated with cancer prognosis. We believe that cell deconvolution by MEnet has the potential for use in clinical settings.

BCAT1, IKZF1 and SEPT9: methylated DNA biomarkers for detection of pan-gastrointestinal adenocarcinomas

Introduction Methylated circulating tumour DNA (ctDNA) blood tests for BCAT1/IKZF1 (COLVERA) and SEPT9 (Epi proColon) are used to detect colorectal cancer (CRC). However, there are no ctDNA assays approved for other gastrointestinal adenocarcinomas. We aimed to characterize BCAT1, IKZF1 and SEPT9 methylation in different gastrointestinal adenocarcinoma and non-gastrointestinal tumours to determine if these validated CRC biomarkers might be useful for pan-gastrointestinal adenocarcinoma detection. Methods Tissue DNA methylation data from colorectal (COAD, READ), gastroesophageal (ESCA, STAD), pancreatic (PAAD) and cholangiocarcinoma (CHOL) adenocarcinoma cohorts within The Cancer Genome Atlas were used for differential methylation analyses. Clinicodemographic predictors of BCAT1, IKZF1 and SEPT9 methylation, and the selectivity of hypermethylated BCAT1, IKZF1 and SEPT9 for colorectal adenocarcinomas in comparison to other cancers were each explored with beta regression. Results Hypermethylated BCAT1, IKZF1 and SEPT9 were each differentially methylated in colorectal and gastroesophageal adenocarcinomas. IKZF1 was differentially methylated in pancreatic adenocarcinoma. Hypermethylated DNA biomarkers BCAT1, IKZF1 and SEPT9 were largely stable across different stages of disease and were highly selective for gastrointestinal adenocarcinomas relative to other cancer types. Discussion Existing CRC methylated ctDNA blood tests for BCAT1/IKZF1 and SEPT9 might be usefully repurposed for use in other gastrointestinal adenocarcinomas and warrant further prospective ctDNA studies.

Abstract LB422: Specific activation of the Epstein-Barr Virus BGLF4 locus via lowered DNA methylation in EBV-associated CNS lymphoproliferative disease

Abstract Introduction: Epstein-Barr virus positive (EBV+) central nervous system lymphoproliferative diseases (CNS-LPD) are aggressive clinical conditions with poor prognosis. We have previously reported use of a ganciclovir (GCV), zidovudine (AZT), rituximab and dexamethasone regimen (GARD) that induced complete and durable responses in patients with primary CNS post-transplant lymphoproliferative disease (PTLD). Responses were associated with detection of two GCV/AZT viral drug targets, the EBV protein kinases BGLF4 and BXLF1. These viral proteins are normally expressed in lytic phase EBV, and the molecular basis for expression in latently infected EBV+ CNS-LPD is unknown. Methods: RNA expression of EBV genes LMP1, BZLF1, BXLF1, and BGLF4 was measured in 12 CNS-LPD biopsies (n=8/12 DLBCL-type) via qRT-PCR. Mass spectrometry-based EpiTYPER assay was used to quantify DNA methylation in EBV gene promoters. Biopsies from 25 patients with systemic PTLD were used for comparison. The BGLF4 locus was investigated in detail using luciferase reporters and 5’RACE mapping of transcription start sites in 3 EBV-infected cell lines and 4 CNS-LPD biopsy samples. Overall EBV methylation was assessed in a screening assay covering 31 EBV loci. Mechanistic involvement of active demethylation via TET2 was studied in a HEK293 CRISPR-Cas9 gene knockout model under EBV (strain M81) infection. Results: Gene expression analysis of biopsy specimens showed expression of LMP1, BXLF1, and BGLF4, but not BZLF1. Expression of BXLF1 and BGLF4 was significantly higher in CNS-LPD when compared to systemic EBV+ PTLD (p=0.043 and p=0.00022, respectively). EpiTYPER data showed significantly decreased promoter methylation at BGLF4 (p=0.00013). Luciferase reporter analysis of the BGLF4 upstream sequence revealed 3 regions of promoter activity which were silenced by in vitro methylation of reporters. 5’RACE identified transcription start sites at all 3 BGLF4 promoters and detected their activity in EBV+ cell lines and CNS-LPD biopsies. By overlaying transcriptional activity, 5’RACE and DNA methylation data, we identified DNA methylation loss at single CpG dinucleotides correlated with BGLF4 expression (r2=0.73). These sites were specifically demethylated in CNS-LPD, while surrounding EBV methylation remained high. Knockout of TET2 followed by EBV infection revealed that active demethylation is required for demethylation at BGLF4 promoters. Conclusions: The molecular determinant for success of the GARD regimen in CNS-LPD has remained poorly understood, with immunohistochemical detection of BGLF4/BXLF1 as the only predictor of treatment response. Our results provide the molecular basis for expression of BGLF4 in CNS-LPD by showing decreased promotor DNA methylation. We identified specific BGLF4 methylation sites that hold promise as potential DNA biomarkers for EBV+ CNS-LPD. Citation Format: Christoph Weigel, Haley L. Klimaszewski, Selamawit Addissie, Sarah Schlotter, Fode Tounkara, James P. Dugan, Bradley M. Haverkos, Lynda Villagomez, Mark Lustberg, Pierluigi Porcu, Timothy Voorhees, Michael A. Caligiuri, Ginny Bumgardner, Christopher C. Oakes, Robert Baiocchi. Specific activation of the Epstein-Barr Virus BGLF4 locus via lowered DNA methylation in EBV-associated CNS lymphoproliferative disease [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB422.

