Irinotecan pharmacogenomics

Irinotecan: clinical background and pharmacology

Irinotecan, a topoisomerase I inhibitor, is a semisynthetic derivative of camptothecin, a plant alkaloid isolated from Camptotheca acuminata (Iyer and Ratain, 1998). It has demonstrated activity against various malignancies (Fuchs et al., 2003; Noda et al., 2002; Reardon et al., 2003), and is currently approved for the treatment of colorectal carcinoma in the United States (Vanhoefer et al., 2001). Irinotecan is a prodrug that needs to be activated by carboxylesterases to the active metabolite 7-ethyl-10-hydroxycamptothecin (SN-38) (Figure 1) (Humerickhouse et al., 2000). Irinotecan also undergoes oxidative metabolism catalysed by CYP3A4 and CYP3A5 to various predominantly inactive metabolites, and the oxidative metabolism of CPT-11 reduces the CPT-11 available for carboxylesterase-catalysed activation to SN-38 (Santos et al., 2000). The major pathway for SN-38 elimination is via glucuronidation to the inactive SN-38 glucuronide (SN-38G) (Iyer et al., 1998). Although the overall pharmacology of irinotecan is quite complex, the clinical importance of irinotecan metabolism is well established. In particular, reduced glucuronidation of the active metabolite SN-38 is associated with increase in the two severe toxicities associated with irinotecan: diarrhea and neutropenia (Gupta et al., 1994; Iyer et al., 2002). The late-onset diarrhea associated with irinotecan is believed to occur as a result of direct enteric injury caused by the biliary excretion of SN-38 (Araki et al., 1993), and in previous studies conducted using the weekly schedule of irinotecan administration, SN-38 glucuronidation was observed to protect against the intestinal toxicities of SN-38 by decreasing its biliary excretion (Gupta et al., 1994). Hence, polymorphic glucuronidation of SN-38 has appropriately been the focus of the early clinical pharmacogenetic studies involving irinotecan.

Figure 1.

Irinotecan metabolism – Irinotecan is inactivated by CYP3A4/5 to oxidative metabolites with minimal activity. Irinotecan is activated to SN-38 by carboxylesterase 2, and the active moiety eventually gets glucuronidated by the polymorphic UGT1A1, thus far the pathway that has been the focus of irinotecan pharmacogenetic studies

Full figure and legend (17K)

Pharmacogenetics of SN-38 glucuronidation

Uridine diphosphate glucuronosyltransferases (UGTs) are microsomal enzymes that catalyse the glucuronidation of numerous endogenous and exogenous substrates (Burchell et al., 1995). The human UGTs are classified into UGT1 and UGT2 families, and further subclassified into subfamilies and isoforms, based on sequence homology (Mackenzie et al., 1997). SN-38 glucuronidation is catalysed by the polymorphic UDP-glucuronosyltransferase 1A1 (UGT1A1), the enzyme responsible for bilirubin glucuronidation (Iyer et al., 1998). UGT1A1 is the most extensively studied UGT isoform, and is known to contain over 30 genetic variants, many of which influence its expression and functional properties (Burchell and Hume, 1999; Tukey and Strassburg, 2000). The UGT1A1 genetic polymorphisms are associated with several hyperbilirubinemic syndromes resulting from genetic traits characterized by absent or very low UGT1A1 activity (Burchell and Hume, 1999). Gilbert's syndrome characterized by mild unconjugated hyperbilirubinemia is relatively common (Sampietro and Iolascon, 1999). The syndrome is typically associated with homozygosity for a dinucleotide (TA) insertion in the (TA)₆TAA element in the UGT1A1 promoter resulting in variant allele (TA)₇ (UGT1A1^*28), which leads to a 70% reduction in UGT1A1 gene expression as compared to the more common (TA)₆ allele (Bosma et al., 1995; Monaghan et al., 1996; Beutler et al., 1998). The UGT1A1 activity appears to be inversely related to the number of TA repeats, since the transcriptional activity of the promoter decreases with an increase in the number of TA repeats (Bosma et al., 1995; Monaghan et al., 1996; Beutler et al., 1998). The frequency of (TA)₇ homozygotes has been reported to be about 0.5–23% in various populations (Monaghan et al., 1997; Beutler et al., 1998; Burchell and Hume 1999). The frequency of (TA)₇ homozygosity is much lower in the Asians compared to Caucasians and Africans; however, other missense mutations in the coding region of UGT1A1 have been described as a common cause of Gilbert's syndrome in the Asians (Figure 2). For instance, UGT1A1^*6 (211G>A) is the most common amino-acid change reported in Asians, with reported allele frequencies in the range of 13–23% (Sato et al., 1996; Akaba et al., 1998), and results in a 30% (heterozygotes) and 60% (homozygotes) decrease in bilirubin-glucuronidating activity. On the other hand, the frequency of missense mutations is low among Caucasians. Compound heterozygosity for the promoter polymorphism and the coding region mutations has also been reported to cause Gilbert's syndrome in various Asian populations (Akaba et al., 1998; Huang et al., 2000). Evidence of the fact that individuals with Gilbert's syndrome might be at an increased risk for drug toxicity has been illustrated by pharmacogenetic studies involving irinotecan, which will be described below (Ando et al., 2000; Iyer et al., 2002).

Figure 2.

