Aberrant Expression of the Chondroitin Sulfate Proteoglycan NG2 on Pediatric AML/abn(11q23) and ALL/t(4;11) Does Not Allow Flow Cytometric Detection of Leukemic Blasts with a Stem Cell-Like Immunophenotype.

Monitoring of minimal residual disease in childhood leukemia is widely used to determine treatment response and to detect persisting leukemic blasts for relapse prediction. In this context, it is increasingly recognized that only few leukemic blasts possess self-renewal capability and that only these self-renewing leukemic stem cells are able to initiate relapses. Therefore, the leukemic stem cells as the relapse-initiating cells should be the target cells of minimal residual disease detection. The human homologue of the rat chondroitin sulfate proteoglycan NG2 protein (detected by the monoclonal antibody 7.1) is aberrantly expressed in a high percentage of leukemias harboring 11q23 abnormalities and its expression has been proposed as a potential marker for MRD detection. To evaluate NG2 as a marker for leukemic stem cells, bone marrow or peripheral blood samples of 4 AML/t(9;11), 1 AML/t(10;11), 1 AML/t(?;11) and 2 ALL/t(4;11) patients with NG2-positive leukemia were analyzed by four color flow cytometry for NG2 expression on the cell surface of primitive cell populations. Immature cells were defined by the antigen profiles CD34+CD38− (CD33−) in AML and CD34+(CD117+)CD19− in ALL. AML stem cells are known to be CD34+CD38−. In the ALL patients, 32% resp. 81% of CD34+CD19− cells were leukemic by FISH analysis for MLL/AF4. For this rare event analysis, 5 x 10e6 cells were incubated with saturating amounts of the respective antibodies. Data acquisition and analysis was performed on a 2-laser FACSCalibur (BDIS). 1 x 10e6 events were acquired but only CD34+CD38− (in AML) or CD34+CD19− cells (in ALL) were stored (storage gate). Matched isotype controls were analyzed using the same storage gate to exclude amplification of unspecific events. In-between each acquisition, the cytometer was vigorously flushed to prevent carry-over of cells from previous acquisitions. In addition, 10.000 ungated events were acquired to compare the CD34+Lin− stem/progenitor cells with the dominant leukemic cell clone.

In all eight patients with NG2-positve leukemia, cells with a stem cell-like immunophenotype were shown to lack expression of NG2. Instead, flow cytometry analyses revealed a close correlation between expression of the myeloid differentiation marker CD33 and increasing levels of NG2 on maturing AML cells. Similarly, by LightCycler RT-PCR CD34+CD19+ cells showed higher expression of NG2-RNA compared with CD34+CD19− cell in both ALL samples. Our results demonstrate that NG2 is not expressed on primitive leukemic cells but is upregulated with differentiation within the leukemic cell clone. This hampers the use of NG2 expression as a sensitive marker for MRD detection as expression of NG2 does not allow detection of relapse-initiating leukemic stem cells.