The long-standing model of human haematopoiesis postulates that myeloid and lymphoid lineages branch separately at very early stages, producing myeloid or erythroid cells and T or B cells, respectively. Conversely, a revised scheme of haematopoietic hierarchy was recently proposed, in which myeloid cells represent a prototype of blood cells, while erythroid, T and B cells are specialized cell types. The validity of these models has been mainly tested in vivo in the mouse, and in vitro through clonal assays on human haemopoietic stem cells (HSC). However, a clear definitive elucidation of the real nature of human haemopoiesis should ideally involve the ability to track the dynamics, survival and differentiation potential of haemopoietic progenitor clones for a long period of time directly in vivo in humans. Upon retroviral gene transfer, transduced cells are univocally tagged by vector insertions allowing them to be distinguished and tracked in vivo by integration profiling. We previously showed that gene therapy (GT) for adenosine deaminase (ADA) deficient SCID based on infusion of transduced CD34+ cells and reduced intensity conditioning, resulted in full multilineage engraftment, in the absence of aberrant expansions. Therefore, long-term studies in these patients provide a unique human model to study in depth haemopoietic clonal dynamics by retroviral tagging. For this reason, we performed a comprehensive multilineage longitudinal insertion profile of bone marrow (BM) (CD34+, CD15+, CD19+, Glycophorin+) and peripheral blood (PB) (CD15+, CD19+, CD4+, CD8+ cells, naïve and memory T cell subpopulations) cells in 4 patients 3–6 years after GT, retrieving to date 1055 and 1999 insertions from BM and PB cell lineages respectively. We could shape the insertional landscape of each lineage through a tri-factorial analysis based on the number of integrations retrieved, the percentage of vector positive cells and the number of insertion shared with other lineages. We were able to uncover the effects of selective advantages of gene-corrected cells in periphery and the frequency of identical integrants in different haematopoietic compartments. BM cells displayed the highest proportion of shared integrants (up to 58.1%), reflecting the real-time repopulating activity of gene-corrected progenitors. On the other hand, PB samples carried in general a higher frequency of vector positive cells, with the exception of PB CD15+ cells showing insertional landscapes very similar to the one of BM lineages. Interestingly, the detection of exclusively shared myeloid-T\B or myeloid-erythroid integrants may be supportive of a myeloid-based haemopoiesis model. We also uncovered “core integrants”, shared between CD34+ cells and both lymphoid and myeloid lineages, stably tagging active long-term multipotent progenitors overtime. Strikingly, one of these progenitor clones carried an insertion inside one of the two fragile sites of MLLT3 gene, involved by translocation events in mixed lineage leukemia. We were able to track this and another integrant (downstream the LRRC30 gene) by specific PCRs, confirming the multilineage contribution to haematopoiesis of the relative progenitor clones and their fluctuating lineage outputs over 4 years, without showing aberrant expansions. We also retrieved 170 and 174 integrations from 4 T cell subtypes (Naive, TEMRA, Central and Effector memory) in two patients under PBL-GT and HSC-GT respectively. We found evidences that single naive T cell clones may survive in patients for up to 10 years after last infusion while maintaining their differentiation capacity into different T cell subpopulations. Interestingly, a cluster of 4 insertions (one of them shared among all T cell subtypes) was found in proximity of the interferon regulatory factor 2 binding protein 2 (IRF2BP2) gene in naive T cells from PBL-GT patient, thus suggesting an influence of transcriptional activity of this region on selective advantage of gene-corrected lymphocytes. In conclusion, through retroviral tagging, we can uniquely track single transduced haemopoietic cells directly in vivo in humans. The application of mathematical models to our insertion datasets is allowing to uncover new information on the fate and activity of haematopoietic progenitors and their differentiated progeny years after transplantation in GT patients.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.