Development of a sensitive radioimmunoassay for the potential anti-migraine drug GR151004: including cross-validation to liquid chromatography—mass spectrometry—mass spectrometry and application to pharmacokinetic studies

The development of a sensitive radioimmunoassay (RIA) for the determination of the potential anti-migraine drug GR151004 in human plasma is described. The validated method was used for the determination of the drug in plasma samples collected during a Phase I safety tolerability study in healthy volunteers. Synthetic routes for the hapten and radioligand are presented. Immunogen synthesis was performed using the 2-chloro-1-methylpyridinium iodide coupling agent (Mukaiyama reaction); this method gave the highest hapten incorporation ratios during a comparison of possible conjugation reactions. Antisera were generated in Soay sheep using two different immunisation regimens. One of these regimens was a ‘fast-track’ method that was also used to compare antisera quality using immunogens prepared with three different carrier proteins. The antiserum used for the assay was produced with a GR151004 hapten — bovine thyroglobulin immunogen during a 19 week immunisation schedule. The potential specificity of the RIA was assessed by performing HPLC-RIA analysis on urine samples collected during toxicokinetic studies, and subsequently cross-reactivity measurements with a major GR151004 metabolite in man. Final confirmation of method specificity was achieved by cross-validating the RIA to an established method employing HPLC with tandem mass spectrometry (LC-MSMS). This latter study established that the two techniques afforded equivalent results. The reliable quantification range of the RIA is from 0.1 to 25.6 ng GR151004 ml−1 in human plasma, using a 0.1 ml undiluted sample. Over this concentration range the intra- and inter-assay bias is < ± 9% and imprecision is <12%. The assay drift, measured as the difference in concentration values for quality control samples assayed immediately before and after the sequence of test plasma samples, is < ± 3% (batch size 30–40 samples).