The bone marrow provides an essential regulatory microenvironment for adult hematopoiesis, however the relationship between the bone marrow microenvironment and malignant hematopoiesis remains poorly understood. To investigate the interactions between leukemia and the bone marrow microenvironment we utilized a mouse model of blast-crisis chronic myelogenous leukemia (BC-CML), in which primitive normal murine hematopoietic cells are modified to leukemic cells by expressing the translocation products BCR/ABL and Nup98/HoxA9. The presence of each translocation was confirmed by their co-expression of Green Fluorescent Protein (GFP) and Yellow Fluorescent Protein (YFP) respectively. Ten days after injection of GFP+/YFP+ leukemic cells into strain-matched immunocompetent, non-myeloablated recipient mice, 50% of the bone marrow was composed of leukemic cells as determined by flow cytometric analysis. Histologic analysis of the contralateral tibiae and femora demonstrated not only progressive replacement of the bone marrow by leukemic cells, but also a significant bone loss. Histomorphometric analysis confirmed 50% decreased trabecular bone volume in leukemic mice compared to control mice that were not injected with leukemic cells (bone volume/total volume (%): 12±2 vs 26±2 p=0.01). Interestingly, numerous multi-nucleated osteoclasts were observed in the bone marrow of leukemic mice and were localized adjacent to leukemic cells, suggesting that leukemic cells may affect osteoclastogenesis and result in massive bone loss. To test this hypothesis, we first measured the expression of known regulators of osteoclastogenesis, including RANKL, in our leukemic cells by quantitative RT-PCR analysis. Compared to GFP−/YFP− cells, GFP+/YFP+ cells have 3-fold increased expression of RANKL, a major osteoclastogenic cytokine. We then examined if leukemic cells can give rise to osteoclasts in the presence of RANKL and M-CSF in vitro and found that these cells were unable to differentiate into osteoclasts themselves. To determine if leukemic cells can induce osteoclastogenesis of normal osteoclast progenitors, we cocultured spleen-derived osteoclast precursors from wild-type mice with GFP+/YFP+ leukemic cells or GFP−/YFP− non-leukemic cells in osteoclastic differentiating media containing optimal concentrations of M-CSF and RANKL. As expected, there was abundant formation of mature osteoclasts, identified as TRAP+ multinucleated cells, in control cultures containing non-leukemic cells and osteoclast precursors. Leukemic cells significantly increased TRAP+ mono-nucleated osteoclast precursors (No. TRAP+ mononucleated cells/well: 34±3.3 vs 20±6.0 in non-leukemic cells, p=0.0136). Under this culture condition, we did not observe increased mature osteoclast formation by leukemic cells. Surprisingly, we found that osteoclast precursors strongly prolonged the survival of leukemia cells. In control cultures without a feeder layer of osteoclast precursors there were no viable leukemia cells present after 6 days in culture while in the co-culture system viable leukemia cells were still abundant after 6 days in culture, identifiable by their expression of GFP/YFP (No. GFP+/YFP+/high power field: 0 vs 142±6.4, p<0.01). In summary, in a murine model of BC-CML, there was a rapid loss of trabecular bone and an increase in the number of osteoclasts. Expression of osteoclast-regulating cytokines in leukemic cells favored osteoclastogenesis, however modified cells did not themselves give rise to osteoclasts, while increasing the population of normal immature osteoclasts. In turn, osteoclast progenitors prolonged survival of leukemic cells in vitro. Taken together, these data strongly suggest reciprocal synergistic interactions of leukemic cells with osteoclast progenitors in the bone microenvironment. These previously unrecognized interactions could be exploited to modify disease progression, providing a novel approach for leukemia treatment.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author