A single-cell multi-omic and spatial atlas of B cell lymphomas reveals differentiation drives intratumor heterogeneity

Abstract

Intratumor heterogeneity is intrinsic to cancer pathogenesis and evolution, although little is known about how it relates to the differentiation trajectories of the tumor’s cell-of-origin. Nodal B-cell non-Hodgkin lymphomas are a diverse set of cancers thought to originate from distinct stages of B-cell maturation. Through a single-cell multi-omic and spatial atlas of diffuse large B-cell, mantle cell, follicular, and marginal zone lymphomas along with reactive lymph nodes (n=51), we found multiple B-cell maturation states coexist within the same tumors. Intratumor maturation states emerged from the same cell-of-origin, revealing that maturation remains plastic in malignancy. The maturation state composition varied across entities and tumors, which included mixtures of cell-of-origin subtypes. Intratumor maturation states inhabited unique spatial niches, which typically retained their maturation-associated cellular interactions and regulatory networks. Intratumor maturation states showed distinct expression patterns of genetic variants, suggesting that maturation and genetic aberrations are intertwined. Our findings put forward a transformative model for cancer pathogenesis, where differentiation continues in malignancy and is central to tumor heterogeneity and evolution.

Functional phenotyping of genomic variants using multiomic scDNA-scRNA-seq

“…We leveraged a data integration approach based on mutual nearest neighbors and canonical correlation analysis to map B-cell maturation states from a published single-cell transcriptomic dataset in non-malignant reactive lymph nodes to tumor samples (Extended Data Fig. 9a-b) 33,35 . Restriction of the immunoglobulin light chain among the vast majority of B-cells in each tumor to either kappa or lambda confirmed their monoclonality and therefore malignancy (Extended Data Fig.…”

“…Here, naïve B-cells (Naïve) undergoing T-cell dependent activation migrate through the dark zone (DZ) and light zone (LZ) of the germinal center where they undergo somatic hypermutation of the B-cell receptor and selection by antigen-presenting cells, followed by maturation into memory B-cells (Mem IgM or Mem IgG) and plasma cells (Plasma) 29–32 . While the cell of origin is central to the classification of B-cell lymphomas and other cancers, it was recently shown that cancer cells can retain their ability to differentiate along their innate differentiation trajectory, whereby tumors evolve multiple maturation states from the same cell of origin 33 . Meanwhile, clonal evolution posits that tumors also evolve through the accumulation of heterogenous genetic variants over time resulting in the generation of treatment resistant sub-clones 34 .…”

“…B-cell maturation states were mapped to each tumor sample from a published reactive lymph node single-cell RNA sequencing dataset through shared gene expression features using previously described methods 33,35 . Gene expression was used to determine Ig light chain restriction.…”

“… a, Prediction scores of cell types assignment using a reference dataset 33,35 . Each cell is assigned a state, higher score indicates better assignment of this cell to the respective state.…”