Detailed immune profiling in pediatric Crohn's disease using methylation cytometry

DNA methylation has been extensively utilized to study epigenetic patterns across many diseases as well as to deconvolve blood cell type proportions. This study builds upon previous studies examining methylation patterns in paediatric patients with varying stages of Crohn's disease to extend the immune profiling of these patients using a novel deconvolution approach. Compared with control subjects, we observed significantly decreased levels of CD4 memory and naive, CD8 naive, and natural killer cells and elevated neutrophil levels in Crohn's disease. In addition, Crohn's patients had a significantly elevated neutrophil-to-lymphocyte ratio. Using an epigenome-wide association approach and adjusting for potential confounders, including cell type, we observed 397 differentially methylated CpG (DMC) sites associated with Crohn's disease. The top genetic pathway associated with the DMCs was the regulation of arginine metabolic processes which are involved in the regulation of T cells.

Molecular subtypes of high-grade serous ovarian cancer across racial groups and gene expression platforms

BACKGROUND

High-grade serous carcinoma (HGSC) gene expression subtypes are associated with differential survival. We characterized HGSC gene expression in Black individuals and considered whether gene expression differences by self-identified race may contribute to poorer HGSC survival among Black versus White individuals.

METHODS

We included newly generated RNA-Seq data from Black and White individuals, and array-based genotyping data from four existing studies of White and Japanese individuals. We used K-means clustering, a method with no predefined number of clusters or dataset-specific features, to assign subtypes. Cluster- and dataset-specific gene expression patterns were summarized by moderated t-scores. We compared cluster-specific gene expression patterns across datasets by calculating the correlation between the summarized vectors of moderated t-scores. Following mapping to The Cancer Genome Atlas (TCGA)-derived HGSC subtypes, we used Cox proportional hazards models to estimate subtype-specific survival by dataset.

RESULTS

Cluster-specific gene expression was similar across gene expression platforms and racial groups. Comparing the Black population to the White and Japanese populations, the immunoreactive subtype was more common (39% versus 23%-28%) and the differentiated subtype less common (7% versus 22%-31%). Patterns of subtype-specific survival were similar between the Black and White populations with RNA-Seq data; compared to mesenchymal cases, the risk of death was similar for proliferative and differentiated cases and suggestively lower for immunoreactive cases (Black population HR=0.79 [0.55, 1.13], White population HR=0.86 [0.62, 1.19]).

CONCLUSIONS

While the prevalence of HGSC subtypes varied by race, subtype-specific survival was similar.

IMPACT

HGSC subtypes can be consistently assigned across platforms and self-identified racial groups.

Associations in cell type-specific hydroxymethylation and transcriptional alterations of pediatric central nervous system tumors

Although intratumoral heterogeneity has been established in pediatric central nervous system tumors, epigenomic alterations at the cell type level have largely remained unresolved. To identify cell type-specific alterations to cytosine modifications in pediatric central nervous system tumors, we utilize a multi-omic approach that integrated bulk DNA cytosine modification data (methylation and hydroxymethylation) with both bulk and single-cell RNA-sequencing data. We demonstrate a large reduction in the scope of significantly differentially modified cytosines in tumors when accounting for tumor cell type composition. In the progenitor-like cell types of tumors, we identify a preponderance differential Cytosine-phosphate-Guanine site hydroxymethylation rather than methylation. Genes with differential hydroxymethylation, like histone deacetylase 4 and insulin-like growth factor 1 receptor, are associated with cell type-specific changes in gene expression in tumors. Our results highlight the importance of epigenomic alterations in the progenitor-like cell types and its role in cell type-specific transcriptional regulation in pediatric central nervous system tumors.

Deciphering the role of immune cell composition in epigenetic age acceleration: Insights from cell-type deconvolution applied to human blood epigenetic clocks

Aging is a significant risk factor for various human disorders, and DNA methylation clocks have emerged as powerful tools for estimating biological age and predicting health-related outcomes. Methylation data from blood DNA has been a focus of more recently developed DNA methylation clocks. However, the impact of immune cell composition on epigenetic age acceleration (EAA) remains unclear as only some clocks incorporate partial cell type composition information when analyzing EAA. We investigated associations of 12 immune cell types measured by cell-type deconvolution with EAA predicted by six widely-used DNA methylation clocks in data from >10,000 blood samples. We observed significant associations of immune cell composition with EAA for all six clocks tested. Across the clocks, nine or more of the 12 cell types tested exhibited significant associations with EAA. Higher memory lymphocyte subtype proportions were associated with increased EAA, and naïve lymphocyte subtypes were associated with decreased EAA. To demonstrate the potential confounding of EAA by immune cell composition, we applied EAA in rheumatoid arthritis. Our research maps immune cell type contributions to EAA in human blood and offers opportunities to adjust for immune cell composition in EAA studies to a significantly more granular level. Understanding associations of EAA with immune profiles has implications for the interpretation of epigenetic age and its relevance in aging and disease research. Our detailed map of immune cell type contributions serves as a resource for studies utilizing epigenetic clocks across diverse research fields, including aging-related diseases, precision medicine, and therapeutic interventions.

Immunological shifts during early-stage Parkinson's disease identified with DNA methylation data on longitudinally collected blood samples

Parkinson's disease (PD) is the second most common neurodegenerative disease in the United States. Decades before motor symptoms manifest, non-motor symptoms such as hyposmia and rapid eye movement (REM) sleep behavior disorder are highly predictive of PD. Previous immune profiling studies have identified alterations to the proportions of immune cells in the blood of clinically defined PD patients. However, it remains unclear if these phenotypes manifest before the clinical diagnosis of PD. We utilized longitudinal DNA methylation (DNAm) microarray data from the Parkinson's Progression Marker's Initiative (PPMI) to perform immune profiling in clinically defined PD and prodromal PD patients (Prod). We identified previously reported changes in neutrophil, monocyte, and T cell numbers in PD patients. Additionally, we noted previously unrecognized decreases in the naive B cell compartment in the defined PD and Prod patient group. Over time, we observed the proportion of innate immune cells in PD blood increased, but the proportion of adaptive immune cells decreased. We identified decreases in T and B cell subsets associated with REM sleep disturbances and early cognitive decline. Lastly, we identified increases in B memory cells associated with both genetic (LRRK2 genotype) and infectious (cytomegalovirus seropositivity) risk factors of PD. Our analysis shows that the peripheral immune system is dynamic as the disease progresses. The study provides a platform to understand how and when peripheral immune alterations occur in PD and whether intervention at particular stages may be therapeutically advantageous.

Tumor microenvironment deconvolution identifies cell-type-independent aberrant DNA methylation and gene expression in prostate cancer

BACKGROUND

Among men, prostate cancer (PCa) is the second most common cancer and the second leading cause of cancer death. Etiologic factors associated with both prostate carcinogenesis and somatic alterations in tumors are incompletely understood. While genetic variants associated with PCa have been identified, epigenetic alterations in PCa are relatively understudied. To date, DNA methylation (DNAm) and gene expression (GE) in PCa have been investigated; however, these studies did not correct for cell-type proportions of the tumor microenvironment (TME), which could confound results.

METHODS

The data (GSE183040) consisted of DNAm and GE data from both tumor and adjacent non-tumor prostate tissue of 56 patients who underwent radical prostatectomies prior to any treatment. This study builds upon previous studies that examined methylation patterns and GE in PCa patients by using a novel tumor deconvolution approach to identify and correct for cell-type proportions of the TME in its epigenome-wide association study (EWAS) and differential expression analysis (DEA).

RESULTS

The inclusion of cell-type proportions in EWASs and DEAs reduced the scope of significant alterations associated with PCa. We identified 2,093 significantly differentially methylated CpGs (DMC), and 51 genes associated with PCa, including PCA3, SPINK1, and AMACR.

CONCLUSIONS

This work illustrates the importance of correcting for cell types of the TME when performing EWASs and DEAs on PCa samples, and establishes a more confounding-adverse methodology. We identified a more tumor-cell-specific set of altered genes and epigenetic marks that can be further investigated as potential biomarkers of disease or potential therapeutic targets.

Matched analysis of detailed peripheral blood and tumor immune microenvironment profiles in bladder cancer

Background: Bladder cancer and therapy responses hinge on immune profiles in the tumor microenvironment (TME) and blood, yet studies linking tumor-infiltrating immune cells to peripheral immune profiles are limited. Methods: DNA methylation cytometry quantified TME and matched peripheral blood immune cell proportions. With tumor immune profile data as the input, subjects were grouped by immune infiltration status and consensus clustering. Results: Immune hot and cold groups had different immune compositions in the TME but not in circulating blood. Two clusters of patients identified with consensus clustering had different immune compositions not only in the TME but also in blood. Conclusion: Detailed immune profiling via methylation cytometry reveals the significance of understanding tumor and systemic immune relationships in cancer patients.

