Estimating the Replication Rate of Hematopoietic Stem Cells (HSC) In Vivo.

Abstract 2522

Poster Board II-499

The replication kinetics of HSC in vivo is difficult to assess because HSC are infrequent, reside in marrow niches, and are regulated by extrinsic as well as intrinsic signals. Determining the replication (self-renewal) rate of human HSC in vivo is especially difficult because limiting dilution competitive transplantation studies and studies with cell division-sensitive markers are not feasible. Therefore, we developed three surrogate methods by extending observations in other species. These approaches use different data and different assumptions yet yield overlapping estimates and together suggest that human HSC replicate on average once per 40 weeks. This is thus much slower than replication rates of HSC in mouse (∼ once per 2.5 wks), cat (∼ once per 8.3 -10 wks) and nonhuman primate (∼ once per 25 wks) (reviewed in ¹).

Specifically, we analyzed the drift in the X-chromosome phenotype of blood cells from 1219 females (ages 18 -100, mean 56±22; Montreal cohort) assuming that HSC expressing X-alleles from one parent might divide slightly faster than HSC expressing X-alleles from the other parent and that over time this subtle growth advantage would lead to the phenotypic skewing of HSC and progeny cells. We calculated that the drift that occurs with aging in the X-chromosome phenotype of granulocytes in female Safari cats (F1 offspring of Geoffroyi (G) x domestic (d) cats) can be explained by a 5% difference in the replication rates of HSC expressing G vs. d X-chromosomes. Knowing that G and d cats evolved independently for > 9 million yrs, we simulated human hematopoiesis using a Markovian description of HSC differentiation and the single constraint that the differences in the replication rates of HSC expressing paternal vs. maternal-derived X-chromosomes in individual human females (150,000 yrs of genetic distance) would be less than 5%. For each replication rate (λ), we generated 100 sets of 1219 virtual females (the replication rates of HSC expressing maternal and paternal X-chromosomes in each individual were randomly drawn from a distribution with median λ and variance as observed in Safari cats) and determined if the pattern of X-chromosome phenotype in their blood cells with aging resembled the Montreal data. If yes, λ was included as a possible human HSC replication rate and if no, it was excluded. Using this approach, we defined plausible values for λ, then confirmed the results by a comparable analysis of second dataset (London cohort; 117 females, ages 18–96, mean 66±24). Remarkably, this value (1 per 40 wks; range 1 per 25–50 wks) is also consistent with estimates derived by two independent methods: analysis of granulocyte telomere length with aging² and application of the concept that the total number of HSC divisions during a mammal's lifetime is evolutionarily conserved (data not shown).

We next demonstrated that the estimate was reasonable by simulating human marrow transplantation. When 100 HSC are transplanted (corresponds to 1.9 ×10⁸ marrow cells (MC)/kg, 70 kg recipient), all virtual recipients engraft, consistent with the clinical recommendation that >2–3 × 10⁸ MC/kg be transplanted. Also, when 20 HSC are transplanted (3.9 ×10⁷ MC/kg), graft failure is common (occurs in 50% of simulations). Of interest, these virtual recipients run out of short-term repopulating cells (STRC) within 30 wks, though HSC persist. There are several clinically relevant corollaries to this latter observation. First, children (small size, fewer mature blood cells) tolerate transplantations of fewer HSCs better than adults because few STRC can produce safe numbers of granulocytes, red cells and platelets. Second, supplementing transplantations with infusions of multipotent progenitor cells, such as present in cytokine or Notch ligand Delta1-expanded cord blood, should be an efficacious method to insure adequate hematopoiesis when the numbers of HSC are low and STRC are especially low. The data also raise the possibility that human marrow failure syndromes such as aplastic anemia and myelodysplasia could result from defective HSC commitment or defective expansion of differentiating STRC clones, and not only from the destruction or depletion of HSC. These events may be difficult to study in mice where simulations predict graft failure occurs from HSC depletion (data not shown), justifying the need for larger animal (cat, dog, primate) or human investigation.

¹Blood 110:1806,2007. ²Exp Hematol 32:1014,2004.