Clonal Tracking of the Geographic Distribution of Hematopoiesis in Nonhuman Primates Provides New Insights into HSPC Migration and Differentiation

Hematopoietic stem and progenitor cells (HSPC) primarily reside in bone marrow (BM) niches, but are also found in peripheral blood (PB) in small numbers. Large numbers of HSPC can be pharmacologically mobilized from the BM into the PB and home back to BM niches following transplantation. We used the rhesus macaque to study the process of hematopoiesis in both space and time following autologous transplantation, as a model with great relevance to humans. We labeled individual CD34+ HSPC and their progeny with genetic "barcodes" via lentiviral transduction with a very high diversity barcode library, allowing tracking of the output from individual HSPC clones in vivo following autologous transplantation, in a quantitative and sensitive manner (Wu et al., 2014). In this study we investigated the pattern of HSPC clonal output in various anatomic niches, sampling left (L) vs right (R) iliac crest BM, PB and lymph node (LN) over time post ablative autologous transplantation, for all major hematopoietic lineages and CD34+ HSPC.

We tracked thousands of HSPC clones in 6 rhesus macaques from 3.5 months (m) to up to 18.5m post transplantation. L and R iliac crest BM and PB were serially collected and CD34+ HSPC along with CD3+CD4+ T, CD3+CD8+ T, CD20+ B, CD14+ monocytes (Mo), CD33+ granulocytes (Gr), and two NK cell subsets (CD3-CD20-CD14-CD16+/ or CD56+) were purified. In all animals, there was marked geographic segregation of CD34+ HSPC for at least 6m post-transplant, with individual clones localized only to the L or R side but not both (Pearson correlation at 3.5-6m: r=0.019±0.08 for CD34+ L vs R, n=6), despite rapid expansion of CD34+ HSPC during early engraftment. This result suggests that during this phase of recovery, HSPC spread contiguously in the BM, and there is little geographic "mixing" via egress into and re-entry from the PB. With time between 7m and 18.5m post-transplant, the distribution of CD34+ HSPC became more homogeneous, with clones detected on both L and R (r=0.52±0.22 at 12-18.5m, n=3). The geographic restriction of clones at earlier time points suggests that release and homing may be dependent on local contiguous niche occupancy status, with migration only occurring following hematopoietic recovery. We next examined the clonal distribution of the more mature lineage-committed cells in the L vs R BM and in PB. Gr, Mo and B cells were produced locally in BM, with the clonal pattern of each lineage matching the CD34+ cells collected from the same side BM time point through 6m (r>0.80 for each lineage vs same side CD34+ cells), and distinct from same lineages (r<0.16 for each lineage in L vs in R BM) and CD34+ cells on the other side (r<-0.01 for each lineage vs other side CD34+ cells). CD16+CD56^-/dim NK clones were completely shared between PB and both L and R BM, suggesting they were not produced locally in BM, but instead in other sites with homogenous lodging or homing back to BM. Most surprisingly, we found population of CD3+CD8+ T cells that appeared to be produced locally, with barcodes matching the CD34+ HSPC at the same location, suggesting a novel T cell development pathway within the BM for this subset during early hematopoietic reconstitution. In contrast, CD3+CD4+ BM T cells had similar clonal constitution on the L vs R, and matched the PB, suggesting they had re-circulated back to the BM following maturation elsewhere, such as the thymus. T, B and NK cells from two LNs obtained simultaneously were also analyzed. Clonal contributions to T, B, and NK cells from L vs R LNs were highly correlated (r=0.95, r=0.88, and r=0.89 respectively). The clonal composition of T or B cells in LN were shared with circulating T or B cells (r=0.88, r=0.85 respectively), while LN NK cells (which are primarily CD56+/CD16-) shared barcodes with circulating CD16-CD56+ NK (r=0.81), not with PB CD16+CD56- cells(r=0.26), suggesting a non-precursor/progeny relationship.

Our model for the first time documents the dynamics of HSPC geographic distribution and migration in primates following transplantation, findings with direct clinical relevance, and provides new insights into hematopoietic lineage development, including a potential novel T cell development pathway in the BM. Our findings may also help explain the extremely patchy distribution of hematopoiesis in humans following transplantation or in the setting of marrow failure or aging, and suggest that analysis of individual BM samples may not fully reflect ongoing global hematopoiesis.