Potential to Enhance Large Scale Molecular Assessments of Skin Photoaging through Virtual Inference of Spatial Transcriptomics from Routine Staining

The advent of spatial transcriptomics technologies has heralded a renaissance in research to advance our understanding of the spatial cellular and transcriptional heterogeneity within tissues. Spatial transcriptomics allows investigation of the interplay between cells, molecular pathways, and the surrounding tissue architecture and can help elucidate developmental trajectories, disease pathogenesis, and various niches in the tumor microenvironment. Photoaging is the histological and molecular skin damage resulting from chronic/acute sun exposure and is a major risk factor for skin cancer. Spatial transcriptomics technologies hold promise for improving the reliability of evaluating photoaging and developing new therapeutics. Challenges to current methods include limited focus on dermal elastosis variations and reliance on self-reported measures, which can introduce subjectivity and inconsistency. Spatial transcriptomics offers an opportunity to assess photoaging objectively and reproducibly in studies of carcinogenesis and discern the effectiveness of therapies that intervene in photoaging and preventing cancer. Evaluation of distinct histological architectures using highly-multiplexed spatial technologies can identify specific cell lineages that have been understudied due to their location beyond the depth of UV penetration. However, the cost and interpatient variability using state-of-the-art assays such as the 10x Genomics Spatial Transcriptomics assays limits the scope and scale of large-scale molecular epidemiologic studies. Here, we investigate the inference of spatial transcriptomics information from routine hematoxylin and eosin-stained (H&E) tissue slides. We employed the Visium CytAssist spatial transcriptomics assay to analyze over 18,000 genes at a 50-micron resolution for four patients from a cohort of 261 skin specimens collected adjacent to surgical resection sites for basal cell and squamous cell keratinocyte tumors. The spatial transcriptomics data was co-registered with 40x resolution whole slide imaging (WSI) information. We developed machine learning models that achieved a macro-averaged median AUC and F1 score of 0.80 and 0.61 and Spearman coefficient of 0.60 in inferring transcriptomic profiles across the slides, and accurately captured biological pathways across various tissue architectures.

Spatial Omics Driven Crossmodal Pretraining Applied to Graph-based Deep Learning for Cancer Pathology Analysis

Graph-based deep learning has shown great promise in cancer histopathology image analysis by contextualizing complex morphology and structure across whole slide images to make high quality downstream outcome predictions (ex: prognostication). These methods rely on informative representations (i.e., embeddings) of image patches comprising larger slides, which are used as node attributes in slide graphs. Spatial omics data, including spatial transcriptomics, is a novel paradigm offering a wealth of detailed information. Pairing this data with corresponding histological imaging localized at 50-micron resolution, may facilitate the development of algorithms which better appreciate the morphological and molecular underpinnings of carcinogenesis. Here, we explore the utility of leveraging spatial transcriptomics data with a contrastive crossmodal pretraining mechanism to generate deep learning models that can extract molecular and histological information for graph-based learning tasks. Performance on cancer staging, lymph node metastasis prediction, survival prediction, and tissue clustering analyses indicate that the proposed methods bring improvement to graph based deep learning models for histopathological slides compared to leveraging histological information from existing schemes, demonstrating the promise of mining spatial omics data to enhance deep learning for pathology workflows.

Enrichment of a neutrophil-like monocyte transcriptional state in glioblastoma myeloid suppressor cells

Glioblastomas (GBM) are lethal central nervous system cancers associated with tumor and systemic immunosuppression. Heterogeneous monocyte myeloid-derived suppressor cells (M-MDSC) are implicated in the altered immune response in GBM, but M-MDSC ontogeny and definitive phenotypic markers are unknown. Using single-cell transcriptomics, we revealed heterogeneity in blood M-MDSC from GBM subjects and an enrichment in a transcriptional state reminiscent of neutrophil-like monocytes (NeuMo), a newly described pathway of monopoiesis in mice. Human NeuMo gene expression and Neu-like deconvolution fraction algorithms were created to quantitate the enrichment of this transcriptional state in GBM subjects. NeuMo populations were also observed in M-MDSCs from lung and head and neck cancer subjects. Dexamethasone (DEX) and prednisone exposures increased the usage of Neu-like states, which were inversely associated with tumor purity and survival in isocitrate dehydrogenase wildtype (IDH WT) gliomas. Anti-inflammatory ZC3HA12/Regnase-1 transcripts were highly correlated with NeuMo expression in tumors and in blood M-MDSC from GBM, lung, and head and neck cancer subjects. Additional novel transcripts of immune-modulating proteins were identified. Collectively, these findings provide a framework for understanding the heterogeneity of M-MDSCs in GBM as cells with different clonal histories and may reshape approaches to study and therapeutically target these cells.

Molecular subtypes of high-grade serous ovarian cancer across racial groups and gene expression platforms

Introduction

High-grade serous carcinoma (HGSC) gene expression subtypes are associated with differential survival. We characterized HGSC gene expression in Black individuals and considered whether gene expression differences by race may contribute to poorer HGSC survival among Black versus non-Hispanic White individuals.

Methods

We included newly generated RNA-Seq data from Black and White individuals, and array-based genotyping data from four existing studies of White and Japanese individuals. We assigned subtypes using K-means clustering. Cluster- and dataset-specific gene expression patterns were summarized by moderated t-scores. We compared cluster-specific gene expression patterns across datasets by calculating the correlation between the summarized vectors of moderated t-scores. Following mapping to The Cancer Genome Atlas (TCGA)-derived HGSC subtypes, we used Cox proportional hazards models to estimate subtype-specific survival by dataset.

Results

Cluster-specific gene expression was similar across gene expression platforms. Comparing the Black study population to the White and Japanese study populations, the immunoreactive subtype was more common (39% versus 23%-28%) and the differentiated subtype less common (7% versus 22%-31%). Patterns of subtype-specific survival were similar between the Black and White populations with RNA-Seq data; compared to mesenchymal cases, the risk of death was similar for proliferative and differentiated cases and suggestively lower for immunoreactive cases (Black population HR=0.79 [0.55, 1.13], White population HR=0.86 [0.62, 1.19]).

Conclusions

A single, platform-agnostic pipeline can be used to assign HGSC gene expression subtypes. While the observed prevalence of HGSC subtypes varied by race, subtype-specific survival was similar.

Feasibility of Inferring Spatial Transcriptomics from Single-Cell Histological Patterns for Studying Colon Cancer Tumor Heterogeneity

Background

Spatial transcriptomics involves studying the spatial organization of gene expression within tissues, offering insights into the molecular diversity of tumors. While spatial gene expression is commonly amalgamated from 1-10 cells across 50-micron spots, recent methods have demonstrated the capability to disaggregate this information at subspot resolution by leveraging both expression and histological patterns. However, elucidating such information from histology alone presents a significant challenge but if solved can better permit spatial molecular analysis at cellular resolution for instances where Visium data is not available, reducing study costs. This study explores integrating single-cell histological and transcriptomic data to infer spatial mRNA expression patterns in whole slide images collected from a cohort of stage pT3 colorectal cancer patients. A cell graph neural network algorithm was developed to align histological information extracted from detected cells with single cell RNA patterns through optimal transport methods, facilitating the analysis of cellular groupings and gene relationships. This approach leveraged spot-level expression as an intermediary to co-map histological and transcriptomic information at the single-cell level.

Results

Our study demonstrated that single-cell transcriptional heterogeneity within a spot could be predicted from histological markers extracted from cells detected within a spot. Furthermore, our model exhibited proficiency in delineating overarching gene expression patterns across whole-slide images. This approach compared favorably to traditional patch-based computer vision methods as well as other methods which did not incorporate single cell expression during the model fitting procedures. Topological nuances of single-cell expression within a Visium spot were preserved using the developed methodology.

Conclusion

This innovative approach augments the resolution of spatial molecular assays utilizing histology as a sole input through synergistic co-mapping of histological and transcriptomic datasets at the single-cell level, anchored by spatial transcriptomics. While initial results are promising, they warrant rigorous validation. This includes collaborating with pathologists for precise spatial identification of distinct cell types and utilizing sophisticated assays, such as Xenium, to attain deeper subcellular insights.

The Overlooked Role of Specimen Preparation in Bolstering Deep Learning-Enhanced Spatial Transcriptomics Workflows

The application of deep learning methods to spatial transcriptomics has shown promise in unraveling the complex relationships between gene expression patterns and tissue architecture as they pertain to various pathological conditions. Deep learning methods that can infer gene expression patterns directly from tissue histomorphology can expand the capability to discern spatial molecular markers within tissue slides. However, current methods utilizing these techniques are plagued by substantial variability in tissue preparation and characteristics, which can hinder the broader adoption of these tools. Furthermore, training deep learning models using spatial transcriptomics on small study cohorts remains a costly endeavor. Necessitating novel tissue preparation processes enhance assay reliability, resolution, and scalability. This study investigated the impact of an enhanced specimen processing workflow for facilitating a deep learning-based spatial transcriptomics assessment. The enhanced workflow leveraged the flexibility of the Visium CytAssist assay to permit automated H&E staining (e.g., Leica Bond) of tissue slides, whole-slide imaging at 40x-resolution, and multiplexing of tissue sections from multiple patients within individual capture areas for spatial transcriptomics profiling. Using a cohort of thirteen pT3 stage colorectal cancer (CRC) patients, we compared the efficacy of deep learning models trained on slide prepared using an enhanced workflow as compared to the traditional workflow which leverages manual tissue staining and standard imaging of tissue slides. Leveraging Inceptionv3 neural networks, we aimed to predict gene expression patterns across matched serial tissue sections, each stemming from a distinct workflow but aligned based on persistent histological structures. Findings indicate that the enhanced workflow considerably outperformed the traditional spatial transcriptomics workflow. Gene expression profiles predicted from enhanced tissue slides also yielded expression patterns more topologically consistent with the ground truth. This led to enhanced statistical precision in pinpointing biomarkers associated with distinct spatial structures. These insights can potentially elevate diagnostic and prognostic biomarker detection by broadening the range of spatial molecular markers linked to metastasis and recurrence. Future endeavors will further explore these findings to enrich our comprehension of various diseases and uncover molecular pathways with greater nuance. Combining deep learning with spatial transcriptomics provides a compelling avenue to enrich our understanding of tumor biology and improve clinical outcomes. For results of the highest fidelity, however, effective specimen processing is crucial, and fostering collaboration between histotechnicians, pathologists, and genomics specialists is essential to herald this new era in spatial transcriptomics-driven cancer research.

