Determination of free follistatin levels in sera of normal subjects and patients with various diseases.

We developed an assay system for measuring free follistatin by using an anti-follistatin mouse monoclonal antibody and [125I]activin A. The sensitivity of this assay was 0.5 microgram/l and cross-reactivities with inhibin, luteinizing hormone, follicle-stimulating hormone and growth hormone were all less than 0.5%. The dose-response curves of human sera and follicular fluid were parallel to the standard curve, and the follicular fluid contained a large amount of follistatin (6.4 +/- 0.5 mg/l, mean +/- SEM; N = 13). The within- and between-assay coefficients of variation calculated from the analysis of serum samples of four different concentrations were 3.3-7.8% and 3.9-11.0%, respectively. The recovery rates of free follistatin at five different doses were 86.4 - 102.4%. When activin A was added to the same sample, free follistatin recovery rate declined dose-dependently. Gel filtration analyses of human serum and follicular fluid resulted in a single peak corresponding to authentic follistatin. Using this assay, free follistatin concentrations in sera were measured in normal, pregnant and diseased subjects. The free follistatin level in serum of normal adults was 3.5 +/- 0.2 micrograms/l (N = 60), which was significantly elevated in pregnant women (16.7 +/- 1.3 micrograms/l, N = 56), and in patients with chronic liver disease (8.1 +/- 1.1 micrograms/l, N = 20), chronic renal failure (6.7 +/- 0.9 micrograms/l, N = 42), advanced solid cancer (8.5 +/- 1.0 micrograms/l, N = 39) and hematological malignancies (6.8 +/- 1.0 micrograms/l, N = 18). These data indicated that the free follistatin concentration in serum is detectable and varies during pregnancy and in various diseased states.