Panel a depicts the promoter variant, UGT1A1^*28, which results from a TA insertion in the (TA)6TAA element of the UGT1A1 promoter region. This alteration leads to decreased gene expression and resultant decrease in function. Panel b depicts the various missense mutations occurring in the coding region of UGT1A1. The promoter variant is more common in Caucasians and African-Americans, whereas the missense mutations are more common in Asians

Full figure and legend (47K)

Irinotecan pharmacogenetic trials

The clinical studies of UGT1A1 polymorphisms strongly suggest that patients deficient in SN-38 glucuronidation could experience increased irinotecan-mediated toxicity. Ando et al. (2000) retrospectively analysed the association between UGT1A1 variants and irinotecan toxicity. Their study was conducted in a Japanese population, and hence they attempted to elucidate the clinical significance of the UGT1A^*28 mutation as well as the missense mutations, which are more common in the Asian populations. They evaluated the relationship between the severe toxicity of irinotecan (grade 4 leukopenia and/or grade 3/4 diarrhea) and UGT1A1 mutations (UGT1A1^*28 in the promoter and UGT1A1^*6, UGT1A1^*27 in the first exon) in the study. The frequencies of UGT1A1 mutations in 26 patients with severe toxicity were compared to those in 92 patients who did not experience toxicity after irinotecan. The frequency of UGT1A1^*28 was 3.5-fold higher in patients with toxicity compared to patients without toxicity (P<0.0001). Although UGT1A1^*6 is highly prevalent among Asian patients with Gilbert's syndrome, UGT1A1^*6 frequency was not significantly different between the two groups. All the three patients who were heterozygous for UGT1A1^*27 experienced severe toxicity. This study also showed that combined mutations (i.e., presence of both UGT1A1^*28 and one of the coding mutations) might increase the risk of toxicity (four out of five patients carrying both UGT1A1^*28 and either UGT1A1^*6 or UGT1A1^*27 suffered life-threatening toxicities). Multivariate analysis demonstrated that the presence of UGT1A1^*28 allele was a risk factor for severe toxicity (P<0.001, odds ratio of 7.23, 95% CI of 2.5–22.3).

There is also evidence to suggest that a -3279G>T change identified in the phenobarbital-responsive enhancer module (PBREM) of UGT1A might contribute to hyperbilirubinemia in Asians (Sugatani et al., 2002). Innocenti et al. (2002) have recently reported that the common promoter variants are in linkage disequilibrium, and that the haplotype structure of UGT1A promoter is likely to be different among various races. These results demonstrated significant linkage disequilibrium between two variant sites in the PBREM (-3279G>T and -3156G>A) and the (TA)_n polymorphism (P<0.0001) in the Caucasian population, whereas the linkage disequilibrium was not as significant in the African-Americans. This information suggests that future analysis should also consider including the PBREM variants for data analysis. Acuna et al. (2002) have recently reported the haplotype structure for UGT1A, and their study demonstrates the potential benefit of using haplotypes for performing genotype–phenotype association studies. Generally, it appears that for most genotype–phenotype association studies data analysis might be more robust if the association studies are carried out based on haplotype structures of the genes of interest. Undertaking association studies based on individual single-nucleotide polymorphisms (SNPs) might miss true correlation, as information provided by such studies has the risk of overlooking variants in linkage disequilibrium with other genes or regulatory regions of interest.

Next, the preliminary results of a prospective pharmacogenetic trial that recently completed patient accrual at the University of Chicago will be discussed. The trial was based on the initial in vitro studies that demonstrated a good concordance between the UGT1A1 TA indel genotype and SN-38 glucuronidation (Iyer et al., 1999). The trial was designed to evaluate the role of UGT1A1^*28 polymorphism in the toxicity and disposition profiles of irinotecan, given at 300 mg/m² every 3 weeks (Iyer et al., 2002). In 20 cancer patients, the pharmacokinetics of SN-38 relative to SN-38G were markedly altered by UGT1A1^*28; UGT1A1^*28 carriers had significantly lower SN-38G/SN-38 AUC ratios compared to UGT1A1^*1 subjects. As SN-38 AUC was inversely related to ANC nadir (r=–0.81, P<0.0001), more severe neutropenia was reported in UGT1A1^*28 carriers compared to UGT1A1^*1 carriers (nonparametric trend analysis, P=0.04). All UGT1A1^*1 patients had grade 0–1 diarrhea and neutropenia, and higher grades of toxicity were observed in patients with the UGT1A1^*28 allele. Moreover, a trend towards worse diarrhea was observed in patients with the UGT1A1^*28 allele. This trial was subsequently amended to study the pharmacogenetics of irinotecan, given at the indicated dose of irinotecan of 350 mg/m² on the 3-week schedule. The study was designed to enroll 60 patients, to investigate a relationship between UGT1A1 genotype and grade 3+ diarrhea. However, the incidence of grade 3+ diarrhea on the trial was much lower (5%) than expected based on the literature (Vanhoefer et al., 2001; Fuchs et al., 2003). The observed incidence of grade 4 neutropenia was 8%, and preliminary analysis of the data demonstrates a significant association between the UGT1A1 genotype and grade 4 neutropenia (P<0.001). Among the 28 patients who were homozygotes for the common allele (UGT1A1^*1), none experienced grade 4 neutropenia; whereas 3/6 patients homozygous for the variant allele (UGT1A1^*28) experienced grade 4 neutropenia (Innocenti et al., 2003).