Genome-Scale Methylation Analysis Identifies Immune Profiles and Age Acceleration Associations with Bladder Cancer Outcomes

BACKGROUND

Immune profiles have been associated with bladder cancer outcomes and may have clinical applications for prognosis. However, associations of detailed immune cell subtypes with patient outcomes remain underexplored and may contribute crucial prognostic information for better managing bladder cancer recurrence and survival.

METHODS

Bladder cancer case peripheral blood DNA methylation was measured using the Illumina HumanMethylationEPIC array. Extended cell-type deconvolution quantified 12 immune cell-type proportions, including memory, naïve T and B cells, and granulocyte subtypes. DNA methylation clocks determined biological age. Cox proportional hazards models tested associations of immune cell profiles and age acceleration with bladder cancer outcomes. The partDSA algorithm discriminated 10-year overall survival groups from clinical variables and immune cell profiles, and a semi-supervised recursively partitioned mixture model (SS-RPMM) with DNA methylation data was applied to identify a classifier for 10-year overall survival.

RESULTS

Higher CD8T memory cell proportions were associated with better overall survival [HR = 0.95, 95% confidence interval (CI) = 0.93-0.98], while higher neutrophil-to-lymphocyte ratio (HR = 1.36, 95% CI = 1.23-1.50), CD8T naïve (HR = 1.21, 95% CI = 1.04-1.41), neutrophil (HR = 1.04, 95% CI = 1.03-1.06) proportions, and age acceleration (HR = 1.06, 95% CI = 1.03-1.08) were associated with worse overall survival in patient with bladder cancer. partDSA and SS-RPMM classified five groups of subjects with significant differences in overall survival.

CONCLUSIONS

We identified associations between immune cell subtypes and age acceleration with bladder cancer outcomes.

IMPACT

The findings of this study suggest that bladder cancer outcomes are associated with specific methylation-derived immune cell-type proportions and age acceleration, and these factors could be potential prognostic biomarkers.

Analysis of DNA methylation at birth and in childhood reveals changes associated with season of birth and latitude

BACKGROUND

Seasonal variations in environmental exposures at birth or during gestation are associated with numerous adult traits and health outcomes later in life. Whether DNA methylation (DNAm) plays a role in the molecular mechanisms underlying the associations between birth season and lifelong phenotypes remains unclear.

METHODS

We carried out epigenome-wide meta-analyses within the Pregnancy And Childhood Epigenetic Consortium to identify associations of DNAm with birth season, both at differentially methylated probes (DMPs) and regions (DMRs). Associations were examined at two time points: at birth (21 cohorts, N = 9358) and in children aged 1-11 years (12 cohorts, N = 3610). We conducted meta-analyses to assess the impact of latitude on birth season-specific associations at both time points.

RESULTS

We identified associations between birth season and DNAm (False Discovery Rate-adjusted p values < 0.05) at two CpGs at birth (winter-born) and four in the childhood (summer-born) analyses when compared to children born in autumn. Furthermore, we identified twenty-six differentially methylated regions (DMR) at birth (winter-born: 8, spring-born: 15, summer-born: 3) and thirty-two in childhood (winter-born: 12, spring and summer: 10 each) meta-analyses with few overlapping DMRs between the birth seasons or the two time points. The DMRs were associated with genes of known functions in tumorigenesis, psychiatric/neurological disorders, inflammation, or immunity, amongst others. Latitude-stratified meta-analyses [higher (≥ 50°N), lower (< 50°N, northern hemisphere only)] revealed differences in associations between birth season and DNAm by birth latitude. DMR analysis implicated genes with previously reported links to schizophrenia (LAX1), skin disorders (PSORS1C, LTB4R), and airway inflammation including asthma (LTB4R), present only at birth in the higher latitudes (≥ 50°N).

CONCLUSIONS

In this large epigenome-wide meta-analysis study, we provide evidence for (i) associations between DNAm and season of birth that are unique for the seasons of the year (temporal effect) and (ii) latitude-dependent variations in the seasonal associations (spatial effect). DNAm could play a role in the molecular mechanisms underlying the effect of birth season on adult health outcomes.

Assessment of emerging pretraining strategies in interpretable multimodal deep learning for cancer prognostication

BACKGROUND

Deep learning models can infer cancer patient prognosis from molecular and anatomic pathology information. Recent studies that leveraged information from complementary multimodal data improved prognostication, further illustrating the potential utility of such methods. However, current approaches: 1) do not comprehensively leverage biological and histomorphological relationships and 2) make use of emerging strategies to "pretrain" models (i.e., train models on a slightly orthogonal dataset/modeling objective) which may aid prognostication by reducing the amount of information required for achieving optimal performance. In addition, model interpretation is crucial for facilitating the clinical adoption of deep learning methods by fostering practitioner understanding and trust in the technology.

METHODS

Here, we develop an interpretable multimodal modeling framework that combines DNA methylation, gene expression, and histopathology (i.e., tissue slides) data, and we compare performance of crossmodal pretraining, contrastive learning, and transfer learning versus the standard procedure.

RESULTS

Our models outperform the existing state-of-the-art method (average 11.54% C-index increase), and baseline clinically driven models (average 11.7% C-index increase). Model interpretations elucidate consideration of biologically meaningful factors in making prognosis predictions.

DISCUSSION

Our results demonstrate that the selection of pretraining strategies is crucial for obtaining highly accurate prognostication models, even more so than devising an innovative model architecture, and further emphasize the all-important role of the tumor microenvironment on disease progression.

Normal gastric tissue Helicobacter pylori infection is associated with epigenetic age acceleration, increased mitotic tick rate, tissue cell composition, and Natural Killer cell methylation alterations

Background

Gastric adenocarcinomas are a leading cause of global mortality, associated with chronic infection with Helicobacter pylori. The mechanisms by which infection with H. pylori contributes to carcinogenesis are not well understood. Recent studies from subjects with and without gastric cancer have identified significant DNA methylation alterations in normal gastric mucosa associated with H. pylori infection and gastric cancer risk. Here we further investigated DNA methylation alterations in normal gastric mucosa in gastric cancer cases (n = 42) and control subjects (n = 42) with H. pylori infection data. We assessed tissue cell type composition, DNA methylation alterations within cell populations, epigenetic aging, and repetitive element methylation.

Results

In normal gastric mucosa of both gastric cancer cases and control subjects, we observed increased epigenetic age acceleration associated with H. pylori infection. We also observed an increased mitotic tick rate associated with H. pylori infection in both gastric cancer cases and controls. Significant differences in immune cell populations associated with H. pylori infection in normal tissue from cancer cases and controls were identified using DNA methylation cell type deconvolution. We also found natural killer cell-specific methylation alterations in normal mucosa from gastric cancer patients with H. pylori infection.

Conclusions

Our findings from normal gastric mucosa provide insight into underlying cellular composition and epigenetic aspects of H. pylori associated gastric cancer etiology.

Hierarchical deconvolution for extensive cell type resolution in the human brain using DNA methylation

Introduction

The human brain comprises heterogeneous cell types whose composition can be altered with physiological and pathological conditions. New approaches to discern the diversity and distribution of brain cells associated with neurological conditions would significantly advance the study of brain-related pathophysiology and neuroscience. Unlike single-nuclei approaches, DNA methylation-based deconvolution does not require special sample handling or processing, is cost-effective, and easily scales to large study designs. Existing DNA methylation-based methods for brain cell deconvolution are limited in the number of cell types deconvolved.

Methods

Using DNA methylation profiles of the top cell-type-specific differentially methylated CpGs, we employed a hierarchical modeling approach to deconvolve GABAergic neurons, glutamatergic neurons, astrocytes, microglial cells, oligodendrocytes, endothelial cells, and stromal cells.

Results

We demonstrate the utility of our method by applying it to data on normal tissues from various brain regions and in aging and diseased tissues, including Alzheimer's disease, autism, Huntington's disease, epilepsy, and schizophrenia.

Discussion

We expect that the ability to determine the cellular composition in the brain using only DNA from bulk samples will accelerate understanding brain cell type composition and cell-type-specific epigenetic states in normal and diseased brain tissues.