Upregulation of GPR133 expression impaired the phagocytosis of macrophages in recurrent spontaneous miscarriage

ABSTRACT Decidual macrophages are the second-largest immune cell group at the maternal-foetal interface. They participate in apoptotic cell removal, and protect the foetus from microorganisms or pathogens. Dysfunction of decidual macrophages gives rise to pregnancy complications such as preeclampsia and recurrent spontaneous miscarriage (RSM). However, the mechanisms by which decidual macrophages are involved in the occurrence of adverse pregnancy outcomes have not been elucidated. Here we integrated DNA methylation and gene expression data from decidua macrophages to identify potential risk factors related to RSM. GPR133 was significantly hypomethylated and upregulated in decidual macrophages from RSM patients. Further demethylation analysis demonstrated that GPR133 expression in decidual macrophages was significantly increased by 5-Aza-dC treatment. In addition, the influence of GPR133 on the phagocytic ability of macrophages was explored. Phagocytosis was impaired in the decidual macrophages of RSM patients with increased GPR133 expression. Increased GPR133 expression induced by demethylation treatment in the decidual macrophages of healthy control patients led to a significant decrease in phagocytic function. Importantly, knockdown of GPR133 resulted in a significant improvement in the phagocytic function of THP-1 macrophages. In conclusion, the existing studies have shown the influence of GPR133 on the phagocytic function of decidual macrophages and pregnancy outcomes, providing new data and ideas for future research on the role of decidual macrophages in RSM.

Unveiling Gene Interactions in Alzheimer's Disease by Integrating Genetic and Epigenetic Data with a Network-Based Approach.

Alzheimer's Disease (AD) is a complex disease and the leading cause of dementia in older people. We aimed to uncover aspects of AD's pathogenesis that may contribute to drug repurposing efforts by integrating DNA methylation and genetic data. Implementing the network-based tool, a dense module search of genome-wide association studies (dmGWAS), we integrated a large-scale GWAS dataset with DNA methylation data to identify gene network modules associated with AD. Our analysis yielded 286 significant gene network modules. Notably, the foremost module included the BIN1 gene, showing the largest GWAS signal, and the GNAS gene, the most significantly hypermethylated. We conducted Web-based Cell-type-Specific Enrichment Analysis (WebCSEA) on genes within the top 10% of dmGWAS modules, highlighting monocyte as the most significant cell type (p < 5 × 10-12). Functional enrichment analysis revealed Gene Ontology Biological Process terms relevant to AD pathology (adjusted p < 0.05). Additionally, drug target enrichment identified five FDA-approved targets (p-value = 0.03) for further research. In summary, dmGWAS integration of genetic and epigenetic signals unveiled new gene interactions related to AD, offering promising avenues for future studies.

Potential value of expression of receptor accessory protein 4 for evaluating the prognosis of lower-grade glioma patients.

REEP4 is involved in the regulation of the biological process of mitosis. Lower grade glioma (LGG), as a malignant tumor, is accompanied by abnormalities in mitosis, but there have been no reports of REEP4 so far. We collected transcriptome data, DNA methylation data and the clinical characteristics of thousands of patients with LGG. Various big data analysis methods and molecular biology experiments were employed to reveal the impact of REEP4 on the pathological process of LGG. It was found that the expression of REEP4 was significantly elevated and negatively regulated by its methylation site. Therefore, both the high expression of REEP4 and low methylation state of cg16311504 showed that the patients are correlated with lower patient survival rate. In addition, high REEP4 expression participates in the regulation of various cancer-related cellular signaling pathways, such as the cell cycle, MAPK signaling pathway, NOD-like receptor signaling pathway, etc. More importantly, the level of immune cell infiltration significantly increased in the high expression group of REEP4 in the LGG tumor microenvironment and REEP4 has a high positive correlation with PD-L1 and other immune checkpoints. In brief, this study is the first to introduce REEP4 in malignant tumors, which can be used as an independent risk factor that participates in the malignant process of LGG. More importantly, REEP4 has the potential to become a new star in the field of anti-tumor treatment.

