INTRODUCTION: Normal hematopoietic stem cells (HSC) are characterized by their ability to self-renew, to generate multiple cell-lineages, and show slow divisional kinetics. Leukemic stem cells (LSC) have been reported to show similar characteristics but their identification has been elusive. We have studied various methods and have identified aldehyde dehydrogenase (ALDH) staining as an optimal method for the enrichment of primary human LSC.

MATERIAL&METHODS: Bone marrow samples were obtained from patients with newly diagnosed AML after informed consent. Mononuclear cells were stained with Aldefluor and sorted by flow cytometry according to their forward/side scatter characteristics and ALDH activity (ALDH+/ALDH−). Alternatively, primary AML samples were being enriched for CD34+ cells by magnetic column, then double-stained with CD34-antibodies and Aldefluor and sorted for the co-expression of CD34+ and ALDH+, respectively for CD34+ alone. Human Mesenchymal Stromal Cells (MSC), isolated from human bone marrow, were used as a surrogate model for the cellular microenvironment of the hematopoietic niche. Adhesion of various AML cell lines and subpopulations of primary leukemic cells (ALDH+, ALDH−, CD34+, CD34+/ALDH+, all blasts) to MSC was tested in the adhesion chamber assay. Semi-quantitative RT-PCR was used to analyze the gene expression of various adhesion molecules and Western- Blot analysis was performed to validate the PCR-results on protein level. The generation of secondary leukemic colonies was evaluated in a semi-solid methylcellulose medium, as well as in a long term co-culture system (LSC-IC assay; in analogy to the LTC-IC assay).

RESULTS: The percentage of ALDH+ cells ranged from 0.01% to 13.2% with a median of 1.47% (n=55). Adhesion significantly differed in the ALDH+ and ALDH− subpopulations: 85±4% of ALDH+ cells but only 61±8% of ALDH− cells were adherent (n=11, p<0.001). Adhesion molecules, such as CXCR4 and CD44, were highly expressed on the ALDH+ subpopulation both on mRNA level and protein level, in contrast to the ALDH− subpopulation. Analysis of the initial divisional kinetics on single cell base showed that the ALDH+ subpopulation contained more slow dividing cells whereas the majority of the ALDH− subpopulation consisted of fast-dividing cells (n=3; p<0.01). The frequency of long term leukemic colony initiating cells (LSC-IC) was 3.82% in the ALDH+ but only 0.01% in the ALDH− (n=21; p<0.01). In the CD34+ the LSC-IC frequency was 1.96% versus 0.01% in the CD34− (n=5, p<0.01). The highest LSC-IC frequency could be monitored in ALDH+/CD34+ cells: 6.1% generated secondary leukemic colonies (n=5). These colonies, harvested after 7 weeks of cultivation, were examined for their immune phenotype and screened for cytogenetic aberrations by fluorescent in situ hybridization (FISH) and the chromosomal aberrations were consistent with the AML samples taken at diagnosis. Furthermore, the frequency of ALDH+ cells correlated significantly with adverse prognostic factors: patients with a high-risk karyotype had a mean of 2.9% ALDH+ cells (n=21); in contrast, patients with a normal karyotype had a mean of 0.4% ALDH+ cells in their bone marrow (n=34; p<0.001). The ability of ALDH+ versus ALDH− subsets to generate secondary leukemia in the animal model is concurrently examined.

DISCUSSION: In summary, measurement of the ALDH activity provides a useful tool for the isolation of a distinct AML-blast subpopulation with stem-cell like features (LSC). The ALDH+ subsets showed higher affinity to the surrogate niche (MSC), elevated expression of CD44, Cadherin-2, and CXCR4 and were associated with an increased frequency of secondary leukemic colonies in vitro (LSC-IC). Above all, the frequency of ALDH+ blasts correlated with clinical prognostic factors, which substanciates LSC as a relevant therapeutic target.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author

Sign in via your Institution