Identification of Spatial Proteomic Signatures of Colon Tumor Metastasis: A Digital Spatial Profiling Approach

Over 150,000 Americans are diagnosed with colorectal cancer (CRC) every year, and annually >50,000 individuals are estimated to die of CRC, necessitating improvements in screening, prognostication, disease management, and therapeutic options. CRC tumors are removed en bloc with surrounding vasculature and lymphatics. Examination of regional lymph nodes at the time of surgical resection is essential for prognostication. Developing alternative approaches to indirectly assess recurrence risk would have utility in cases where lymph node yield is incomplete or inadequate. Spatially dependent, immune cell-specific (eg, tumor-infiltrating lymphocytes), proteomic, and transcriptomic expression patterns inside and around the tumor-the tumor immune microenvironment-can predict nodal/distant metastasis and probe the coordinated immune response from the primary tumor site. The comprehensive characterization of tumor-infiltrating lymphocytes and other immune infiltrates is possible using highly multiplexed spatial omics technologies, such as the GeoMX Digital Spatial Profiler. In this study, machine learning and differential co-expression analyses helped identify biomarkers from Digital Spatial Profiler-assayed protein expression patterns inside, at the invasive margin, and away from the tumor, associated with extracellular matrix remodeling (eg, granzyme B and fibronectin), immune suppression (eg, forkhead box P3), exhaustion and cytotoxicity (eg, CD8), Programmed death ligand 1-expressing dendritic cells, and neutrophil proliferation, among other concomitant alterations. Further investigation of these biomarkers may reveal independent risk factors of CRC metastasis that can be formulated into low-cost, widely available assays.

Assessment of immune cell profiles among post-menopausal women in the Women's Health Initiative using DNA methylation-based methods

BACKGROUND

Over the past decade, DNA methylation (DNAm)-based deconvolution methods that leverage cell-specific DNAm markers of immune cell types have been developed to provide accurate estimates of the proportions of leukocytes in peripheral blood. Immune cell phenotyping using DNAm markers, termed immunomethylomics or methylation cytometry, offers a solution for determining the body's immune cell landscape that does not require fresh blood and is scalable to large sample sizes. Despite significant advances in DNAm-based deconvolution, references at the population level are needed for clinical and research interpretation of these additional immune layers. Here we aim to provide some references for immune populations in a group of multi-ethnic post-menopausal American women.

RESULTS

We applied DNAm-based deconvolution to a large sample of post-menopausal women enrolled in the Women's Health Initiative (baseline, N = 58) or the ancillary Long Life Study (WHI-LLS, N = 1237) to determine the reference ranges of 58 immune parameters, including proportions and absolute counts for 19 leukocyte subsets and 20 derived cell ratios. Participants were 50-94 years old at the time of blood draw, and N = 898 (69.3%) self-identified as White. Using linear regression models, we observed significant associations between age at blood draw and absolute counts and proportions of naïve B, memory CD4+, naïve CD4+, naïve CD8+, memory CD8+ memory, neutrophils, and natural killer cells. We also assessed the same immune profiles in a subset of paired longitudinal samples collected 14-18 years apart across N = 52 participants. Our results demonstrate high inter-individual variability in rates of change of leukocyte subsets over this time. And, when conducting paired t tests to test the difference in counts and proportions between the baseline visit and LLS visit, there were significant changes in naïve B, memory CD4+, naïve CD4+, naïve CD8+, memory CD8+ cells and neutrophils, similar to the results seen when analyzing the association with age in the entire cohort.

CONCLUSIONS

Here, we show that derived cell counts largely reflect the immune profile associated with proportions and that these novel methods replicate the known immune profiles associated with age. Further, we demonstrate the value this methylation cytometry approach can add as a potential application in epidemiological studies.

Down Syndrome Altered Cell Composition in Blood, Brain, and Buccal Swab Samples Profiled by DNA-Methylation-Based Cell-Type Deconvolution

Down syndrome (DS) is a genetic disorder caused by an extra copy of chromosome 21 that presents developmental dysfunction and intellectual disability. To better understand the cellular changes associated with DS, we investigated the cell composition in blood, brain, and buccal swab samples from DS patients and controls using DNA methylation-based cell-type deconvolution. We used genome-scale DNA methylation data from Illumina HumanMethylation450k and HumanMethylationEPIC arrays to profile cell composition and trace fetal lineage cells in blood samples (DS N = 46; control N = 1469), brain samples from various regions (DS N = 71; control N = 101), and buccal swab samples (DS N = 10; control N = 10). In early development, the number of cells from the fetal lineage in the blood is drastically lower in DS patients (Δ = 17.5%), indicating an epigenetically dysregulated maturation process for DS patients. Across sample types, we observed significant alterations in relative cell-type proportions for DS subjects compared with the controls. Cell-type proportion alterations were present in samples from early development and adulthood. Our findings provide insight into DS cellular biology and suggest potential cellular interventional targets for DS.

Abstract 6675: Integration of associations of immune profiles in peripheral blood and tumor microenvironment with bladder cancer outcomes

Immune cell profiles in peripheral blood have been associated with bladder cancer outcomes, however, their association with response to immunotherapy and the tumor microenvironment is a major unresolved issue. Although tumor growth can be attenuated via the activation of tumor-infiltrating effector T cells, the relationship between tumor infiltration and immune activation remains unclear. This study explored the interaction between bladder cancer outcomes and immune profiles within peripheral blood and a tumor microenvironment (TME) based on DNA methylation profiles. Peripheral blood and the matched tumor FFPE DNA methylation profiles of 60 non-muscle-invasive bladder cancer (NMIBC) and 12 muscle-invasive bladder cancer (MIBC) patients. Cell-type deconvolution approaches were applied to estimate 12 peripheral immune cell-type proportions and 17 cell-type proportions within TME. We found a positive correlation between dendritic cell proportions in the TME with peripheral CD8T memory cell proportions (r = 0.35, P = 0.003) and a negative correlation between dendritic cell proportions in the TME with peripheral regulatory T cell proportions (r = -0.28, P = 0.021). In addition, monocyte cell proportions in TME had a positive correlation with peripheral B memory (r = 0.37, P = 0.002) and CD8T memory cell proportions (r = 0.43, P = 0.0002). To investigate associations of bladder cancer outcomes with immune cell profiles, using Cox proportional hazard models, we observed an association between the fraction of dendritic cells and the hazard of death (HR = 1.27, 95% CI = 1.06-1.53). Further, a high endothelial cell proportion was significantly associated with an increased hazard of death and tumor recurrence (HR = 1.06, 95% CI = 1.01-1.13) in TME. In addition, the peripheral neutrophil-to-lymphocyte ratio (HR = 1.49, 95% CI = 1.01-2.22), monocyte (HR = 1.17, 95% CI = 1.05-1.31), neutrophil (HR = 1.04, 95% CI = 1.01-1.07), and basophil (HR = 1.35, 95% CI = 1.01-1.81) cell proportions were associated with an increased hazard of death and tumor recurrence. Our results integrated the information on bladder cancer outcomes and cell profiles in TME and peripheral blood, providing biomarkers for estimating bladder cancer prognosis using genome-scale DNA methylation measures.

Citation Format: Ji-Qing Chen, Lucas A. Salas, John K. Wiencke, Devin C. Koestler, Annette M. Molinaro, Angeline S. Andrew, John D. Seigne, Margaret R. Karagas, Karl T. Kelsey, Brock C. Christensen. Integration of associations of immune profiles in peripheral blood and tumor microenvironment with bladder cancer outcomes. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6675.

Abstract 3506: High-grade serous ovarian cancer somatic mutational signatures in Black and White individuals

Causes of racial disparities in ovarian cancer survival are likely multifactorial, including socio-environmental, structural, and biological factors. Mutational signatures reflect endogenous and exogenous exposures, which may differ by race and ethnicity. This study aims to characterize mutational signatures and their associated etiologies among Black high-grade serous ovarian cancer (HGSC) cases, and to compare the occurrence and distribution of signatures with White TCGA cases. Mutational profiling with whole exome sequencing in matched blood and tumor formalin-fixed paraffin-embedded specimens was performed in 271 HGSC cases identifying as Black or African American and participating in the African American Cancer Epidemiology Study (AACES) or the North Carolina Ovarian Cancer Study (NCOCS). After filtering, we identified 28,601 high-confidence coding mutations in 256 samples that passed quality control assessments. The median number of mutations per sample was 67, similar to that in TCGA at 66. Mutation frequencies of the major genes identified in TCGA were nearly identical in Black AACES/NCOCS cases; mutations in TP53 were present in 96% and 89%, respectively, and mutation frequencies in BRCA1/2, CSMD3, NF1, CDK12, FAT3, GABRA6, and RB1 differed by only 0-2%. We performed mutational signature analysis using SigProfiler and refitting to the Catalogue of Somatic Mutations in Cancer (COSMIC) previously published single base-pair substitution (SBS) signatures reflecting mutational processes resulting from known endogenous and exogenous exposures. SBS signatures are defined by 96 mutation features including the single nucleotide variant class (e.g., C&gt;A, C&gt;G, C&gt;T, T&gt;A, T&gt;C, and T&gt;G) and the identity of the 5’/3’ flanking bases. We observed much higher frequencies of the mismatch repair deficiency, homologous recombination deficiency, ultraviolet light, and treatment exposure signatures in Black AACES/NCOCS cases (26%, 22%, 9%, and 8%, respectively) than in White TCGA cases (3%, 10%, 3%, and 2%, respectively). The frequencies of the remaining signatures with known etiologies (clock-like, reactive oxygen species, base excision repair, mutagen exposure, DNA polymerase epsilon, and tobacco signatures) differed by less than 5% between Black and White cases. Many mismatch repair deficiency signatures identified in Black AACES/NCOCS cases are dominated by C&gt;T or T&gt;C mutations and may represent an artifact from formalin fixation. While we observed that the somatic mutation frequencies in major genes associated with HGSC were similar in Black and White cases, the frequencies of some known mutational signatures were considerably higher in Black HGSC. In particular, despite similar frequencies of BRCA1/2 mutations, the homologous recombination deficiency signature was considerably more common in Black than White HGSC.