Thus, several observations from different institutions have pointed out that UGT1A1 mutations can change patients' exposure to SN-38, and, hence, their susceptibility to toxicity. UGT1A1 promoter polymorphism is clinically relevant, due to the more severe toxicity observed in patients with UGT1A1^*28 allele. At the present time, we do not know whether irinotecan dose modification based on genotype is appropriate, since dose reduction in patients at risk for toxicity may result in a decreased response. The only clinical recommendation that can be made is that the drug should be used with caution in patients with Gilbert's syndrome, based on either UGT1A1^*28 genotyping or an elevated unconjugated bilirubin level. Currently, the role of other genetic determinants in irinotecan pharmacology is also unknown. To address this issue, the prospective trial described earlier will also evaluate the functional role of CYP3A, carboxylesterase and MDR1 variants as well. Based on the irinotecan pharmacology, it is certainly plausible that these additional variants might also contribute to the altered drug response of irinotecan. The planned multivariate analysis for the trial will attempt to elucidate the clinical importance of different environmental and genetic factors affecting the safety and efficacy of irinotecan. The prospective clinical trial also demonstrates that even meticulously planned studies might not demonstrate clinically relevant associations as a result of unpredictable changes in the incidence of assumed clinical end points, either due to selection bias associated with nonrandomized studies or unpredictable changes in prospectively assumed clinical outcomes for uncertain reasons. Nonetheless, conservatively planned and adequately powered studies that utilize multiple end points can minimize the chances of failure.

TS Pharmacogenomics

TS is a folate-dependent enzyme that catalyses the reaction of deoxyuridine-5'-monophosphate to deoxythymidine-5'-monophosphate. It is widely regarded that inhibition of TS is one of the major targets for fluorouracil (5-FU) and its anticancer activity (Johnston et al., 1992, 1995; Leichman et al., 1997). Overexpression of TS has been linked with resistance to 5-FU-based therapies. TS overexpression has also been linked to poor prognosis of patients with various malignancies, and is inversely related to clinical response (Johnston et al., 1994; Pestalozzi et al., 1995; Lenz et al., 1996; Kornmann et al., 1997; Aschele et al., 1999). The control mechanism of TS regulation is quite complex, and involves various regulatory steps from the level of transcription to post-translation (Chu et al., 1991, 1999; Dolnick, 2000). The TS expression has been reported as being partially controlled by a polymorphism in the thymidylate synthase enhancer region (TSER) consisting of variable number of copies of a 28-base pair (bp) tandem-repeat sequence (Horie et al., 1995). In the TSER, two, three, four and nine copies of the 28-bp repeats (TSER^*2, TSER^*3, TSER^*4 and TSER^*9) have been described, with TSER^*2 and TSER^*3 being the common alleles in all the populations studied (Kaneda et al., 1987; Horie et al., 1995; Marsh et al., 2000; Luo et al., 2002). In vitro studies have demonstrated that increasing number of tandem repeats leads to an increase in TS gene expression and TS enzyme activity (Kawakami et al., 1999). The relative probability of response to therapy has also been correlated with the TSER genotypes in several studies. Pullarkat et al. (2001) reported that individuals homozygous for the TSER^*3 allele had a 3.6-fold higher TS mRNA levels compared to individuals homozygous for TSER^*2 allele. TSER^*2 carriers were noted to have a relatively higher probability of pathological downstaging after neoadjuvant chemoradiotherapy, compared to TSER^*3 homozygotes with rectal cancer (Villafranca et al., 2001). In another study, TSER variants were evaluated in the liver metastases of patients, and the analysis suggested that TSER^*3 homozygotes demonstrated higher overall TS expression and a lower response rate compared to TSER^*2 homozygotes (Marsh et al., 2001). There are other similar examples in the literature as well (Grau et al., 2003; Kawabe et al., 2003); however, most of the studies have been small retrospective studies; hence, their applicability in clinical practice is currently limited. Also, different studies have utilized different methods (ranging from real-time PCR to immunohistochemistry) and tissues (tumor versus surrogate peripheral tissues) to quantify TS levels, thus making direct comparisons of the results problematic. Adding to the argument are reports suggesting that TS genotype might not necessarily correlate with protein levels or mRNA levels in some instances (Soong et al., 2003). A recent report by Mandola et al. (2003) might change the landscape of this science if confirmed clinically. The investigators have identified two and one USF family E-box consensus element(s) within the tandem repeats of TSER^*3 and TSER^*2 alleles, respectively (Figure 3). In vitro, these consensus elements were demonstrated to bind USF proteins, and USF-1 binding enhanced TS transcription. Furthermore, the additional USF consensus element found only with TSER^*3 was shown to confirm transcriptional advantage to TSER^*3 over TSER^*2. The authors have also identified a novel G>C SNP in the second of the three 28-bp repeats of TSER^*3. The novel G>C variant is within the USF consensus element; thus, it alters the USF protein binding to the element, and leads to a decrease in TS transcription. The new variant was prevalent at a high frequency (range 28–56%) in the four populations (non-Hispanic whites, Hispanic whites, African Americans and Singapore Chinese) that were studied. The authors also propose that this novel SNP should be genotyped in future pharmacogenomic trials of TS inhibitors, as they expect the new SNP to be more predictive of drug response. These new data are quite provocative; however, they need to be confirmed in the clinical arena. A simple starting point for testing the new hypothesis would be to utilize available tissue from the previous retrospective trials that have investigated TSER variants, and examine the relevance of the new SNP using the existing drug response data. Eventually, prior to routine clinical use, the proposed hypothesis will have to be proven in prospective clinical studies.

