Soluble TRAIL does not impair the anti-osteoclastic activity of osteoprotegerin: Letter to the editor

It has been extensively documented that osteoprotegerin (OPG), a soluble member of the TNF receptor superfamily, inhibits osteoclastogenesis by binding to receptor activator of NF-kB ligand (RANKL) and preventing interaction with its cognate transmembrane receptor RANK. However, OPG can also interact with, and neutralize, TNF-related apoptosis inducing ligand (TRAIL), whose extracellular domain shares a 30% homology with the extracellular domain of RANKL. Although some inconsistencies on the differential binding affinity of OPG for RANKL versus TRAIL were present in initial studies, it has been recently demonstrated that the affinity of native OPG for native TRAIL is comparable to that for RANKL (45 nM versus 23 nM, respectively) at 37°C, as determined by plasmon surface resonance analysis. Consistently with this biochemical study, OPG has been shown to act in a paracrine and autocrine manner by binding TRAIL and promoting the survival of multiple myeloma, prostate cancer, ameloblastoma cells and synovial fibroblasts. Interestingly, previous data from different groups have shown that recombinant TRAIL promote monocytic differentiation, while inhibits both human and mouse osteoclastogenesis when added to pre-osteoclast cultures induced to differentiate with recombinant M-CSF+RANKL as well as to mature osteoclasts. On the other hand, a couple of studies suggested that recombinant soluble TRAIL might promote osteoclastogenesis, and the proposed molecular mechanism to explain such observation was a competition between TRAIL and RANKL for OPG binding. However, it should be noticed that Vitosky et al. used much higher concentrations of TRAIL (500 ng/ml) than RANKL (30 ng/ml) or OPG (50 ng/ml) and more importantly used mouse bone marrow pre-osteoclasts. In this respect, it has been clearly shown that mouse pre-osteoclasts only express TRAIL-R2, while human peripheral blood-derived pre-osteoclasts express both death receptors TRAIL-R1 and TRAIL-R2 as well as TRAIL-R4. Therefore, to further elucidate the important issue of the interply between RANKL, TRAIL and OPG in human osteoclastogenesis, we have cultured adherent PBMC with M-CSF+RANKL for 14 days in the absence or presence of recombinant OPG and recombinant TRAIL, used alone or in combination. Importantly, all cytokines were used at the same concentration (50 ng/ml). As expected [10-13], the addition of TRAIL to M-CSF+RANKL significantly (p<0.05) inhibited osteoclast formation (Figure 1A-B). Of note, recombinant OPG completely abrograted (p<0.05) osteoclast formation irrespectively of the presence of recombinant TRAIL in culture. The anti-osteoclastic activity of OPG could not be ascribed to a low affinity of OPG for TRAIL since OPG (50 ng/ml) efficiently inhibited the apoptosis induced by TRAIL (50 ng/ml) in HL-60 leukemic cells. Our current observations on one hand confirm that TRAIL has anti-osteoclastic activity and on the other hand indicate that it does not affect the potent anti-osteoclastic activity of OPG at least in the simplified model of osteoclastogenesis represented by human PBMC induced to differentiate by M-CSF+RANKL. Taken together with previous studies, these data also suggest that the relative concentrations of TRAIL, RANKL and OPG in the local microenvironment are likely key determinant for the regulation of osteoclastogenesis.