Citation Format: Katherine A. Lawson-Michod, David Nix, Lindsay Collin, Natalie Davidson, Ariel Hippen, Chad Huff, Aaron Atkinson, Lucas A. Salas, Lauren Peres, Casey Greene, Joellen Schildkraut, Jeffrey Marks, Jennifer A. Doherty. High-grade serous ovarian cancer somatic mutational signatures in Black and White individuals [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3506.

Calculating detection limits and uncertainty of reference-based deconvolution of whole-blood DNA methylation data

DNA methylation (DNAm)-based cell mixture deconvolution (CMD) has become a quintessential part of epigenome-wide association studies where DNAm is profiled in heterogeneous tissue types. Despite being introduced over a decade ago, detection limits, which represent the smallest fraction of a cell type in a mixed biospecimen that can be reliably detected, have yet to be determined in the context of DNAm-based CMD. Moreover, there has been little attention given to approaches for quantifying the uncertainty associated with DNAm-based CMD. Here, analytical frameworks for determining both cell-specific limits of detection and quantification of uncertainty associated with DNAm-based CMD are described. This work may contribute to improved rigor, reproducibility and replicability of epigenome-wide association studies involving CMD.

Inferring spatial transcriptomics markers from whole slide images to characterize metastasis-related spatial heterogeneity of colorectal tumors: A pilot study

Over 150 000 Americans are diagnosed with colorectal cancer (CRC) every year, and annually over 50 000 individuals will die from CRC, necessitating improvements in screening, prognostication, disease management, and therapeutic options. Tumor metastasis is the primary factor related to the risk of recurrence and mortality. Yet, screening for nodal and distant metastasis is costly, and invasive and incomplete resection may hamper adequate assessment. Signatures of the tumor-immune microenvironment (TIME) at the primary site can provide valuable insights into the aggressiveness of the tumor and the effectiveness of various treatment options. Spatially resolved transcriptomics technologies offer an unprecedented characterization of TIME through high multiplexing, yet their scope is constrained by cost. Meanwhile, it has long been suspected that histological, cytological, and macroarchitectural tissue characteristics correlate well with molecular information (e.g., gene expression). Thus, a method for predicting transcriptomics data through inference of RNA patterns from whole slide images (WSI) is a key step in studying metastasis at scale. In this work, we collected tissue from 4 stage-III (pT3) matched colorectal cancer patients for spatial transcriptomics profiling. The Visium spatial transcriptomics (ST) assay was used to measure transcript abundance for 17 943 genes at up to 5000 55-micron (i.e., 1-10 cells) spots per patient sampled in a honeycomb pattern, co-registered with hematoxylin and eosin (H&E) stained WSI. The Visium ST assay can measure expression at these spots through tissue permeabilization of mRNAs, which are captured through spatially (i.e., x-y positional coordinates) barcoded, gene specific oligo probes. WSI subimages were extracted around each co-registered Visium spot and were used to predict the expression at these spots using machine learning models. We prototyped and compared several convolutional, transformer, and graph convolutional neural networks to predict spatial RNA patterns at the Visium spots under the hypothesis that the transformer- and graph-based approaches better capture relevant spatial tissue architecture. We further analyzed the model's ability to recapitulate spatial autocorrelation statistics using SPARK and SpatialDE. Overall, the results indicate that the transformer- and graph-based approaches were unable to outperform the convolutional neural network architecture, though they exhibited optimal performance for relevant disease-associated genes. Initial findings suggest that different neural networks that operate on different scales are relevant for capturing distinct disease pathways (e.g., epithelial to mesenchymal transition). We add further evidence that deep learning models can accurately predict gene expression in whole slide images and comment on understudied factors which may increase its external applicability (e.g., tissue context). Our preliminary work will motivate further investigation of inference for molecular patterns from whole slide images as metastasis predictors and in other applications.

Hydroxymethylation alterations in progenitor-like cell types of pediatric central nervous system tumors are associated with cell type-specific transcriptional changes

Although intratumoral heterogeneity has been established in pediatric central nervous system tumors, epigenomic alterations at the cell type level have largely remained unresolved. To identify cell type-specific alterations to cytosine modifications in pediatric central nervous system tumors we utilized a multi-omic approach that integrated bulk DNA cytosine modification data (methylation and hydroxymethylation) with both bulk and single-cell RNA-sequencing data. We demonstrate a large reduction in the scope of significantly differentially modified cytosines in tumors when accounting for tumor cell type composition. In the progenitor-like cell types of tumors, we identified a preponderance differential CpG hydroxymethylation rather than methylation. Genes with differential hydroxymethylation, like HDAC4 and IGF1R, were associated with cell type-specific changes in gene expression in tumors. Our results highlight the importance of epigenomic alterations in the progenitor-like cell types and its role in cell type-specific transcriptional regulation in pediatric CNS tumors.

Application of novel breast biospecimen cell type adjustment identifies shared DNA methylation alterations in breast tissue and milk with breast cancer risk factors

BACKGROUND

DNA methylation patterning is cell-type-specific and altered DNA methylation is well established to occur early in breast carcinogenesis, affecting non-cancerous, histopathologically normal breast tissue. Previous work assessing risk factor-associated alterations to DNA methylation in breast tissue has been limited, with even less published research in breast milk, a non-invasively obtained biospecimen containing sloughed mammary epithelial cells that may identify early alterations indicative of cancer risk.

METHODS

Here, we present a novel library for the estimation of the cellular composition of breast tissue and milk and subsequent assessment of cell type-independent alterations to DNA methylation associated with established breast cancer risk factors in solid breast tissue (n = 95) and breast milk (n = 48) samples using genome-scale DNA methylation measures from the Illumina HumanMethylation450 array.

RESULTS

We identified 772 hypermethylated CpGs (P < 0.01) associated with age consistent between breast tissue and breast milk samples. Age-associated hypermethylated CpG loci were significantly enriched for CpG island shore regions known to be important for regulating gene expression. Among the overlapping hypermethylated loci mapping to genes, a differentially methylated region was identified in the promoter region of SFRP2, a gene observed to undergo promoter hypermethylation in breast cancer.

CONCLUSIONS

Our findings suggest the potential to identify epigenetic biomarkers of breast cancer risk in noninvasively obtained, tissue-specific breast milk specimens.

IMPACT

This work demonstrates the potential of using breast milk as a non-invasive biomarker of breast cancer risk, improving our ability to detect early stage disease and lowering the overall disease burden.

Umbilical cord blood immune cell profiles in relation to the infant gut microbiome

During infancy, the interplay between the developing immune system and the microbiome is critical. We examined whether blood immune cell composition at birth in the umbilical cord (inferred by DNA methylation profiling) related to the early infant gut microbiome (assessed by 16S rRNA gene sequencing) among 73 infants in the New Hampshire Birth Cohort Study. We used generalized estimating equations and controlled for false discovery rate to select microbial taxa associated with immune cells. We found associations between the infant gut microbiome and immune cells, including a positive association between B cells and Enterobacter, a negative association between natural killer cells and Bifidobacterium, and a positive association between granulocytes and Bifidobacterium. Our findings give clues that immune profiles at the time of birth as measured in umbilical cord blood are associated with the development of the gut microbiome in early life.

Comparative analysis of the DNA methylation landscape in CD4, CD8, and B memory lineages

BACKGROUND

There is considerable evidence that epigenetic mechanisms and DNA methylation are critical drivers of immune cell lineage differentiation and activation. However, there has been limited coordinated investigation of common epigenetic pathways among cell lineages. Further, it remains unclear if long-lived memory cell subtypes differentiate distinctly by cell lineages.

RESULTS

We used the Illumina EPIC array to investigate the consistency of DNA methylation in B cell, CD4 T, and CD8 T naïve and memory cells states. In the process of naïve to memory activation across the three lineages, we identify considerable shared epigenetic regulation at the DNA level for immune memory generation. Further, in central to effector memory differentiation, our analyses revealed specific CpG dinucleotides and genes in CD4 T and CD8 T cells with DNA methylation changes. Finally, we identified unique DNA methylation patterns in terminally differentiated effector memory (TEMRA) CD8 T cells compared to other CD8 T memory cell subtypes.

CONCLUSIONS

Our data suggest that epigenetic alterations are widespread and essential in generating human lymphocyte memory. Unique profiles are involved in methylation changes that accompany memory genesis in the three subtypes of lymphocytes.

Evaluation of cross-platform compatibility of a DNA methylation-based glucocorticoid response biomarker

Background

Identifying blood-based DNA methylation patterns is a minimally invasive way to detect biomarkers in predicting age, characteristics of certain diseases and conditions, as well as responses to immunotherapies. As microarray platforms continue to evolve and increase the scope of CpGs measured, new discoveries based on the most recent platform version and how they compare to available data from the previous versions of the platform are unknown. The neutrophil dexamethasone methylation index (NDMI 850) is a blood-based DNA methylation biomarker built on the Illumina MethylationEPIC (850K) array that measures epigenetic responses to dexamethasone (DEX), a synthetic glucocorticoid often administered for inflammation. Here, we compare the NDMI 850 to one we built using data from the Illumina Methylation 450K (NDMI 450).

