Plasticity of Stromal and Hematopoietic Bone Marrow Cells in a Murine Model of Myelodysplastic Syndrome

Bone marrow (BM) contains both hematopoietic cells and non-hematopoietic, stromal cells, which support hematopoiesis. In mice and humans, CD45 is a pan-hematopoietic marker universally expressed on nucleated hematopoietic cells. Recent studies have shown that CD45-, non-hematopoietic stromal cells present in the BM of patients with MDS can display the same cytogenetic abnormalities found in the CD45+ MDS clone, suggesting that stromal cells may be part of the malignant clone in patients with MDS. Although CD45- , non-hematopoietic BM cells (such as multipotent adult progenitor cells) can produce CD45+, hematopoietic stem cells, there is little evidence that CD45+ hematopoietic cells can de-differentiate into CD45-, non-hematopoietic stromal cells. To study the possible conversion of hematopoietic (CD45+) to non-hematopoietic (CD45-) cells, we used the NUP98-HOXD13 (NHD13) mouse model of myelodysplastic syndrome (MDS).

Two immortal cell lines (251 and 63B) were established from NHD13 mice with AML. These cell lines grow as a combination of adherent and suspension cells; although the suspension cells are almost entirely (>98%) CD45+, only 50% of the adherent cells are CD45+. Surprisingly, sub-culture of CD45+ suspension cells, generates CD45- adherent cells, leading to the suggestion that CD45+ hematopoietic cells can de-differentiate and produce CD45-, non-hematopoietic cells. Further immunophenotype analysis demonstrates that the CD45+ cells are also positive for myelo-monocytic markers such as CD11b, CD16/32, and F4/80, whereas the CD45- fraction is negative for these markers. To verify the conversion of CD45+ to CD45- cells at a clonal level, we sorted highly purified (>99.99%) CD45+ cells and sub-cultured CD45+ cells at a density of 1 or 10 cells per well in 96 well plates. Although there was no growth in the wells seeded at 1 cell/well, 23 out of 96 wells seeded at 10 cells/well expanded; at least 5 of these clones produced CD45- cells. Expression of the NHD13 transgene (driven by the pan-hematopoietic Vav1 promoter) was 100-fold higher in the CD45+ fraction compared to the CD45- fraction, and gene expression arrays identified differential expression of a number of hematopoietic markers, such as Pu.1 and Csf2rb. Finally, preliminary bisulfite sequencing experiments suggest that the CD45 promoter is heavily methylated in the CD45- fraction, but un-methylated in the CD45+ fraction. Taken together, these results indicate that the cell lines can cycle between non-hematopoietic and hematopoietic cells.

To determine if the cell lines have retained malignant potential, we transplanted lethally irradiated recipients via intravenous injection. At 6 and 16 weeks post-transplant, there was no engraftment detected in the peripheral blood (PB). However, at 24 weeks, one of three recipients showed low level engraftment (1.7% of PB). Remarkably, this mouse became frankly leukemic 2 weeks later, with splenomegaly, anemia, monocytosis, and donor engraftment of 17-53% in the PB, BM, and spleen.

To determine which cell fraction was responsible for the disease phenotype, we transplanted purified CD45+ and CD45- cells via intrafemoral (IF) injection. Interestingly, all recipients of the CD45+ cells died due to profound pancytopenia (hgb, 3.9±1.0 g/dL; WBC, 0.24±0.30 K/uL) by day 13 post-transplant; in contrast recipients of CD45- cells were healthy, with normal CBCs. Although we could not detect donor cells in recipients of CD45- cells by FACS at any time point up to 34 weeks post-transplant, engraftment of donor cells in the BM of these recipients could be detected by transgene specific genomic DNA PCR. These results indicate that the disease entity resides in the CD45+ fraction, and that long term engraftment of quiescent CD45- cells occurs. In addition, these results suggest the intriguing possibility that quiescent CD45- cells, which do not express the NHD13 transgene, can be transformed by demethylation of the CD45 promoter (and subsequent activation of the Vav promoter, which drives the NHD13 transgene). Experiments to test this hypothesis are currently in progress. Taken together, these findings suggest that malignant cells can cycle between stromal and hematopoietic cells, based at least in part on epigenetic events such as cytosine methylation, and support the hypothesis that non-hematopoietic BM cells may be an important part of the malignant clone in patients with MDS and AML.