Author:

Abstract:

Research on the development of biologic therapeutics to treat rheumatoid arthritis (RA)has been ongoing for several decades. A number of drugs have been brought onthe market and several of these have been found to improve the prognosis of thepatients. A major breakthrough was the development of inhibitors of tumor necrosis factor-alfa in the nineties. However, many patients today still do not respond or become resistant to the availablemedication. Especially bone destruction by osteoclasts is an aspect of thedisease that should receive more attention, since it is an irreversible processand can lead to significant loss of life quality.In addition to the development of new therapies, diagnosis and monitoring of RAusing non-invasive imaging is also characterized by major improvements over thelast ten years. Imaging techniques have evolved to allow more rapid anddetailed visualization of the disease process. The goal at this point is thecharacterization of new molecular markers that are specifically expressed inthe inflamed joints.In this study we investigated several molecules with a potential to interfere withosteoclastogenesis and/or to serve as markers of macrophages for in vivoimaging of mice with collagen-induced arthritis (CIA), a well-defined mousemodel for RA. We investigated the expression of the markers in different cellsand organs by qPCR and flow cytometry, evaluated their potential to interferewith osteoclast formation in cell cultures and assessed their use in in vivosingle-photon emission computed tomography (SPECT) of mice with CIA.The macrophage mannose receptor (MMR; CD206) is a C-type lectin responsible forendocytosis and phagocytosis. It is predominantly expressed by maturemacrophages, endothelial cells and dendritic cells and can be detected inspleen, liver and lymph nodes. Interestingly, MMR was described to be involvedin the process of osteoclast formation. We demonstrated that MMR was highlyinduced upon stimulation with macrophage colony-stimulating factor (M-CSF) thatstimulates the formation of macrophages, but less with receptor activator ofnuclear factor κB (RANKL), which is the main inducer of osteoclasts. On theother hand, we could not show differences in osteoclast formation betweenwildtype and MMR-deficient bone marrow cells, indicating that MMR is mostlikely not important in osteoclastogenesis or bone destruction. We detected MMRin the rheumatic joints of CIA mice in which expression could be predominantlytraced to CD11b+F4/80+ macrophages of the synovial fluid. SPECT imaging of CIAmice with radioactively labeled MMR-targeting nanobodies visualized theinflamed joints and could even detect incipient inflammation in symptom-freepaws of mice with CIA. We concluded that MMR is not a preferred target toinhibit osteoclasts, but is very well suited as a marker for visualization of macrophagesin the inflamed joints.A recent paper reported the expression of Forkhead box p3 (Foxp3), conventionallydescribed to be restricted to regulatory T cells, in macrophages withimmunosuppressive properties. In another publication, it was described that apopulation of osteoclast precursors with suppressive properties expands duringinflammatory arthritis. These data led us to investigate the expression ofFoxp3 in osteoclasts generated in vitro and in macrophages of mice with CIA. Wecould, however, not confirm the expression of Foxp3 in macrophages. In fact,with the use of two transgenic mouse strains, i.e. Foxp3-GFP or Foxp3CreRosa-YFP mice, we did not detect any expression of Foxp3 in monocytes ormacrophages either in vitro generated or in inflammatory conditions ex vivo andconcluded that Foxp3 is not expressed on cells of the monocyte/macrophagelineage.Dendritic cell-specific transmembrane protein (DC-STAMP) and osteoclast-stimulatorytransmembrane protein (OC-STAMP) have both been described to be essential forfusion of macrophages to osteoclasts and giant cells. We confirmed theirupregulation during the formation of osteoclasts and giant cells and byproducing antibodies against both proteins we could inhibit osteoclastformation in vitro. Furthermore, we showed that DC-STAMP and OC-STAMP areexpressed in the synovium and synovial fluid of mice with CIA. The effect ofthe DC-STAMP and OC-STAMP-targeting antibodies was assessed in two in vivomodels. sRANKL-induced bone loss could not be altered, but in CIA, theanti-OC-STAMP antibodies reduced inflammation and parameters of bonedestruction. Unfortunately, we were not able to produce nanobodies that boundwith sufficient affinity to be used in in vivo SPECT imaging of CIA mice. Inconclusion, the expression pattern of both proteins shows a lot of promise fortheir use as markers in imaging and further research is required in order todraw a firm conclusion about the potential of DC-STAMP and OC-STAMP as targetsfor therapy in autoimmune arthritis.V-set and immunoglobulin domain-containing 4 (VSIG4) was investigated since it wasdescribed to be expressed in the synovium of RA patients. A VSIG4-Fc fusionprotein was capable of suppressing inflammation and bone loss in murinearthritis models. Our studies showed that VSIG4 is expressed on macrophages butbarely or not on osteoclasts. When expression was investigated in differentorgans of mice with CIA, we concluded that VSIG4 was rather specificallyexpressed in the synovium and synovial fluid, in addition to the liver where itis known to be expressed on Kupffer cells. Nanobodies against VSIG4 located tothe inflamed joints in SPECT imaging. By performing fine dissection afterimaging, the signal was shown to originate from the synovial and synovialfluid.Programmed death-ligand 1 (PD-L1; CD274), a membrane protein of the immunoglobulin family,was previously shown to be expressed in arthritic joints and administration ofa PD-L1-Fc fusion protein to mice with CIA ameliorated disease severity. Inaddition, the receptor of PD-L1, programmed death-1 (PD-1; CD279) was shown tobe involved in osteoclastogenesis. Mice with a deficiency in PD-1 showed mildosteopetrosis due to a decrease in osteoclasts. Therefore we investigated theexpression and role of PD-L1 in osteoclastogenesis. We found that PD-L1 washighly up-regulated after RANKL stimulation and, in a murine macrophage cellline, neutralizing antibodies against PD-L1 dose-dependently inhibited theformation of osteoclasts. This inhibition was, however, not confirmed by theuse of primary bone marrow or spleen cells. PD-L1 deficient mice showed lessosteoclast precursors in blood, but not in spleen and bone marrow.Consequently, fewer osteoclasts from blood cells of PD-L1 deficient mice were generatedas compared to those of wildtype counterparts. We confirmed the expression ofPD-L1 on a variety of inflammatory cells in the synovium and synovial fluid ofmice with CIA. Unfortunately, radioactively labeled nanobodies against PD-L1could not specifically visualize the cells in the inflamed synovium during invivo SPECT imaging, presumably due to their binding to blood cells. Weconcluded that this protein was not suitable as a marker for inflammatory cellsin the rheumatic joint and that further research is needed to better define therole of PD-L1 in osteoclastogenesis.Summarizing our investigations, we showed that Foxp3 was not expressed in cells of themonocyte/macrophage lineage. PD-L1 was expressed with high intensity onosteoclasts, but our nanobodies targeting this protein were not suitable for invivo SPECT imaging of CIA mice. DC-STAMP and OC-STAMP were the most promisingtargets for development of a therapy counteracting RA-related bone loss, whileMMR and VSIG4 were most successful in visualization of inflammatory cells inthe joints of mice with CIA.