Development of Valid Methods to Measure Insulin-Like Growth Factors-I and -II in Bone Cell-Conditioned Medium*

Abstract

A variety of human cells and biological fluids have been shown to produce or contain insulin-like growth factor (IGF)-specific binding proteins (BPs). The existence of these BPs in serum and conditioned medium of cell and organ cultures has complicated radioligand assays for measurement of IGFs. Various strategies have been proposed to avoid interference of BPs with these assays, including acid-ethanol precipitation of BPs and acid-gel filtration. Many of these procedures are time consuming, exhibit low recoveries, and do not completely eliminate BP artifacts. In this study we have investigated interference of inhibitory IGF-BP (In-IGF-BP) purified from bone cell-conditioned medium in an IGF-II RRA and IGF-I RIA and developed methods to neutralize In-IGF-BP artifacts in IGF assays. In the IGF-II RRA, purified In-IGF-BP competed for [125I]IGF-II binding to H-35 cells in a dose-dependent manner and, thus, increased the apparent value for IGF-II in the medium. Fifty percent inhibition of [125I]IGF-II binding to H-35 cells was seen at 12.2 and 5.7 ng/ml unlabeled IGF-II and In-IGF-BP, respectively. In-IGF-BP also competed for [125I]IGF-I in the IGF-I RIA; however, the interference was much less in the IGF-I RIA compared to the IGF-II RRA. Fifty percent displacement of [125I]IGF-I binding was seen at 0.25 and 5 ng/ml unlabeled IGF-I and In-IGF-BP, respectively. Our approach to eliminate BP artifacts was as follows. We knew that In-IGF-BP showed comparatively equal binding affinities with both IGF-I and IGF-II, and binding of these ligands to cell receptors (IGF-II) and antibodies (IGF-I) was very specific (2% and 0.5% cross-reactivity for IGF-I and IGF-II, respectively). Therefore, in the IGF-I RIA we blocked the In-IGF-BP artifacts by adding an excess of IGF-II, and in the IGF-II RRA we blocked the In-IGF-BP artifacts by adding an excess of IGF-I. By incubating purified In-IGF-BP with different amounts of IGF-I, we found that 30-min preincubation of 5 ng In-IGF-BP with 10 ng IGF-I completely blocked BP artifacts in the IGF-II RRA. Similarly, preincubation of In-IGF-BP with IGF-II blocked BP artifacts in the IGF-I RIA. By using this strategy to inhibit BP artifacts with IGFs, we determined the relative levels of total IGF-I and IGF-II (free and bound to BP) in the conditioned medium derived from a number of osteosarcoma and nonosteosarcoma cell lines. The data obtained with this method were validated by an alternative method in which IGFs were separated from BPs by Amicon filtration with a YM10 membrane in 1 M acetic acid. Studies on the relative rates of production of IGF-I, IGF-II, and In-IGF-BP revealed differences between different osteosarcoma cell lines.