Results

The NDMI 450 consisted of 22 loci, 15 of which were present on the NDMI 850. In adult whole blood samples, the linear composite scores from NDMI 450 and NDMI 850 were highly correlated and had equivalent predictive accuracy for detecting DEX exposure among adult glioma patients and non-glioma adult controls. However, the NDMI 450 scores of newborn cord blood were significantly lower than NDMI 850 in samples measured with both assays.

Conclusions

We developed an algorithm that reproduces the DNA methylation glucocorticoid response score using 450K data, increasing the accessibility for researchers to assess this biomarker in archived or publicly available datasets that use the 450K version of the Illumina BeadChip array. However, the NDMI850 and NDMI450 do not give similar results in cord blood, and due to data availability limitations, results from sample types of newborn cord blood should be interpreted with care.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13148-022-01352-1.

HiTIMED: hierarchical tumor immune microenvironment epigenetic deconvolution for accurate cell type resolution in the tumor microenvironment using tumor-type-specific DNA methylation data

Background

Cellular compositions of solid tumor microenvironments are heterogeneous, varying across patients and tumor types. High-resolution profiling of the tumor microenvironment cell composition is crucial to understanding its biological and clinical implications. Previously, tumor microenvironment gene expression and DNA methylation-based deconvolution approaches have been shown to deconvolve major cell types. However, existing methods lack accuracy and specificity to tumor type and include limited identification of individual cell types.

Results

We employed a novel tumor-type-specific hierarchical model using DNA methylation data to deconvolve the tumor microenvironment with high resolution, accuracy, and specificity. The deconvolution algorithm is named HiTIMED. Seventeen cell types from three major tumor microenvironment components can be profiled (tumor, immune, angiogenic) by HiTIMED, and it provides tumor-type-specific models for twenty carcinoma types. We demonstrate the prognostic significance of cell types that other tumor microenvironment deconvolution methods do not capture.

Conclusion

We developed HiTIMED, a DNA methylation-based algorithm, to estimate cell proportions in the tumor microenvironment with high resolution and accuracy. HiTIMED deconvolution is amenable to archival biospecimens providing high-resolution profiles enabling to study of clinical and biological implications of variation and composition of the tumor microenvironment.

DNA methylation as a pharmacodynamic marker of glucocorticoid response and glioma survival

Assessing individual responses to glucocorticoid drug therapies that compromise immune status and affect survival outcomes in neuro-oncology is a great challenge. Here we introduce a blood-based neutrophil dexamethasone methylation index (NDMI) that provides a measure of the epigenetic response of subjects to dexamethasone. This marker outperforms conventional approaches based on leukocyte composition as a marker of glucocorticoid response. The NDMI is associated with low CD4 T cells and the accumulation of monocytic myeloid-derived suppressor cells and also serves as prognostic factor in glioma survival. In a non-glioma population, the NDMI increases with a history of prednisone use. Therefore, it may also be informative in other conditions where glucocorticoids are employed. We conclude that DNA methylation remodeling within the peripheral immune compartment is a rich source of clinically relevant markers of glucocorticoid response.

Abstract 1212: ExTIME: Extended tumor immune micro-environment cell mixture deconvolution using DNA methylation and a novel tumor-site-specific hierarchical approach

Background: The solid tumor microenvironment is heterogeneous and varies in composition by tumor type. Previous gene expression and DNA methylation deconvolution approaches for tumor micro-environment have had some success for major cell types. However, existing methods lack specificity to tumor type and detailed cell types. We developed 21 tumor-specific DNA methylation-based libraries. We employed a novel hierarchical approach in 3 major tumor microenvironment components (tumor, angiogenic, immune) to profile 17 cell types (see methods below).

Methods: DNA methylation data on tumor samples (n=6183) and normal control samples (n=689) for 21 tumor sites were downloaded from GEO and TCGA to develop tumor-type-specific libraries. The top 1000 most informative differentially methylated CpG (DMC) sites were identified using InfiniumPurify for 21 tumor types to project tumor cell proportion. Epithelial, endothelial, and stromal cell samples were used to identify DMCs to profile the cells in the angiogenic environment. Basophil, eosinophil, neutrophil, dendritic cell, monocyte, B naïve, B memory, CD4T naïve, CD4T memory, CD8T naïve, CD8T memory, T regulatory, and natural killer cells were used to identify DMCs to deconvolve the immune environment. In conjunction with the constrained projection/quadratic programming approach, a novel hierarchical approach was employed with six layers and 12 libraries per tumor type to project cell proportions in first, tumor, second, angiogenic, and third, immune micro-environments. The method was validated using purified samples and experimental artificial mixtures.

Results: 12 libraries were developed per tumor site to deconvolve 17 cell types in 21 tumors. A preliminary application of the method on TCGA data investigating the association between angiogenic cells and survival revealed worse survival outcomes with a higher proportion of angiogenic cell proportions in BLCA (p&lt;0.01) and HNSC (p=0.02), a higher endothelial cell proportion in CESC (p=0.04), a higher epithelial cell proportion in COAD (p=0.02), a lower endothelial proportion in KIRC (p&lt;0.01), and a lower epithelial proportion in LUAD (p=0.04). Further analyses will be done to investigate the angiogenic and immune microenvironments with prognosis across tumor sites.

Conclusion: We developed a DNA methylation-based algorithm, ExTIME, to estimate cell proportions in the tumor microenvironments. This novel approach increased the specificity and accuracy of cell projection by employing a tumor-site-specific hierarchical model. Furthermore, the ExTIME profiles the tumor microenvironment to the most granular level compared to the existing methods. ExTIME’s capability of depicting the cellular composition in tumors promises a better understanding of the cell heterogeneity and its relationship with prognosis across cancers.

Citation Format: Ze Zhang, Brock C. Christensen, Lucas A. Salas. ExTIME: Extended tumor immune micro-environment cell mixture deconvolution using DNA methylation and a novel tumor-site-specific hierarchical approach [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1212.

MethylMasteR: A Comparison and Customization of Methylation-Based Copy Number Variation Calling Software in Cancers Harboring Large Scale Chromosomal Deletions

DNA methylation-based copy number variation (CNV) calling software offers the advantages of providing both genetic (copy-number) and epigenetic (methylation) state information from a single genomic library. This method is advantageous when looking at large-scale chromosomal rearrangements such as the loss of the short arm of chromosome 3 (3p) in renal cell carcinoma and the codeletion of the short arm of chromosome 1 and the long arm of chromosome 19 (1p/19q) commonly seen in histologically defined oligodendrogliomas. Herein, we present MethylMasteR: a software framework that facilitates the standardization and customization of methylation-based CNV calling algorithms in a single R package deployed using the Docker software framework. This framework allows for the easy comparison of the performance and the large-scale CNV event identification capability of four common methylation-based CNV callers. Additionally, we incorporated our custom routine, which was among the best performing routines. We employed the Affymetrix 6.0 SNP Chip results as a gold standard against which to compare large-scale event recall. As there are disparities within the software calling algorithms themselves, no single software is likely to perform best for all samples and all combinations of parameters. The employment of a standardized software framework via creating a Docker image and its subsequent deployment as a Docker container allows researchers to efficiently compare algorithms and lends itself to the development of modified workflows such as the custom workflow we have developed. Researchers can now use the MethylMasteR software for their methylation-based CNV calling needs and follow our software deployment framework. We will continue to refine our methodology in the future with a specific focus on identifying large-scale chromosomal rearrangements in cancer methylation data.

A Novel Framework for the Identification of Reference DNA Methylation Libraries for Reference-Based Deconvolution of Cellular Mixtures

Reference-based deconvolution methods use reference libraries of cell-specific DNA methylation (DNAm) measurements as a means toward deconvoluting cell proportions in heterogeneous biospecimens (e.g., whole-blood). As the accuracy of such methods depends highly on the CpG loci comprising the reference library, recent research efforts have focused on the selection of libraries to optimize deconvolution accuracy. While existing approaches for library selection work extremely well, the best performing approaches require a training data set consisting of both DNAm profiles over a heterogeneous cell population and gold-standard measurements of cell composition (e.g., flow cytometry) in the same samples. Here, we present a framework for reference library selection without a training dataset (RESET) and benchmark it against the Legacy method (minfi:pickCompProbes), where libraries are constructed based on a pre-specified number of cell-specific differentially methylated loci (DML). RESET uses a modified version of the Dispersion Separability Criteria (DSC) for comparing different libraries and has four main steps: 1) identify a candidate set of cell-specific DMLs, 2) randomly sample DMLs from the candidate set, 3) compute the Modified DSC of the selected DMLs, and 4) update the selection probabilities of DMLs based on their contribution to the Modified DSC. Steps 2–4 are repeated many times and the library with the largest Modified DSC is selected for subsequent reference-based deconvolution. We evaluated RESET using several publicly available datasets consisting of whole-blood DNAm measurements with corresponding measurements of cell composition. We computed the RMSE and R² between the predicted cell proportions and their measured values. RESET outperformed the Legacy approach in selecting libraries that improve the accuracy of deconvolution estimates. Additionally, reference libraries constructed using RESET resulted in cellular composition estimates that explained more variation in DNAm as compared to the Legacy approach when evaluated in the context of epigenome-wide association studies (EWAS) of several publicly available data sets. This finding has implications for the statistical power of EWAS. RESET combats potential challenges associated with existing approaches for reference library assembly and thus, may serve as a viable strategy for library construction in the absence of a training data set.