Figure 3.

The 28-bp tandem repeat regions of the two most common TS variants TSER^*2 and TSER^*3 (consisting of either two or three repeats, respectively) are depicted (adapted from Mandola et al., 2003). The shaded regions represent the newly identified putative E-box binding site for upstream stimulatory factor (USF). The USF consensus elements are present in repeats 1 and 2 of TSER^*3, and only in repeat 1 of TSER^*2. The last repeat in both TSER^*2 and TSER^*3 contains a G>C change (arrows) that disrupts this site. A novel G>C SNP (asterisk), only identified within the USF consensus element of second 28-bp repeat of TSER^*3, decreases the in vitro transcriptional activation of TS

Full figure and legend (29K)

EGFR pharmacogenomics: paradigm for targeted therapeutics

The improved understanding of different aspects of cancer biology over the past two decades has led to identification of several new putative therapeutic targets, which has also led to an increased focus on development of 'targeted' therapeutics. One of the important distinctions in the development of the new targeted agents is that the use of traditional dose-finding studies based on maximum tolerated dose (MTD) as an approach has been questioned (Stadler and Ratain, 2000; Korn et al., 2001). The paradigm of defining 'maximum effective dose' based on nonclinical and clinical pharmacodynamic studies to identify therapeutic concentrations sufficient for inhibition of the desired targets is being considered more relevant. Furthermore, some of the novel agents are also being thought of as disease-stabilizing agents that would require prolonged treatment durations (Kohn et al., 1997; Hidalgo et al., 2001). In such instances, even the traditional 'moderate' toxicities considered tolerable for cytotoxic agents administered over relatively short durations can become 'dose- or compliance-limiting' toxicities. Therefore, identifying the pharmacodynamic consequences of genetic variation and minimizing the unpredictability becomes critical during early development of the agents.

A significant amount of work over the past 10 years has suggested that various growth factors and growth factor receptors play crucial roles in the pathogenesis of human malignancies, and the epidermal growth factor receptor (EGFR) is one of the more extensively studied growth factor receptors. EGFR, a membrane spanning 170-kDa glycoprotein, stimulates cell proliferation after ligand binding and receptor dimerization (Savage and Cohen, 1972; Adamson and Rees, 1981; Carpenter, 1987). Extensive data have implicated the EGFR signaling pathway in both carcinogenesis and prognosis of various solid tumors, and activation of EGFR tyrosine kinase (TK) has been identified as a key event that initiates intracellular signaling, which regulates cell proliferation, differentiation and survival (Salomon et al., 1995; Akimoto et al., 1999; Chen et al., 2000; Porebska et al., 2000; Sung et al., 2000; Ciardiello and Giampaolo, 2001). Overexpression of EGFR is a prevalent alteration of the EGFR pathway in human tumors, and the level of EGFR expression is predominantly regulated by the abundance of its mRNA (Merlino et al., 1985a, 1985b). The transcription initiation of EGFR gene derives from multiple initiation sites within a GC-rich promoter region (Ishii et al., 1985). The first intron of EGFR has been shown to regulate gene transcription, and it contains a CA dinucleotide repeat sequence that is highly polymorphic and is located in proximity to a transcriptional enhancer element (Maekawa et al., 1989; Haley and Waterfield, 1991; Chrysogelos, 1993) (Figure 4). The CA repeat lengths varied from 14 to 21 in a Caucasian population, with the most common alleles containing 16 (42%), 18 (20%) and 20 (26%) repeats (Chi et al., 1992). Gebhardt et al. (1999) have reported an inverse correlation between the number of CA repeats present and the EGFR gene transcriptional activity. Recently, several new anticancer agents have been developed to either cause extracellular or intracellular blockade of EGFR function and/or its cellular signaling pathway (Peus et al., 1997; Fry et al., 1999; Ciardiello and Giampaolo, 2001), and EGFR inhibitors have demonstrated promising activity against various malignancies (Kris et al., 2002; Saltz et al., 2002; Cohen et al., 2003). Skin rash has been the most commonly observed toxicity associated with the various EGFR inhibitors; interindividual differences in the onset, duration and severity of the rash have been observed, and no threshold plasma levels have been linked to the occurrence of the rash (Hidalgo et al., 2001; Bruno et al., 2003). Most intriguing are emerging data demonstrating a significant correlation between skin rash and survival among various patients treated with different anti-EGFR therapies (Clark et al., 2003; Saltz et al., 2003, Cohen et al., 2003). There are several potential hypotheses being put forward to explain both the variable toxicity and efficacy of EGFR inhibitors. One such hypothesis proposes that variability in clinical observations is related to variable drug exposure. For example, the small-molecule EGFR tyrosine kinase inhibitors geftinib and erlotinib are metabolized by CYP3A, and it is certainly plausible that individuals with variant CYP3A alleles might have differences in drug exposure. This hypothesis will be investigated in clinical trials with the strategy of dose escalating the EGFR inhibitor of interest to maximize the number of patients with skin rash in order to increase the efficacy. This approach reverts the drug development process to the more empirical approach of dose escalation to the toxicity ceiling, the strategy that targeted therapies were supposed to avoid. One example of an alternative strategy to understand the drug response discrepancies of the EGFR inhibitors is a recently initiated prospective clinical trial at the University of Chicago, in which all the patients are being treated at the recommended phase II dose of erlotinib, while pharmacokinetic, clinical, pharmacodynamic (skin biopsies) and genotyping data are being prospectively collected to elucidate the underlying determinants of variable drug response. The previously described CA dinucleotide repeat polymorphism might influence the drug response due to differences in target expression. Data that indirectly lend support to this hypothesis come from a higher response rate observed in Japanese patients compared to Caucasian patients (when treated with geftinib) two populations with a difference in the frequencies of the EGFR dinucleotide repeat variants (Fukuoka et al., 2002; Liu et al., 2003). The prospective trial of erlotonib being conducted at the University of Chicago is also designed to collect samples for genotyping the EGFR polymorphisms of interest, and correlating it with EGFR expression along with the other prospectively collected data for multivariate analysis to better understand the interindividual differences in EGFR inhibitor drug response. One limitation of the trial is that the data being collected will not directly capture genotypic or phenotypic information regarding the tumor tissue itself.

