Bone tissue remodeling results from the coordinated and balanced activities of osteoblasts and osteoclasts. Osteoblast function is intimately linked to osteoclast activity via the osteoblast production of cytokines, growth factors and prostaglandins (PGs). The production of some of these factors is controlled by mechanical strains. Recently, a numbers of in vitro models attempted to screen genes and signalling pathways involved in this mechanism, mainly by stretching osteoblasts or by submitting them to a fluid shear stress. Osteoblasts possess mechanosensors which activate intracellular signals including ion channels, integrins, calveolar membrane structure and cytoskeleton. Nevertheless, response to physical signals may be quite different according to the type of mechanical stress applied. Fluid shear stress have been shown to elicit multiple intracellular signalling pathways involving intracellular calcium rise, extracellular signal-regulated kinase (ERK)1/2 activation of c-Fos and nuclear factor (NF)-κB translocation [2,9]. Downstream of such signalling events, various gene expression are induced, including type I collagen (COL1), osteopontin (OPN), insulin-like growth factor-I (IGF-1) and cyclooxygenase (COX)-2 [2]. Cyclic tensile stresses are also potent activator of the signalling cascade formed by ERK/c-fos/ NF-κB [11]. Stretching increases the production of vascular endothelial growth factor (VEGF), transforming growth factor (TGF)-β1 [18], alkaline phosphatase activity (ALP), osteocalcin (OC), osteoprotegerin (OPG), matrix metalloproteinases (MMP)-1 and -3 [10], COX-1 and -2, prostaglandin (PG)D2 synthase, peroxisome proliferator-activated receptor (PPAR) gamma-1 [17], but decreases the release of the soluble receptor activator of nuclear factor ligand (sRANKL) by osteoblasts [19]. In contrast, no significant effect has been reported on MMP-2, tissue inhibitor of metalloproteinases (TIMP)-1 and -2, and PPARgamma-2 synthesis [16,17]. One major barrier to understanding bone physiology at a cellular level is the lack of models to study cells in their native environment. Usually, compression is generated by a bending system [11] or a glass cylinder and is applied on osteoblasts cultured in monolayer on flat surfaces [14]. Herein, we propose an original model of 3D-osteoblast culture, allowing the study of compression on osteoblasts embedded in their own produced extracellular matrix. In this model, cell/matrix interactions are conserved and fluid flow through a three-dimensional extracellular matrix is allowed. In our study, loading was applied at a large amplitude (6–10% or 1–1.67 MPa) and at a frequency of 1 Hz. These loading conditions are included in the physiological range of amplitude and frequency