BIOACTIVITY OF HUMAN GROWTH-HORMONE IN SERUM - VALIDATION OF AN INVITRO BIOASSAY

GH, in clinical practice, is determined by RIA, but RIA estimates may not accurately reflect serum GH bioactivity. The available measures of GH bioactivity lack either sensitivity, specificity, or a physiologically relevant end point. The objective of this research was to develop a physiologically relevant GH bioassay which would not only measure the bioactivity of purified GH preparations, but would also have sufficient sensitivity to measure GH bioactivity in human serum. The method consisted of incubating murine 3T3-F442A adipocytes in serum-free medium containing BSA, C-14-glucose, and increasing concentrations of GH or test materials for 24 h, followed by measurement of conversion of glucose to lipid. Interference by nonspecific serum factors was reduced by the addition of 10 mug/liter insulin, 25 nm dexamethasone, and 37 nm estradiol to the medium. In the presence of 10 mug/liter insulin, 50 mug/liter insulin-like growth factor-1 did not alter the ability of GH to suppress lipid accumulation. Epinephrine and glucagon could suppress lipid accumulation but only at concentrations greatly in excess of the physiological range in serum. Twenty two thousand dalton hGH produced dose-dependent suppression of lipid accumulation which was linear between 0.625 and 10 mug/liter (r = 0.926; P = 0.0001) with a half-maximal response of 3.0 +/- 0.2 mug/liter (n = six experiments). The intra- and interassay coefficients of variation were 7% and 19%, respectively. The assay was specific for GH since addition of human PRL produced suppression of lipid accumulation only at concentrations where contamination of the preparation by GH became a significant factor. ACTH also suppressed lipid accumulation but only at doses of 1000 mug/liter or greater. Human placental lactogen and hLH, hFSH, and hTSH did not cross-react with GH in this assay. Addition of human serum did not alter the slope or ED50 of the GH dose-response curve. Pools of serum from prepubertal and pubertal boys and girls, subjects treated with arginine or insulin, a diabetic girl, and a boy with gigantism who had a serum GH content of 80 mug/liter by RIA and 40 mug/liter by bioassay, produced dose response curves parallel to that of the GH standard curve. Serum from patients with hypopituitarism did not produce significant suppression of lipid accumulation in any assay. Recovery of 5 mug/liter GH added to human serum was 94%. Twenty thousand dalton GH also suppressed lipid accumulation in this assay, but was 2-fold less potent than 22,000 dalton GH. Both bioactive GH and GH measured by RIA (RIA GH) increased in serum samples obtained after arginine stimulation in seven short subjects. The correlation between bioassay and RIA GH measurements was 0.704 (P = 0.005). Bioactive GH was greater in nocturnal peak samples of 3 girls with Turner syndrome than in pre-peak pools, where the bioactive GH values were near assay sensitivity. Exogenous GHRH administration also increased GH bioactivity in these girls. This new GH bioassay, based on suppression of lipid accumulation in 3T3-F442A adipocytes, is specific and sufficiently sensitive to allow insights into the bioavailability of GH in humans and should allow more precise determination of the role of GH isoforms in human physiology.