Spatially decoding genotype-associated epigenetic landscapes in human lymphoma FFPE tissues via epi-Patho-DBiT

Abstract Chromatin organization within the cell nucleus governs the epigenetic regulation of diverse cellular functions and is tightly controlled within the spatial context of tissues. In human non-Hodgkin B-cell lymphoma (B-NHL), epigenetic regulation is critical to disease pathogenesis, yet their underlying mechanisms remain poorly understood. While formalin-fixed paraffin-embedded (FFPE) samples are plentiful and invaluable for B-NHL translational research, effective methods for spatial profiling of chromatin accessibility and histone modifications in these tissues remain limited, hindering efforts to unravel epigenetic heterogeneity and regulatory mechanisms in B-NHL. Here, we introduce epi-Patho-DBiT, a novel platform integrating reverse crosslinking in FFPE tissues with spatially resolved epigenomic assays, including transposase-accessible chromatin using sequencing (spatial-FFPE-ATAC) and cleavage under targets and tagmentation (spatial-FFPE-CUT&Tag). We applied epi-Patho-DBiT to archival B-NHL FFPE samples to investigate tumor development, progression, and transformation. First, using spatial-FFPE-ATAC, we mapped genotype-specific epigenetic landscapes in human mucosa-associated lymphoid tissue (MALT) and follicular lymphoma (FL), identifying chromatin variants linked to B-cell malignancy. Spatially resolved chromatin accessibility profiles enabled inference of genomic copy number variations (CNVs) in MALT and FL lymphomas, resolving tissue heterogeneity, tumor karyotypes, and evolutionary relationships between tumor clones. We identified a chromosome 7 locus (7q21.12–7q31.31) with copy number amplifications (CNAs) in both MALT and FL tumor regions. In addition, we performed mitotic age inference analysis to trace tumor dynamics in MALT and FL, which revealed intra-to-extracellular migration of tumor cells, supported by multiomics velocity and pseudotemporal ordering analysis. Spatial mitotic age inference further identified a strong correlation between mitotic age and cholesterol metabolism gene activity, uncovering cholesterol-mediated cell cycle regulation in lymphoma development. Next, we performed spatial-FFPE-CUT&Tag targeting H3K4me3 and H3K27me3 histone modifications in low-grade FL, diffuse large B-cell lymphoma (DLBCL), and concurrent FL-DLBCL transformation samples, which revealed cell proliferation dynamics within the tumor microenvironment and identified DIP2C as a novel upregulated tumor marker. Unexpectedly, in the FL-DLBCL transformation sample, a chromosome 2 locus (2p15–2p16.1) exhibited high expression of multiple tumor-promoting genes alongside elevated levels of the repressive histone mark H3K27me3, presenting a paradox of “high gene expression; high repressor H3K27me3 levels”. This genomic locus covers genes including PAPOLG, REL, KIAA1841 and XPO1. Through whole-genome sequencing and CNV inference from spatial transcriptomics data on the same tissue sample, we demonstrated that this paradoxical observation was attributed to CNAs in transformed DLBCL, highlighting complex interplays between genomic and epigenomic regulation in large cell transformation. Overall, epi-Patho-DBiT enables high-resolution, spatially resolved epigenomic profiling of clinical archival FFPE samples, uncovering lymphoma tumor heterogeneity, chromatin variants, and genome-epigenome interactions. This approach provides a powerful tool for decoding the epigenetic mechanisms of lymphoma evolution and transformation.