Refining risk prediction in pediatric acute lymphoblastic leukemia through DNA methylation profiling

Acute lymphoblastic leukemia (ALL) is the most prevalent cancer in children, and despite considerable progress in treatment outcomes, relapses still pose significant risks of mortality and long-term complications. To address this challenge, we employed a supervised machine learning technique, specifically random survival forests, to predict the risk of relapse and mortality using array-based DNA methylation data from a cohort of 763 pediatric ALL patients treated in Nordic countries. The relapse risk predictor (RRP) was constructed based on 16 CpG sites, demonstrating c-indexes of 0.667 and 0.677 in the training and test sets, respectively. The mortality risk predictor (MRP), comprising 53 CpG sites, exhibited c-indexes of 0.751 and 0.754 in the training and test sets, respectively. To validate the prognostic value of the predictors, we further analyzed two independent cohorts of Canadian (n = 42) and Nordic (n = 384) ALL patients. The external validation confirmed our findings, with the RRP achieving a c-index of 0.667 in the Canadian cohort, and the RRP and MRP achieving c-indexes of 0.529 and 0.621, respectively, in an independent Nordic cohort. The precision of the RRP and MRP models improved when incorporating traditional risk group data, underscoring the potential for synergistic integration of clinical prognostic factors. The MRP model also enabled the definition of a risk group with high rates of relapse and mortality. Our results demonstrate the potential of DNA methylation as a prognostic factor and a tool to refine risk stratification in pediatric ALL. This may lead to personalized treatment strategies based on epigenetic profiling.

Inference of Developmental Hierarchy and Functional States of Exhausted T Cells from Epigenetic Profiles with Deep Learning.

Exhausted T cells are a key component of immune cells that play a crucial role in the immune response against cancer and influence the efficacy of immunotherapy. Accurate assessment and measurement of T-cell exhaustion (TEX) are critical for understanding the heterogeneity of TEX in the tumor microenvironment (TME) and tailoring individualized immunotherapeutic strategies. In this study, we introduced DeepEpiTEX, a novel computational framework based on deep neural networks, for inferring the developmental hierarchy and functional states of exhausted T cells in the TME from epigenetic profiles. DeepEpiTEX was trained using various modalities of epigenetic data, including DNA methylation data, microRNA expression data, and long non-coding RNA expression data from 30 bulk solid cancer types in the TCGA pan-cancer cohort, and identified five optimal TEX subsets with significant survival differences across the majority of cancer types. The performance of DeepEpiTEX was further evaluated and validated in external multi-center and multi-type cancer cohorts, consistently demonstrating its generalizability and applicability in different experimental settings. In addition, we discovered the potential relationship between TEX subsets identified by DeepEpiTEX and the response to immune checkpoint blockade therapy, indicating that individuals with immune-favorable TEX subsets may experience the greatest benefits. In conclusion, our study sheds light on the role of epigenetic regulation in TEX and provides a powerful and promising tool for categorizing TEX in different disease settings.

Computational deconvolution of DNA methylation data from mixed DNA samples.

In this review, we provide a comprehensive overview of the different computational tools that have been published for the deconvolution of bulk DNA methylation (DNAm) data. Here, deconvolution refers to the estimation of cell-type proportions that constitute a mixed sample. The paper reviews and compares 25 deconvolution methods (supervised, unsupervised or hybrid) developed between 2012 and 2023 and compares the strengths and limitations of each approach. Moreover, in this study, we describe the impact of the platform used for the generation of methylation data (including microarrays and sequencing), the applied data pre-processing steps and the used reference dataset on the deconvolution performance. Next to reference-based methods, we also examine methods that require only partial reference datasets or require no reference set at all. In this review, we provide guidelines for the use of specific methods dependent on the DNA methylation data type and data availability.

Abstract 7010: A pan-tissue investigation of epigenetic aging in association with genetic risk for cancer among European descendants

