Toxicity of xanthene food dyes by inhibition of human drug-metabolizing enzymes in a noncompetitive manner.

The synthetic food dyes studied were rose bengal (RB), phroxine (PL), amaranth, erythrosine B (ET), allura red, new coccine, acid red (AR), tartrazine, sunset yellow FCF, brilliant blue FCF, and indigo carmine. First, data confirmed that these dyes were not substrates for CYP2A6, UGT1A6, and UGT2B7. ET inhibited UGT1A6 (glucuronidation of p-nitrophenol) and UGT2B7 (glucuronidation of androsterone). We showed the inhibitory effect of xanthene dye on human UGT1A6 activity. Basic ET, PL, and RB in those food dyes strongly inhibited UGT1A6 activity, with IC(50) values = 0.05, 0.04, and 0.015 mM, respectively. Meanwhile, AR of an acidic xanthene food dye showed no inhibition. Next, we studied the inhibition of CYP3A4 of a major phase I drug-metabolizing enzyme and P-glycoprotein of a major transporter by synthetic food dyes. Human CYP3A4 and P-glycoprotein were also inhibited by basic xanthene food dyes. The IC(50) values of these dyes to inhibit CYP3A4 and P-glycoprotein were the same as the inhibition level of UGT1A6 by three halogenated xanthene food dyes (ET, PL, and RB) described above, except AR, like the results with UGT1A6 and UGT2B7. We also confirmed the noninhibition of CYP3A4 and P-gp by other synthetic food dyes. Part of this inhibition depended upon the reaction of (1)O(2) originating on xanthene dyes by light irradiation, because inhibition was prevented by (1)O(2) quenchers. We studied the influence of superoxide dismutase and catalase on this inhibition by dyes and we found prevention of inhibition by superoxide dismutase but not catalase. This result suggests that superoxide anions, originating on dyes by light irradiation, must attack drug-metabolizing enzymes. It is possible that red cosmetics containing phloxine, erythrosine, or rose bengal react with proteins on skin under lighting and may lead to rough skin.