Author(s): Tarek Magdy 1 , Rudolf Arlanov 1 , Stefan Winter 1 , Thomas Lang 1 , Kathrin Klein 1 , Yu Toyoda 2 , Toshihisa Ishikawa 3 , Matthias Schwab 1 4 , Ulrich M Zanger [*] 5
5-fluorouracil; ABCC11; drug-metabolizing enzyme; drug transporter; haplotype; leukopenia; pharmacogenetics; pharmacogenomics; polymorphism; toxicity
For over 50 years, 5-fluorouracil (5-FU) has been in use as the backbone of chemotherapy treatment regimens for a wide range of cancers with approximately two million patients treated worldwide each year  . However, drug resistance and severe toxicity, including mucositis, diarrhea, neutropenia, and nausea and vomiting, in up to 15-20% of treated patients, often leading to dose limitation or treatment discontinuation, are still major problems in the clinical use of this drug [2,3] . The development of individualized treatment strategies including the use of predictive genetic host factors to manage toxicity and improve outcome remains an option that has so far been of limited success, presumably because important determinants are still unknown  .
Following 5-FU administration more than 85% of the administered dose is metabolized to dihydrofluorouracil by DPD, the initial and rate-limiting enzyme for 5-FU catabolism [5,6] . Only 1-3% of the administered dose is anabolized intracellularly into three main active metabolites: 5-fluoro-2´-deoxyuridine monophosphate (5-FdUMP), which forms a stable ternary complex with TS and 5,10-methylenetetrahydrofolate, thereby inhibiting TS and DNA synthesis; fluorouridine triphosphate, which is incorporated into RNA, thus disrupting RNA synthesis; and fluorodeoxyuridine triphosphate, which is incorporated into and damages cellular DNA  .
Pharmacogenomic research has so far focused on the genes encoding DPD ( DPYD ), TS (TYMS ) and MTHFR , providing overall strong evidence for their involvement in toxicity and/or drug effect [3,8-20] . In addition to these most-studied genes, few other candidate genes were studied; however, they have been studied in diverse settings and less intensely [21-26] . By contrast, the only genome-wide association study carried out in patients receiving 5-FU alone or in combination with oxaliplatin, presumably owing to lack of power, identified only one association signal in an intergenic gene region on chromosome 12 that was consistently associated with diarrhea  . These data emphasize the need for investigating new candidate genes as novel biomarkers for 5-FU toxicity.
An important class of candidate genes are drug transporters responsible for cellular uptake or efflux of anticancer drugs and/or their metabolites, as they can confer resistance to 6-mercaptopurine  , methotrexate  , vinorelbine  , pemetrexed  and other anticancer drugs. An interesting novel candidate gene in relation to 5-FU toxicity is ABCC11, encoding the transporter MRP8  . MRP8 catalyzes the extrusion of different anionic lipophilic compounds from the cell, including bile acids, sulfated steroids, glucuronides and cyclic nucleotides  . Importantly, MRP8 acts as an efflux pump for 5-FdUMP, the main active intracellular metabolite of 5-FU, thus conferring resistance to fluoropyrimidines including 5-FU in cell culture models [34,35] .
ABCC11 appears to be a functionally polymorphic gene and ethnically distinct genotype-phenotype correlations have been reported. Thus a SNP, rs17822931 (538G>A, Gly180Arg), determines dry earwax type, a Mendelian trait that is more frequent in Asians than in other populations (earwax SNP)  . Furthermore, we...