The enzyme, xanthine oxidase, has been implicated as a major source of destructive free radicals in ischemia-reperfusion injury associated with a range of pathological states [l]. Enzymic activity has been detected in many mammalian tissues [4] by using, variously, spectrophotometric [51, radiometric[6] and fluorimetric [71 assays. Apart from bovine milk fat globule membrane, which has long provided the major source of purified enzyme [a], liver and intestine generally show the highest levels of activity [4]. This is also true for humans, for whom enzyme activity in other tissues is barely detectable by conventional spectrophotometric assay [9]. Particular attention has, in recent years, has been focussed on human heart, in which the majority of workers have failed to detect significant enzyme activity [lo]. This failure contrasts with the immunofluorescence microscopy studies of Jarasch and co-workers [9] who clearly detected xanthine oxidase in capillary endothelial cells of human heart in common with such endothelial cells in many other bovine and human tissues. We have recently described the purification of xanthine oxidase from human milk [ll]. The human enzyme, surprisingly, shows specific xanthine oxidase activity that is only about 1% that of bovine milk enzyme, an observation that could reflect a predominance of 'inactive', possibly 'desulpho' form in human milk [ll]. We have also succeeded in isolating xanthine oxidase from human heart [11,12]. This enzyme too has very low activity toward conventional substrates which could explain previous failures to detect it [lo]. These wide variations in conventional xanthine oxidase activity, particularly of the human enzyme, point out a need for an immunoassay for detection of xanthine oxidase in tissue. The basis for such an assay is now described. Rabbits were immunised with bovine milk xanthine oxidase (Biozyme Laboratories, Blaenavon, UK, 1OOpg) and the resulting antiserum [13] was affinity purified on a column of bovine xanthine oxidase (Biozyme) coupled to CNBr-activated Sepharose 4B (Sigma Chemical Co., Poole, U.K.). A portion of the resulting polyclonal anti(bovine xanthine oxidase) antibodies were biotinylated as described by Kemeny 1141, for use, together with non-biotinylated antibodies, in the ELISA described below. Non-biotinylated, affinity-purified antibodies [lOOpl, lpl/ml in coating buffer (5OmM carbonate-bicarbonate buffer, pH 9.6, containing 0.01% thiomersal)] were added to each well of a 96-well microtitre plate (Becton Dickinson Labware, Plymouth, UK) and allowed to stand overnight at 4°C. The wells were then washed with phosphate buffered saline containing 0.1% Tween 20 (PBS-T) and incubated with 1% (w/v) casein in PBS-T (PBS-T/casein) for lh at room temperature. Serial two-fold dilutions of a standard solution of bovine xanthine oxidase (Biozyme, 5pg/ml in PBS-T/casein) were added to successive wells and allowed to stand for 2h at room temperature before washing with PBS-T/casein and incubation with biotinylated affinity-purified antibodies (lOOpl, 0.7Fg/ml in PBS-T/casein) f o r 2h at room temperature. The wells were again washed with PBS-T/casein and streptavidinhorse radish peroxidase (Sigma, 100~1, 0.25~g/ml in PBS-T/casein) was added to each well. The plates were allowed to stand overnight at 4C, washed extensively with PBS and incubated with peroxidase substrate (1OOp1) [substrate solution comprises tetramethylbenzidine (Sigma) in dimethylsulphoxide (10mg/ml, 0.25ml) in 0.1M sodium acetate buffer, pH 6.0, (24.75ml) containing 30% hydrogen peroxide (5pl)I. Colour was allowed to develop before stopping the reaction by addition of 1.84M sulphuric acid and measuring A,,, in a Multiscan MCC plate reader (Labsystems, Basingstoke, UK) . A plot of h9, against dilution of stock xanthine oxidase solution gave a standard curve, use of which allowed levels of enzyme as low as 0.5ng/ml consistently to be detected . We thank the Medical Research Council for a studentship to LJP.