Fig. 5.
Levels of the proinflammatory cytokines GM-CSF and TNF-α.
Introduction of the L-selectin mutation into the E/P-selectin double-mutant mice lowered the levels of GM-CSF and TNF-α in triple-selectin–null mice. (A) GM-CSF levels were elevated by a factor of 6 in E/P double-selectin–null mice over values in wild-type and L-selectin–deficient mice (P < .001 for both; mice 22-24 weeks of age; n = 6 for all). GM-CSF was also significantly increased (3 fold) in E/P −/− mice compared with triple-mutant mice (P < .0013; n = 6). GM-CSF levels in triple-mutant mice were not significantly higher than those in wild-type mice (P < .159) or L-selectin −/− mice (P < .156). (B) TNF-α levels in E/P double-null mice were significantly higher than those in wild-type or L-selectin–deficient mice (P < .0001 andP < .0016, respectively). Serum TNF-α levels in triple-null mice, though lower than those in E/P double mutants, were also significantly higher than those in either wild-type or L-selectin–deficient mice (P < .02 for both). Although the levels of TNF-α in E/P double-null mice were nearly twice those in triple-null mice, the difference was not significant (P < .16). The error bars in the histograms indicate means ± SD. The asterisk indicates a significant difference from wild-type and L-selectin–deficient mice (P < .001), and the number sign indicates a significant difference from E/P double-selectin mutant mice (P < .002).