The relationships between normal and leukemic stem/progenitor cells are unclear. We show that in ∼80% of primary human CD34+ acute myeloid leukemia (AML), two expanded populations with hemopoietic progenitor immunophenotype coexist in most patients. Both populations have leukemic stem cell (LSC) activity and are hierarchically ordered; one LSC population gives rise to the other. Global gene expression profiling shows the LSC populations are molecularly distinct and resemble normal progenitors but not stem cells. The more mature LSC population most closely mirrors normal granulocyte-macrophage progenitors (GMP) and the immature LSC population a previously uncharacterized progenitor functionally similar to lymphoid-primed multipotential progenitors (LMPPs). This suggests that in most cases primary CD34+ AML is a progenitor disease where LSCs acquire abnormal self-renewal potential.

The blood system is maintained by a small pool of haematopoietic stem cells (HSCs), which are required and sufficient for replenishing all human blood cell lineages at millions of cells per second throughout life. Megakaryocytes in the bone marrow are responsible for the continuous production of platelets in the blood, crucial for preventing bleeding--a common and life-threatening side effect of many cancer therapies--and major efforts are focused at identifying the most suitable cellular and molecular targets to enhance platelet production after bone marrow transplantation or chemotherapy. Although it has become clear that distinct HSC subsets exist that are stably biased towards the generation of lymphoid or myeloid blood cells, we are yet to learn whether other types of lineage-biased HSC exist or understand their inter-relationships and how differently lineage-biased HSCs are generated and maintained. The functional relevance of notable phenotypic and molecular similarities between megakaryocytes and bone marrow cells with an HSC cell-surface phenotype remains unclear. Here we identify and prospectively isolate a molecularly and functionally distinct mouse HSC subset primed for platelet-specific gene expression, with enhanced propensity for short- and long-term reconstitution of platelets. Maintenance of platelet-biased HSCs crucially depends on thrombopoietin, the primary extrinsic regulator of platelet development. Platelet-primed HSCs also frequently have a long-term myeloid lineage bias, can self-renew and give rise to lymphoid-biased HSCs. These findings show that HSC subtypes can be organized into a cellular hierarchy, with platelet-primed HSCs at the apex. They also demonstrate that molecular and functional priming for platelet development initiates already in a distinct HSC population. The identification of a platelet-primed HSC population should enable the rational design of therapies enhancing platelet output.

An internal tandem duplication in the fms-like tyrosine kinase 3 gene (FLT3/ITD) is associated with poor prognosis in acute myeloid leukemia (AML), but the impact of mutant level, size, and interaction with nucleophosmin 1 (NPM1) mutations remains controversial. We evaluated these characteristics in a large cohort of young adult AML patients. There was a highly significant trend for worsening in relapse risk (RR) and overall survival (OS) with increasing FLT3/ITD mutant level (P < .001 for both), and even in the low level mutant group (1%-24% of total FLT3 alleles), RR was significantly worse than in the FLT3 wild-type (WT) group (P < .001). In multivariate analysis, mutant level was the most powerful prognostic factor for RR. Mutant size and number had no significant impact on outcome. The beneficial impact of an NPM1 mutation on RR and OS was seen in FLT3/ITD(+) as well as FLT3/WT patients; both markers were highly significant independent predictors of outcome (P < .001). Stratification using both markers identified 3 prognostic groups: good (FLT3/ITD(-)NPM1(+)), intermediate (FLT3/ITD(-)NPM1(-) or FLT3/ITD(+)NPM1(+)), and poor (FLT3/ITD(+)NPM1(-)). Patients with high FLT3/ITD mutant level (greater than 50%) or FLT3/ITD(+) in the absence of an NPM1 mutation may be good candidates for more experimental therapeutic approaches.