Review | Published:

Cytochrome P450 pharmacogenetics and cancer

Oncogenevolume 25pages16791691 (2006) | Download Citation

Subjects

Abstract

The cytochromes P450 (CYPs) are key enzymes in cancer formation and cancer treatment. They mediate the metabolic activation of numerous precarcinogens and participate in the inactivation and activation of anticancer drugs. Since all CYPs that metabolize xenobiotics are polymorphic, much emphasis has been put on the investigation of a relationship between the distribution of specific variant CYP alleles and risk for different types of cancer, but a consistent view does not yet exist. This is to a great extent explained by the fact that the CYPs involved in activation of precarcinogens are in general not functionally polymorphic. This is in contrast to CYPs that are active in drug biotransformation where large interindividual differences in the capacity to metabolize therapeutic drugs are seen as a consequence of polymorphic alleles with altered function. This includes also some anticancer drugs like tamoxifen and cyclophosphamide metabolized by CYP2D6, CYP2C19 and CYP2B6. Some P450 forms are also selectively expressed in tumours, and this could provide a mechanism for drug resistance, but also future therapies using these enzymes as drug targets can be envisioned. This review gives an up-to-date description of our current knowledge in these areas.

Introduction to cytochrome P450 and cancer

Cytochrome P450 (CYP) enzymes (http://drnelson.utmem.edu/cytochromeP450.html) are key players in the phase I-dependent metabolism of drugs and other xenobiotics, mostly catalysing oxidations of the substrate, but occasionally also reduction reactions. As a result of the CYP-dependent metabolism, intermediates that often exert toxicity or carcinogenicity, but which also are targets for phase II enzyme dependent conjugation reactions are formed, rendering them inactive polar products suitable for excretion via the kidneys. Several exceptions are of course at hand where the phase II-dependent metabolism could produce more harmful products than the parent compounds, although this is not common. Many different cytotoxic drugs are inactivated by the action of CYP, whereas several prodrugs are activated by the action of CYP, rendering them cytotoxic and effective in cancer chemotherapy (McFadyen et al., 2004). Therefore, because of the important role of the CYPs in the bioactivation and inactivation of carcinogens and their participation in the activation and inactivation of anticancer drugs, they play an important role both in the aetiology of cancer diseases and as determinants of cancer therapy (Oyama et al., 2004; Rooseboom et al., 2004). These processes are mainly hepatic, but the activity of P450s in extrahepatic tissues might also be critical.

At present more than 57 active human P450 genes and 58 pseudogenes are known (Ingelman-Sundberg, 2004a; Nelson et al., 2004). The majority of genes are polymorphic and at the human CYP allele home page (http://www.imm.ki.se/cypalleles/) updated information is presented regarding the nomenclature and properties of the variant alleles with links to the dbSNP database (http://www.ncbi.nlm.nih.gov/projects/SNP/) and relevant literature references. At present, more than 434 different alleles of the genes encoding xenobiotic metabolizing P450 enzymes are presented on the page, as well as several SNPs with functional consequences, but where the corresponding allele has not yet been identified. The most polymorphic CYPs on the Web site are CYP2B6 (48 alleles), CYP2C9 (32), CYP2D6 (92) and CYP3A4 (34). Most of the functional polymorphisms are seen regarding the variability in the CYP2A6, CYP2B6, CYP2C9, CYP2C19 and CYP2D6 genes.

The mutations in the CYP genes may cause absence of enzyme, diminished enzyme expression, enzyme with altered substrate specificity or increased enzyme expression. Based on the composition of the alleles, the affected individuals might be divided into four major phenotypes: poor metabolizers (PMs), having two nonfunctional genes, intermediate metabolizers (IMs) being deficient on one allele, extensive metabolizers (EMs) having two copies of normal genes and ultrarapid metabolizers (UMs) having three or more functional active gene copies (see Ingelman-Sundberg, 2004b; Ingelman-Sundberg and Rodriguez-Antona, 2005). In the CYP gene family, the most penetrant genetic alterations are gene deletions, missense mutations and mutations creating splicing defects and premature stop codons. In only few examples mutations in the 5′- or 3′-untranslated regulatory regions affect the CYP phenotype. Despite this, a huge amount of literature reports association studies linking such low-penetrance polymorphisms to the incidence of severe diseases, among them various types of cancer (Agundez, 2004; Ingelman-Sundberg, 2004a).

The polymorphic xenobiotic metabolizing CYP enzymes can be mainly divided into two classes:

Class I, composed of CYP1A1, CYP1A2, CYP2E1 and CYP3A4, which are well conserved, do not have important functional polymorphisms, and are active in the metabolism of precarcinogens and drugs.

Class II, composed of CYP2B6, CYP2C9, CYP2C19 and CYP2D6, which are highly polymorphic and active in the metabolism of drugs, but not of precarcinogens.

No common polymorphic variant with a mutation in the open reading frame has been described for the Class I group of enzymes (Table 1). This is surprising in view of the lack of any important phenotype in the knockout mice for these CYP enzymes (Gonzalez, 2003). However, transgenic CYP3A4 knockout mice develop endocrine alterations (Yu et al., 2005) and specific functions of possible endocrine character during some phase of human development, which could explain the degree of conservation, cannot be excluded. At present only one subject with a true defective CYP3A4 allele has been documented, where the capacity for midazolam hydroxylation was severely decreased (Westlind-Johnsson et al., 2006). CYP1B1 represents a special case where several rare defective alleles have been identified and their occurrence associated to glaucoma, and, in addition, many common variant haplotypes with missense mutations are distributed in the population, but their functional consequences are less pronounced. The Cyp1b1 knockout mice develop ocular drainage structure abnormalities resembling those reported in patients having primary congenital glaucoma (Libby et al., 2003).

Table 1 Polymorphic cytochromes P450 of importance for the metabolism of drugs and carcinogens

Association studies of CYP polymorphisms with cancer incidence

Owing to the important role of CYPs in the metabolic activation of precarcinogens (see Table 2), hundreds of studies aimed at finding genetic variants that could predispose to certain types of cancer have been carried out in the past. In essence, no major conclusions from these association studies can be drawn at present. This is to a great extent due to many negative studies, relatively small risk factors obtained requiring large number of cases and controls, lack of proper control of confounding factors, relatively small functional alterations between the variant alleles studied, low penetrance of the P450 reaction in question for the development of the cancer and strong environmental factors in the etiology of the type of cancer. It is beyond the aim of this review to summarize this literature.

Table 2 Precarcinogens metabolized by cytochromes P450

A simple case clearly illustrating the role of a specific CYP in cancer formation is the null mice for Cyp1b1. Dimethylbenzanthracene (DMBA) is activated by CYP1B1, and carcinogenicity and adduct formation following DMBA injections are more frequent in mice carrying the enzyme than in the corresponding (−/−) mice (Buters et al., 1999, 2003). This system is simple in that: (i) only one gene product is studied, (ii) the genetic variation studied is drastic, that is, wt or null variants and (iii) one single carcinogen at high doses is studied. Furthermore, the mice are genetically identical except for the Cyp1b1 gene variation, and exposed to the same type of environment and food, which minimizes the effect of uncontrolled factors in the study. In contrast, such clear studies relating the influence of one specific P450 gene on cancer risk cannot be done in humans. This is because: (i) in contrast to the clear phenotypes of PMs vs EMs frequently affecting drug metabolism, such distinct phenotypes cannot be identified for most genes encoding precarcinogen-activating P450s and most of the SNPs studied have only subtle functional effects (see above); (ii) the polygenic influence on genetic susceptibility to cancer is often pronounced and the effect of P450 variation, representing a low penetrance genetic factor, is difficult to detect; (iii) the environmental factors differ a lot between individuals and are difficult to compensate for; (iv) diet, age, hormonal status, previous diseases, etc represent confounding factors that have to be taken into account and (v) the frequencies of the polymorphisms studied are often relatively low, which would require very large well-phenotyped cohorts in order to get reliable data.

