Doubts concerning the recently reported human neutrophil lifespan of 5.4 days

To the editor:

Using orally administered deuterium-labeled water to label neutrophils in vivo, Pillay et al measured urinary and blood deuterium-to-proton enrichment ratios at intervals of ∼ 12 days over several weeks before and after termination of intake.¹  From mathematical modeling they obtained a median peripheral blood (PB) human neutrophil lifespan of 5.4 days.

This result is highly surprising in view of the range of neutrophil labeling techniques, including whole blood labeling with DFP-32² ; in vivo H-3 labeling followed by blood transfusion³ ; and in vitro labeling of purified neutrophils with Cr-51,⁴  In-111,⁵  or Tc-99m,⁵  which have all previously given a PB lifespan of ∼ 10 hours. It is also starkly at odds with the clinical observations of rapid neutrophil depletion after myeloablative chemotherapy (within 3-5 days) and short-lived normalization of neutrophil counts following therapeutic granulocyte infusion (1-2 days).

We wish to raise several objections to the work of Pillay et al.

1. Bone marrow is a major site of neutrophil destruction as well as development,⁶  but there is no discussion on the availability of deuterium-labeled adenosine for salvage that might result in reutilization in newly dividing neutrophil precursors.

2. The authors claim their previous work on human lymphocytes,⁷  based on similar sampling intervals, validates their modeling, but lymphocytes had a reported lifespan of > 100 days, long enough to be measurable from the sampling interval used.

3. The reported peak (63-day) enrichment in blood (4.3× urine value) is unexplained. Urine enrichment is consistent with the deuterium dose (see supplemental Data, available on the Blood Web site; see the Supplemental Materials link at the top of the online letter), but blood enrichment is inexplicably high. The authors introduced an “amplification factor” c between urine and blood enrichment ratios, attributing this to the multiple hydrogen atoms in a single adenosine deoxyribose moiety. However, the deuterium enrichment ratio at a particular hydrogen location in a molecule is unlikely to depend on the number of hydrogen locations in a molecule, so the high blood enrichment remains unexplained.

4. A possible explanation for the high deuterium enrichment in blood may lie in the fact that reactions involving the cleavage of bonds to hydrogen occur at a faster rate compared with deuterium (∼ 7× faster for C-H compared with C-D),⁸  which results in a longer residence time for deuterium.⁹  This is known as the kinetic isotope effect.⁸  If this effect is important during neutrophil development, it may explain the elevated blood deuterium enrichment, but at the same time would cast doubt over almost any kind of modeling, unless the mechanism and location of deuterium enrichment were well understood.

We present an alternative, simpler model that critically assumes a neutrophil PB lifespan that is negligible relative to the development time of neutrophils in marrow. Thus, instead of using 3 parameters (ie, the scaling factor [a] for the difference between urine and blood deuterium enrichments, the time of neutrophil development in marrow and the PB lifespan), we found that the data in the authors' published figures could be equally well fitted using only 2 parameters (a and a simple delay between urine and blood enrichment curves; see their supplemental data). This model gave small residuals, no evidence of systematic deviation from the data, and a delay of 11.2 days (Figure 1). Interestingly, the fit was improved using a distribution of delays, which, although a more likely scenario, appears not to have been considered by the authors.

In conclusion, a PB neutrophil lifespan ∼ 13× longer than the currently accepted value is highly unlikely. The authors' conclusion is based on a mathematical model that attempts to resolve total neutrophil lifespan into bone marrow transit time and PB lifespan. Our alternative, simpler model, which assumes a negligible PB lifespan, fits their data just as well as their own model does. The authors' deuterium technique may be sufficient to measure total neutrophil lifespan, and indeed the authors' mean value of 11.1 days (4.8 days in marrow and 6.3 days in PB) is similar to the value given by our model and in line with published values for neutrophil development time in marrow.¹⁰  It does not, however, have the temporal resolution to measure a PB lifespan of 10 hours.