Expression of C-Type Lectin-Like Molecule-1 (CLL-1) on Stem Cells Might Discriminate De Novo Acute Myeloid Leukemia from Acute Myeloid Leukemia Originating from Myelodysplastic Syndromes

It is generally accepted that myelodysplastic syndromes (MDS) most often originate in a multipotent, myelorestricted progenitor population, although primary transformation may occur at the hematopoietic stem cell level. MDS can be classified into low risk and high risk with evolution to acute myeloid leukemia (AML) predominantly in the latter cases. In AML, survival of leukemia-initiating cells, often referred to as leukemic stem cells, after chemotherapy is thought to lead to minimal residual disease and relapse. Hence, in de novo AML a larger size of the stem cell compartment is predictive for poor survival. [Van Rhenen et al.,Clin Cancer Res 2005,11] The monoclonal antibody against the cell surface antigen C-type lectin-like molecule-1, CLL-1, together with lineage infidelity markers enables discrimination of normal and malignant stem cells. [Van Rhenen et al.,Blood 2007,110; Van Rhenen et al.,Leukemia 2007,21] It could be hypothesized that CLL-1 and aberrant marker expression on MDS stem cells together with size of the stem cell compartment may predict leukemic evolution. Therefore, stem cells, defined as CD45^dimCD34⁺CD38⁻, were analyzed for expression of CLL-1 and aberrant lineage markers in bone marrow samples from 88 MDS patients classified by WHO as 16 RA w/o RS, 42 RCMD w/o RS, 3 MDS-U, 5 hypoplastic MDS, 6 MDS/MPD and CMML, 15 RAEB-1 and 2, 20 AML patients with a known MDS history and 26 healthy controls. Analysis of the CD34⁺CD38⁻ frequency in all MDS patients and normal controls revealed no significant differences (median 0.0061% vs. 0.0074%, respectively), whereas the frequency of CD34⁺CD38⁻ cells was 17-fold higher in high risk MDS (RAEB-1 and 2, median: 0.076%) as compared to low risk MDS (median: 0.0046%, p<0.001). Similar as in AML, stem cells were significantly more prevalent within the blast cell fraction (CD45^dimSSC^int/low) of high risk MDS as compared to low risk MDS (median 0.77% and 0.25%, respectively), reflecting the differences in clinical course in these patients (p=0.040). Regarding CLL-1 expression, a reliable number of stem cells (>20) could be tested in 11/15 high risk RAEB-1 and 2 cases and in 16/73 of the remaining low risk MDS cases. In these cases, median CLL-1 expression on the CD34⁺CD38⁻ cells was 1.6% (range 0–50) in low risk and 2.0% (range 0–27) in high risk MDS. Median CLL-1 expression on stem cells was 0.0% (range 0–4.7) in normal controls. Nevertheless, expression of lineage infidelity markers, such as CD5, CD7 and CD56, on CD34⁺CD38⁻ stem cells in MDS strongly suggests that a considerable part of these stem cells is malignant (median 35% in 7/16 patients tested). Our data show that CLL-1 is virtually absent on stem cells in MDS. Remarkably, median CLL-1 expression on stem cells in AML cases that evolved from MDS (7%, range 0–53, n=9) was manifold lower than in de novo AML (median 45% when excluding non de novo AML [Van Rhenen et al.,Blood 2007,110], p=0.034). Detailed analysis of CLL-1 expression in AML had already revealed that CLL-1 expression increases with differentiation (CD34⁻ > CD34⁺CD38⁻ > CD34⁺CD38⁺). [Bakker et al.,Cancer Res 2004,64;Van Rhenen et al.,Blood 2007,110] Thus, our data suggest that the CD34⁺CD38⁻ cells in high risk MDS and AML with antecedent MDS are more immature than in most de novo AML, which might explain poor prognosis of AML cases with MDS history. To conclude, our data indicate that CLL-1 is a specific marker of de novo AML, while CLL-1-negative AML may have been evolved from a MDS pre-phase that is further characterized by an increasing size of the stem cell compartment upon progression towards AML.