Among the CYPs studied in relation to cancer, CYP1A1 and CYP2E1 have been the most commonly investigated. Although these enzymes are involved in the activation of several different precarcinogens, the allelic variants have not shown any consistent functional effect. Thus, in 16 studies investigating CYP1A1 polymorphisms and breast cancer, no associations have been found (Masson et al., 2005). In the case of CYP2E1, the c1/c2 polymorphism, an SNP at −1053 bp in the 5′-upstream regulatory region has been much studied because of one early study reporting a higher expression of the c2 allele in a gene reporter system (Hayashi et al., 1991). However, this result has not been reproduced by any other group and may indeed just represent an incidental finding. In our opinion this is a common tendency in the field of cancer association studies, where unfortunately sometimes little emphasis is given to the functionality of the genetic variations and association studies are performed without rigorous pre-validation. In addition, as mentioned, it is difficult to study such polymorphisms in case–control studies because of their relatively small role in the overall risk for cancer (cf. Vineis, 2002).

Cancer incidence, cancer therapy and CYP polymorphisms

CYP1A1/2

CYP1A1 is expressed extrahepatically and CYP1A2 is mainly expressed in the liver, indicating a very different basal regulation, but they share induction via the aryl hydrocarbon receptor (AhR), similarly to CYP1B1 (Hankinson, 1995). The CYP1A enzymes activate and detoxify numerous environmental polycyclic aromatic hydrocarbons (PAHs) and aromatic and heterocyclic amines present in combustion products such as cigarette smoke and charcoal-grilled foods. Thus, interindividual differences in CYP1A1/2 activity may influence individual susceptibility to cancer risk. CYP1A1 activity does not seem to be very variable, but there are large interindividual differences in CYP1A2 activity (Guengerich et al., 1999) and, interestingly, a study with twins phenotyped for CYP1A2 with caffeine activity showed that CYP1A2 activity was governed mainly by genetic factors, but also showed that induction by smoking is a powerful environmental factor that influences activity (Rasmussen et al., 2002). Common polymorphisms with important functional effects in CYP1A2 activity have not been identified, and only a couple of very rare genetic variants, CYP1A2*7 and CYP1A2*11, have been described. However, two common putatively important genetic variants are CYP1A2*1F and CYP1A2*1K. The former, with a −163C≥A change in intron 1, has been associated to a higher induction by smoking (Sachse et al., 1999) and omeprazole (Han et al., 2002) using caffeine as a probe drug. However, no molecular mechanism for this increased inducibility has been provided and other studies regarding the altered inducibility of the allele have been negative. On the other hand, CYP1A2*1F has been associated with increased risk for colorectal cancer (Landi et al., 2005). The CYP1A2*1K allele (−739G, −729T and −163A) results in lower constitutive CYP1A2 activity and the mutation at −729 abolishes the binding of nuclear proteins, presumably of the Ets family (Aklillu et al., 2003). Further studies are needed in order to evaluate the functional consequences of these two variant alleles and their possible influence on carcinogen-induced cancers.

CYP1B1

CYP1B1 is predominantly extra-hepatic and is frequently overexpressed in tumour tissue. Similarly to the CYP1A enzymes, CYP1B1 expression is induced by the AhR and the enzyme has an important role in the metabolism of polyaromatic carcinogens. CYP1B1 also metabolizes steroid hormones and may play a role in susceptibility to hormone-dependent cancers such as those from the breast and prostate. Five common missense mutations causing amino-acid substitutions in CYP1B1 have been identified and seven haplotypes carrying one or more of these SNPs have been characterized. For one haplotype, the corresponding enzyme CYP1B1.7 was shown to exhibit a significantly decreased capacity to metabolize estradiol (Aklillu et al., 2002) and benzo[a]-pyrene (Aklillu et al., 2005), which suggests a potential role in the interindividual differences in cancer risk or in hormone therapy. With respect to the metabolism of anticancer drugs, McFadyen et al. (2001b) showed that a cell line overexpressing CYP1B1 had a significantly decreased sensitivity towards docetaxel and Bournique and Lemarie (2002) showed that the underlying mechanism was the binding of docetaxel to CYP1B1 and an effector action of this enzyme. CYP1B1 is also involved in the 2-hydroxylation of flutamide (Rochat et al., 2001). In general, the CYP1B1 enzyme is not believed to play any major role for the overall clearance of drugs because of its extrahepatic localization, but it may play critical roles in the tissue-specific metabolism of certain drugs and physiological compounds.

CYP2A6

In the human CYP2A family, three genes, CYP2A6, CYP2A7 and CYP2A13, have been reported, but CYP2A7 is a pseudogene and CYP2A13 mainly expressed in olfactory mucosa. The CYP2A6 gene is highly polymorphic (Oscarson, 2001) and the variant genes of highest importance are CYP2A6*4, representing a gene deletion mainly present in Asian populations, CYP2A6*9, having a mutation in the TATA box which causes a decreased expression of the enzyme, and CYP2A6*1B, where a gene conversion event with CYP2A7 creates a 3′-UTR that stabilizes the corresponding mRNA (Wang et al., 2006), resulting in higher metabolism in vivo of, for example, nicotine (Nakajima et al., 2001; Gambier et al., 2005). The most important functionally altered allele, CYP2A6*4, has a 7–22% allele frequency in Asians, but only 0.5–1% in Caucasians (Oscarson et al., 1999). Another defective allele in Caucasians is the CYP2A6*2, but it is very rare. Thus, CYP2A6 does not exhibit very important polymorphism in Caucasians. As with many CYP genes, genotyping for the various alleles is difficult due to the risk of amplifying the homologous CYP2A7 pseudogene, and careful controls of the primary PCR products are necessary. The higher expression of the CYP2A6 enzyme among carriers of CYP2A6*1B apparently affects smoking behavior (Malaiyandi et al., 2005), and, for example, Gambier et al. (2005) reported that subjects homozygous for CYP2A6*1B smoked more cigarettes per day as compared to subjects homozygous for CYP2A6*1A. In Japan, where the defective CYP2A6*4 allele is very common, carriers of this genotype have been shown to have less risk of tobacco-induced lung cancer (Ariyoshi et al., 2002). This can possibly be explained by higher cigarette consumption in carriers of active CYP2A6 alleles and/or by a higher rate of formation of carcinogenic products by the action of the active CYP2A6 enzyme.

CYP2A6 metabolizes a number of tobacco-related precarcinogens (Table 2), as well as clinically important drugs such as nicotine, coumarin, methoxyflurane, halothane, valproic acid and disulfiram. Concerning anticancer drugs, CYP2A6 catalyses the activation of tegafur to 5-fluorouracil, a drug commonly used for colorectal cancer. In one study, a patient having a poor tegafur-metabolizing phenotype was found to be heterozygous for CYP2A6*4 and CYP2A6*11 (Daigo et al., 2002). However, CYP2C8 and CYP1A2 also catalyse the activation of tegafur (Komatsu et al., 2000) and further investigations are needed to clarify the impact of CYP2A6 polymorphisms on anticancer drug metabolism.

CYP2B6

The functional CYP2B6 gene and the pseudogene CYP2B7P are located in the middle of the chromosome-19 cluster, which also contains the CYP2A and CYP2F subfamilies. CYP2B6 is mainly expressed in liver, where it constitutes about 3–5% of the total microsomal P450 pool (Gervot et al., 1999; Lang et al., 2001), but it is also detected at lower levels in extrahepatic tissues, including intestine, kidney, lung, skin and the brain (Gervot et al., 1999; Miksys et al., 2003; Yengi et al., 2003). CYP2B6 activity in liver microsomes varies more than 100-fold and a broad inter-individual variability of in vivo pharmacokinetic parameters of several CYP2B6 drug substrates suggests significant interindividual differences in the systemic exposure to a variety of drugs that are metabolized by CYP2B6 (Ekins et al., 1998). CYP2B6 expression is induced through proximal and distal response elements at −1.7 and −8.5 kb via constitutive androstane receptor (CAR) (Goodwin et al., 2001; Wang et al., 2003). Well-known inducers include phenobarbital and cyclophosphamide, which will cause auto-induction. CYP2B6 is highly polymorphic and presently more than 48 different alleles have been described (http://www.imm.ki.se/CYPalleles/cyp2b6.htm); this number is increasing rapidly, indicating that the polymorphism is even higher than previously thought. No common defective CYP2B6 allele has been described but, instead, many variant alleles with amino-acid substitutions causing functional alterations, at least as assessed in heterologous expression systems, have been described. The role of the different CYP2B6 alleles for the in vivo metabolism of drugs is still largely unknown. CYP2B6*5 (R487C) and CYP2B6*7 (Q172H, K262R and R487C) variants have been suggested to cause significantly reduced protein expression levels in human liver (Lang et al., 2001), but other studies have not confirmed this (Xie et al., 2003; Hesse et al., 2004). The CYP2B6*6 allele (Q172H and K262R) has been associated with a decreased protein expression, but higher activity using cyclophosphamide as substrate (Xie et al., 2003). On the other hand, two studies showed that CYP2B6*6 carriers have a reduced in vivo capacity to metabolize efavirenz (Tsuchiya et al., 2004; Wang et al., 2005) and a lower activity using bupropion as probe drug (Hesse et al., 2004). CYP2B6*16 with K262R and I328T substitutions has a decreased stability that influences the in vivo rate of efavirenz metabolism (Wang et al., 2005). Additionally, there are several rare nonsynonymous SNPs, resulting in absent or nonfunctional proteins (Lang et al., 2004; Klein et al., 2005), but their role in vivo is not known. Further studies are thus needed in order to characterize the clinical impact of the polymorphisms identified.

