Fibrous dysplasia of bone (FD) is caused by somatic mutations of the GNAS1 gene, which lead to constitutive activation of adenylyl cyclase and overproduction of cAMP in osteogenic cells. Previous in vitro studies using nonclonal, heterogeneous strains of FD-derived cells suggested that IL-6 might play a critical role in promoting excess osteoclastogenesis in FD. In this study, we investigated IL-6 expression in FD in situ and its relationship to the actual patterns of osteoclastogenesis within the abnormal tissue. We found that osteoclastogenesis is not spatially restricted to bone surfaces in FD but occurs to a large extent ectopicly in the fibrous tissue, where stromal cells diffusely express IL-6 mRNA and exhibit a characteristic cell morphology. We also observed specific expression of IL-6 mRNA in a proportion of osteoclasts, suggesting that an autocrine/paracrine loop may contribute to osteoclastogenesis in vivo in FD, as in some other bone diseases, including Paget's disease. We also generated homogeneous, clonally derived strains of wild-type and GNAS1-mutated stromal cells from the same individual, parent FD lesions. In this way, we could show that mutated stromal cells produce IL-6 at a basal magnitude and rate that are significantly higher than in the cognate wild-type cells. Conversely, wild-type cells respond to db-cAMP with a severalfold increase in magnitude and rate of IL-6 production, whereas mutant strains remain essentially unresponsive. Our data establish a direct link between GNAS1 mutations in stromal cells and IL-6 production but also define the complexity of the role of IL-6 in regulating osteoclastogenesis in FD in vivo. Here, patterns of osteoclastogenesis and bone resorption reflect not only the cell-autonomous effects of GNAS1 mutations in osteogenic cells (including IL-6 production) but also the local and systemic context to which non-osteogenic cells, local proportions of wild-type vs mutated cells, and systemic hormones contribute.