Single-Cell Revelation of Extensive Heterogeneity in the Human Lymphocyte and Neutrophil/Monocyte Lineage Restriction Process Despite Highly Co-Regulated Changes in Cell Cycle Control

The production of mature blood lineages in humans has been viewed as a hierarchical process in which the primitive cells with multi-lineage potential generate a series of progenitors with increasingly reduced self-renewal abilities and restricted lineage potentials. This concept has been largely supported by phenotypic characterization of starting CD34+ subsets of human CD45+ cells isolated from different hematopoietic tissues and assayed for their mature cell outputs in supportive systems in vitro or post-transplant and more recently by tracking changes in their molecular composition. Such experiments have suggested that cells that give rise to the lymphocyte lineages (L) share a common origin almost exclusively with the neutrophil and monocyte (macrophage) lineages (NM). However, the mechanisms that regulate the subsequent separation of these lineages starting from cells with presumed restricted L+NM potential have remained elusive.

To address this question, we first developed a culture system that supports outputs of differentiated B (CD19+) and NM (CD15+/CD14+) from ≥50% of single CD34+ cells plated on a stromal layer of MS-5 feeders in a medium containing human IL-7, SCF and FLT3L. Use of this clonal assay revealed that cells with B+NM bi-potential constitute 20% of the CD45RA-CLEC12A- (RA-C-) cell subset of the CB “P-Mix” phenotype (CD34+CD38medCD71-CD10-, Blood 133:927, 2019); with the CD45RA+CLEC12A- (RA+C-) and CLEC12A+ (C+) cells displaying largely restricted outputs of only B- (30%) or only NM- (50%) progeny phenotypes, respectively. These same 3 “progenitor” phenotypes (RA-C-, RA+C-, C+) were also found to be regenerated in xenografted immunodeficient mice as well as in a further optimized in vitro system and their clonal output types and numbers consistent with the outputs of those isolated from fresh CB. To analyze T-cell lineage (T) restriction in parallel, we then further modified this culture system to include adherent DLL4, that we found allows B, T, N and M progeny to be produced from single input CB cells at a 40-60% efficiency, with numerous clones derived from single RA+C- cells containing B+T+M, or B+M, or B+T or T-only; a few with B- or M- only, and none with N. In contrast, under the same conditions, half of the documented C+ outputs were exclusively N with much lower lymphoid outputs of T-only or T+N/M. Together, these findings reveal the unanticipated importance of the external stimuli in revealing lineage potentials when these are undergoing restrictive changes.

We then asked if and how these identified early phenotypes might change their cycling properties during the course of their progression. For this, the rate and progeny content of sequential cell divisions during the stimulated expansion of the CFSE labelled RA-C- input cells were measured under a supportive liquid culture condition. Throughout a 6-day period, almost all of the C+ progeny and a slightly lower proportion of the RA+C- outputs were generated from cells that transit through the cell cycle within 24 hours, whereas cells maintaining an undifferentiated RA-C- phenotype divided at significantly slower rates, suggesting a simultaneously initiated shortening of the cell cycle transit time during both L- and NM- restriction. In agreement with these functional results, a recent single-cell CITE-seq dataset generated in parallel showed higher expression of genes involved in cycling activities in the early L- and NM-restricted subsets as compared to the unrestricted RA-C- cells.

Taken together, these results reveal important and disconnected features in the cell cycle control and previously unrecognized complex patterns in the biological outputs and molecular changes during the steps of L and NM restriction. In addition, they set the stage to enable preferred sequences and molecular changes responsible for normal and regenerative needs to be identified and their potential relevance to the formation of human lymphoblastic and myeloid leukemias to be characterized.