CYP2B6 participates in the metabolism of a few precarcinogens and some important therapeutic drugs such as artemisinin, ketamine, propofol, bupropion and the HIV-1 reverse transcriptase inhibitors nevirapine and efavirenz. Several potent and specific inhibitors have been described, including the anticancer agent N,N′,N″-triethylene thiophosphoramide (thiotepa) (Rae et al., 2002). With respect to the metabolism of anticancer drugs, CYP2B6 is involved in the metabolic activation of the cytotoxic prodrugs cyclophosphamide, ifosfamide, thiotepa and procarbazine (Table 3). Despite the structural similarities between cyclophosphamide and ifosfamide, they have important differences in their metabolism, toxicity and therapeutic spectrum. About 45% of a therapeutic dose of ifosfamide is typically metabolized via N-dechloroethylation to the toxic chloroacetaldehyde, whereas only 10% of cyclophosphamide is converted to chloroacetaldehyde (Kaijser et al., 1993). The activation through 4-hydroxylation is mediated mainly by CYP2B6, but also by CYP3A4, CYP2C19 and CYP2C9 for cyclophosphamide and by CYP3A4 for ifosfamide (Huang et al., 2000b). The 4-hydroxy-derivative is in chemical equilibrium with aldophosphamide, which can undergo chemical decomposition into phosphoramide mustard and acrolein (Figure 1). Phosphoramide mustard is an active DNA-alkylating metabolite and acrolein is a toxic byproduct. Owing to the important role of CYP2B6 for cyclophosphamide activation, polymorphisms of this enzyme would likely affect cyclophosphamide pharmacokinetics. In this respect, Takada et al. (2004) found that patients with proliferative lupus nephritis homozygous for CYP2B6*5 and treated with pulse cyclophosphamide had a significantly higher probability of reaching end-stage renal disease with double serum creatine levels (Takada et al., 2004). Thiotepa undergoes oxidative desulfuration, by the action of CYP2B6 and CYP3A4, to a pharmacologically active compound with a longer plasma half-life, N,N′,N″-triethylene phosphoramide (TEPA) (Jacobson et al., 2002). CYP2B6 is also involved in the metabolism of tamoxifen; however, CYP2D6 and CYP3A4 appear to have a higher impact in its metabolism (Crewe et al., 2002). With respect to procarbazine, CYP2B6 and CYP1A enzymes are both involved in its activation (Goria-Gatti et al., 1992).

Table 3 Anticancer agents that are substrates for cytochromes P450 and their medical use
Figure 1
Figure 1

Chemical structure of cyclophosphamide and major biotransformation pathways. Cyclophosphamide (CPA) is activated by CYP2B6, CYP2C19, CYP2C9 and CYP3A4/5 in the liver to 4-hydroxycyclophosphamide (4-OH-CPA), which is in tautomeric equilibrium with aldophosphamide. Aldophosphamide spontaneously decompose to the alkylating metabolite phosphoramide mustard and the toxic byproduct acrolein. In a minor pathway, the dechloroacetylation of CPA mediated by CYP3A4/5 results in the toxic agent chloroacetaldehyde. In other alternative pathways the inactive 4-keto-CPA, carboxyphosphamide, alcophosphamide and 4-glutathionyl-CPA can be formed. CYP2C19*2 allele has been shown to result in a decreased rate of CPA activation. ALDH, aldehyde dehydrogenase; ADH, alcohol dehydrogenase; GST, glutathione S-transferase.

CYP2Cs

In humans, there are four CYP2C genes that have been mapped to chromosome 10q24 in the order 2C18–2C19–2C9–2C8. The corresponding enzymes are all primarily expressed in the liver, with the exception of CYP2C18, which probably represents an inactive gene. The CYP2C enzymes are involved in the metabolism of about 20% of the currently used drugs. CYP2C19 catalyses the metabolism of, for example, citalopram, diazepam and omeprazole, whereas CYP2C9 is active in the metabolism of warfarin, phenytoin and nonsteroidal anti-inflammatory drugs. CYP2C8 plays an important role in about 5% of used drugs, especially antidiabetics and antimalarials. The polymorphic behaviour of CYP2C9 and CYP2C19 has been thoroughly studied and is of very high clinical significance, whereas the clinical relevance of the CYP2C8 polymorphisms is more controversial. CYP2C8*3 (R139K and K399R) is mainly present in Caucasians (13% frequency) and, interestingly, the allele is in linkage disequilibrium with CYP2C9*2 (Yasar et al., 2002). CYP2C8.3 has been shown to exhibit a lower paclitaxel 6α-hydroxylase activity in heterologous expression systems. However, in vivo the data are contradictory and opposite results have been obtained using different substrates (Dai et al., 2001; Niemi et al., 2003; Martinez et al., 2005). Other variant alleles with missense mutations are CYP2C8*2 and CYP2C8*4, but their impact on enzyme activity is unclear. CYP2C8*5 produces a frameshift and a truncated inactive enzyme, but is very rare. The most relevant CYP2C9 genetic variations, which result in a decreased activity, are CYP2C9*2, mainly present in Caucasians with a 10–15% frequency, but almost absent in Africans and Asians, and CYP2C9*3, which has an allele frequency of 4–10% in Caucasians, 4–7% in Asians and 1–2% in Africans (Kirchheiner and Brockmoller, 2005). Concerning CYP2C19, the common CYP2C19*2 and CYP2C19*3 alleles result in enzyme deficiency. Together, they are most common in Asians (35% allele frequency), but less common in Africans (17%) and Caucasians (15%) (Desta et al., 2002). One common allele (CYP2C19*17) causes a higher expression of the enzyme due to a mutation in the 5′-upstream region recruiting nuclear factor binding resulting in increased transcription (Sim et al., 2005a).

All four human CYP2C enzymes, but to a major extent CYP2C19, catalyse the activation of cyclophosphamide and ifosfamide, but also CYP2B6 and other P450s are also implicated in these reactions (Huang et al., 2000b). In a study encompassing 60 cancer patients, data by Timm et al. (2005) indicated that those carrying the inactive CYP2C19*2 allele had a significant decreased cyclophosphamide elimination, while no differences in elimination rates of cyclophosphamide were found between subjects of different CYP2B6, CYP2C9 and CYP3A5 genotypes (Timm et al., 2005). Accordingly, Takada et al. (2004) found that, in pulse cyclophosphamide treatment of proliferative lupus nephritis, heterozygous or homozygous CYP2C19*2 patients had a significantly lower risk of developing premature ovarian failure, and in survival analysis patients homozygous for CYP2C19*2 had a higher probability of a poor renal response (Takada et al., 2004). These studies suggest that the presence of the inactive CYP2C19*2 causes a reduction in the metabolic activation of cyclophosphamide, thereby lowering the risk of toxicity but worsening the therapeutic response. Similarly, it could be envisioned that the rapid CYP2C19*17 allele with an allele frequency of about 18% in Caucasians (Sim et al., 2005a) would cause a more efficient treatment with cyclophosphamide. It could, therefore, be suggested that predictive genotyping for CYP2C19 would increase the success of cyclophosphamide treatment. In addition, CYP2C9 participates in the metabolism of idarubicin and both CYP2C9 and CYP2C19 are active in tamoxifen metabolism.

