Letter by Secchiero and Zauli regarding article, "Osteoprotegerin inhibits vascular calcification without affecting atherosclerosis in ldlr(-/-) mice".

In their study, Morony and colleagues have shown that treatment of diet-fed atherogenic ldlr(-/-) mice with human osteoprotegerin (OPG)-Fc (10 mg/kg) injected subcutaneously 3 times per week for 2 to 5 months did not affect the degree of atherosclerosis and rather significantly reduced the atherosclerotic calcified lesion area. Endogenous plasma OPG levels started to increase after as early as 2 weeks from the beginning of the diet, suggesting that OPG elevation was related to the onset of atherosclerosis.1 We likewise previously demonstrated in an apolipoprotein E (ApoE)–deficient model of atherosclerosis that the rise of serum OPG occurred early after induction of diabetes mellitus, a condition associated with the induction of atherosclerosis in these mice. Like Morony et al, we documented the existence of an inverse relationship between serum OPG and receptor activator of nuclear factor-kB ligand (RANKL) levels in ApoE-/- diabetic mice, but this phenomenon is likely due to the fact that the ELISA kit (R&D Systems, Minneapolis, Minn) used in both studies does not recognize mouse RANKL masked by OPG. The major concern raised by the interpretation of the interesting study by Morony et al, however, is the assumption that the observed rise in native plasma OPG is not causally related to atherosclerosis and that it might rather represent a marker of atherosclerosis or a compensatory mechanism. Because Morony et al used human OPG-Fc rather than native human OPG, it should be emphasized that in OPG-Fc the signal-peptide, heparin-binding domain and death domain–homologous regions are removed and the remaining peptide is fused to the Fc domain of human immunoglobulin IgG1 in order to enhance the pharmacological activity of native OPG. Although OPG-Fc maintains the potent dimeric nature of OPG and significantly increases its circulating half-life, it shows considerably lower affinity for tumor necrosis factor–related apoptosisinducing ligand (TRAIL) than native OPG, whereas its affinity for RANKL does not appear to be significantly impaired. This consideration is particularly relevant because we have previously demonstrated that repeated intraperitoneal injections of human recombinant TRAIL counteract the extent of atherosclerotic lesions in an ApoE-/- mouse model of atherosclerosis, which renders unlikely the possibility mentioned by Morony et al that the antiatherosclerotic activity of OPG-Fc might be mediated by endogenous TRAIL inhibition in the atherogenic ldlr(-/-) mice. Moreover, it has been recently shown that recombinant human OPG is able to strongly induce the in vitro proliferation of human vascular smooth muscle cells, a phenomenon that represents a key step in atherosclerosis, and we have independently observed that human recombinant OPG is able to promote proliferation of rodent vascular smooth musclecells (P.S. and G.Z., unpublished observations, 2008). Clearly, OPG has a fundamental role in preventing vascular calcification, and the findings of Morony et al add the important information that OPG-Fc might directly act at the vascular level by suppressing the ability of RANKL to promote vascular calcification and that the use of OPG-Fc is safe in a relevant animal model of atherosclerosis. However, this does not exclude the possibility that plasma elevation of native OPG might play a pathogenic role in atherosclerosis