Cytotoxicity is predicted by unbound and not total bilirubin concentration.

Although it has been suggested that the unbound, free, (B(f)) rather than total (B(T)) bilirubin level correlates with cell toxicity, direct experimental evidence supporting this conclusion is limited. In addition, previous studies never included a direct measurement of B(f), using newer, accurate methods. To test "the free bilirubin hypothesis", in vitro cytotoxicity was assessed in four cell lines exposed to different B(f) concentrations obtained by varying B(T)/Albumin ratio, using serum albumins with different binding affinities, and/or displacing unconjugated bilirubin (UCB) from albumin with a sulphonamide. B(f) was assessed by the modified, minimally diluted peroxidase method. Cytotoxicity varied among cell lines but was invariably related to B(f) and not B(T). Light exposure decreased toxicity parallel to a decrease in B(f). In the absence of albumin, no cytotoxicity was found at a B(f) of 150 nM whereas in the presence of albumin a similar B(f) resulted in a 40% reduction of viability indicating the importance of total cellular uptake of UCB in eliciting toxic effect. In the presence of albumin-bound UCB, bilirubin-induced cytotoxicity in a given cell line is accurately predicted by B(f) irrespective of the source and concentration of albumin, or total bilirubin level.