Effective strategies for tumors affecting chemopreventive metabolism.

In this issue, Charles et al. (1) showed in vivo that cytokines secreted by tumors, such as interleukin-6, down-regulate the transcription of cytochrome P450 CYP3A4 in liver and consequently the metabolism of CYP3A4 substrates. Reduced metabolism of chemotherapeutic agents could be responsible for side effects observed for these agents. The investigators postulate that treatment of the inflammation evoked by the tumor could normalize the expression of CYP3A4 and perhaps other cytochromes P450 and hence reduce the side effects of drugs used for the treatment of cancers. Chemotherapeutic treatment of cancers has a narrow therapeutic window and can be accompanied by life-threatening side effects. It would be desirable to dose patients individually with a dose high enough to kill tumor cells but below a concentration that yields life-threatening side effects. In cancer chemotherapies, dosing is commonly done according to body surface area for lack of a better variable and dosing is reduced when leukocyte counts decline (i.e., patients are dosed to the maximum tolerable dose under the axiom ‘‘more is better’’). Dosing to body surface area is understandable as it gives an exact dimension to the intuitive notion that one size dosing does not fit all and that larger patients require a higher dose than smaller patients do. This is based on an allometric approach that size predicts the speed of physiologic processes, including the rate of elimination of drugs. Although this approach is applicable for drugs that are excreted unchanged, it does not apply for drugs where metabolism determines the rate of excretion. Many chemotherapeutic agents are detoxified by metabolism, and the activities of metabolizing enzymes vary between individuals. This explains why a standard allometric approach, such as dosing according to body surface area, is not very suited for drugs that are detoxified by metabolism (2). Variations in activity of metabolizing enzymes and thus the pharmacokinetics of the drugs they metabolize are influenced genetically but also environmentally and, as shown by Charles et al., possibly by the presence of inflammation caused by tumor cells. Predicting deviating pharmacokinetics in individuals from their genome has required years of intense research, and success has been slow. Nevertheless, polymorphisms in thiopurine methyltransferase (TPMT*2; 3A; 3C) for 6-mercaptopurine, UDP-glucuronyltransferase (UGT1A1*28 and *6) for irinotecan, and dihydropyrimidine dehydrogenase (DPYD*2A) for 5-fluorouracil predict slow detoxifying capacity and consequently more side effects when individuals carrying these mutations receive standard doses (3). These examples represent the extremes of the spectrum: absent catalytic activity of the major metabolizing enzyme. For many drugs, a single enzyme does not govern metabolism, and slow metabolism by one major enzyme stays unnoticed because the other enzymes involved in the metabolism of a compound are active enough to guarantee elimination within the reference range. Pharmaceutical companies deliberately try to avoid, if possible, bringing compounds to the market in which a single enzyme dominates elimination because of the possibility of unexpected side effects in a subpopulation of poor metabolizers. To date, however, sufficient clinically used chemotherapeutics remain in which a single enzyme dominates the major detoxifying or activating pathway. On top of the genetic variation, environmental factors play a dominant role in the metabolic variation between individuals. Inducers and inhibitors in food, concomitant viral infections (4), inflammation evoked by tumors, and inhibitory and inducing drugs ensure that metabolic enzymes are not a static unity, but vary according to the situation. These environmental factors influence the farnesoid X-activated receptor, constitutive androstane receptor, and nuclear factor-nB pathways of CYP3A4 regulation (5). Because of the many factors influencing absorption and metabolism, measuring blood levels of a chemotherapeutic would be ideal. This needs expensive equipment, however, and in addition to technical complexity, we lack defined concentrations at which therapy will be effective and also concentrations at which side effects will occur. Despite these difficulties, Yamamoto et al. (6) have shown that using a probe for CYP3A4 activity to aid dosing reduced side effects in patients receiving docetaxel. The hunt for probes representing metabolic capacity, parallel to creatinine for kidney clearance, has been on for decades. CYP3A4 is the main liver and intestinal cytochrome P450 and is involved in the metabolism of a large number of drugs (7), including chemotherapeutics, such as paclitaxel, docetaxel, vincristine, vinblastine, imatinib, gefitinib, etoposide, and ifosphamide. The activity of CYP3A4 shows a 20-fold interpatient variation (8). The activity of CYP3A4 can be predicted by the specific probes midazolam (elimination from blood), [C-N-methyl]erythromycin breath test (a radioactive-labeled methyl group is cleaved of erythromycin by CYP3A4 and appears in exhaled air as carbon dioxide), or 6h-hydroxycortisol to cortisol ratio in urine. All Editorial