Single-Cell Transcriptional Profiling of Zebrafish Hematopoiesis Offers Insight into Early Lymphocyte Development and Reveals Novel Immune Cell Populations

The advent of single-cell RNA sequencing (scRNA-seq) has greatly expanded our appreciation for cell state diversity beyond classical developmental hierarchies and simple population subsets. In particular, rich transcriptional heterogeneity has been observed within immune cell populations leading to the identification of novel cell types. Except for a few notable exceptions, intensive work in this area has largely been confined to mammals. Much of the transcriptional profiling of blood cell development in zebrafish has failed to capture early stages of lymphocyte development as the majority of research has not included thymus datasets or captured sufficient marrow B cells to explore developmental trajectories. To gain insight into T and B cell development in the zebrafish and immune cell diversity more broadly, we performed scRNA-seq using 10x Genomics Next GEM technology on adult zebrafish kidney marrows (n = 5 biological replicates) in addition to whole juvenile thymi at 4 weeks post-fertilization (wpf) (n = 4 technical replicates) and whole adult thymi at 3-4 months post-fertilization (n = 4 biological replicates). After filtering out low quality cellular barcodes, 34,492 kidney marrow cells and 35,268 thymus cells remained for analysis. With respect to T cell development, we identified putative early thymic progenitors from their clustering with hematopoietic stem and progenitor cells and shared transcriptional signatures, including the enrichment of CD34, CSF1R, FLI1, and DNMT3B human orthologs. Multiple subsets of thymic and marrow T cells were characterized, including a subset of gamma delta T cells readily identified by their expression of T-cell receptor gamma and delta chain components and expression of a SOX13 ortholog in addition to a Th2-like population expressing IL4, IL13, and GATA3 orthologs. Among other immune cell populations, rich transcriptional diversity was present. Two distinct populations of B cells, largely mutually exclusive for ighd and ighz expression (dual detection <1%), were present across all datasets, including the 4 wpf thymi, a surprisingly early time point in zebrafish B cell ontogeny. A clustering of the adult datasets demonstrated ighz predominance in the thymus (63% of thymic B cells), whereas the majority of marrow B cells (85%) fell within the ighd cluster. Stages of B cell development were also clearly evident, with the earliest B cell progenitors expressing orthologs of human PAX5, DNTT, RAG1, and RAG2, in addition to sid1, an understudied gene proposed to be orthologous to VPREB1. The expression of dntt was notably absent from more mature subsets of rag1 and rag2+ B cell progenitors, analogous to its expression in mammals. Transcriptional signatures unique but reminiscent of mammalian dendritic cell subsets were identified: Plasmacytoid-like dendritic cells characterized by high expression of TLR7, TLR9, and IRF8 orthologs and conventional-like dendritic cells characterized by high expression of CKB, BATF3, and ZNF366 were present in both marrow and thymus datasets, suggestive of greater dendritic-like cell diversity in the zebrafish than previously appreciated. These findings illustrate the power of single-cell transcriptional profiling for illuminating immune cell development and heterogeneity in the zebrafish, demonstrating increasing parallels to the mammalian system.