The Clinical Relevance of Acute Myeloid Leukemia Stem Cells

Abstract 240

Relapse of acute myeloid leukemia (AML) is hypothesized to reflect the failure of current therapies to adequately target leukemia stem cells (LSCs) - the rare, resistant cells presumed responsible for maintenance of the leukemia. These cells have generally been reported to have a similar phenotype to normal hematopoietic stem cells (HSCs). However, despite the considerable research on LSCs over the past two decades, the clinical significance of these cells remains uncertain. We hypothesized that any minimal residual disease (MRD) present after therapy would be enriched for LSCs and that such persistence of LSCs would predict relapse.

CD34⁺ cell populations from 27 AML patients and 10 normal donors were analyzed by flow cytometry for CD38 expression and for aldehyde dehydrogenase (ALDH) activity by Aldefluor. A total of 16 AML patients who achieved morphologic complete remission (CR) after induction chemotherapy were followed throughout their treatment course, and the flow cytometric staining patterns of their CD34⁺ cells at follow-up intervals were correlated with clinical outcomes. Cell subpopulations were sorted and then analyzed by fluorescence in situ hybridization (FISH) for leukemia-specific cytogenetic abnormalities (when applicable) and by transplantation into NOD/SCID-IL2Rg^null (NSG) mice to determine their in vivo self-renewal capacity.

Normal bone marrow CD34⁺CD38⁻ cells consistently exhibited two, discrete subpopulations by ALDH activity: one with low ALDH activity levels (CD34⁺CD38⁻ALDH^low) and another with high levels (CD34⁺CD38⁻ALDH^high). As few as 1000 CD34⁺CD38⁻ALDH^high cells generated normal hematopoiesis after transplantation into NSG mice. An additional population of CD34⁺CD38⁻ cells with intermediate (int) levels of ALDH activity was found in the AML patients, but not in any of the normal donors studied. When present, even in patients in cytogenetic CR, this CD34⁺CD38⁻ALDH^int population was at least 89% leukemic by FISH; and 1000 of these cells generated AML when transplanted into NSG mice. The CD34⁺CD38⁻ALDH^high cells were invariably present in small numbers in newly-diagnosed AML patients. These cells did not harbor the AML-specific FISH abnormality, and 1000 generated normal hematopoiesis when transplanted into NSG mice, consistent with a residual population of normal HSCs. In those AML patients who achieved CR, any detectable MRD was enriched for the CD34⁺CD38⁻ALDH^int leukemic cells. This population comprised 34% (range 9–51%) of the total leukemic burden when detectable in patients in cytogenetic CR, as compared to just 3% (range 0.5–4%) at initial diagnosis (p=0.02). Six of the seven CR patients with a detectable CD34⁺CD38⁻ALDH^int population ultimately relapsed; the lone exception underwent allogeneic transplantation in first CR. Conversely, all nine of the patients with a consistently undetectable CD34⁺CD38⁻ALDH^int population have remained in CR (p<0.01), with a median follow-up of 688 days.

Putative AML LSCs have a unique CD34⁺CD38⁻ALDH^int phenotype, which distinguishes them from normal (CD34⁺CD38⁻ALDH^high) HSCs. These LSCs appear to be more resistant to therapy than the bulk leukemic cells, as reflected by their relative enrichment in MRD after therapy. Furthermore, the persistence of these putative LSCs in patients after therapy was highly predictive of subsequent clinical relapse. Analysis of the ALDH activity of CD34⁺CD38⁻ cells potentially offers a patient-specific means by which to clinically assess response at the level of the LSC. The ability to separate LSCs from normal HSCs within the same individual may also facilitate better identification of therapeutic targets and resistance mechanisms. These data provide some of the first evidence supporting the clinical relevance of LSCs.