Paclitaxel undergoes extensive hepatic oxidative metabolism through 6α- and 3′-p-hydroxylations catalysed by CYP2C8 and CYP3A4, respectively. As the formation of 6α-hydroxypaclitaxel, approximately 30 times less toxic than the parent compound (Harris et al., 1994), has been shown to be the primary metabolic pathway of paclitaxel metabolism (Taniguchi et al., 2005), it would be reasonable to hypothesize that CYP2C8 polymorphisms could influence the efficacy of paclitaxel treatment. This question has been dealt with by Nakajima et al. (2005), who found a 16-fold interindividual variation in the 6α-hydroxypaclitaxel area under the curve (AUC) among 23 female ovarian cancer patients, but apparently no CYP2C8 variant alleles, due to the small study or low frequency of the alleles in Japanese. Recently, Henningsson et al. (2005), using 97 patients, found a 13-fold variation in paclitaxel clearance but no significant influence of the CYP2C8*2, CYP2C8*3 and CYP2C8*4 variant alleles. This issue, however, warrants further investigations. CYP2C8 is also involved in the metabolism of all-trans retinoic acid, which is given alongside chemotherapy in several cancers.

CYP2D6

There is an extensive interindividual variation in CYP2D6 activity, which to some extent is likely determined by the adaptation to the environment through metabolism of natural compounds such as alkaloids. The CYP2D6 gene is one of the best studied human P450 genes and correlations between the phenotype and genotype have been extensively studied for various drugs, providing a rather well-understood molecular basis for the variation in CYP2D6 activity (Ingelman-Sundberg, 2005). The polymorphisms can result in defective or increased enzyme activity and CYP2D6 genotypes usually exhibit large inter-ethnic differences. Increased activity results from gene duplication/amplification and individuals carrying up to 13 functional CYP2D6 copies in one allele have been found (Johansson et al., 1993; Aklillu et al., 1996). Defective CYP2D6 allelic variants carry gene deletions, stop codons or splicing defects, and the most common functionally altered variants are CYP2D6*4 (15–21% in Caucasians), CYP2D6*5 (about 3–6% in the different populations), CYP2D6*10 (38–70% in Asians, 3–9% in Africans) and CYP2D6*17 (20–34% in Africans). Furthermore, a large number of CYP2D6 polymorphisms with lower frequencies, but resulting in a defective enzyme, also contribute to the extensive interindividual variation in CYP2D6 activity (http://www.imm.ki.se/CYPalleles/cyp2d6.htm).

CYP2D6 is involved in the metabolism of 20–25% of all drugs in clinical use, and it has a special impact on the treatment of psychiatric and cardiovascular diseases. By contrast, the role of CYP2D6 in the metabolism of precarcinogens is minor and the polymorphism of the enzyme is apparently without importance for interindividual differences in susceptibility for cancer. CYP2D6 has been shown to play a crucial role in the metabolism of tamoxifen, which is an estrogen receptor modulator widely used for the endocrine treatment of all stages of hormone receptor-positive breast cancer. Tamoxifen is activated by the CYP system to antioestrogenic metabolites that are more potent than the parent compound (Jin et al., 2005). In vitro studies implicated many CYP isoforms such as CYP3A, CYP2D6, CYP2C9, CYP2C19, CYP2B6 and CYP1A2, in the biotransformation of tamoxifen. However, the key metabolites of tamoxifen seem to be 4-hydroxytamoxifen and endoxifen, formed primarily by CYP2D6, and N-desmethyltamoxifen, formed primarily by CYP3A4 (Figure 2). In patients receiving tamoxifen, the most abundant compounds in plasma are N-desmethyltamoxifen and endoxifen, and it has been shown that endoxifen has approximately 100 times greater affinity for the oestrogen receptor than tamoxifen and N-desmethyltamoxifen (Jordan et al., 1977; Clarke et al., 2003; Jin et al., 2005). As endoxifen is mainly formed by the action of CYP2D6, patients with defective CYP2D6 alleles would obtain less benefit from tamoxifen therapy than those carrying functional copies of CYP2D6. Thus, in a study of 80 women with breast cancer starting tamoxifen adjuvant therapy, the plasma concentrations of endoxifen after 4 months of therapy were significantly lower in patients being homozygous or heterozygous for defective CYP2D6 genes as compared to those with two functional alleles (Jin et al., 2005). Additionally, those subjects using CYP2D6 inhibitors had 58% reduction in the plasma concentration of endoxifen. The CYP2D6 genotype is also relevant for cancer patients with respect to the action of the antiemetic drugs tropisetron and ondasetron. Lower plasma levels and higher frequency and intensity of vomiting were found in subjects carrying a higher number of active CYP2D6 gene copies (Kaiser et al., 2002).

Figure 2
Figure 2

Chemical structure of tamoxifen and major biotransformation pathways. CYP3A4/5 are the more efficient enzymes responsible for the N-demethylation of tamoxifen (TAM), whereas the generations of endoxifen and 4-hydroxytamoxifen (4-OH-TAM) are predominantly catalysed by CYP2D6. Other CYP isoforms, including CYP2C19, CYP2C9, CYP2B6 and CYP1A2, have also been shown to participate in the metabolism of tamoxifen. The most abundant compounds in plasma are N-desmethyltamoxifen and endoxifen, and endoxifen has approximately 100 times greater affinity for the oestrogen receptor than tamoxifen and N-desmethyltamoxifen. CYP2D6 polymorphisms have been shown to affect the plasma concentrations of endofixen.

CYP2E1

CYP2E1 is responsible for the metabolism and activation of a large number of low-molecular-weight chemicals, solvents, cancer suspect agents and a few drugs (Table 2). Thus, CYP2E1 might be an important determinant of human susceptibility to toxicity and carcinogenicity of industrial and environmental chemicals. However, polymorphisms affecting CYP2E1 expression or activity have not been found, probably because of high conservation due to a critical role of the enzyme in gluconeogenesis during conditions of starvation. By contrast, induction of the enzyme by, for example, alcohol might provide a more important factor for interindividual susceptibility to cancer in reactions mediated by CYP2E1.

CYP3A4/5

The human CYP3A locus carries four genes, but only CYP3A4, CYP3A5 and CYP3A7 encode active enzymes relevant for the metabolism of a wide range of structurally different xenobiotics. The expression of these enzymes is regulated in a tissue-specific manner, the P450s being predominant in the liver and gastrointestinal tract. During fetal stages CYP3A4 is absent, while CYP3A7 expression is maximum. In adult life the predominant hepatic P450 is CYP3A4, with some individuals also exhibiting a significant expression of the ‘fetal’ CYP3A7 (Lacroix et al., 1997; Sim et al., 2005b). CYP3A5 is mainly absent from Caucasian livers but contributes to CYP3A activity in Africans (Kuehl et al., 2001). The substrate specificities of the CYP3A enzymes are overlapping, but CYP3A4 usually exhibits a higher specific activity towards many CYP3A substrates when compared to CYP3A5 and CYP3A7 (Williams et al., 2002). The CYP3A enzymes are involved in the metabolism of about 50% of all drugs currently on the market (Li et al., 1995) and they participate in the metabolic activation and metabolism of several carcinogens such as aflatoxin B and also of anticancer drugs (see Tables 2 and 3). Interindividual variation in CYP3A activity, thus, has a major impact on pharmacokinetics and metabolism of a majority of different drugs. Generally, a five-fold interindividual variability in clearance of CYP3A substrates in vivo has been found with some scarce ‘outliers’. This variation can be caused by environmental factors or drugs that inhibit or induce CYP3A enzymes but, additionally, it has been shown that the variation is determined to a high extent by genetic factors (Ozdemir et al., 2000). Important genetic polymorphisms that severely decrease the expression of CYP3A5 protein have been described, that is, CYP3A5*3, CYP3A5*6 and CYP3A5*7 (see Kuehl et al., 2001; Lee et al., 2003). However, this is not true for CYP3A4 since, despite the analysis of thousands of subjects, no major functionally variant allele has been found at an allele frequency higher than 0.1%. The only allele that appears to influence the CYP3A4 expression is CYP3A4*1B, common in Africans and present at 5% frequency in Caucasians, through alteration of nuclear proteins binding to the polymorphic element (Rodriguez-Antona et al., 2005). The distribution of the CYP3A4*1B allele has been associated to prostate and lung cancer, although the data are generally conflicting (Rebbeck et al., 1998; Paris et al., 1999; Spurdle et al., 2002; Tayeb et al., 2002, 2003; Dally et al., 2003). The basis for any genetic background for the interindividual variation in CYP3A4 expression remains a challenge.

