Pain is the foremost clinical symptom of osteoarthritis (OA), however the processes that lead to pain in the joint are poorly understood. I have used two different murine models of OA, induced by surgical destabilisation of the joint. Upon destabilisation, two episodes of painful behaviour are observed, an initial post-operative phase, lasting a few days, and a late phase that starts between 8-11 weeks post-surgery, detected by differences in weight bearing measured by the Linton incapacitance tester. The later phase is associated with significant cartilage damage, but no overt synovitis. A screen of pain sensitising molecules expressed in the whole joint at the time animals developed late pain was performed. Statistically significantly regulated molecules included nerve growth factor (Ngf), bradykinin receptors 1 and 2 (Bdkrb1 and Bdkrb2), Tachykinin 1 (Tac1) and its receptor (Tacr1), determined by real-time PCR. When a similar analysis was performed in the microdissected tissues of the joint (cartilage, bone and meniscus), all but one (Bdkrb2), of the previously regulated molecules were significantly regulated in the articular cartilage. Ngf, Bdkrb1, Bdkrb2 and Tac1 were also statistically significantly regulated upon in vitro cartilage injury, using the murine hip and porcine metacarpophalangeal explantation models, suggesting that in vivo mechanical damage may be sufficient to drive pain molecules. One mechanism previously described by our group, by which chondrocytes respond to injury, is via release of FGF2 from the pericellular matrix. When in vivo cartilage injury studies were performed in Fgf2-/- tissues, or in the presence of an FGF2 receptor inhibitor, induction of Ngf was significantly blunted, indicating that FGF2 drives Ngf induction upon in vitro cartilage injury. Induction of Bdkrb1, Bdkrb2 and Tac1 were not FGF-dependent in vitro. To test the significance of FGF2-driven Ngf induction in vivo, pain was assessed in Fgf2-/- mice following joint destabilisation. Rather than reducing or delaying the development of pain in these animals, painful behaviour developed earlier than wild type mice, and correlated with cartilage damage scores. Pain could also be significantly neutralised by anti NGF antibody, indicating that FGF2 is unlikely to be driving Ngf expression in vivo. Female mice are relatively protected from OA following joint destabilisation, a phenomenon that may be related to differences in their molecular response to acute joint destabilisation. Despite having little in the way of cartilage damage, they develop painful behaviour at a similar stage as male mice, associated with the same pain sensitising molecules.