Human neutrophil kinetics: modeling of stable isotope labeling data supports short blood neutrophil half-lives

Abstract

Human neutrophils have traditionally been thought to have a short half-life in blood; estimates vary from 4 to 18 hours. This dogma was recently challenged by stable isotope labeling studies with heavy water, which yielded estimates in excess of 3 days. To investigate this disparity, we generated new stable isotope labeling data in healthy adult subjects using both heavy water (n = 4) and deuterium-labeled glucose (n = 9), a compound with more rapid labeling kinetics. To interpret results, we developed a novel mechanistic model and applied it to previously published (n = 5) and newly generated data. We initially constrained the ratio of the blood neutrophil pool to the marrow precursor pool (ratio = 0.26; from published values). Analysis of heavy water data sets yielded turnover rates consistent with a short blood half-life, but parameters, particularly marrow transit time, were poorly defined. Analysis of glucose-labeling data yielded more precise estimates of half-life (0.79 ± 0.25 days; 19 hours) and marrow transit time (5.80 ± 0.42 days). Substitution of this marrow transit time in the heavy water analysis gave a better-defined blood half-life of 0.77 ± 0.14 days (18.5 hours), close to glucose-derived values. Allowing the ratio of blood neutrophils to mitotic neutrophil precursors (R) to vary yielded a best-fit value of 0.19. Reanalysis of the previously published model and data also revealed the origin of their long estimates for neutrophil half-life: an implicit assumption that R is very large, which is physiologically untenable. We conclude that stable isotope labeling in healthy humans is consistent with a blood neutrophil half-life of less than 1 day.

Figure 1

Physiology of granulocyte turnover translated to a model. We consider mitotic neutrophil precursors as a single pool, Np (proliferating neutrophils). The Np pool consists mainly of myelocytes, promyelocytes, and myeloblasts but also contains earlier precursors. Cells in the Np population proliferate at a mean rate p. After the last mitosis, cells enter the postmitotic maturation/transit pool at a rate q. Transit neutrophils remain in the postmitotic pool for a period of 4 to 6 days,^,, referred to as the transit time (Δ), before egressing from bone marrow into the blood pool. Most egressing cells are segmented neutrophils, but some cells leave as band neutrophils. In the blood, neutrophils exist both in a freely circulating pool and a marginal pool (cells retained in proximity to the endothelium). Because the circulating and marginal pool are considered to be in rapid dynamic equilibrium,^, we consider them a single kinetically homogeneous pool, N_B. Blood neutrophils are lost from the circulation at a rate z primarily to the bone marrow, liver, and spleen.^- This loss is generally considered to be a random, irreversible process.^,,,, GMP, granulocyte-monocyte progenitor cells; HSC, hematopoietic stem cell.

Figure 2

Best model fits to neutrophil DNA enrichment data in subjects labeled with deuterium-labeled glucose. Subjects were labeled with deuterium-labeled glucose (6,6-²H₂-glucose) for 10 hours (C36R and C41), 4 hours (C42), 3 hours (C46, C48, C49, and C50), or as a single oral dose (C59 and C60). Circles represent measured fractional enrichment of deoxyadenosine in blood neutrophil DNA; lines represent best fits of the model (Equations 5 and 6) to the data.

Figure 3

Best model fits to neutrophil DNA enrichment data in subjects labeled with heavy water. Subjects were labeled with heavy water (²H₂O) for 9 weeks (subjects A, B, C, D, E from Vrisekoop et al) or 7 weeks (subjects DW04, DW09, DW10, and DW11 from the current study’s data). Circles represent measured fractional enrichment of deoxyadenosine in blood neutrophil DNA; lines represent best fits of the model (Equations 5 and 6) to the data.