Function of Fos proteins in bone cell differentiation

Abstract

c-Fos, a component of the dimeric transcription factor AP-1, is a key regulator of bone cell growth and differentiation, affecting both osteoblast and osteoclast lineages during normal development and in bone diseases. Overexpression of c-Fos in transgenic mice results in osteoblastic or chondroblastic tumors. Osteosarcoma formation in c-Fos transgenic mice is enhanced in the p53-/-background, where higher numbers of tumors per animal develop with an earlier onset, thus providing a valuable tumor model (Jochum, W. and Wagner, E. F., unpublished observations). Unlike c-Fos, overexpression of Fra-1, a c-Fos-related protein, in transgenic mice did not induce bone tumors. Lack of tumorigenicity in Fra-1 transgenic mice is consistent with the notion that Fra-1 lacks transcriptional activation domains required for cellular transformation. Unexpectedly, however, Fra-1 transgenic mice showed a progressive increase in bone mass. This osteosclerotic phenotype is characterized by an increase in osteoblast surface and osteoid volume. We are currently analyzing the direct involvement of Fra-1 in osteoblast proliferation and differentiation. In contrast, mice lacking c-Fos develop osteopetrosis due to a complete block in osteoclast differentiation. To have an insight into the mechanism by which c-Fos is required for osteoclastogenesis, an in vitro c-fos-/-rescue system was established whereby c-Fos-deficient splenocytes were infected with retroviral vectors expressing various AP-1 components. After culturing the infected cells in osteoclastogenic conditions, rescued osteoclast formation was quantitated by counting cell numbers expressing osteoclast markers such as tartrate-resistant acid phosphatase (TRAP) and calcitonin receptor. In addition, bone resorption was measured on bovine bone slices. Surprisingly, Fra-1, which lacks transactivation domains detectable in c-Fos, rescued osteoclast differentiation even more efficiently than c-Fos in vitro. Fra-2 and FosB rescued to a lesser extent than c-Fos. Unlike Fos proteins, however, none of the Jun proteins (c-Jun, JunB, and JunD) rescued the differentiation block. Furthermore, the above-mentioned fra-1 transgene, when crossed into mice lacking c-Fos, allowed osteoclast differentiation to proceed and rescued the osteopetrotic phenotype. Regarding the specific role of Fos proteins in osteoclast formation, a number of conclusions can be drawn from our present analyses. First, Fra-1 is the most potent osteoclastogenic Fos protein; introduction of Fra-1 into osteoclastogenic cell lines further increased their differentiation potentials. Second, the fra-1 gene carries AP-1 sites and is transcriptionally activated by c-Fos in the macrophage/osteoclast lineage. Although the fra-1 gene can be activated in a c-Fos-independent manner in many other cell types, transcriptional activation of fra-1 in osteoclast precursors seems to require c-Fos. This implies that c-Fos knock-out mice may be equivalent to c-Fos and Fra-1 double knock-out in the macrophage/osteoclast lineage, leading to an osteopetrotic phenotype. Note that Fra-1 knock-out mice are embryonic lethal due to a placental defect. (Schreiber, M. and Wagner, E. F., unpublished observations). Osteoclastogenic activity of c-Fos in wild-type mice may be either partially or fully mediated by Fra-1. Third, c-Fos is known to carry a major transactivation domain in its COOH-terminal part, which responds to Ras signaling and is critical for transformation. A detailed structure-function analysis showed that COOH-terminal parts of Fos proteins are dispensable for osteoclast differentiation. Further analyses are necessary to identify the mechanisms by which Fra-1, or c-Fos lacking transactivation domains, efficiently activate target genes required for osteoclast differentiation in the absence of c-Fos. One mechanism by which Fos proteins regulate osteoclastspecific gene expression is revealed by the analysis of the TRAP promoter region. Transient transfection and DNA binding experiments revealed juxtaposed functional binding sites for transcription factors NF-AT (nuclear factor of activated T cells) and AP-1. Overexpression of full-length NF-AT3 and Fra-1 activated TRAP promoter-reporter constructs in nonhematopoietic cell lines treated with calcium ionophore and phorbol 12-myristate 13-acetate (PMA). Moreover, ectopic expression of NF-AT3 in peritoneal macrophages and in a macrophage cell line induced expression of the endogenous TRAP gene, suggesting that NF-AT is indeed a key player in osteoclast-specific TRAP gene regulation. Identification of additional AP-1 target genes during osteoclastogenesis as well as further experiments aiming to define a causal role of the transcription factor NF-AT in osteoclast-specific gene expression are important avenues for future research.