Abstract Epigenetic age (EA) is recognized as a valuable biomarker of aging. The discrepancy between epigenetic and chronological age (CA), termed epigenetic age acceleration (EAA), has been associated with increased mortality from various age-related diseases, including cancer. However, previous studies primarily focused on the EA in blood cells, resulting in a notable gap in understanding the EA in cancer-relevant tissues and its relationship to cancer development. To examine if genetic risks for seven distinct cancers are associated with EAA in cancer-relevant tissues and blood cells, we utilized Illumina EPIC-array-based DNA methylation data and whole genome sequencing (WGS)-derived genotype data of 659 normal tissue and 42 whole blood samples from 315 European-ancestry subjects in the Genotype-Tissue Expression (GTEx) project. EA was estimated across 30 female breast, 167 colon, 41 kidney, 163 lung, 118 ovary, 96 prostate, 44 testis, and 42 whole blood samples via seven established epigenetic clocks: HorvathAge, GrimAge, GrimAge2, HannumAge, PhenoAge, Zhang-BLUP-Age, and Zhang-EN-Age. EAA was calculated as the residual of EA regressed on CA. For each participant, a polygenic risk score (PRS) was computed to assess the genetic risk for each cancer, using validated cancer-specific PRSs for European populations from the Polygenic Score Catalogue. Linear regression analysis was performed to evaluate the association between cancer PRS and EAA, adjusting for body mass index, sex (where applicable), cigarette smoking, and WGS platform and library construction protocol. All EA estimates of cancer-relevant tissues and blood cells exhibited significant correlations with CA (maximum P=2.44 × 10−3). Among them, GrimAge, known for its strong predictive capacity for mortality, had the most pronounced correlation with CA, with Spearman coefficients ranging from 0.80 in blood cells to 0.98 in breast tissues. A higher prostate cancer PRS was significantly associated with increased EAA in normal prostate tissues measured by PhenoAge (β=2.49; P=0.01), Zhang-BLUP-Age (β=1.71; P=0.02), HannumAge (β=1.36; P=0.04), and GrimAge (β=0.73; P=0.04). In addition, a higher PRS for renal cell carcinoma showed a significant association with increased GrimAge acceleration in normal kidney tissues (β=0.82; P=0.02). Moreover, breast and lung cancer PRSs were significantly associated with increased EAA in blood cells. Specifically, breast cancer PRS was associated with Zhang-BLUP-Age acceleration (β=0.99; P=0.04), while lung cancer PRS showed associations with HannumAge (β=1.78; P=0.01) and Zhang-EN-Age acceleration (β=0.94; P=0.02). Our results suggest a potential connection between genetic susceptibility to certain cancers and accelerated aging in relevant tissues and/or blood cells before cancer onset. Future large-scale studies are needed to substantiate and expand upon our findings. Citation Format: Yaxin Chen, Yawen Qi, Gang Wang, Yufang Xie, Yaohua Yang, Weimin Li. A pan-tissue investigation of epigenetic aging in association with genetic risk for cancer among European descendants [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7010.

Abstract 1733: SET-NUP214 rearranges the DNA-methylation landscape to upregulate the HOX-gene cluster in acute myeloid leukemia

Abstract The SET-NUP214 (S/N) fusion gene, which results from either cryptic t(9:9) (q34;q34) or del(9) (q34.11q34.13), is found in a very narrow spectrum of patients diagnosed with Acute Myeloid Leukemia (AML) or T-cell acute lymphoblastic leukemia (T-ALL). AML patients with S/N fusion were reported to have relatively poor prognostic outcomes despite receiving myeloablative conditioning. While previously, the fusion has been described to overrepresent HOX cluster genes, the underlying mechanism of gene regulation that contributes to the pathology remains elusive due to the rare incidence of the disease. The current study used the MEGAL (ACC719, DSMZ) cell line, identified as the solo megakaryocytic AML line expressing the S/N-fusion, for determining genome-wide DNA methylation (DNAm) and gene expression using 935K Infinium MethylationEPIC v2.0 BeadChip array and RNA-sequencing, respectively. The differential DNAm and expression data were analyzed and compared to the hematopoietic stem progenitor cells (HSPC) or normal megakaryocytes (MK). The ChIP-sequencing data from the S/N positive T-ALL cell line LOUCY has been used to represent the coincidence of regulatory chromatin marks and differentially methylated CpG (mC) sites. We observed n=1934 genes upregulated and n=2396 downregulated in MEGAL, compared to both HSPC and MK. The direction of expression changes between MEGAL and HSPC and MEGAL and MK showed a high correlation (r=0.92, p&amp;lt; 0.01). Per the previous literature, we observed an overrepresentation of HOX genes (n=19) in MEGAL. Based on DNAm profiling, we noticed a propensity of hypermethylation in MEGAL across the promoters (93%±0.6), bodies (93%±0.8), and intergenic (85%±0.3) regions, compared to HSPC/MK. Next, we integrated the mC and expression at the promoters and identified that the majority (75%±3) of the genes belong to the hyper-down (n=2563 against HPSC; n=1907 against MK) or hyper-up (n=965 against HPSC; n=923 against MK) clusters. We further investigated the HOXB (HOXB7, HOXB8, HOXB9) and HOXC (HOXC10, HOXC12, HOXC13, HOXC13-AS, HOXC4, HOXC5, HOC6, HOXC8, HOXC9, HOXC-AS1, HOXC-AS2, and HOXC-AS3) cluster genes, where hypermethylated CpGs mainly were concentrated at the CpG islands and overlapped with active promoter (H3K4me3, H3K4me1) signatures or enhancer-like (CTCF bound) elements. We also observed frequent overlaps between mC and H3K36me3 across the observed genes. The differentially expressed genes in this AML subtype were enriched for the hippo-signaling pathway, calcium signaling, focal adhesion, and PI3-Akt pathways. Collectively, our data showed that S/N favors genome-wide hypermethylation, which, in conjunction with the cis-regulatory elements, induces overrepresentation of the HOX cluster genes. These novel epigenetic vulnerabilities are worth pursuing for investigating the cellular function and pathophysiology of the disease. Citation Format: Samrat Roy Choudhury, Akhilesh Kaushal. SET-NUP214 rearranges the DNA-methylation landscape to upregulate the HOX-gene cluster in acute myeloid leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 1733.