A core of differentially methylated CpG loci in gMDSCs isolated from neonatal and adult sources

BACKGROUND

Myeloid-derived suppressor cells (MDSCs), which include monocytic (mMDSCs) and granulocytic (gMDSCs) cells, are an immunosuppressive, heterogeneous population of cells upregulated in cancer and other pathologic conditions, in addition to normal conditions of stress. The origin of MDSCs is debated, and the regulatory pattern responsible for gMDSC differentiation remains unknown. Since DNA methylation (DNAm) contributes to lineage differentiation, we have investigated whether it contributes to the acquisition of the gMDSC phenotype.

RESULTS

Using the Illumina EPIC array to measure DNAm of gMDSCs and neutrophils from diverse neonatal and adult blood sources, we found 189 differentially methylated CpGs between gMDSCs and neutrophils with a core of ten differentially methylated CpGs that were consistent across both sources of cells. Genes associated with these loci that are involved in immune responses include VCL, FATS, YAP1, KREMEN2, UBTF, MCC-1, and EFCC1. In two cancer patient groups that reflected those used to develop the methylation markers (head and neck squamous cell carcinoma (HNSCC) and glioma), all of the CpG loci were differentially methylated, reaching statistical significance in glioma cases and controls, while one was significantly different in the smaller HNSCC group.

CONCLUSIONS

Our findings indicate that gMDSCs have a core of distinct DNAm alterations, informing future research on gMDSC differentiation and function.

Enhanced cell deconvolution of peripheral blood using DNA methylation for high-resolution immune profiling

DNA methylation microarrays can be employed to interrogate cell-type composition in complex tissues. Here, we expand reference-based deconvolution of blood DNA methylation to include 12 leukocyte subtypes (neutrophils, eosinophils, basophils, monocytes, naïve and memory B cells, naïve and memory CD4 + and CD8 + T cells, natural killer, and T regulatory cells). Including derived variables, our method provides 56 immune profile variables. The IDOL (IDentifying Optimal Libraries) algorithm was used to identify libraries for deconvolution of DNA methylation data for current and previous platforms. The accuracy of deconvolution estimates obtained using our enhanced libraries was validated using artificial mixtures and whole-blood DNA methylation with known cellular composition from flow cytometry. We applied our libraries to deconvolve cancer, aging, and autoimmune disease datasets. In conclusion, these libraries enable a detailed representation of immune-cell profiles in blood using only DNA and facilitate a standardized, thorough investigation of immune profiles in human health and disease.

Navigating the hydroxymethylome: experimental biases and quality control tools for the tandem bisulfite and oxidative bisulfite Illumina microarrays

Aim: Tandem bisulfite (BS) and oxidative bisulfite (oxBS) conversion on DNA followed by hybridization to Infinium HumanMethylation BeadChips allows nucleotide resolution of 5-hydroxymethylcytosine genome-wide. Here, the authors compared data quality acquired from BS-treated and oxBS-treated samples. Materials & methods: Raw BeadArray data from 417 pairs of samples across 12 independent datasets were included in the study. Probe call rates were compared between paired BS and oxBS treatments controlling for technical variables. Results: oxBS-treated samples had a significantly lower call rate. Among technical variables, DNA-specific extraction kits performed better with higher call rates after oxBS conversion. Conclusion: The authors emphasize the importance of quality control during oxBS conversion to minimize information loss and recommend using a DNA-specific extraction kit for DNA extraction and an oxBSQC package for data preprocessing.

Immune profiles and DNA methylation alterations related with non-muscle-invasive bladder cancer outcomes

BACKGROUND

Non-muscle-invasive bladder cancer (NMIBC) patients receive frequent monitoring because ≥ 70% will have recurrent disease. However, screening is invasive, expensive, and associated with significant morbidity making bladder cancer the most expensive cancer to treat per capita. There is an urgent need to expand the understanding of markers related to recurrence and survival outcomes of NMIBC.

METHODS AND RESULTS

We used the Illumina HumanMethylationEPIC array to measure peripheral blood DNA methylation profiles of NMIBC patients (N = 603) enrolled in a population-based cohort study in New Hampshire and applied cell type deconvolution to estimate immune cell-type proportions. Using Cox proportional hazard models, we identified that increasing CD4T and CD8T cell proportions were associated with a statistically significant decreased hazard of tumor recurrence or death (CD4T: HR = 0.98, 95% CI = 0.97-1.00; CD8T: HR = 0.97, 95% CI = 0.95-1.00), whereas increasing monocyte proportion and methylation-derived neutrophil-to-lymphocyte ratio (mdNLR) were associated with the increased hazard of tumor recurrence or death (monocyte: HR = 1.04, 95% CI = 1.00-1.07; mdNLR: HR = 1.12, 95% CI = 1.04-1.20). Then, using an epigenome-wide association study (EWAS) approach adjusting for age, sex, smoking status, BCG treatment status, and immune cell profiles, we identified 2528 CpGs associated with the hazard of tumor recurrence or death (P < 0.005). Among these CpGs, the 1572 were associated with an increased hazard and were significantly enriched in open sea regions; the 956 remaining CpGs were associated with a decreased hazard and were significantly enriched in enhancer regions and DNase hypersensitive sites.

CONCLUSIONS

Our results expand on the knowledge of immune profiles and methylation alteration associated with NMIBC outcomes and represent a first step toward the development of DNA methylation-based biomarkers of tumor recurrence.

Mixed Effects Machine Learning Models for Colon Cancer Metastasis Prediction using Spatially Localized Immuno-Oncology Markers

Spatially resolved characterization of the transcriptome and proteome promises to provide further clarity on cancer pathogenesis and etiology, which may inform future clinical practice through classifier development for clinical outcomes. However, batch effects may potentially obscure the ability of machine learning methods to derive complex associations within spatial omics data. Profiling thirty-five stage three colon cancer patients using the GeoMX Digital Spatial Profiler, we found that mixed-effects machine learning (MEML) methods† may provide utility for overcoming significant batch effects to communicate key and complex disease associations from spatial information. These results point to further exploration and application of MEML methods within the spatial omics algorithm development life cycle for clinical deployment.

Phase I Study of High-Dose L-methylfolate in Combination with Temozolomide and Bevacizumab in Recurrent IDH wild-type High-Grade Glioma

Purpose

IDH mutations in low-grade gliomas (LGGs) results in improved survival and DNA hypermethylation compared to IDH wild-type LGGs. IDH-mutant LGGs become hypomethylated during progression. It's uncertain if methylation changes occur during IDH wild-type GBM progression and if the methylome can be reprogrammed. This phase I study evaluated the safety, tolerability, efficacy and methylome changes after L-methylfolate (LMF) treatment, in combination with temozolomide and bevacizumab in patients with recurrent high-grade glioma.

Patients and Methods

Fourteen patients total, 13 with GBM, one with anaplastic astrocytoma, all IDH wild-type were enrolled in the study. All patients received LMF at either 15, 30, 60, or 90 mg daily plus temozolomide (75mg/m² 5 days per month) and bevacizumab (10mg/kg every two weeks).

Results

No MTD was identified. LMF treated had mOS of 9.5 months (95% CI, 9.1-35.4) comparable to bevacizumab historical control 8.6 months (95% CI, 6.8-10.8). Six patients treated with LMF survived more than 650 days. Across all treatment doses the most adverse events were diarrhea (7%, 1 patient, grade 2), reflux (7%, 1 patient, grade 2), and dysgeusia (7%, 1 patient, grade 2). In the six brains donated at death, there was a 25% increase in DNA methylated CpGs compared to the paired initial tumor.

Conclusions

LMF in combination with temozolomide and bevacizumab was well tolerated in patients with recurrent IDH wild-type high-grade glioma. This small study did not establish a superior efficacy with addition of LMF compared to standard bevacizumab therapy, however, this study did show methylome reprogramming in high-grade glioma.

Methylation-derived inflammatory measures and lung cancer risk and survival

BACKGROUND

Examining immunity-related DNA methylation alterations in blood could help elucidate the role of the immune response in lung cancer etiology and aid in discovering factors that are key to lung cancer development and progression. In a nested, matched case-control study, we estimated methylation-derived NLR (mdNLR) and quantified DNA methylation levels at loci previously linked with circulating concentrations of C-reactive protein (CRP). We examined associations between these measures and lung cancer risk and survival.

RESULTS

Using conditional logistic regression and further adjusting for BMI, batch effects, and a smoking-based methylation score, we observed a 47% increased risk of non-small cell lung cancer (NSCLC) for one standard deviation (SD) increase in mdNLR (n = 150 pairs; OR: 1.47, 95% CI 1.08, 2.02). Using a similar model, the estimated CRP Scores were inversely associated with risk of NSCLC (e.g., Score 1 OR: 0.57, 95% CI: 0.40, 0.81). Using Cox proportional hazards models adjusting for age, sex, smoking status, methylation-predicted pack-years, BMI, batch effect, and stage, we observed a 28% increased risk of dying from lung cancer (n = 145 deaths in 205 cases; HR: 1.28, 95% CI: 1.09, 1.50) for one SD increase in mdNLR.