Figure 4.

EGFR 5'-region including the promoter, exon 1 and part of intron 1 are depicted. Also shown are three enhancer (open boxes) elements E1, E2 and E3, identified to play an important role in EGFR transcription. The EGFR promoter lacks consensus sequences such as a TATA or CAAT box. A highly polymorphic CA dinucleotide repeat sequence (repeats ranging from 14 to 21) is located in proximity to E2. In vitro, EGFR transcription activity has been demonstrated to inversely correlate with numbers of CA repeats in intron 1

Full figure and legend (17K)

However, given the abundant EGFR expression in skin tissue, and the observed association between skin toxicity and tumor response; the use of surrogate tissue in this instance might be justified. Nonetheless, this issue highlights an important hurdle in conducting translational work in this field, since obtaining tumor biopsies in prospective trials for hypotheses generation is not a trivial matter for obvious ethical and practical concerns. A reasonable approach is to try and validate relevant hypotheses using nonclinical and surrogate clinical tissues, prior to embarking upon more ambitious prospective trials requiring extensive procedures. Improvement in available techniques such as tissue microarray technology that would allow extraction of relevant pharmacogenomic data more reliably from archived tissues might also resolve some of the current problems.

The dialogue regarding variable drug response of the targeted agents is in its infancy, and the debate will certainly carry forward perhaps at a more sophisticated level as the science relevant to their development matures. However, despite all the current limitations, it is still prudent to intensively pursue plausible mechanisms of unpredictable drug response in the early development of these agents for their efficient and optimal development.

Conclusion

The intrinsic potency of cytotoxic agents, their narrow therapeutic index and use at maximally tolerated doses render anticancer agents a high-risk treatment for patients who deviate from the average population. Hence, identifying the heritable differences responsible for either the occurrence of toxicity or lack of efficacy will reduce the unpredictability of cancer treatment, in particular because adjusting the dose by body surface area does not correct for interindividual differences in drug disposition. The availability of effective, straightforward and reliable genetic techniques can change the way patients will receive anticancer therapy in the future. Thus, it becomes essential that clinical pharmacologists and oncologists include pharmacogenetic investigation and DNA collection into early phases of clinical drug development. The future anticancer trials should include strategies designed to understand the role of genomic covariates, in order to realize the goal of individualized anticancer therapy. This brief and ostensible discussion pertaining to pharmacogenomics also highlights the need of extensive multidisciplinary efforts that will be required to make the field clinically relevant. The task will continue to become more feasible in the future with the rapid advances in high-throughput genotyping methods, increasing the understanding of pharmacogenomic trial methodology and continued expansion of bioinformatics.

