Failure of an indirect competitive enzyme-linked immunosorbent assay to measure prolactin in rat plasma samples

An indirect competitive enzyme-linked immunosorbent assay (ELISA) for rat prolactin was applied to the measurement of prolactin (PRL) in plasma. The assay involved PRL adsorbed to a 96-well plate competing with soluble PRL in the sample for rabbit anti-rat PRL antibody-binding sites. Antibody bound by the PRL on the well was detected by a goat anti-rabbit immunoglobulin antibody conjugated to the enzyme horseradish peroxidase. Dilutions of PRL in plasma-free assay buffer produced accurate and reliable standard curves over a range of 1-500 ng/ml. Plasma or serum samples, however, gave very low absorbance values, suggesting falsely high levels of PRL in the samples. For example, hypophysectomized rat plasma and fetal calf serum gave readings equivalent to over 250 ng/ml rat PRL by ELISA, whereas radioimmunoassay confirmed the expected PRL values of less than 5 ng/ml. This result indicated considerable nonspecific interference by the plasma. Inclusion of 20% or 50% hypophysectomized rat plasma in the assay significantly lowered the absorbance values and the slope of the standard curve (P less than 0.05). Dilution of plasma samples 1:1 and 1:4 with assay buffer caused nonlinear changes in absorbance and resulted in inaccurate ELISA estimates of plasma PRL concentrations compared with radioimmunoassay. Several treatments were attempted to remove the interference from plasma. Preexposure of plasma to 56 degrees C for 30 min, dialysis for 24 hr, fractional precipitation with 5% polyethylene glycol, increased ionic strength of the buffer, preadsorption of plasma with rabbit immunoglobulins bound to a solid phase support, and covalent attachment of the PRL to the wells with 2% glutaraldehyde all failed to remove the interference from the plasma samples in the ELISA. Thus, while the assay worked effectively for plasma-free applications, it was unsuitable for measurement of PRL in plasma or serum samples.