Clonal Tracking of the Source of Red Cell and Platelet Production in Rhesus Macaques

The classical model of hematopoietic hierarchies is being reconsidered, based on data from in vitro assays and single cell expression profiling. Recent experiments suggest that erythroid and megakaryocytic lineages might differentiate directly from multipotent hematopoietic stem/progenitor cells (HSPC) or from a highly biased subpopulations of HSPC, rather than transiting through a common MEP or CMP. We examined the clonal ontogeny of the erythroid lineage using genetic barcoding of rhesus macaque HSPC (Wu Cell Stem Cell, 2014; Koelle Blood, 2017), allowing quantitative and sensitive tracking of the in vivo clonal output of thousands of individual HSPC over time following autologous transplantation. CD34+ HSPC were lentivirally-transduced with a high diversity barcode library, with the barcode in an expressed region of the provirus, allowing barcode retrieval from DNA or RNA, with each barcode representing an individual HSPC clone.

CD34+ HSPC were purified from bone marrow(BM) of 3 macaques at 3-45 months post-transplant, and plated in CFU assays. 240 colonies each of CFU-E, CFU-G and CFU-GM were plucked individually, and each colony type was pooled before DNA extraction for barcode retrieval, along with purification and barcode retrieval from concurrent BM CD34+ cells and both blood and BM T cells (T), B cells (B), granulocytes (Gr), and monocytes (Mono). The majority of barcodes retrieved from pooled CFU-E were also detected in pooled CFU-G and CFU-GM, along with purified T cells, B cells, Mono and Gr, suggesting a shared unbiased precursor pool. A small fraction of clones unique to CFU-E were identified, however, unique clones were also detected in CFU-G and in CFU-GM pools, likely reflecting low frequency clones that were to be represented randomly in the pooled CFU of each lineage. To overcome the sampling bias inherent in colony assays on any reasonable colony number, we FACS purified CD71+/CD45- nucleated maturing erythroid lineage cells (nRBC) from the BM, and compared nRBC to other lineages purified concurrently from the same BM sample. There was very high correlation of barcode contributions between BM nRBC and other BM-produced lineages, with the highest correlation between nRBC and both Gr and Mono (r> 0.9), whether at earlier or later time points.

We investigated whether RNA barcode retrieval could be utilized for clonal tracking, allowing analysis of anucleate circulating RBC and thus a more global analysis of hematopoiesis compared to local BM production at a limited number of sites. We have reported that clonal output from individual HSPC remains highly geographically restricted within the BM for months-years post-transplant. We compared fractional contributions of DNA and RNA barcodes retrieved from the same sample of each lineage. There was very high correlation between DNA and RNA barcode contributions to T, B, NK, Gr and Mono lineages (r= 0.85±0.04), suggesting the differentiation pathway for these lineages does not impact significantly on expression level of barcodes from the proviral promoter, and RNA fractional contributions in these lineages reflect the clonal representation of cells in a sample. However, nRBC DNA and RNA barcode contributions were less closely related (r= 0.62), suggesting that erythroid differentiation was more likely to alter expression from loci in a manner disconnecting RNA barcode expression from clonal representation of cells in a sample, and suggesting that RNA barcode retrieval may not be ideal for comparing erythroid cells to other lineages. However, tracking RNA barcodes can be used to assess clonal stability in circulating RBC over time, and revealed very stable clonal contributions to erythropoiesis for as long as 4 years post-transplant.

Finally, we used RNA barcode retrieval to compare clonal contributions between circulating platelets and other lineages. Whether DNA or RNA was used for T, B, Gr, and Mono clonal mapping, at steady state platelet RNA barcodes were clonally closely related to other lineages. But preliminary data suggests that a unique set of clones is newly recruited to contribute only to platelets following inflammatory stimuli. The presence of a separate pool of platelet-biased HSPC contributing following inflammation has been suggested by prior in vitro assays, but our model may provide the first clonal in vivo confirmation of a unique inflammation-related platelet-biased HSPC pool.