Abstract 6231: Multi-omic characterisation of clear cell renal carcinoma reveals endogenous and exogenous processes related to its aetiology

Abstract Endogenous and exogenous exposures are associated with distinctive molecular marks in somatic tissues, including human tumours. Integrative multi-omic approaches could help improving the resolution to detect the molecular marks associated with disease aetiology. Here, we leveraged the overlapping molecular information of clear cell renal cell carcinomas (ccRCC) tumours from Mutographs cohort by integrating DNA methylation, transcriptome, and whole-genome sequencing-based somatic mutation data to explore and gain new insights into exposures linked to ccRCC aetiology. Our novel framework identified important sources of biological variance across ccRCC tumours that were summarized as molecular components. We inferred these molecular ccRCC components into independent datasets to further explore their relationship with molecular and epidemiological features of ccRCC, normal adjacent kidney tissues, and other cancer types. Our results revealed a major ccRCC molecular component correlating with cellular mitotic age-related features, particularly the mitotic epigenetic clock (age-adjusted epiTOC2), clock-like DNA mutational signatures (SBS1/ID1), and telomere attrition. This molecular component was overrepresented in ccRCC tumours in comparison with normal paired kidney tissues, associating with PBRM1 and SETD2 somatic cancer driver mutations, genome stability, tumor stage, grade and ccRCC patient survival, independently to chronological age. Pan-cancer analysis supported the mitotic-age effect, represented by this molecular component, in other cancer types. Another ccRCC component was associated with tobacco usage, the presence of tobacco related DNA mutational and methylation signatures, increased total mutation burden, and sex. This component was also related to the epigenetic regulation of xenobiotic metabolism-related genes (e.g., GSTP1), further suggesting a relationship with genotoxic compounds. We further identified molecular components related to the ccRCC tumour microenvironment and cell proliferation, including one component related to BAP1 cancer driver mutations, pro-inflammatory immune cells, and ccRCC patient survival. In conclusion, our study reports a novel framework to molecular characterize ccRCC tumours using an integrative multi-omic approach, providing additional insights into the molecular footprints of endogenous and exogenous exposures in ccRCC tumours, including molecular components with prognostic value for patients. Citation Format: Ricardo Cortez Cardoso Penha, Alexandra Sexton Oates, Sergey Senkin, Han La Park, Nicolas Alcala, Matthieu Foll, Karl Smith-Byrne, Paul Brennan, James Mckay. Multi-omic characterisation of clear cell renal carcinoma reveals endogenous and exogenous processes related to its aetiology [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6231.

Abstract 3526: GAIN-BRCA: A graphical explainable AI-net framework for breast cancer subtype classification using multiomics data

Abstract Higher standards for precise distinction and classification among the four breast cancer subtypes, i.e., Luminal A. Luminal B, HER-2-enriched, and Basal, are essential to delivering customized treatment in precision oncology. The effective integration of readily accessible multi-omics datasets from the same patient may improve the accuracy of subtype prediction algorithms and help decipher reliable molecular features for the prediction and prognosis of breast cancer. Previous models for predicting breast cancer subtypes relied mostly on concatenation and autoencoder methods but rarely utilized graph-based integration to leverage the biological associations among omics datatypes. To overcome these limitations, we developed a graph-based multi-omics integration method using features from mRNA, DNA methylation, and miRNA data and synthesized features from their interactions. GAIN-BRCA computes weightage scores from miRNA-mRNA and DNA methylation-mRNA interactions to derive a new transformed feature vector without compromising biological information. The neural network architecture was designed and trained on the transformed features to classify breast cancer subtypes. Neural network prediction accuracies were compared using concatenated, autoencoder, and GAIN-BRCA-based integrated datasets. GAIN-BRCA outperforms the existing similar models MOGONET (Acc: 0.72) and moBRCA-net (Acc: 0.86) with an accuracy of 0.91. These transformed features were then processed using SHAP, and the top 357 features were selected for functional characterization and pathway analysis. We identified specific pathways and genes related to each breast cancer subtype separately. The most enriched pathways found in each subtype are ribonucleotide reductase signaling and glucocorticoid signaling pathways in luminal A, PFKB4 signaling pathway in luminal B, adipogenesis and DNA methylation and transcriptional signaling in HER2+, and the role of OCT4 in mammalian embryonic stem cell pluripotency in basal. Similarly, the most discriminated gene features are BRAF, ESR1, KRAS, MAPK1, SMARCE1, PPP3CA, and TGFBR2 in luminal A, TGFB3 and NCOA3 in luminal B, HNF4A, SIN3A, and SOX9 in HER2+, and POU5F1 and FOXA2 in basal subtypes. The GAIN-BRCA framework, combined with SHAP, weighs independent omics as cross-disciplinary rather than multidisciplinary and presents a distinctive picture for each subtype. As a result, the framework presents the importance of features about the corresponding subtype, which might lead to a greater understanding of the molecular basis of breast cancer subtypes. Citation Format: Jai Chand Patel, Sushil Kumar Shakyawar, Sahil Sethi, Chittibabu Guda. GAIN-BRCA: A graphical explainable AI-net framework for breast cancer subtype classification using multiomics data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3526.

