New Fine Structure within the Adult Hematopoietic Stem Cell Compartment Revealed by Longterm Analysis of Mice Repopulated with Single Cells or Their In Vitro Generated Clonal Progeny.

Assessments of hematopoietic stem cell (HSC) repopulating activity in vivo have historically relied on calculated average longterm (12–16 wk) progeny outputs using non-purified transplants, thereby precluding definitive clonal assignments of donor-derived cells. Viral marking circumvents this problem, but has not been used for large scale surveys. Heterogeneity observed in the repopulation patterns has generally been inferred to reflect stochastic processes. We now report the in vivo repopulation kinetics of 89 individual longterm repopulating cells (LTRCs) before (n=49) and after (n=40) 4 days of clonal growth in vitro. LTRCs were defined here as cells whose WBC progeny could be detected at levels of ≥1% for at least 16 wk in sublethally irradiated Ly5-congenic W⁴¹/W⁴¹ hosts. Recipients were transplanted with either freshly isolated, single lin⁻Rho⁻SP LTRCs or 4-day clones generated from similar cells in serum-free cultures (+ 300 ng/ml SF, 20 ng/ml IL-11 & 1ng/ml Flt3-L). 4, 8, 12, 16, and 24 wk post-transplant, blood samples were stained for donor-derived B, T, and myeloid cells using a procedure that identifies donor/recipient doublets and Ly6g/Mac1^low cells (which have features of lymphoid rather than myeloid WBCs) to exclude false-positive myeloid events. Four distinct patterns of repopulation were revealed. Type 1 showed a delayed production of predominantly myeloid WBCs (low or undetectable before 12 wk) that increased progressively (reaching 0.4–15% of all WBCs by 16 wk). Type 2 showed a robust multilineage repopulation that remained stable or increased over time (6–84% of all WBCs at 16 wk). Type 3 also showed an initially robust pattern of multilineage repopulation (peak numbers of WBCs at 8–12 wk and 1–51% at 16 wk), but the contribution of donor-derived myeloid cells was transient (<0.5% by 16 wk). Type 4 showed a lymphoid-restricted pattern (myeloid contribution <0.5% at all time points), with repopulation levels peaking at 8 wk and decreasing thereafter (1–22% at 16 wk). Persisting granulopoiesis, indicated by a high proportion of donor-derived cells in the Ly6g/Mac1⁺SSC^hi population at 16–24 wk, clearly distinguished the type 1 and 2 patterns from types 3 and 4 which showed near or complete cessation of donor-derived granulopoiesis beyond 12 wk. Preliminary secondary transplant experiments show that donor-derived LTRCs (with and without longterm granulopoietic activity) were exclusively generated in primary recipients with type 1 and 2 repopulation patterns. Amongst the freshly isolated LTRCs, 18% (9/49) were type 1, 41% (20/49) were type 2, 22% (11/49) were type 3, and 18% (9/49) were type 4. In contrast, 4-day clones derived from cells of the same phenotype and containing LTRC activity showed a marked decrease in type 1 and type 2 activity with a corresponding increase in type 3 and type 4 activity: type 1 = 5% (2/41), type 2 = 18% (7/40), type 3 = 28% (11/40) and type 4 = 50% (20/40). Collectively, these data identify a new hierarchy of four biologically discrete states within the compartment of cells currently defined as LTRCs. Proliferation of LTRCs either in vitro or in vivo appears to induce an irreversible transition from one state to another (from Type 1 to 2 to 3 to 4), suggesting the existence of intrinsic molecular correlates for each of these states and specific mechanisms that underlie their sequential appearance.