The objective of this study was to use an in vivo model of periodontitis (mouse calvaria) to quantify the effects of local release of secreted human macrophage products, 17beta-estradiol (E2), and proinflammatory lipopolysaccharide (LPS) on histologic bone resorption. Human THP-1 monocytes (106) were converted to macrophage phenotype by 500 ng/ml phorbol 12-myristate- 13-acetate (PMA) and treated as follows: no stimulation or Escherichia coli LPS (10 microg/ml) alone or in combination with a physiologic dose of E2 (100 pg/ml) for 24 h in RPMI/10% FBS, washed extensively, then incubated for 24 h in serum-free media. Supernatant products were concentrated and incorporated into a 4% (w/v) methylcellulose gel. Separate gels were incorporated with the following: LPS (500 microg/animal) alone, high dose of E2 (10 ng/animal) alone, a combination of LPS + E2, or gel only (controls). Loaded or control gels were placed into a polylactic acid occlusive dome, inserted subcutaneously over the calvaria of mature ovariectomized ICR Swiss mice (8 mice x 7 groups x 2 times [5/14 days] = 112 animals), then calvaria were evaluated histologically. Macrophage stimulation with LPS alone, but not LPS in combination with E2, produced supernatants which upregulated osteoclast numbers in the suture area compared to gel controls at 5 days (p = 0.009). The addition of LPS directly to the local delivery gels significantly upregulated osteoclasts in endosteal surfaces compared to gel controls at 5 days (p = 0.024) and at 14 days (p = 0.025). The addition of E2 to LPS down-regulated resorption to a level not different from gel controls at 14 days. This in vivo model appears effective in studying inflammatory bone resorption, which may be inhibited by E2 directly or through its influence on secreted macrophage products.