With respect to the action of anticancer drugs, the variability of CYP3A4 is expected to influence the outcome of several different treatments. Docetaxel is metabolized by CYP3A4 to inactive hydroxylated derivatives (Figure 3) and, therefore, a high CYP3A4 activity would result in a poor therapeutic outcome of the drug. Accordingly, in cancer patients treated with docetaxel in combination with the potent CYP3A4 inhibitor ketoconazole, a 49% decrease in docetaxel clearance was found (Engels et al., 2004). Similarly, hepatic CYP3A4 activity measured by the erythromycin breath test and midazolam clearance predicted docetaxel clearance, finding the greatest toxicity in patients with the lowest CYP3A4 activity (Hirth et al., 2000; Goh et al., 2002). Furthermore, Yamamoto et al. (2005) phenotyped CYP3A4 in patients with advanced non-small-cell lung cancer by measuring urinary 6-beta-OHF after cortisol administration and found that an individualized dosing method, based on CYP3A4 phenotyping, decreased the pharmacokinetic variability of docetaxel when compared to body-surface area-based dosing (Yamamoto et al., 2005). In addition, CYP3A4 expression in breast tumour tissue has been shown to predict therapeutic response to docetaxel (Miyoshi et al., 2002, 2005). Similarly to docetaxel, irinotecan is inactivated by CYP3A4 and induction of CYP3A4 in patients receiving irinotecan results in a significant decrease in the formation of the toxic metabolite of this drug (Friedman et al., 1999; Mathijssen et al., 2002). Additionally, Mathijssen et al. (2004) showed that CYP3A4 phenotype, as assessed by midazolam clearance, is statistically significantly associated with irinotecan pharmacokinetics. With respect to the already described CYP3A genotypes, a combination of CYP3A4, CYP3A5, GSTM1 and GSTT1 genotypes was shown to influence the probability of treatment failure after high-dose adjuvant chemotherapy for node-positive breast cancer (DeMichele et al., 2005). Table 3 also shows other important anticancer agents metabolized by CYP3A4, including taxanes, vinca-alkaloids and new drugs such as imatinib and gefitinib.

Figure 3
Figure 3

Chemical structure of docetaxel and major biotransformation pathways. Docetaxel (DOC) is inactivated in the liver by CYP3A4/5 through successive oxidations of the tert-butyl ester group of the C13-side chain, mainly the direct hydroxylation to an alcohol (hydroxyl-DOC) and a subsequent oxidation to an unstable aldehyde, which cycles to two stereomeric hydroxyoxazolidinones. DOC can also 7-epimerize to a diasteromer, which undergoes metabolic pathways similar to DOC. CYP3A4 phenotype has been shown to influence DOC pharmacokinetics and therapeutic outcome.

P450 expression in tumours

In addition to an interindividual variability in the pharmacokinetics of anticancer drugs caused by hepatic CYPs, an altered CYP activity in the tumour cells could result in an altered drug efficacy. Cancer cells by means of genetic or epigenetic mechanisms, due to their higher DNA instability and more frequent alterations in chromatin structure than nontumour cells, could alter P450 transcription. The capacity of the tumours to metabolize drugs is a potential means to achieve optimal therapy by activation of prodrugs in the cancer cells; however, it is also a potential mechanism of resistance to therapy by an increased inactivation of anticancer drugs caused by an overexpression of P450s. Many studies have reported the presence of drug-metabolizing enzymes in tumours (Dhaini et al., 2003; Gharavi and El-Kadi, 2004; Oyama et al., 2004; Downie et al., 2005; Kumarakulasingham et al., 2005). However, differences in the quantification and sampling techniques and heterogeneous patient populations have resulted, in some cases, in conflicting data, making it difficult to conclude about any impact on deactivation of anticancer agents or activation of prodrugs.

With respect to the impact of tumour P450s on drug therapy outcome, Tanaka et al. (2004), using 19 human cancer cell lines and eight common anticancer drugs, measured the cytotoxic activity of the drugs and performed cDNA microarray analysis to identify associations between specific gene expression and effect of the drug in question (Tanaka et al., 2004). In all, 12 genes with proven functional significance to drug sensitivity, which included CYP2C8 and CYP3A4, were selected and prediction models to accurately predict the in vitro efficacy of the drugs were developed. The in vivo relevance of the model was tested for 5-fluorouracil treatment in gastric cancer patients. The model of predictive value in terms of survival, time to treatment failure and tumour growth showed that the tumour phenotype was indeed related to the therapeutic response to 5-fluorouracil (Tanaka et al., 2004). Miyoshi et al. (2002, 2005) showed that a low CYP3A4 expression in breast tumours, as determined at mRNA and protein level, resulted in a better response to docetaxel, which is inactivated by CYP3A4. Similarly, Dhaini et al. (2003) showed that a high CYP3A expression in osteosarcoma tumours from 18 patients predicted metastasis and poor prognosis. CYP3A4 is involved in the oxidation of compounds that are usually used as chemotherapeutic agents for the treatment of osteosarcomas such as etoposide, ifosfamide, cyclophosphamide and doxorubicin, suggesting that the response to these drugs could be worse in tumours with high CYP3A expression, increasing the risk of metastasis. Therefore, the main hepatic drug metabolizing P450 enzymes if expressed in the tumour cells could influence the success of drug therapy.

In addition, some extrahepatic P450s, many of which have major roles in the metabolism of endogenous substrates and are not involved in xenobiotic biotransformation, have been found to be overexpressed in tumour tissue. These include CYP1B1, CYP2J2, CYP2W1 and CYP4Z1. The AhR binds several carcinogens that are metabolized by the CYP1 enzymes and regulates the expression of CYP1B1 (Nebert et al., 2004), which has been found to be overexpressed in a large number of tumours, including cancers of the prostate, kidney, ovarian, breast and colon tumours (Murray et al., 1997; McFadyen et al., 1999, 2001a; Gibson et al., 2003; Tokizane et al., 2005). CYP2J2 is able to metabolize arachidonic acid to epoxyeicosatrienoic acids, which have been suggested to play a role in angiogenesis and to exert antiapoptotic effects (Chen et al., 2001; Jiang et al., 2005; Pozzi et al., 2005). CYP2J2 was much overexpressed relative to adjacent normal tissue in the majority of tumours examined, which included esophageal squamous cell carcinoma, esophageal adenocarcinoma, pulmonary squamous cell carcinoma, pulmonary adenocarcinoma, small-cell pulmonary carcinoma, breast carcinoma, stomach carcinoma, liver carcinoma and colon adenocarcinoma (Jiang et al., 2005). CYP2W1 has been shown to be almost exclusively expressed during embryogenesis and in adult humans it is mainly detected in tumour tissue samples, more frequently from colon and adrenal gland (Karlgren et al., 2005, submitted). With respect to CYP4Z1, it is regulated by the glucocorticoid and progesterone receptors and has been shown to be overexpressed preferentially in breast carcinoma tissue and mammary gland (Rieger et al., 2004; Savas et al., 2005).