References

1. Acuna G, Foernzler D, Leong D, Rabbia M, Smit R, Dorflinger E, Gasser R, Hoh J, Ott J, Borroni E, To Z, Thompson A, Li J, Hashimoto L and Lindpaintner K. (2002). Pharmacogenomics J., 2, 327–334. | Article | PubMed | ChemPort |
2. Adamson ED and Rees AR. (1981). Mol. Cell Biochem., 34, 129–152. | PubMed |
3. Akaba K, Kimura T, Sasaki A, Tanabe S, Ikegami T, Hashimoto M, Umeda H, Yoshida H, Umetsu K, Chiba H, Yuasa I and Hayasaka K. (1998). Biochem. Mol. Biol. Int., 46, 21–26. | PubMed | ChemPort |
4. Akimoto T, Hunter NR, Buchmiller L, Mason K, Ang KK and Milas L. (1999). Clin. Cancer Res., 5, 2884–2890. | PubMed | ISI | ChemPort |
5. Ando Y, Saka H, Ando M, Sawa T, Muro K, Ueoka H, Yokoyama A, Saitoh S, Shimokata K and Hasegawa Y. (2000). Cancer Res., 60, 6921–6926. | PubMed | ISI | ChemPort |
6. Araki E, Ishikawa M, Iigo M, Koide T, Itabashi M and Hoshi A. (1993). Jpn. J. Cancer Res., 84, 697–702. | PubMed | ChemPort |
7. Aschele C, Debernardis D, Casazza S, Antonelli G, Tunesi G, Baldo C, Lionetto R, Maley F and Sobrero A. (1999). J. Clin. Oncol., 17, 1760–1770. | PubMed | ChemPort |
8. Beutler E, Gelbart T and Demina A. (1998). Proc. Natl. Acad. Sci. USA, 95, 8170–8174. | Article | PubMed | ChemPort |
9. Bosma PJ, Chowdhury JR, Bakker C, Gantla S, de Boer A, Oostra BA, Lindhout D, Tytgat GNJ, Jansen PLM, Oude Elferink RPJ and Chowdhury NR. (1995). N. Engl. J. Med., 333, 1171–1175. | Article | PubMed | ISI | ChemPort |
10. Bruno R, Mass RD, Jones C, Lu JF and Winer E. (2003). Proc. ASCO., 22, (Abstr. 823).
11. Burchell B, Brierley CH and Rance D. (1995). Life Sci., 57, 1819–1831. | Article | PubMed | ChemPort |
12. Burchell B and Hume R. (1999). J. Gastroenterol. Hepatol., 14, 960–966. | Article | PubMed | ISI | ChemPort |
13. Carpenter G. (1987). Annu. Rev. Biochem., 56, 881–914. | Article | PubMed | ISI | ChemPort |
14. Chen Z, Ke LD and Yuan XH. (2000). Anticancer Res., 20, 899–902. | PubMed | ChemPort |
15. Chi DD, Hing AV, Helms C, Steinbrueck T, Mishra SK and Donis-Keller H. (1992). Hum. Mol. Genet., 1, 35. | PubMed |
16. Chrysogelos SA. (1993). Nucleic Acids Res., 21, 5736–5741. | PubMed |
17. Chu E, Ju J and Schmitz J. (1999). Anticancer Drug Development Guide: Antifolate Drugs in Cancer Therapy Jackman A (ed). Humana Press: Totowa, 397–408.
18. Chu E, Koeller DM, Casey JL, Drake JC, Chabner BA, Elwood PC, Zinn S and Allegra CJ. (1991). Proc. Natl. Acad. Sci. USA, 88, 8977–8981. | Article | PubMed | ChemPort |
19. Ciardiello F and Tortora GL. (2001). Clin. Cancer. Res., 7, 2958–2970. | PubMed | ISI | ChemPort |
20. Clark GM, Perez-Soler R, Siu L, Gordon A and Santabarbara P. (2003). Proc. ASCO, 22, (Abstr. 786).
21. Cohen EE, Rosen F, Stadler WM, Recant W, Stenson K, Huo D and Vokes EE. (2003). J. Clin. Oncol., 21, 1980–1987. | Article | PubMed | ISI | ChemPort |
22. Dolnick BJ. (2000). Cancer J., 6, 215–216. | PubMed | ISI | ChemPort |
23. Evans DA. (1968). Ann. NY Acad. Sci., 151, 723–733. | PubMed |
24. Evans WE and Relling MV. (1999). Science, 286, 487–491. | Article | PubMed | ISI | ChemPort |
25. Fry DW. (1999). Pharmacol. Ther., 82, 207–218. | Article | PubMed | ChemPort |
26. Fuchs CS, Moore MR, Harker G, Villa L, Rinaldi D and Hecht JR. (2003). J. Clin. Oncol., 21, 807–814. | Article | PubMed | ISI | ChemPort |
27. Fukuoka M, Yano S, Giaccone G, Tamura T, Nakagawa K, Douillard JY, Nishiwaki Y, Vansteenkiste JF, Kudo S, Averbuch S, Macleod A, Feyereislova A and Baselga J. (2002). Proc. Am. Soc. Clin. Oncol., 21, 298a.
28. Gebhardt F, Zanker KS and Brandt B. (1999). J. Biol. Chem., 274, 13176–13180. | Article | PubMed | ISI | ChemPort |
29. Grau JJ, Domingo-Domenech J, Montagut C, Mellado B, Pera M, Albanell J and Gascon P. (2003). Proc. ASCO., 22, (Abstr. 1049).
30. Gupta E, Lestingi TM, Mick R, Ramirez J, Yokes EE and Ratain MJ. (1994). Cancer Res., 54, 3723–3725. | PubMed | ISI | ChemPort |
31. Haley JD and Waterfield MD. (1991). J. Biol. Chem., 266, 1746–1753. | PubMed | ChemPort |
32. Hidalgo M, Siu LL, Nemunaitis J, Rizzo J, Hammond LA, Takimoto C, Eckhardt SG, Tolcher A, Britten CD, Denis L, Ferrante K, Von Hoff DD, Silberman S and Rowinsky EK. (2001). J. Clin. Oncol., 19, 3267–3279. | PubMed | ISI | ChemPort |
