AML Cells Alter Bone Homeostasis By Promoting the Formation of Immature Bone and Resorption of Mature Bone That Supports Leukemia Progression

Background: Genetic alterations in the osteoprogenitor cells has been shown to induce acute myeloid leukemia (AML) in several mouse models. Moreover, we have recently reported that AML cells induce osteogenic differentiation in mesenchymal stromal cells (MSC) to gain growth advantage in the bone marrow (BM; Battula et al., JCI Insight, 2017). However, the effect of AML cells on bone homeostasis/turnover and its impact on AML progression is unknown. Here, we hypothesize that AML cells expand osteoprogenitor cells and alter the balance between bone formation and resorption.

Methods: To investigate the effect of AML cells on osteoprogenitor cells and mature osteoblasts, we used triple transgenic reporter mice with the genotype Osx-CreERt2;Ocn-GFP;ROSA-tdTomato. Murine AML cells with MLL-ENL translocation were implanted into these transgenic mice, and Osteoprogenitor (Osx+) cells and mature osteoblasts (Ocn+) in femurs were measured by confocal microscopy. To investigate the effect of human AML cells on bone composition, patient-derived xenograft (PDX) cells were implanted into non-obese diabetic scid interleukin-2Rγ^null (NSG) mice and bone histomorphometry was performed using H&E and Goldner's trichrome staining. Computed tomography (micro-CT) was used to measure bone volume (BV) and mineral density (BMD) in mice. Tartrate-resistant acidic phosphatase (TRAP) staining was performed to measure osteoclast number/activation. Finally, bone density on the vertebral bone (T12) was measured in 263 de-novo AML patients and 23 normal individuals by CT imaging.

Results: In transgenic mice implanted with syngenic AML cells, we found a 3-4 fold increase in Osterix⁺, but not Osteocalcin⁺ cells, suggesting AML cells expand osteoprogenitor cells in the BM during short term exposure (2-3 weeks). To investigate the effects of AML on bone during a long term exposure, we implanted 10 different AML-PDX models in NSG mice (3 mice per model) and analyzed femurs by micro-CT and bone histomorphometry analysis. Interestingly, we observed a dramatic increase in new web-like medullary bone formation in 5/10 of the PDX models tested. Moreover, higher bone volume is associated with less aggressive PDX models (which takes 4-6 months to reach 90% circulating leukemia), but not aggressive PDX models (only 4-8 weeks to reach 90% circulating cells). These findings were also confirmed by micro-CT of mouse femurs. Interestingly, in some PDX models, CT images revealed large cavities in cortical bones close to epiphysis and metaphysis areas in the femur and tibia of mice with AML suggesting bone resorption. To validate bone resorption, we performed TRAP staining and found a significant increase in osteoclast activity on the endosteal surface and massive bone resorption in AML bone compared to normal bone. These data indicate high bone turnover in mice with AML compared to control mice. Next, we measured bone densities in AML patients and normal individuals by chest CT imaging. We found bone densities were gradually decrease with age in both healthy individuals and AML patients. However, compared to healthy individuals, bone densities are significantly higher in majority of AML patients (~70%) (p<0.01). Interestingly, survival analysis revealed that higher bone density is associated with good prognosis in AML patients (p<0.01), suggesting high bone turnover alters patient outcomes.

Conclusion: Our data suggest that AML cells expand osteoprogenitor-rich niche and alter BM microenvironment by high bone turnover. New bone induction by less aggressive AML-PDX models coupled with AML patient CT data suggests that high bone density and volume are associated with favorable prognostic factors in AML. Mechanisms underlying this observation are under investigation.