While many aspects of blood production are now well understood, the spatial organization of myeloid differentiation in the bone marrow (BM) remains unknown. Here, we use imaging to track granulocyte/macrophage progenitor (GMP) behavior during emergency and leukemic myelopoiesis. At steady state, we find individual GMPs scattered throughout the BM. During regeneration, we observe expanding GMP patches forming defined GMP clusters, which, in turn, locally differentiate into granulocytes. We describe how the timed release of important BM niche signals (SCF, IL-1β, G-CSF, TGF-β, CXCL4) and activation of an inducible Irf8/β-catenin progenitor self-renewal network controls the transient formation of regenerating GMP clusters. In leukemia, we show that GMP clusters are constantly produced due to persistent activation of the self-renewal network and lack of termination cytokines that normally restore stem cell quiescence. Our results uncover a previously unrecognized dynamic behavior of GMPs in situ, which tunes emergency myelopoiesis and is hijacked in leukemia.

No abstract

Each recording underwent a second analysis at 2 weeks following the first analysis to evaluate reproducibility. The effect of data sampling (5-min segment/hour), the system sensitivity to detect 5 ms increase in Q-T, and the ability to assess circadian variation were also evaluated. Results: The fully AQA resulted in identical QT for the first and second analyses, but with obvious errors in Q-T measurements. Compared to the complete onscreen MOR, the 24-hour mean Q-T was longer with AQA (416 Ϯ 41 vs 387 Ϯ 30 ms, p < .001, r = .3). The reproducibility of automatic analysis with complete MOR was very good (Q-T: 387 Ϯ 30 vs 387 Ϯ 30 ms), coefficient of variation (CV) = 0.2%, r = .986, p < .001. The 5-minute mean Q-T intervals correlated well with the hourly mean Q-T intervals (r = .994, p < .001, CV = 1 ms) and both showed a similar circadian variation. The system was sensitive to detect a 5 ms change in Q-T intervals (5 Ϯ 2 ms, CV = 0.6%, r = .998, p < .001). Conclusions. The fully automatic Q-T analysis is not an acceptable method, while the automatic analysis with MOR is a highly sensitive and reproducible method. Data sampling by analyzing 5-minute segments per hour is also sensitive and reproducible. The 12-lead digital Holter technique is suitable for Q-T analysis and may have advantage compared to the serial recordings of large number of standard 12-lead ECGs in the evaluation of drug effects on the Q-T interval.