Distinct stage-specific transcriptional states of B cells derived from human tonsillar tissue
Diego A. Espinoza, Carole Le Coz, Emylette Cruz Cabrera, Neil Romberg, Amit Bar-Or, Rui Li
Diego A. Espinoza, Carole Le Coz, Emylette Cruz Cabrera, Neil Romberg, Amit Bar-Or, Rui Li
View: Text | PDF
Resource and Technical Advance Cell biology Immunology

Distinct stage-specific transcriptional states of B cells derived from human tonsillar tissue

Abstract

B cells within secondary lymphoid tissues encompass a diversity of activation states and multiple maturation processes that reflect antigen recognition and transition through the germinal center (GC) reaction, in which mature B cells differentiate into memory and antibody-secreting cells (ASCs). Here, utilizing single-cell RNA-seq, we identify a range of distinct activation and maturation states of tonsil-derived B cells. In particular, we identify what we believe is a previously uncharacterized CCL4/CCL3 chemokine–expressing B cell population with an expression pattern consistent with B cell receptor/CD40 activation. Furthermore, we present a computational method that leverages regulatory network inference and pseudotemporal modeling to identify upstream transcription factor modulation along a GC-to-ASC axis of transcriptional maturation. Our data set provides valuable insight into diverse B cell functional profiles and will be a useful resource for further studies into the B cell immune compartment.

Authors

Diego A. Espinoza, Carole Le Coz, Emylette Cruz Cabrera, Neil Romberg, Amit Bar-Or, Rui Li

×

Figure 3

Trajectory inference of tonsillar B cell scRNA-seq models transcriptional dynamics of a GC-to-ASC transition.

Options: View larger image (or click on image) Download as PowerPoint
Trajectory inference of tonsillar B cell scRNA-seq models transcriptiona...
(A) UMAP plot and cluster labels for the reanalysis (normalization, integration, dimensionality reduction) of the GC and ASC clusters. Slingshot trajectories (black) are overlaid on the UMAP plot. (B) Gene expression levels of XBP1 (log-normalized counts) overlaid on cells in the UMAP plot. (C) Pseudotemporal ordering (pseudotime) results for trajectory of interest. (D) Modeled expression values across increasing pseudotime (left to right) for the cellular trajectory shown in C. Heatmap shows genes deemed associated with pseudotime based on an adjusted P value of less than 0.01 from tradeSeq’s associationTest. Selected genes are labeled. Expression patterns were clustered using Manhattan distance of the modeled gene expression values across pseudotime and hierarchical clustering was performed with a 6-cluster cutoff. (E) Modeled regulon activity values across increasing pseudotime (left to right) for the cellular trajectory shown in C. Regulons with adjusted P values of less than 1 × 10–6 determined by fitting generalized additive models are shown. Expression patterns were clustered using Manhattan distance of the modeled gene expression values across pseudotime and hierarchical clustering was performed with a 6-cluster cutoff. (F) Generalized additive model results alongside regulon activity values are displayed for cluster 1 from E.