Abstract 4772: Metabolic phenotype, age acceleration, and obesity-related cancer risk in Women's Health Initiative and Atherosclerosis Risk in Communities cohorts

Abstract Introduction: While normal physiologic aging causes metabolic decline, metabolic dysregulation owing to excess nutrient intake, physical inactivity, and obesity may drive accelerated aging. However, the relationship of accelerated aging with metabolic dysfunction across different levels of BMI (i.e., metabolic obesity phenotype) and whether it may mediate the relationship with obesity-related cancer (ORC) risk is not well established. Methods: To test this hypothesis, we are leveraging existing blood-based DNA methylation (DNAm) data within the Women’s Health Initiative (WHI) and the Atherosclerosis Risk in Communities (ARIC) cohorts. Weight, height and blood pressure were measured by trained technicians in the respective cohorts and fasting blood glucose, triglycerides and HDL cholesterol were quatified by standard clinican assays. Metabolic obesity phenotypes are classified as metabolically healthy normal weight (MHNW), metabolically unhealthy normal weight (MUNW), metabolically healthy overweight/obese (MHO), and metabolically unhealthy overweight/obese (MUO) when defined by ATP III criteria. Age acceleration was measured as a residual from regressing pan-tissue Horvath’s DNAmAge on chronological age. Results: The mean baseline age was 64 ± 6.6 years and 54 ± 5.7 year among WHI and ARIC participants, respectively and the age acceleration was measured at an mean follow-up time of 16 years from baseline for WHI and 3.5 years from baseline for ARIC cohort. The outcome of interest, ORC incidence was observed among 474 participants of which 387 were women. The median time for cancer diagnosis from baseline was 25 year and from DNAm measurement was 18.6 years. Of the 4341 (70% women) participants included in the analysis, 23.9% were MHNW, 3.2% MUNW, 39% MHO, and 33.5% MUO. A 31.5% of MHNW, 40.4% of MUNW, 28.7% of MHO, and 35.2% of MUO showed age acceleration. MUNW (OR 1.47, 95% CI: 1.03-2.11) and MUO (OR 1.18, 95% CI: 1.00-1.40) individuals at baseline had increased odds of experiencing age acceleration at a follow-up visit. The risk of all ORC combined was significantly elevated among MUO (HR 1.63, 95% CI: 1.27-2.09) compared to MHNW. Age acceleration at the follow-up visit was not significantly associated with ORC risk (HR 0.91, 95% CI: 0.75-1.11). These analyses were not adjusted for confounders. Conclusions: MUNW and MUO individuals had increased odds of age acceleration at a follow-up visit. Furthermore, MUO individuals were at an elevated risk of ORC, compared to individuals with MHNW. Accelerated aging does not appear to mediate this association, as accelerated aging was not associated with ORC. Citation Format: Prasoona Karra, Garnet L. Anderson, Mary C. Playdon, Sheetal Hardikar, Aladdin Shadyab, Kari E. North, Anne E. Justice, Elizabeth A. Platz, Jan Bressler, Anna Prizment, Weihua Guan, Dominique S. Michaud, Eric A. Whitsel, Brock C. Christensen, Lucas A. Salas. Metabolic phenotype, age acceleration, and obesity-related cancer risk in Women's Health Initiative and Atherosclerosis Risk in Communities cohorts [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4772.

Abstract 7029: Trajectory of DNA methylation in blood up to 18 years prior to lung cancer diagnosis

Abstract Purpose: Lung cancer (LC) has the highest mortality rate in South Korea. As prediction and early detection of lung cancer is pivotal to enhance survival rates, we attempted to investigate the trajectory of DNA methylation (DNAm) several years prior to lung cancer diagnosis by using prediagnostic blood samples. Methods: From National Cancer Center Health Examinee Cohort, established in 2003, we selected study participants who became diagnosed with LC several years after health examination, thereby collecting prediagnostic samples from 18 years to 1 year prior to LC diagnosis. Normal controls (n=150) were matched with patients (n=150) by sex, age, and the time of sample collection. Epigenome-wide DNAm data were obtained by using Infinium Human MethylationEPIC v2.0 Beadchip. We assumed non-linear relationship between DNAm changes and the time-to-diagnosis and estimated an optimal threshold point of non-linear relationship with a piecewise linear regression analysis at each CpG sites (CpGs). We used the ratio of M-values of patients and matched controls to establish the non-linear model composed of two linear models with a threshold point where the slope changes. Then, we performed functional enrichment analysis by each year prior to LC diagnosis using ReactomePA. Results: We identified threshold models with a threshold point where the slope changes from zero (from 18 years prior to LC up to the threshold point) to positive or negative (from threshold point to the point of LC diagnosis), which we termed “flat-to-change model”. Out of 790,048 CpGs, we identified a total of 7214 CpGs showing the “flat-to-change model”, of which 944, 789, 480, 413, 687, 542, 516, 1140, 315, 284, 327, 432, 83, 95, 74, 32, and 23 CpGs were found prior to 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, and 17 years before LC diagnosis, respectively. Functional enrichment analysis showed that various pathways were associated at different time point berfore LC diagnosis. CpGs of which DNAm was changed two years before LC diagnosis were associated with telomerase extension by telomerase involving ANKRD28/TERT/TERF2/PPP6R3/WRAP53. Transcriptional regulation by TP53 was associated with DNAm changes before both 5 and 14 years of LC diagnosis via POLR2A. Conclusions: We identified the trajectories of DNA methylation up to 18 years prior to LC diagnosis, and found that a greater number of CpG sites showed DNAm changes adjacent to the time point of LC diagnosis. CpG sites associated with specific time point before LC diagnosis may elucidate the mechanisms of tumorigenesis and also have potential for LC prediction markers. Citation Format: Yoon-Jung Choi, Ki-Jae Hong, Yeon-Su Lee, Woo Jin Kim, Yeol Kim, Sun Ha Jee, Jeongseon Kim. Trajectory of DNA methylation in blood up to 18 years prior to lung cancer diagnosis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7029.