CONCLUSIONS

Our study demonstrates that immunity status measured with DNA methylation markers is associated with lung cancer a decade or more prior to cancer diagnosis. A better understanding of immunity-associated methylation-based biomarkers in lung cancer development could provide insight into critical pathways.

Genome-scale DNA methylation analysis identifies repeat element alterations that modulate the genomic stability of melanocytic nevi

Acquired melanocytic nevi grow and persist in a stable form into adulthood. Using genome-wide methylation profiling, we evaluated 32 histopathologically and dermoscopically characterized nevi, to identify key epigenetic regulatory mechanisms involved in nevogenesis. Benign (69% globular and 31% non-specific dermoscopic pattern) and dysplastic (95% reticular/nonspecific dermoscopic pattern) nevi were dissimilar with only two shared differentially methylated (DM) loci. Benign nevi demonstrated an increase in both genome-scale methylation and methylation of Alu/LINE-1 retrotransposable elements, a marker of genomic stability, as well as global methylation. In contrast, dysplastic nevi showed evidence for genomic instability via hypomethylation of Alu/LINE-1 (Alu; P=0.00019 and LINE-1; P=0.000035). Using dermoscopic classifications, reticular/non-specific patterned nevi had 59,572 CpG DM loci (Q < 0.05), whereas globular nevi had no significant DM loci. In reticular/non-specific patterned nevi, the tumor suppressor PTEN had the greatest proportion of hypermethylated CpG loci in its promoter region compared to all other assayed gene promoters. The relative activity of reticular/non-specific nevi was evidenced by 50,720 hypomethylated loci being enriched for accessible chromatin, and 8,852 hypermethylated loci strongly enriched, for example, marks of active gene promoters, which suggests that gain of DNA methylation observed in these nevus types plays a role in gene regulation.

A transdisciplinary approach to understand the epigenetic basis of race/ethnicity health disparities

Health disparities correspond to differences in disease burden and mortality among socially defined population groups. Such disparities may emerge according to race/ethnicity, socioeconomic status and a variety of other social contexts, and are documented for a wide range of diseases. Here, we provide a transdisciplinary perspective on the contribution of epigenetics to the understanding of health disparities, with a special emphasis on disparities across socially defined racial/ethnic groups. Scientists in the fields of biological anthropology, bioinformatics and molecular epidemiology provide a summary of theoretical, statistical and practical considerations for conducting epigenetic health disparities research, and provide examples of successful applications from cancer research using this approach.

Interactions of Age and Blood Immune Factors and Non-Invasive Prediction of Glioma Survival

BACKGROUND

Tumor-based classification of human glioma portends patient prognosis; however, considerable unexplained survival variability remains. Host factors (eg, age) also strongly influence survival times, partly reflecting a compromised immune system. How blood epigenetic measures of immune characteristics and age augment molecular classifications in glioma survival has not been investigated. We assess the prognostic impact of immune-cell fractions and epigenetic age in archived blood across glioma molecular subtypes for the first time.

METHODS

We evaluated immune-cell fractions and epigenetic age in archived blood from the University of California San Francisco Adult Glioma Study, including a training set of 197 IDH-wildtype, 1p19q intact, TERT wildtype (IDH/1p19q/TERT-WT) glioma patients, an evaluation set of 350 patients with other subtypes of glioma, and 454 subjects without glioma.

RESULTS

IDH/1p19q/TERT-WT patients had lower lymphocyte fractions (CD4+T, CD8+T, natural killer, and B cells) and higher neutrophil fractions than subjects without glioma. Recursive partitioning analysis delineated four statistically significantly different survival groups for IDH/1p19q/TERT-WT patients based on an interaction between chronological age and two blood immune factors, CD4+T cells, and neutrophils with median overall survival ranging from 0.76 years [95% confidence intervaI = 0.55 to 0.99] for the worst survival group (n = 28) to 9.72 years [95% confidence intervaI = 6.18 to NA] for the best (n = 33). The Recursive partitioning analysis also statistically significantly delineated four risk groups in patients with other glioma subtypes.

CONCLUSION

The delineation of different survival groups in the training and evaluation sets based on an interaction between chronological age and blood immune characteristics suggests that common host immune factors among different glioma types may impact survival. The ability of DNA methylation-based markers of immune status to capture diverse, clinically relevant information may facilitate non-invasive personalized patient evaluation in the neuro-oncology clinic.

MethylSPWNet and MethylCapsNet: Biologically Motivated Organization of DNAm Neural Networks, Inspired by Capsule Networks

DNA methylation (DNAm) alterations have been heavily implicated in carcinogenesis and the pathophysiology of diseases through upstream regulation of gene expression. DNAm deep-learning approaches are able to capture features associated with aging, cell type, and disease progression, but lack incorporation of prior biological knowledge. Here, we present modular, user-friendly deep-learning methodology and software, MethylCapsNet and MethylSPWNet, that group CpGs into biologically relevant capsules-such as gene promoter context, CpG island relationship, or user-defined groupings-and relate them to diagnostic and prognostic outcomes. We demonstrate these models' utility on 3,897 individuals in the classification of central nervous system (CNS) tumors. MethylCapsNet and MethylSPWNet provide an opportunity to increase DNAm deep-learning analyses' interpretability by enabling a flexible organization of DNAm data into biologically relevant capsules.

Prediagnostic breast milk DNA methylation alterations in women who develop breast cancer

Prior candidate gene studies have shown tumor suppressor DNA methylation in breast milk related with history of breast biopsy, an established risk factor for breast cancer. To further establish the utility of breast milk as a tissue-specific biospecimen for investigations of breast carcinogenesis, we measured genome-wide DNA methylation in breast milk from women with and without a diagnosis of breast cancer in two independent cohorts. DNA methylation was assessed using Illumina HumanMethylation450k in 87 breast milk samples. Through an epigenome-wide association study we explored CpG sites associated with a breast cancer diagnosis in the prospectively collected milk samples from the breast that would develop cancer compared with women without a diagnosis of breast cancer using linear mixed effects models adjusted for history of breast biopsy, age, RefFreeCellMix cell estimates, time of delivery, array chip and subject as random effect. We identified 58 differentially methylated CpG sites associated with a subsequent breast cancer diagnosis (q-value <0.05). Nearly all CpG sites associated with a breast cancer diagnosis were hypomethylated in cases compared with controls and were enriched for CpG islands. In addition, inferred repeat element methylation was lower in breast milk DNA from cases compared to controls, and cases exhibited increased estimated epigenetic mitotic tick rate as well as DNA methylation age compared with controls. Breast milk has utility as a biospecimen for prospective assessment of disease risk, for understanding the underlying molecular basis of breast cancer risk factors and improving primary and secondary prevention of breast cancer.

[Long-term cognitive and functional status in survivors of an aneurysmal subarachnoid hemorrhage: Analysis of a retrospective cohort]

OBJECTIVE

Little data is available on long-term functional and cognitive outcomes in patients with aneurysmal subarachnoid hemorrhage (ASH). The main objective of this study was to assess cognition, functional state, mood disorders, and quality of life in patients with SAH at least six months following the ASH.

PATIENTS AND METHODS

Cross-sectional study of 40 patients (aged 58.2 [SD 9.9] years) with ASH, discharged from a Neurologic Rehabilitation unit between January 2010 and July 2017.

MAIN OUTCOME VARIABLES

functional status (Barthel index), cognition (Pfeiffer questionnaire), depression (Hamilton scale), and health-related quality of life (European Quality of Life-5 Dimensions [EQ-5D]), as well as type and duration of therapeutic rehabilitation procedures after discharge.

RESULTS

From 35 patients with cognitive disorders, only 12 received cognitive therapy at hospital discharge. In the long-term follow-up, cognitive impairment persisted in 22 patients. When compared with those without cognitive impairment, they presented significantly worse mean differences in the Barthel index (15.5 [95% CI: 1.2-29.7]), Hamilton scale (-0.8 [95% CI: -1.27 to -0.37]), and EQ-5D (27.6 [95% CI: 12.4-19]).

CONCLUSION

The prevalence of long-term cognitive impairments in survivors of a SAH episode is high, and their presence is associated with worse functional status, more depression and worse quality of life. The low percentage of subjects who received cognitive therapies through their recovery process and the clinical implications observed, support the need of including neuropsychological therapies in the rehabilitation programs after an SAH event.

Identification of a foetal epigenetic compartment in adult human kidney

The mammalian kidney has extensive repair capacity; however, identifying adult renal stem cells has proven elusive. We applied an epigenetic marker of foetal cell origin (FCO) in diverse human tissues as a probe for developmental cell persistence, finding a 5.4-fold greater FCO proportion in kidney. Normal kidney FCO proportions averaged 49% with extensive interindividual variation. FCO proportions were significantly negatively correlated with immune-related gene expression and positively correlated with genes expressed in the renal medulla, including those involved in renal organogenesis (e.g., FGF2, PAX8, and HOXB7). FCO associated genes also mapped to medullary nephron segments in mouse and rat, suggesting evolutionary conservation of this cellular compartment. Renal cancer patients whose tumours contained non-zero FCO scores survived longer. The kidney appears unique in possessing substantial foetal epigenetic features. Further study of FCO-related gene methylation may elucidate regenerative regulatory programmes in tissues without apparent discrete stem cell compartments.