33. Horie N, Aiba H, Oguro K, Hojo H and Takeishi K. (1995). Cell Struct. Funct., 20, 191–197. | PubMed | ISI | ChemPort |
34. Huang C-S, Luo G-A, Huang M-J, Yu S-C and Yang S-S. (2000). Pharmacogenetics, 10, 539–544. | Article | PubMed | ISI | ChemPort |
35. Humerickhouse R, Lohrbach K, Li L, Bosron WF and Dolan ME. (2000). Cancer Res., 60, 1189–1192. | PubMed | ISI | ChemPort |
36. Innocenti F, Grimsley C, Das S, Ramirez J, Cheng C, Kuttab-Boulos H, Ratain MJ and Di Rienzo A. (2002a). Pharmacogenetics, 12, 725–733. | Article | PubMed | ISI | ChemPort |
37. Innocenti F, Iyer L and Ratain MJ. (2002b). Pharmacogenomics: The Search for Individualized Therapies Licinio and Wong (eds). Wiley-VCH Verlag GmbH: Weinheim.
38. Innocenti F and Ratain MJ. (2002). Eur. J. Cancer, 38, 639–644. | Article | PubMed | ChemPort |
39. Innocenti F, Undevia S, Iyer L, Das S, Karrison T, Janixh L, Ramirez J, Rudin CM, Yokes EE and Ratain MJ. (2003). Proc. ASCO, 22, (Abstr. 495).
40. Ishii S, Xu YH, Stratton RH, Roe BA, Merlino GT and Pastan I. (1985). Proc. Natl. Acad. Sci. USA, 82, 4920–4924. | PubMed | ChemPort |
41. Iyer L, Das S, Janisch L, Wen M, Ramirez J, Karrison T, Fleming GF, Vokes EE, Schilsky RL and Ratain MJ. (2002). Pharmacogenomics J., 2, 43–47. | Article | PubMed | ChemPort |
42. Iyer L, Hall D, Das S, Mortell MA, Ramirez J, Kim S, Di Rienzo A and Ratain MJ. (1999). Clin. Pharmacol. Ther., 65, 576–582. | Article | PubMed | ISI | ChemPort |
43. Iyer L, King CD, Whitington PF, Green MD, Roy SK, Tephly TR, Coffman BL and Ratain MJ. (1998). J. Clin. Invest., 101, 847–854. | PubMed | ISI | ChemPort |
44. Iyer L and Ratain MJ. (1998). Cancer Chemother. Pharmacol., 42 (suppl), 31–43. | PubMed |
45. Johnston PG, Drake JC, Trepel J and Allegra CJ. (1992). Cancer Res., 52, 4306–4212.
46. Johnston PG, Fisher ER, Rockette HE, Fisher B, Wolmark N, Drake JC, Chabner BA and Allegra CJ. (1994). J. Clin. Oncol., 12, 2640–2647. | PubMed | ISI | ChemPort |
47. Johnston PG, Lenz HJ, Leichman CG, Danenberg KD, Allegra CJ, Danenberg PV and Leichman L. (1995). Cancer Res., 55, 1407–1412. | PubMed | ISI | ChemPort |
48. Kalow W. (1956). Lancet, 2, 576. | Article |
49. Kaneda S, Takeishi K, Ayusawa D, Shimizu K, Seno T and Altman S. (1987). Nucleic Acid. Res., 15, 1259–1270. | PubMed |
50. Kawabe S, Takiuchi H, Gothoh M, Ohta S and Katsu KI. (2003). Proceedings of the American Society of Clinical Oncology Vol. 22, Abstract 1436.
51. Kawakami K, Salonga and Omura K. (1999). Proc. Am. Assoc. Cancer Res., 40, 436–437.
52. Kohn EC, Figg WD, Sarosy GA, Bauer KS, Davis PA, Soltis MJ, Thompkins A, Liotta LA and Reed E. (1997). J. Clin. Oncol., 15, 1985–1993. | PubMed |
53. Korn EL, Arbuck SG, Pluda JM, Simon R, Kaplan RS and Christian MC. (2001). J. Clin. Oncol., 19, 265–272. | PubMed | ChemPort |
54. Kornmann M, Link KH, Lenz HJ, Pillasch J, Metzger R, Butzer U, Leder GH, Weindel M, Safi F, Danenberg KD, Beger HG and Danenberg PV. (1997). Cancer Lett., 118, 29–35. | Article | PubMed | ISI | ChemPort |
55. Kris MG, Natale RB, Herbst R, Lynch TJ, Prager D and Belani CP. (2002). Proc. ASCO, 21, Abstr. 1168.
56. Leichman CG, Lenz HJ, Leichman L, Danenberg K, Baranda J, Groshen S, Boswell W, Metzger R, Tan M and Danenberg PV. (1997). J. Clin. Oncol., 10, 3223–3229.
57. Lenz HJ, Leichman CG, Danenberg KD, Danenberg PV, Groshen S, Cohen H, Laine L, Crookes P, Silberman H, Baranda J, Garcia Y, Li J and Leichman L. (1996). J. Clin. Oncol., 14, 176–182. | PubMed | ISI | ChemPort |
58. Liu W, Innocenti F, Chen P, Das S, Cook EH and Ratain MJ. (2003). Clin. Cancer Res., 3, 1009–1012.
59. Luo HR, Lu XM, Yao YG, Horie N, Takeishi K, Jorde LB and Zhang YP. (2002). Biochem. Genet., 40, 41–51. | Article | PubMed | ISI | ChemPort |
60. Mackenzie PI, Owens IS, Burchell B, Bock KW, Bairoch A, Belanger A, Fournel-Gigleux S, Green M, Hum DW, Iyanagi T, Lancet D, Louisot P, Magdalou J, Chowdhury JR, Ritter JK, Schachter H, Tephly TR, Tipton KF and Nebert DW. (1997). Pharmacogenetics, 7, 255–269. | Article | PubMed | ISI | ChemPort |
