CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 89 NUMBER 5 | MAY 2011 693 ARTICLES nature publishing group Erythromycin is the oldest and best-known antibiotic agent in the class of macrolides, with proven efect against Gram-positive bacteria and modest efects against some Gram-negative bacte- ria as well. 1 he hepatic metabolism of erythromycin by cyto- chrome P450 (CYP) 3A isozymes has been extensively studied and was the basis for the development of the erythromycin breath test (ERMBT) for studying the CYP3A activity pheno- type in patients. 2,3 However, several studies in recent years have demonstrated that membrane transporters may afect the rate of erythromycin metabolism, thereby inluencing the interpreta- tion of the ERMBT results. 4–9 For example, a mouse model that lacks Abcb1a (mdr1a P-glycoprotein) has been shown to have a higher erythromycin metabolism rate, as determined by the ERMBT. 4 Recently, erythromycin was also shown to be a sub- strate for the rabbit transporter Abcc2 (Mrp2; cMoat), localized in the corneal epithelium. 10 Given the abundant expression of human ABCC2 along the bile canalicular membrane of hepato- cytes, 11,12 we hypothesized that impairment of ABCC2 function could also increase the metabolism of erythromycin by reduc- ing the biliary secretion of the drug, increasing hepatocellular residence time and thereby facilitating the interaction of the drug with metabolic enzymes. he goals of this study were (i) to conirm erythromycin as a substrate for human ABCC2 and evaluate the ainity of the drug for other members of the ABCC family that are expressed in the liver and (ii) to use preclinical and clinical models to conirm the in vivo relevance of ABCC2 in the metabolism of erythromycin. RESULTS In vitro transport studies To assess the extent of erythromycin difusion across cell mem- branes, a parallel artiicial membrane permeability assay was performed under steady-state conditions. he extent of transfer of erythromycin was 27 ± 1.7%, as compared with 0.40 ± 0.14% for methotrexate (an agent known to difuse poorly across cell membranes) and 54 ± 5.6% for midazolam (an agent known to difuse freely across cell membranes). his relatively low difu- sion rate supports the possibility that membrane transport of erythromycin is at least partially dependent on active carriers. In order to identify the elux proteins involved in membrane transport of erythromycin, experiments were carried out using validated transfected inside-out vesicles expressing seven diferent ATP-binding cassette (ABC) transporters (Supplementary Figure S1 online). Ater a 5-min incubation period with these vesicles, erythromycin was identiied as a substrate for human ABCC2, for mouse Abcc2, as well as for the human transporter ABCC3 (Figure 1a). his inding is consistent with the prior hypothesis that ABCC2 is closely related to ABCC3 (MRP3) because the two 1 Department of Pharmaceutical Sciences, St Jude Children’s Research Hospital, Memphis, Tennessee, USA; 2 Clinical Pharmacology Program, National Cancer Institute, Bethesda, Maryland, USA; 3 Department of Medical Oncology, Erasmus MC—Daniel den Hoed Cancer Center, Rotterdam, The Netherlands. Correspondence: A Sparreboom (alex.sparreboom@stjude.org) Received 1 November 2010; accepted 24 January 2011; advance online publication 30 March 2011. doi:10.1038/clpt.2011.25 Effect of ABCC2 (MRP2) Transport Function on Erythromycin Metabolism RM Franke 1 , CS Lancaster 1 , CJ Peer 2 , AA Gibson 1 , AM Kosloske 1 , SJ Orwick 1 , RH Mathijssen 3 , WD Figg 2 , SD Baker 1 and A Sparreboom 1 The macrolide antiobiotic erythromycin undergoes extensive hepatic metabolism and is commonly used as a probe for cytochrome P450 (CYP) 3A4 activity. By means of a transporter screen, erythromycin was identified as a substrate for the transporter ABCC2 (MRP2) and its murine ortholog, Abcc2. Because these proteins are highly expressed on the biliary surface of hepatocytes, we hypothesized that impaired Abcc2 function may influence the rate of hepatobiliary excretion and thereby enhance erythromycin metabolism. Using Abcc2 knockout mice, we found that Abcc2 deficiency was associated with a significant increase in erythromycin metabolism, whereas murine Cyp3a protein expression and microsomal Cyp3a activity were not affected. Next, in a cohort of 108 human subjects, we observed that homozygosity for a common reduced-function variant in ABCC2 (rs717620) was also linked to an increase in erythromycin metabolism but was not correlated with the clearance of midazolam. These results suggest that impaired ABCC2 function can alter erythromycin metabolism, independent of changes in CYP3A4 activity.