Abstract 3458: DNA methylation-predicted blood protein levels and breast cancer risk

Abstract Background: Blood DNA methylation (DNAm) profiles can be used to predict a variety of human physiologic traits that are associated with breast cancer incidence and plasma concentrations of over 100 proteins, including those involved in inflammation, coagulation, metabolism, and growth. DNAm predictions of circulating proteins, here referred to as “DNAm proteins,” may be useful in providing molecular insights into breast cancer etiology and could be a valuable resource for personalizing breast cancer risk prediction. Methods: The Sister Study is a nationwide, prospective cohort of 50,884 women with a family history of breast cancer designed to identify novel environmental and biological risk factors for breast cancer. Epigenome-wide DNAm data were previously generated using the HumanMethylation450 or MethylationEPIC arrays on baseline blood samples collected from two case-cohort samples, for a combined sample of 4,479 women. This sample included 2,151 (48%) women who were selected because they were diagnosed with breast cancer (1,673 invasive cancer and 478 ductal carcinomas in situ) in the 15 years (mean: 7.9 years) following blood draw. Blood levels of 109 DNAm proteins were estimated from DNAm profiles using the method developed by Gadd et al. (eLife, 2022). Weighted Cox regression models were used to estimate associations between these DNAm proteins and incident breast cancer; associations were considered statistically significant at a False Discovery Rate (q) ≤ 0.10. Results: In age and platform adjusted models, higher breast cancer rates were associated with higher DNAm-predicted blood concentrations of RARRES2, IGFBP4, and CCL21 and lower predicted concentrations of F7, SELL, CXCL9, CD48, and IL19. The strongest associations were observed for RARRES2 (per 1-SD increase, HR: 1.24, 95% CI: 1.10, 1.40, P= 6 × 10−4, q= 0.03) and F7 (per 1-SD increase, HR: 0.82, 95% CI: 0.74, 0.91, P= 3 × 10−4, q= 0.03). Estimates were similar after adjustment for known breast cancer risk factors and excluding those diagnosed within two years of their blood draw. Associations appeared stronger for invasive breast cancer. Conclusions: Using novel DNAm-based estimates of blood proteins, we identified associations between circulating proteins and breast cancer incidence. IGFBP4 and RARRES2 have previously been reported to be higher in newly-diagnosed breast cancer patients. We also identify new potential protein biomarkers related to immune response (CCL21, SELL, CXCL9, CD48, and IL19). Citation Format: Jacob K. Kresovich, Brett Reid, Doratha A. Byrd, Katie M. O'Brien, Zongli Xu, Clarice R. Weinberg, Dale P. Sandler, Jack A. Taylor. DNA methylation-predicted blood protein levels and breast cancer risk [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3458.

Abstract 2235: Epigenetic inflammation scores for C-Reactive protein and interleukin-6 associate with adverse cardiometabolic and cardiac outcomes among survivors of childhood cancer

