Functional Inhibition of Osteoblastic Cells in An In Vivo Mouse Model of Myeloid Leukemia

Abstract 243

Disruption of normal hematopoietic development is a major problem in acute myeloid leukemia (AML). Osteoblastic cells have been shown to support hematopoiesis. We hypothesized that myeloid leukemia inhibits osteoblastic cells contributing to the loss of normal hematopoiesis. To define effects of leukemia on osteoblastic cells, we employed a previously described syngeneic murine model of AML where transplant of leukemic cells does not require irradiation of recipients (Neering et al. Blood, 2007). Leukemic mice had decreased serum levels of the bone formation marker osteocalcin (OC) (92.6 ± 11.6 vs 25.9 ± 4.4 ng/ml Normal (N) vs Leukemic (L) p<0.0001 n=8–12 mice/group). OC mRNA in CD45-marrow cells was reduced in leukemic mice 6 days after transplantation of leukemia cells when disease burden was low, and severely diminished 11 days after transplant when disease burden was high (1290 ± 124.7 vs 440.1 ± 34 and 3.3 ± 0.8 relative expression N vs L day6 p<0.001 and N vs L day11 p<0.001 n=5 mice/group). In situ hybridization revealed almost complete loss of OC and collagen 1 expression at the endosteal bone surface suggesting a severe disruption of bone formation. Immunohistochemical analysis showed a reduction in endosteal-lining osteopontin+ cells in leukemic animals further confirming a phenotype of osteoblastic dysfunction. Osteoprogenitor cells were also reduced, determined by CFU-OB assays (1.45 ± 0.31 vs 0.21 ± 0.09 CFU-OBs/1×10⁶ marrow cells N vs L p<0.01 n=7–10 mice/group). In leukemic mice TRAP+ osteoclastic cells were transiently increased (N: 50.8 ± 3.6 vs L day6: 64.4 ± 3.2; L day10: 33.8 ± 4.7 cells/ histological section N vs L p<0.05 n=5 mice/group) and the bone resorption marker serum CTX did not change. Despite the lack of a robust increase in osteoclast number or bone resorption, leukemic mice exhibited trabecular and cortical bone loss by micro-CT analysis (0.23 ± 0.01 vs 0.13 ± 0.01 tibia trabecular BV/TV, N vs L p<0.001 n=5 mice/group). Treatment with the bisphosphonate zoledronic acid (ZA) inhibited bone resorption in leukemic animals (21.79 ± 0.97 vs 8.28 ± 1.41 serum CTX ng/ml, Vehicle vs ZA n=5 mice/group), and rescued trabecular bone loss, but failed to rescue cortical bone loss (0.24 ± 0.017 vs 0.10 ± 0.011 cortical BV/TV, N vs L p<0.001 n=4 mice/group). ZA treatment did not affect the course of disease. Thus, we suggest that in our model leukemic cells functionally inhibit osteoblastic cells and uncouple bone formation from bone resorption. To determine if osteoblastic inhibition was mediated by cell contact dependent or secreted factors we performed cultures with osteoblastic cells from normal mice co-cultured with leukemic cells or treated with conditioned media from leukemic cell cultures. Osteoblastic differentiation and ability to form alkaline phosphatase positive colonies were absent in both culture conditions suggesting that a secreted factor is sufficient to inhibit osteoblasts in our model. The chemokine CCL-3 was recently reported to inhibit osteoblastic function in multiple myeloma. CCL3 protein was present at very high concentrations in both co-culture and conditioned media cultures (Osteoblasts: 14.08 ± 6.6 vs co-culture: 222.0 ± 14.2 p<0.0001; conditioned media: 130.2 ± 10.4 ng/ml p<0.0001 n=4 wells/group). CCL3 protein was also present at much higher levels in both the marrow plasma and blood serum of leukemic mice (Marrow: 4.53 ± 0.57 vs 122.0 ± 13.06 pg/ml N vs L p<0.0001, Blood: 0.103 ± 0.003 vs 26.12 ± 2.51 pg/ml N vs L p<0.0001 n=5 mice/group). CCL-3 mRNA which is expressed at very low levels in normal marrow was significantly elevated in malignant hematopoietic marrow cells compared to normal hematopoietic marrow cells from the same mice (0.47 ± 0.07 vs 3.70 ± 0.46 relative expression N vs L p<0.01 n=3 mice/group). To confirm that CCL3 signaling may play a role in human myeloid leukemias we analyzed CCL3 expression and protein levels in primary AML samples. Primitive CD34+CD38-CD123+ cells from 3 patients with AML expressed CCL3 mRNA at levels 54.7, 19.83, and 26.5 fold higher than 5 normal CD34+CD38- primary samples. CCL3 protein was also increased in marrow plasma from primary AML patients (10.76 ± 1.42 vs 18.97 ± 3.43 pg/ml N vs AML p<0.05 n=6–7 primary samples/group). Based on these results, we propose that therapeutic mitigation of leukemia-induced osteoblastic dysfunction may represent a novel approach to promote normal hematopoiesis in patients with myeloid neoplasms.