Effect of Estrogen versus Testosterone on Circulating Osteoprotegerin and Other Cytokine Levels in Normal Elderly Men

S Khosla, EJ Atkinson, CR Dunstan… - The Journal of Clinical …, 2002 - academic.oup.com
The Journal of Clinical Endocrinology & Metabolism, 2002academic.oup.com
Recent studies have shown that estrogen (E) likely plays a dominant role in inhibiting bone
resorption in normal elderly men. Because both E and T inhibit osteoclast development and
activity, stimulate osteoclast apoptosis, and inhibit osteoblast production of IL-6, it is unclear
why T is less potent than E in inhibiting bone resorption in vivo. Osteoprotegerin (OPG)
binds to and inactivates RANKL, the final mediator of osteoclastogenesis. In vitro, OPG
production is stimulated by E, and preliminary data suggest that T has the opposite effect …
Abstract
Recent studies have shown that estrogen (E) likely plays a dominant role in inhibiting bone resorption in normal elderly men. Because both E and T inhibit osteoclast development and activity, stimulate osteoclast apoptosis, and inhibit osteoblast production of IL-6, it is unclear why T is less potent than E in inhibiting bone resorption in vivo. Osteoprotegerin (OPG) binds to and inactivates RANKL, the final mediator of osteoclastogenesis. In vitro, OPG production is stimulated by E, and preliminary data suggest that T has the opposite effect. Thus, we analyzed serum for OPG levels from a study in which 59 elderly men (mean age, 68 yr) were made acutely hypogonadal using a GnRH agonist and were also placed on an aromatase inhibitor to block conversion of androgens to estrogens. They were studied first under conditions of physiologic E and T replacement, and then randomized to no replacement, replacement with E alone, T alone, or both E and T. E alone resulted in an 18.6 ± 7.9% (mean ± sem) increase in serum OPG levels (P < 0.05), whereas T alone tended to decrease OPG levels (by 10.0 ± 8.5%; P < 0.05 compared with E alone). Using a two-factor ANOVA model, there was a highly significant T effect (P = 0.006) on decreasing serum OPG levels. Serum TNF-α, IL-6, and IL-6 soluble receptor levels increased significantly in the men who had both E and T withdrawn, and the increases in TNF-α and IL-6sR were absent in the men treated with either E or T. However, due to the variability in these cytokine measurements, the ANOVA models were not significant for E or T effects. Taken together, these data suggest that in vivo, T decreases OPG levels, whereas E tends to have the opposite effect. These differential effects of E vs. T on OPG production may explain, at least in part, why T has weaker effects than E on inhibiting bone resorption in vivo in humans.
Oxford University Press