Abstract The tumor microenvironment (TME) comprises diverse cell types whose interactions are crucial for cancer progression. Traditional bulk tissue analyses mask cell-type-specific epigenetic changes that may drive disease progression. We used DNA methylation data and a reference- based cell-type deconvolution algorithm — hierarchical tumor immune microenvironment epigenetic deconvolution (HiTIMED) — to quantify cell types in the TME. Then, we employed CellDMC, a statistical interaction testing framework that can use cell type proportions in conjunction with differential methylation analysis to identify cell-specific differential methylation. Genome-scale DNA methylation profiles from breast cancer tissue (n=609) and normal breast tissue (n=230) samples were accessed in GEO and GTEx. Our novel approach leverages bulk DNA methylation data to infer cell-type-specific epigenetic landscapes, circumventing challenges and limitations associated with direct single-cell methylation profiling, such as high costs and technical variability. Reference-based computational tools allow us to dissect complex tissue compositions and reveal distinct methylation patterns and pathways that may be unique to individual cell types, providing insights that are not discernible from traditional bulk tumor analysis approaches. Compared with nontumor normal samples, we identified significant DNA methylation alterations (FDR≤0.05) across tumor, stromal, endothelial, epithelial, lymphocytes, and myeloid cells in breast tumors. Tumor cells exhibited the highest number, 8,510, differentially methylated cytosines (DMCs) with 1,966 of these unique to tumor cells. Stromal, endothelial, and epithelial cells also showed considerable methylation alterations, with a total of 530, 440, 339 DMCs, respectively, with 5 DMCs unique to epithelial cells. Despite lower immune cell proportions in our samples, we identified 104 lymphocyte specific DMCs and 185 myeloid specific DMCs, with 14 and 20 DMCs unique to these cell types respectively. The unique DMCs identified in each cell type suggest distinct epigenetic landscapes that may drive specific cellular functions and interactions within the TME. Importantly, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of the annotated cell-type-specific DMCs revealed distinct pathways that were not apparent in the pathway analysis of bulk tumor versus normal tissue, underscoring the value of examining epigenetic alterations at the cell-type level. Our findings highlight the opportunity to leverage DNA-based cell typing with DNA methylation data to discern cell-specific molecular alterations. By focusing on cell-type-specific alterations, we can better understand the epigenetic mechanisms driving TME crosstalk, such as the activation of stromal cells and the modulation of immune responses by tumor cells. Our findings underscore the need for targeted therapeutic strategies that consider the unique epigenetic landscapes of individual cell types within the TME, offering new avenues for improving cancer treatment outcomes. Citation Format: Barbara Karakyriakou, Brock C. Christensen, Lucas A. Salas. Identification of cell-specific DNA methylation alterations in the breast tumor microenvironment [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Tumor-body Interactions: The Roles of Micro- and Macroenvironment in Cancer; 2024 Nov 17-20; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(22_Suppl):Abstract nr B033.
Read full abstract