Abstract Background: Inflammaging, a chronic low-grade inflammation as individuals age, is measured by elevated levels of circulating markers of inflammation (e.g., C-reactive protein [CRP] and interleukin-6 [IL-6]) and is associated with age-related chronic health conditions (CHCs). We aim to evaluate epigenetic inflammation scores (EIS) for CRP and IL-6 and their associations with cardiometabolic and cardiac-based CHCs and epigenetic age acceleration (EAA) in childhood cancer survivors (CCS). Methods: Among 2,052 CCS (47.2% female; mean [SD] = 33.7 [9.3] years of age at blood draw) from the St Jude Lifetime Cohort, EIS of CRP and IL-6 were calculated as the weighted sum of the DNA methylation (DNAm) levels of CpGs that are associated with circulating levels of CRP or IL-6, where blood DNAm data were generated with Infinium EPIC BeadChip, and weights are regression coefficients from previously published EWAS studies. CHCs with a CTCAE (Common Terminology Criteria for Adverse Events) grade ≥2 were considered except for myocardial infarction (MI) where a minimum CTCAE grade of 3 was applied. Residuals of EIS were calculated by regressing EIS on age at blood draw and sex. A multivariable Cox proportional hazard regression model evaluated the associations between the EIS residuals and each CHC adjusting for sex and cancer treatments. The timeframe was years since age at blood draw. All participants alive at the last follow-up were censored. EAA was calculated as residuals from regressing epigenetic age (both GrimAge and PhenoAge) on age at blood draw. A linear regression model evaluated the association between the EIS residual and EAA adjusting for sex and treatments. Results: The incidence of cardiometabolic conditions (i.e., after the blood draw) included: abnormal glucose metabolism (AGM, 5.1%), hypertension (HTN, 16.6%), hypercholesterolemia (HCL, 3.9%), hypertriglyceridemia (HTG, 6.0%), and cardiomyopathy (CMP, 9.0%), MI (4.4%), and obesity (24.2%). CRP-EIS and IL6-EIS were correlated (r=0.62). The CRP-EIS residual was significantly associated with increased risk of AGM (Hazard Ratio [95% CI] = 1.33 [1.01-1.75], P=0.04) and CMP (1.44 [1.19-1.74], P=1.58 × 10−4) and suggestively associated with HTN (1.16 [0.99-1.36], P=0.07) and MI (1.55 [0.94-2.57], P=0.09). IL6-EIS residual was associated with CMP (1.25 [1.04-1.49], P=0.02). In addition, both CRP- and IL6-EIS were positively associated with PhenoAge-EAA (β = 0.53 and 0.50 years per SD increase of EIS) and GrimAge-EAA (β = 0.30 and 0.36 per SD increase of EIS). Conclusion: The CRP- and IL6-EIS are associated with increased risk of cardiometabolic and cardiac-based CHCs and EAA among CCS, suggesting that EIS may be used as clinical tools to identify survivors at higher risk of such CHCs and inflammation may be targeted for intervention to reduce the burden of CHCs. Citation Format: Qian Dong, Yan Chen, Xijun Zhang, John Easton, Heather Mulder, Emily Walker, Geoffrey Neale, Kyla Shelton, Stephanie B. Dixon, Jinghui Zhang, Gregory T. Armstrong, Melissa M. Hudson, Kirsten K. Ness, Zhaoming Wang. Epigenetic inflammation scores for C-Reactive protein and interleukin-6 associate with adverse cardiometabolic and cardiac outcomes among survivors of childhood cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2235.

Abstract 3496: EpigenPlot: A tool for the gene-level methylation analysis of colorectal tumors

Abstract Introduction. Aberrant DNA methylation patterns are frequent in colorectal adenocarcinoma (CRC) tissues. The investigation of this layer of epigenetic regulation holds the promise of a better understanding of tumour progression and the identification of novel biomarker candidates. Purpose. Our goal was to collect and analyse publicly available DNA methylation data of colorectal samples and create a user-friendly tool for exploring and visualizing the CRC methylome. Methods. We used Illumina Infinium Methylation Arrays, which is a popular platform for genome-wide methylation analysis of large cohorts. Gene Expression Omnibus database (GEO) and TCGA-COAD (The Cancer Genome Atlas Colon Adenocarcinoma) datasets including normal colorectal mucosa, adenoma, and adenocarcinoma samples were systematically collected and filtered. Differentially methylated regions were identified using the Kruskal-Wallis test. Genes with significant methylation changes were further analysed using ROC (Receiver Operator Characteristic) analysis. Finally, we established a web application using the shiny R package. Results. We assembled a database including methylation data of 2,586 samples. Adenoma and adenocarcinoma samples were globally hypomethylated. Hypermethylation (p &amp;lt; 0.05, Δβ &amp;gt;= 0.2) in adenocarcinoma samples was predominantly present in regions near proximal promoters. Out of the 74 previously proposed biomarkers examined, 50 had a cross-validated AUC of over 0.8. FDA-approved biomarkers BMP3, NDRG4, and SEPT9 have shown the highest performance in the TSS200 region (cvAUC = 0.8, 0.83, and 0.76, respectively). Genes with the highest performance were ITGA4 (5’UTR, first exon: cvAUC = 0.9), MDFI (5’UTR: cvAUC = 0.9), CNRIP1 (5’UTR, first exon: cvAUC = 0.89). The established web application can visualize methylation at CpG sites and gene regions, and KEGG pathway genes. The platform is available at https://epigenplot.com. Conclusion. We collected and analysed a sizeable database of colorectal tissue data. Gene regions with the best performance include previously proposed and novel diagnostic biomarker candidates. The web platform provides an tool for the efficient exploration of the assembled data. Citation Format: Dalma Müller, Balazs Gyorffy. EpigenPlot: A tool for the gene-level methylation analysis of colorectal tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3496.