Lineage-affiliated specification events and molecular characterization during early stages of haematopoietic development

The well-established “classical” model of haematopoiesis reflects a hierarchical decision-making process where early multipotent progenitors make an irrevocable decision to differentiate towards either the lymphoid or myeloid lineages through so-called Common Lymphoid and Common Myeloid progenitor intermediates respectively. However, the proposals for alternative differentiation pathways and the description of progenitor cells that contradict the lympho-myeloid dichotomy have prompted multiple revisions of the strict compartmentalized classical model. We have previously characterized a B220+ CD117int CD19- NK1.1- uncommitted and multipotent haematopoietic progenitor with combined lymphoid and myeloid potential that we called Early Progenitor with Lymphoid and Myeloid potential (EPLM). The emergence of high throughput methods enabling the investigation of single-cell whole-transcriptome profiles generates data that enhances the active debate regarding the heterogeneity of apparently phenotypically homogenous progenitors having different multiple lineage potentials. This thesis provides a detailed analysis of EPLM heterogeneity by combining the alternative and complementary “top-down” and “bottom-up” experimental designs. Using the “top-down” approach based on the expression of the cell-surface markers Ly6D, SiglecH and CD11c, we could subdivide EPLM into four subpopulations with distinct lineage biases. As revealed by the subsequent functional experiments, the Ly6D+ EPLM fraction was lymphoid restricted and contained most B-cell potential whereas the so-called triple negative (TN) EPLM expressing none of the above markers remained as a lympho-myeloid multipotent fraction and the potential precursor of the Ly6D+ subset. Subsequently, single-cell RNA sequencing (“bottom-up” approach) of 152 Ly6D+ and 213 TN single cells revealed that in fact TN are composed of a mixture of cells where the myeloid potential is mainly due to the contribution of the G3 TN subset whereas the lymphoid potential resides in the G1 TN clustered group of cells. This heterogeneity was masked in previous bulk molecular and functional experiments, thus demonstrating the power of single-cell RNA-sequencing technology to study heterogeneity in haematopoietic progenitors at an unprecedented resolution. Moreover, single-cell transcriptome profiles enabled the detection in an unbiased manner of markers that better define cellular identity. Here we redefined the “top down” EPLM classification by identifying Terminal deoxynucleotide Transferase (TdT) as a potential marker with which to discriminate the lymphoid and myeloid potential of EPLM since, in addition to the previously identified lymphoid primed Ly6D+ cells, TdT is also expressed in the G1 TN fraction, which turned to be molecularly indistinguishable from the G2 Ly6D+ fraction. The use of other candidate markers such as Ebf1 and CD115 enabled us to prospectively isolate cells from different newly identified subgroups of EPLM and to confirm their genetic signatures with functional assays, thus supporting the increasing belief that the repertoire of genes expressed reflects the immediate lineage bias of that cell. Finally, within the Ly6D+ cells, we found a B-cell priming gradient and propose that the G1 Ly6D+ fraction is the direct precursor of the first B-cell committed stage, the CD117+ CD19+ Pro-B cell. Therefore, we favour the concept that haematopoiesis occurs through a process of graded commitment where molecular priming is initiated earlier than previously anticipated. Overall, this study provides a valuable model demonstrating that previously characterized, phenotypically homogeneous, multipotent progenitor cells are in fact composed of mixtures of cells with differently restricted differentiation capacities.