P450 as a drug target in cancer therapy

A major objective of cancer research is the development of therapeutic agents specifically targeted to tumour cells. P450s expressed at higher levels in the tumour cells than in the surrounding normal tissue offer therapeutic options by the activation of prodrugs specifically in the cancer cells and avoiding undesirable systemic effects (see Riddick et al., 2005). In this respect, there are therapeutic options and opportunities arising from both the enhanced endogenous expression of CYP in tumours and CYP-mediated gene therapy. Concerning endogenous overexpression of individual forms of P450 enzymes in tumour cells, CYP1B1 is the best studied example, because although several CYP1As, CYP2Cs and CYP3As exhibit enhanced expression in some tumour cells, these enzymes display considerable expression in normal tissue, mainly in the liver. On the other hand, CYP1B1 mRNA and protein expression has been found in a wide range of malignant tumours and in metastatic disease (McFadyen et al., 2001a), but the CYP1B1 protein is generally not detected in normal tissue at important levels (Gibson et al., 2003). Taking advantage of this, several agents activated by CYP1B1 are currently in preclinical evaluation, such as resveratrol and phortress (Potter et al., 2002; Leong et al., 2003); in addition, there is a CYP1B1 vaccine (Zyc300) in phase I/II trials, aimed to destroy cancer cells through induction of T-cell response (Gribben et al., 2005). Similar strategies could be initiated with other P450s mainly identified in tumour cells, such as CYP2W1, CYP2J2 and CYP4Z1, after identification of an appropriate prodrug. The polymorphism of these genes in relation to the success of P450-based cancer therapy remains to be elucidated.

Gene therapy offers another approach to get a differential P450 expression between tumour/normal tissue, where an exogenous P450 gene and a prodrug activated by that P450 are delivered to the tumour. The enzyme expression can be genetically controlled or its delivery targeted to ensure tumour selectivity. The gene-directed enzyme prodrug therapy systems with CYP have been mainly based on cyclophosphamides, which needs to be activated mainly by CYP2B6. Expression of CYP enzymes has been shown to sensitize cells to both cyclophosphamide in a range of cell lines in vitro and the bystander effect is mediated through the soluble derivative 4-hydroxycyclophosphamide. Hepatic P450 enzymes like CYP2B6 or CYP2B1 as well as P450 reductase have been inserted into 9L gliosarcoma cells by viral transfection in order to facilitate tumour growth suppression in cultured cells and in xenogaft models upon treatment with anticancer agents (Huang et al., 2000a; Roy and Waxman, 2005). In addition, a combinatory treatment of cyclophosphamide and another drug impairing the hepatic expression of P450 reductase and hence minimizing the hepatic activation of cyclophosphamide can be used (Huang et al., 2000a). In vitro and animal models showed promising effects of this approach (McFadyen et al., 2004; Dachs et al., 2005) and, for example, in a trial of 14 patients with inoperable pancreatic cancer, the median survival was doubled in the treatment group compared to historical controls and 1-year survival improved three-fold (Lohr et al., 2001; Salmons et al., 2003).

Conclusions

CYPs have important roles in activation and inactivation of both precarcinogens and of anticancer drugs (see Figure 4). Interindividual differences in the P450-mediated actions are caused both by environmental and genetic factors. Due to the relatively high extent of conservation of genes encoding CYPs participating in the activation of precarcinogens, the genetic factors are less important determinants of individual susceptibility, whereas inducers of P450s like smoking, ethanol, etc., appear to be more relevant factors for such variability. An exception might be CYP2A6 in Asia, where the functional polymorphism is pronounced. The metabolism of several anticancer drugs is catalysed by specific polymorphic forms of CYP, like CYP2B6, CYP2C19 and CYP2D6. Here the knowledge about the different CYP alleles distributed in the populations and their functional consequences is relatively well known, whereas the impact of the polymorphism for in vivo treatment with anticancer drugs remains largely to be elucidated. The recent achievements in using the polymorphic P450 as drug targets in cancer therapy are promising and could provide a novel and effective alternative of future cancer therapy.

Figure 4
Figure 4

Xenobiotic metabolizing CYP enzymes and cancer. The CYP enzymes can be divided into two different groups: Class I composed of CYP1A1, CYP1A2, CYP2E1 and CYP3A4 and Class II composed of CYP2B6, CYP2C9, CYP2C19 and CYP2D6. Class I enzymes are in general well conserved and active in the metabolism of precarcinogens and drugs, while Class II enzymes have important functional polymorphisms and are active in the metabolism of drugs, but not of precarcinogens. Inducers play an important role in regulating the expression of Class I enzymes, with the exception of CYP2E1, and to a minor extent of Class II enzymes. Class I P450s are important for the aetiology of cancer diseases, while Class II P450s play an important role in cancer therapy.

References

  1. Agundez JA . (2004). Curr Drug Metab 5: 211–224.

  2. Aklillu E, Carrillo JA, Makonnen E, Hellman K, Pitarque M, Bertilsson L et al. (2003). Mol Pharmacol 64: 659–669.

  3. Aklillu E, Oscarson M, Hidestrand M, Leidvik B, Otter C, Ingelman-Sundberg M . (2002). Mol Pharmacol 61: 586–594.

  4. Aklillu E, Ovrebo S, Botnen IV, Otter C, Ingelman-Sundberg M . (2005). Cancer Res 65: 5105–5111.

  5. Aklillu E, Persson I, Bertilsson L, Johansson I, Rodrigues F, Ingelman-Sundberg M . (1996). J Pharmacol Exp Ther 278: 441–446.

  6. Ariyoshi N, Miyamoto M, Umetsu Y, Kunitoh H, Dosaka-Akita H, Sawamura Y et al. (2002). Cancer Epidemiol Biomarkers Prev 11: 890–894.

  7. Bournique B, Lemarie A . (2002). Drug Metab Dispos 30: 1149–1152.

  8. Buters J, Quintanilla-Martinez L, Schober W, Soballa VJ, Hintermair J, Wolff T et al. (2003). Carcinogenesis 24: 327–334.

  9. Buters JT, Sakai S, Richter T, Pineau T, Alexander DL, Savas U et al. (1999). Proc Natl Acad Sci USA 96: 1977–1982.

  10. Chen JK, Capdevila J, Harris RC . (2001). Mol Cell Biol 21: 6322–6331.

  11. Clarke R, Liu MC, Bouker KB, Gu Z, Lee RY, Zhu Y et al. (2003). Oncogene 22: 7316–7339.

  12. Crewe HK, Notley LM, Wunsch RM, Lennard MS, Gillam EM . (2002). Drug Metab Dispos 30: 869–874.

  13. Dachs GU, Tupper J, Tozer GM . (2005). Anticancer Drugs 16: 349–359.

  14. Dai D, Zeldin DC, Blaisdell JA, Chanas B, Coulter SJ, Ghanayem BI et al. (2001). Pharmacogenetics 11: 597–607.

  15. Daigo S, Takahashi Y, Fujieda M, Ariyoshi N, Yamazaki H, Koizumi W et al. (2002). Pharmacogenetics 12: 299–306.

  16. Dally H, Edler L, Jager B, Schmezer P, Spiegelhalder B, Dienemann H et al. (2003). Pharmacogenetics 13: 607–618.

  17. DeMichele A, Aplenc R, Botbyl J, Colligan T, Wray L, Klein-Cabral M et al. (2005). J Clin Oncol 23: 5552–5559.

  18. Desta Z, Zhao X, Shin JG, Flockhart DA . (2002). Clin Pharmacokinet 41: 913–958.

  19. Dhaini HR, Thomas DG, Giordano TJ, Johnson TD, Biermann JS, Leu K et al. (2003). J Clin Oncol 21: 2481–2485.

  20. Downie D, McFadyen MC, Rooney PH, Cruickshank ME, Parkin DE, Miller ID et al. (2005). Clin Cancer Res 11: 7369–7375.

  21. Ekins S, Vandenbranden M, Ring BJ, Gillespie JS, Yang TJ, Gelboin HV et al. (1998). J Pharmacol Exp Ther 286: 1253–1259.

  22. Engels FK, Ten Tije AJ, Baker SD, Lee CK, Loos WJ, Vulto AG et al. (2004). Clin Pharmacol Ther 75: 448–454.

  23. Friedman HS, Petros WP, Friedman AH, Schaaf LJ, Kerby T, Lawyer J et al. (1999). J Clin Oncol 17: 1516–1525.

  24. Gambier N, Batt AM, Marie B, Pfister M, Siest G, Visvikis-Siest S . (2005). Pharmacogenomics J 5: 271–275.

  25. Gervot L, Rochat B, Gautier JC, Bohnenstengel F, Kroemer H, de Berardinis V et al. (1999). Pharmacogenetics 9: 295–306.

