Assessment of macrophage migration inhibitory factor in humans: protocol for accurate and reproducible levels

The analytical validation of a possible biomarker is the first step in the long translational process from basic science to clinical routine. Although the chemokine-like cytokine macrophage migration inhibitory factor (MIF) has been investigated intensively in experimental approaches to various disease conditions, its transition into clinical research is just at the very beginning. Because of its presence in preformed storage pools, MIF is the first cytokine to be released under various stimulation conditions. In the first proof-of-concept studies, MIF levels correlated with the severity and outcome of various disease states. In a recent small study with acute coronary syndrome patients, elevation of MIF was described as a new factor for risk assessment. When these studies are compared, not only MIF levels in diseased patients differ, but also MIF levels in healthy control groups are inconsistent. Blood MIF concentrations in control groups vary between 0.56 and 95.6 ng/ml, corresponding to a 170-fold difference. MIF concentrations in blood were analyzed by ELISA. Other than the influence of this approach due to method-based variations, the impact of preanalytical processing on MIF concentrations is unclear and has not been systematically studied yet. Before large randomized studies are performed to determine the impact of circulating MIF on prognosis and outcome and before MIF is characterized as a diagnostic marker, an accurate protocol for the determination of reproducible MIF levels needs to be validated. In this study, the measurement of MIF in the blood of healthy volunteers was investigated focusing on the potential influence of critical preanalytical factors such as anticoagulants, storage conditions, freeze/thaw stability, hemolysis, and dilution. We show how to avoid pitfalls in the measurement of MIF and that MIF concentrations are highly susceptible to preanalytical factors. MIF serum concentrations are higher than plasma concentrations and show broader ranges. MIF concentrations are higher in samples processed with latency than in those processed directly and strongly correlate with hemoglobin in plasma. Neither storage temperature nor storage length or dilution or repeated freezing and thawing influenced MIF concentrations in plasma. Preanalytical validation of MIF is essential. In summary, we suggest using plasma and not serum samples when determining circulating MIF and avoiding hemolysis by processing samples immediately after blood drawing.