61. Maekawa T, Imamoto F, Merlino GT, Pastan I and Ishii S. (1989). J. Biol. Chem., 264, 5488–5494. | PubMed | ChemPort |
62. Mandola MV, Stoehlmacher J, Muller-Weeks S, Cesarone G, Yu MC, Lenz HJ and Ladner RD. (2003). Cancer Res., 63, 2898–2904. | PubMed | ISI | ChemPort |
63. Marsh S, Ameyaw MM, Githang'a J, Indalo A, Ofori-Adjei D and McLeod HL. (2000). Human Mutat., 16, 528.
64. Marsh S, McKay JA, Cassidy J and McLeod HL. (2001). Int. J. Oncol., 19, 383–386. | PubMed | ISI | ChemPort |
65. Merlino GT, Ishii S, Whang-Peng J, Knutsen T, Xu YH, Clark AJ, Stratton RH, Wilson RK, Ma DP and Roe BA. (1985a). Mol. Cell Biol., 5, 1722–1734. | PubMed | ISI | ChemPort |
66. Merlino GT, Xu YH, Richert N, Clark AJ, Ishii S, Banks-Schlegel S and Pastan I. (1985b). J. Clin. Invest., 75, 1077–1079. | PubMed |
67. Monaghan G, Foster B, Jurima-Romet M, Hume R, Burchell B and Owens IS. (1997). Pharmacogenetics, 7, 153–156. | Article | PubMed | ChemPort |
68. Monaghan G, Ryan M, Seddon R, Hume R and Burchell B. (1996). Lancet, 347, 578–581. | Article | PubMed | ISI | ChemPort |
69. Nagasubramanian R, Innocenti F and Ratain MJ. (2003). Ann. Rev. Med., 54, 437–452. | PubMed |
70. Noda K, Nishiwaki Y, Kawahara M, Negro S, Sugiura T, Yokoyoma A, Fukuoka M, Mori M, Watanabe K, Tamura T, Yamamoto S and Saijo N. (2002). N. Engl. J. Med, 346, 85–91. | Article | PubMed | ISI | ChemPort |
71. Pestalozzi BC, McGinn CJ, Kinsella TJ, Drake JC, Glennon MC, Allegra CJ and Johnston PG. (1995). Br. J. Cancer, 71, 1151–1157. | PubMed | ISI | ChemPort |
72. Peus D, Hamacher L and Pittelkow MR. (1997). J. Invest. Dermatol., 109, 751–756. | Article | PubMed | ISI | ChemPort |
73. Porebska I, Harlozinska A and Bojarowski T. (2000). Tumor Biol., 21, 105–115.
74. Pullarkat ST, Stoehlmacher J, Ghaderi V, Xiong YP, Ingles SA, Sherrod A, Warren R, Tsao-Wei D, Groshen S and Lenz HJ. (2001). Pharmacogenomics J., 1, 65–70. | Article | PubMed | ChemPort |
75. Ratain MJ and Relling MV. (2001). Nat. Med., 7, 283–285. | Article | PubMed | ISI | ChemPort |
76. Reardon DA, Friedman HS, Powell JB, Gilbert M and Yung WK. (2003). Oncology, 17, 9–14. | PubMed |
77. Salomon DS, Brandt R, Ciardiello F and Normanno N. (1995). Crit. Rev. Oncol. Haematol, 19, 183–232. | ChemPort |
78. Saltz L, Kies M, Abbruzzese JL, Azarnia N and Needle M. (2003). Proc. ASCO., 22, (Abstr. 817).
79. Saltz L, Meropol NJ, Loehrer PJ, Waskal H, Needle MN and Mayer RJ. (2002). Proc. ASCO, 21, (Abstr. 504).
80. Sampietro M and Iolascon A. (1999). Haematologica, 84, 150–157. | PubMed |
81. Santos A, Zanetta S, Cresteil T, Deroussent A, Pein F, Raymond E, Vernillet L, Risse ML, Boige V, Gouyette A and Vassal G. (2000). Clin. Cancer Res., 6, 2012–2020. | PubMed | ISI | ChemPort |
82. Sato H, Adachi Y and Koiwai O. (1996). Lancet, 347, 557–558. | Article | PubMed | ISI | ChemPort |
83. Savage CR and Cohen S. (1972). J. Biol. Chem., 247, 7609–7611. | PubMed | ISI | ChemPort |
84. Soong R, Boedfeld WM, Heslin MJ, Wang K, Johnson MR and Diasio RB. (2003). Proc. ASCO, 22, (Abstract 493).
85. Stadler WM and Ratain MJ. (2000). Invest. New Drugs, 18, 7–16. | Article | PubMed | ChemPort |
86. Sugatani J, Yamakawa K, Yoshinari K, Machida T, Takagi H, Mori M, Kakizaki S, Sueyoshi T, Negishi M and Miwa M. (2002). Biochem. Biophys. Res. Common., 292, 492–497.
87. Sung T, Miller DC, Hayes RL, Alonso M, Yee H and Newcomb EW. (2000). Brain Pathol., 10, 249–259. | PubMed | ISI | ChemPort |
88. Tukey RH and Strassburg CP. (2000). Annu. Rev. Pharmacol. Toxicol., 40, 581–616. | Article | PubMed | ISI | ChemPort |
89. Vanhoefer U, Harstrick A, Achterrath W, Cao S, Seeber S and Rustum YM. (2001). J. Clin. Oncol., 19, 1501–1518. | PubMed | ISI | ChemPort |
90. Villafranca E, Okruzhnov Y, Dominguez MA, Garcia-Foncillas J, Azinovic I, Martinez E, Illarramendi JJ, Arias F, Martinez Monge R, Salgado E, Angeletti S and Brugarolas A. (2001). J. Clin. Oncol., 19, 1779–1786. | PubMed | ISI | ChemPort |

Acknowledgements

This work was supported in part by the Pharmacogenetics of Anticancer Agents Research (PAAR) Group (http://pharmacogenetics.org) by the NIH/NIGMS grant U01GM61393.