  26. Gharavi N, El-Kadi AO . (2004). Curr Drug Metab 5: 203–210.

  27. Gibson P, Gill JH, Khan PA, Seargent JM, Martin SW, Batman PA et al. (2003). Mol Cancer Ther 2: 527–534.

  28. Goh BC, Lee SC, Wang LZ, Fan L, Guo JY, Lamba J et al. (2002). J Clin Oncol 20: 3683–3690.

  29. Gonzalez FJ . (2003). Drug Metab Rev 35: 319–335.

  30. Goodwin B, Moore LB, Stoltz CM, McKee DD, Kliewer SA . (2001). Mol Pharmacol 60: 427–431.

  31. Goria-Gatti L, Iannone A, Tomasi A, Poli G, Albano E . (1992). Carcinogenesis 13: 799–805.

  32. Gribben JG, Ryan DP, Boyajian R, Urban RG, Hedley ML, Beach K et al. (2005). Clin Cancer Res 11: 4430–4436.

  33. Guengerich FP, Parikh A, Turesky RJ, Josephy PD . (1999). Mutat Res 428: 115–124.

  34. Han XM, Ouyang DS, Chen XP, Shu Y, Jiang CH, Tan ZR et al. (2002). Br J Clin Pharmacol 54: 540–543.

  35. Hankinson O . (1995). Annu Rev Pharmacol Toxicol 35: 307–340.

  36. Harris JW, Katki A, Anderson LW, Chmurny GN, Paukstelis JV, Collins JM . (1994). J Med Chem 37: 706–709.

  37. Hayashi S, Watanabe J, Kawajiri K . (1991). J Biochem (Tokyo) 110: 559–565.

  38. Henningsson A, Marsh S, Loos WJ, Karlsson MO, Garsa A, Mross K et al. (2005). Clin Cancer Res 11: 8097–8104.

  39. Hesse LM, He P, Krishnaswamy S, Hao Q, Hogan K, von Moltke LL et al. (2004). Pharmacogenetics 14: 225–238.

  40. Hirth J, Watkins PB, Strawderman M, Schott A, Bruno R, Baker LH . (2000). Clin Cancer Res 6: 1255–1258.

  41. Huang Z, Raychowdhury MK, Waxman DJ . (2000a). Cancer Gene Ther 7: 1034–1042.

  42. Huang Z, Roy P, Waxman DJ . (2000b). Biochem Pharmacol 59: 961–972.

  43. Ingelman-Sundberg M . (2004a). Naunyn-Schmiedeberg's Arch Pharmacol 369: 89–104.

  44. Ingelman-Sundberg M . (2004b). Trends Pharmacol Sci 25: 193–200.

  45. Ingelman-Sundberg M . (2005). Pharmacogenomics J 5: 6–13.

  46. Ingelman-Sundberg M, Rodriguez-Antona C . (2005). Philos Trans R Soc Lond B Biol Sci 360: 1563–1570.

  47. Jacobson PA, Green K, Birnbaum A, Remmel RP . (2002). Cancer Chemother Pharmacol 49: 461–467.

  48. Jiang JG, Chen CL, Card JW, Yang S, Chen JX, Fu XN et al. (2005). Cancer Res 65: 4707–4715.

  49. Jin Y, Desta Z, Stearns V, Ward B, Ho H, Lee KH et al. (2005). J Natl Cancer Inst 97: 30–39.

  50. Johansson I, Lundqvist E, Bertilsson L, Dahl ML, Sjoqvist F, Ingelman-Sundberg M . (1993). Proc Natl Acad Sci USA 90: 11825–11829.

  51. Jordan VC, Collins MM, Rowsby L, Prestwich G . (1977). J Endocrinol 76: 305–316.

  52. Kaijser GP, Korst A, Beijnen JH, Bult A, Underberg WJ . (1993). Anticancer Res 13: 1311–1324.

  53. Kaiser R, Sezer O, Papies A, Bauer S, Schelenz C, Tremblay PB et al. (2002). J Clin Oncol 20: 2805–2811.

  54. Karlgren M, Gomez A, Stark K, Svärd J, Rodriguez-Antona C, Oliw E et al. (submitted).

  55. Karlgren M, Miura S, Ingelman-Sundberg M . (2005). Toxicol Appl Pharmacol 207: 57–61.

  56. Kirchheiner J, Brockmoller J . (2005). Clin Pharmacol Ther 77: 1–16.

  57. Klein K, Lang T, Saussele T, Barbosa-Sicard E, Schunck WH, Eichelbaum M et al. (2005). Pharmacogenet Genomics 15: 861–873.

  58. Komatsu T, Yamazaki H, Shimada N, Nakajima M, Yokoi T . (2000). Drug Metab Dispos 28: 1457–1463.

  59. Kuehl P, Zhang J, Lin Y, Lamba J, Assem M, Schuetz J et al. (2001). Nat Genet 27: 383–391.

  60. Kumarakulasingham M, Rooney PH, Dundas SR, Telfer C, Melvin WT, Curran S et al. (2005). Clin Cancer Res 11: 3758–3765.

  61. Lacroix D, Sonnier M, Moncion A, Cheron G, Cresteil T . (1997). Eur J Biochem 247: 625–634.

  62. Landi S, Gemignani F, Moreno V, Gioia-Patricola L, Chabrier A, Guino E et al. (2005). Pharmacogenet Genomics 15: 535–546.

  63. Lang T, Klein K, Fischer J, Nussler AK, Neuhaus P, Hofmann U et al. (2001). Pharmacogenetics 11: 399–415.

  64. Lang T, Klein K, Richter T, Zibat A, Kerb R, Eichelbaum M et al. (2004). J Pharmacol Exp Ther 311: 34–43.

  65. Lee SJ, Usmani KA, Chanas B, Ghanayem B, Xi T, Hodgson E et al. (2003). Pharmacogenetics 13: 461–472.

  66. Leong CO, Gaskell M, Martin EA, Heydon RT, Farmer PB, Bibby MC et al. (2003). Br J Cancer 88: 470–477.

  67. Li AP, Kaminski DL, Rasmussen A . (1995). Toxicology 104: 1–8.

  68. Libby RT, Smith RS, Savinova OV, Zabaleta A, Martin JE, Gonzalez FJ et al. (2003). Science 299: 1578–1581.

  69. Lohr M, Hoffmeyer A, Kroger J, Freund M, Hain J, Holle A et al. (2001). Lancet 357: 1591–1592.

  70. Malaiyandi V, Sellers EM, Tyndale RF . (2005). Clin Pharmacol Ther 77: 145–158.

  71. Martinez C, Garcia-Martin E, Blanco G, Gamito FJ, Ladero JM, Agundez JA . (2005). Br J Clin Pharmacol 59: 62–69.

  72. Masson LF, Sharp L, Cotton SC, Little J . (2005). Am J Epidemiol 161: 901–915.

  73. Mathijssen RH, de Jong FA, van Schaik RH, Lepper ER, Friberg LE, Rietveld T et al. (2004). J Natl Cancer Inst 96: 1585–1592.

