It has been hypothesized that bone resorption during tumor osteolysis is performed by osteoclasts. Data supporting this hypothesis have been provided from analysis of human biopsy specimens obtained from sites of tumor osteolysis, as well as from experimentation with in vivo animal models. Experiments in this report take this concept one step further by testing the hypothesis that osteoclasts are required for bone tumors to grow and destroy bone. To test this hypothesis, the influence of an osteolytic sarcoma tumor, NCTC clone 2472 (2472), on bone was studied in animals that are osteoclast deficient (microphthalmic, strain B6C3Fe-ala-Mitf(mi)) but whose osteoclast deficiency can be reversed following bone marrow transplantation. Femora of these mice and unaffected wild-type siblings were injected with 105 2472 cells, and after 14 days the femora were analyzed by radiographic and histomorphometric analysis. Macroscopic tumor, tumorinduced osteolysis, and increased osteoclast number were noted in femora of normal mice but not in femora of osteoclast-deficient mice (p < 0.001). Bone marrow transplantation converted osteoclast-deficient mice to mice with femora that contained osteoclasts in 4 weeks. Femora of these mice were then injected with 105 2472 tumor cells; after 14 days, in contrast to the findings in the original osteoclast-deficient mice, macroscopic tumor was present, tumor-induced osteolysis was noted on roentgenograms, and osteoclast number was increased when tumor-bearing limbs were compared with sham-injected limbs (p < 0.001). These data prove the hypothesis that osteoclasts are required for 2472 tumor- induced osteolysis, and they introduce the exciting possibility that osteoclasts are also required for tumors to grow in bone.