Differences in circulating concentrations of total, free and bound leptin relate to gender and body composition in adult humans

We describe a radioimmunoassay (RIA) for total leptin and a gel filtration procedure for the separation of free and bound leptin in human serum. The RIA, based on a locally prepared antibody, has a minimum detection limit of 0.9 ng/mL, a working range (CV < 10%) of 25-50 ng/mL, inter-assay precision of 10.2, 7.2 and 8.9%CV at 7.9, 15.4 and 30.0 ng/mL, respectively, 94% recovery of exogenous leptin (range 81.1-120.6%), exhibited parallelism and demonstrated no significant cross-reactivity or interferences. A difference plot of results from this method and those from a commercially available kit (Linco Research) demonstrated satisfactory agreement up to concentrations of 50 ng/mL total leptin, with no significant bias. A gender-dependent correlation was obtained between body mass index (BMI) and total leptin (r = 0.91, P<0.001, n = 75 for men; r = 0.79, P<0.001, n = 72 for women), with women having higher leptin concentrations than men for any given BMI. Gel filtration studies (inter-assay precision: 4.7%CV, n = 18) demonstrated that a variable fraction (between 10% and 40%) of total leptin in serum was bound with high affinity (K-eq = 1.0-1.45 x 10(9) L/mol) to a non-albumin, non-lipid macromolecule. Binding affinities were found to be similar irrespective of gender or fat mass. A significant positive correlation between free or bound leptin concentrations and BMI was obtained for both men and women (r = 0.87-0.94); free and bound leptin concentrations were also significantly higher in women (P<0.01) than in men for any given BMI, and higher in obese (P<0.01) than in lean individuals. We conclude that leptin 'resistance' associated with obesity cannot be accounted for by reduced free leptin concentrations in serum and that the methods described are suitable for the investigation of total, free and bound leptin for both clinical and research purposes.