  74. Mathijssen RH, Verweij J, de Bruijn P, Loos WJ, Sparreboom A . (2002). J Natl Cancer Inst 94: 1247–1249.

  75. McFadyen MC, Breeman S, Payne S, Stirk C, Miller ID, Melvin WT et al. (1999). J Histochem Cytochem 47: 1457–1464.

  76. McFadyen MC, Cruickshank ME, Miller ID, McLeod HL, Melvin WT, Haites NE et al. (2001a). Br J Cancer 85: 242–246.

  77. McFadyen MC, McLeod HL, Jackson FC, Melvin WT, Doehmer J, Murray GI . (2001b). Biochem Pharmacol 62: 207–212.

  78. McFadyen MC, Melvin WT, Murray GI . (2004). Mol Cancer Ther 3: 363–371.

  79. Miksys S, Lerman C, Shields PG, Mash DC, Tyndale RF . (2003). Neuropharmacology 45: 122–132.

  80. Miyoshi Y, Ando A, Takamura Y, Taguchi T, Tamaki Y, Noguchi S . (2002). Int J Cancer 97: 129–132.

  81. Miyoshi Y, Taguchi T, Kim SJ, Tamaki Y, Noguchi S . (2005). Breast Cancer 12: 11–15.

  82. Murray GI, Taylor MC, McFadyen MC, McKay JA, Greenlee WF, Burke MD et al. (1997). Cancer Res 57: 3026–3031.

  83. Nakajima M, Fujiki Y, Kyo S, Kanaya T, Nakamura M, Maida Y et al. (2005). J Clin Pharmacol 45: 674–682.

  84. Nakajima M, Kwon JT, Tanaka N, Zenta T, Yamamoto Y, Yamamoto H et al. (2001). Clin Pharmacol Ther 69: 72–78.

  85. Nebert DW, Dalton TP, Okey AB, Gonzalez FJ . (2004). J Biol Chem 279: 23847–23850.

  86. Nelson DR, Zeldin DC, Hoffman SM, Maltais LJ, Wain HM, Nebert DW . (2004). Pharmacogenetics 14: 1–18.

  87. Niemi M, Leathart JB, Neuvonen M, Backman JT, Daly AK, Neuvonen PJ . (2003). Clin Pharmacol Ther 74: 380–387.

  88. Oscarson M . (2001). Drug Metab Dispos 29: 91–95.

  89. Oscarson M, McLellan RA, Gullsten H, Yue QY, Lang MA, Bernal ML et al. (1999). FEBS Lett 448: 105–110.

  90. Oyama T, Kagawa N, Kunugita N, Kitagawa K, Ogawa M, Yamaguchi T et al. (2004). Front Biosci 9: 1967–1976.

  91. Ozdemir V, Kalowa W, Tang BK, Paterson AD, Walker SE, Endrenyi L et al. (2000). Pharmacogenetics 10: 373–388.

  92. Paris PL, Kupelian PA, Hall JM, Williams TL, Levin H, Klein EA et al. (1999). Cancer Epidemiol Biomarkers Prev 8: 901–905.

  93. Potter GA, Patterson LH, Wanogho E, Perry PJ, Butler PC, Ijaz T et al. (2002). Br J Cancer 86: 774–778.

  94. Pozzi A, Macias-Perez I, Abair T, Wei S, Su Y, Zent R et al. (2005). J Biol Chem 280: 27138–27146.

  95. Rae JM, Soukhova NV, Flockhart DA, Desta Z . (2002). Drug Metab Dispos 30: 525–530.

  96. Rasmussen BB, Brix TH, Kyvik KO, Brosen K . (2002). Pharmacogenetics 12: 473–478.

  97. Rebbeck TR, Jaffe JM, Walker AH, Wein AJ, Malkowicz SB . (1998). J Natl Cancer Inst 90: 1225–1229.

  98. Riddick DS, Lee C, Ramji S, Chinje EC, Cowen RL, Williams KJ et al. (2005). Drug Metab Dispos 33: 1083–1096.

  99. Rieger MA, Ebner R, Bell DR, Kiessling A, Rohayem J, Schmitz M et al. (2004). Cancer Res 64: 2357–2364.

  100. Rochat B, Morsman JM, Murray GI, Figg WD, McLeod HL . (2001). J Pharmacol Exp Ther 296: 537–541.

  101. Rodriguez-Antona C, Sayi JG, Gustafsson LL, Bertilsson L, Ingelman-Sundberg M . (2005). Biochem Biophys Res Commun 338: 299–305.

  102. Rooseboom M, Commandeur JN, Vermeulen NP . (2004). Pharmacol Rev 56: 53–102.

  103. Roy P, Waxman DJ . (2005). Toxicol In Vitro (in press).

  104. Sachse C, Brockmoller J, Bauer S, Roots I . (1999). Br J Clin Pharmacol 47: 445–449.

  105. Salmons B, Lohr M, Gunzburg WH . (2003). J Gastroenterol 38(Suppl 15): 78–84.

  106. Savas U, Hsu MH, Griffin KJ, Bell DR, Johnson EF . (2005). Arch Biochem Biophys 436: 377–385.

  107. Sim SC, Edwards RJ, Boobis AR, Ingelman-Sundberg M . (2005b). Pharmacogenet Genomics 15: 625–631.

  108. Sim SC, Risinger C, Dahl M-L, Aklillu E, Christensen M, Bertilsson L et al. (2005a). Clin Pharm Ther (in press).

  109. Spurdle AB, Goodwin B, Hodgson E, Hopper JL, Chen X, Purdie DM et al. (2002). Pharmacogenetics 12: 355–366.

  110. Takada K, Arefayene M, Desta Z, Yarboro CH, Boumpas DT, Balow JE et al. (2004). Arthritis Rheum 50: 2202–2210.

  111. Tanaka T, Tanimoto K, Otani K, Satoh K, Ohtaki M, Yoshida K et al. (2004). Int J Cancer 111: 617–626.

  112. Taniguchi R, Kumai T, Matsumoto N, Watanabe M, Kamio K, Suzuki S et al. (2005). J Pharmacol Sci 97: 83–90.

  113. Tayeb MT, Clark C, Haites NE, Sharp L, Murray GI, McLeod HL . (2003). Br J Cancer 88: 928–932.

  114. Tayeb MT, Clark C, Sharp L, Haites NE, Rooney PH, Murray GI et al. (2002). Oncol Rep 9: 653–655.

  115. Timm R, Kaiser R, Lotsch J, Heider U, Sezer O, Weisz K et al. (2005). Pharmacogenomics J 5: 365–373.

  116. Tokizane T, Shiina H, Igawa M, Enokida H, Urakami S, Kawakami T et al. (2005). Clin Cancer Res 11: 5793–5801.

  117. Tsuchiya K, Gatanaga H, Tachikawa N, Teruya K, Kikuchi Y, Yoshino M et al. (2004). Biochem Biophys Res Commun 319: 1322–1326.

  118. Vineis P . (2002). Toxicology 181–182: 457–462.

  119. Wang H, Faucette S, Sueyoshi T, Moore R, Ferguson S, Negishi M et al. (2003). J Biol Chem 278: 14146–14152.

  120. Wang J, Pitarque M, Ingelman-Sundberg M . (2006). Biochem Biophys Res Commun 340: 491–497.

  121. Wang J, Sönnerborg A, Rane A, Josephson F, Lundgren S, Ståhle L et al. (2005). Pharmacogenet Genomics (in press).

  122. Westlind-Johnsson A, Bernhard Hauns D, Hermann R, Huennemeyer A, Lahu G, Nassr N et al. (2006). Clin Pharm Ther (in press).

  123. Williams JA, Ring BJ, Cantrell VE, Jones DR, Eckstein J, Ruterbories K et al. (2002). Drug Metab Dispos 30: 883–891.

  124. Xie HJ, Yasar U, Lundgren S, Griskevicius L, Terelius Y, Hassan M et al. (2003). Pharmacogenomics J 3: 53–61.

  125. Yamamoto N, Tamura T, Murakami H, Shimoyama T, Nokihara H, Ueda Y et al. (2005). J Clin Oncol 23: 1061–1069.

  126. Yasar U, Lundgren S, Eliasson E, Bennet A, Wiman B, de Faire U et al. (2002). Biochem Biophys Res Commun 299: 25–28.

  127. Yengi LG, Xiang Q, Pan J, Scatina J, Kao J, Ball SE et al. (2003). Anal Biochem 316: 103–110.

  128. Yu AM, Fukamachi K, Krausz KW, Cheung C, Gonzalez FJ . (2005). Endocrinology 146: 2911–2919.

Download references

Acknowledgements

The research at the author's laboratories is supported by The Swedish Research Council, The Swedish Cancer Foundation, NIH (NIGMS 1-R01 GM60548), by the ‘Ramon y Cajal’ program from the Spanish Ministry of Education and Science, and by a Marie Curie European Reintegration Grants of the European Community programme Structuring the European Research Area under contract number MERG-CG-6-2005-014881.

Author information

Affiliations

  1. Endocrine Cancer Group, Human Cancer Genetics Programme, Spanish National Cancer Center (CNIO), Madrid, Spain

    • C Rodriguez-Antona
  2. Department of Physiology and Pharmacology, Karolinska Institutet, Stockholm, Sweden

    • M Ingelman-Sundberg

Authors

  1. Search for C Rodriguez-Antona in:

  2. Search for M Ingelman-Sundberg in:

Corresponding author

Correspondence to M Ingelman-Sundberg.

About this article

Publication history

Published

Issue Date

DOI

https://doi.org/10.1038/sj.onc.1209377

Further reading