Impact of ABCG2 polymorphisms on the clinical outcome and toxicity of gefitinib in non-small-cell lung cancer patients

ReseaRch aRticle

ISSN 1462-2416Pharmacogenomics (2011) 12(2), 159–17010.2217/PGS.10.172 © 2011 Future Medicine Ltd 159

Impact of ABCG2 polymorphisms on the clinical outcome and toxicity of gefitinib in non-small-cell lung cancer patients

The breast cancer resistance protein (BCRP/ABCG2) is a member of the ATP-binding cas-sette (ABC) superfamily of multidrug efflux transporters. Overexpression of ABCG2 is com-monly associated with multidrug resistance to a wide spectrum of structurally and mechanis-tically unrelated anticancer agents, including mitoxantrone, camptothecins, anthracyclines, flavopiridol and antifolates [1,2].

Emerging data suggest that the novel class of targeted drugs, tyrosine kinase inhibitors (TKIs), which target EGFR and include gefi-tinib, are transport substrates of ABCG2 [3]. Of note, gefitinib is transported by ABCG2 at clinically achievable concentrations (≤1 µM), whereas at higher drug concentrations (>1 µM), gefitinib is an ABCG2 transport inhibitor rather than a substrate [4–6]. Therefore, ABCG2 expression has an important impact on gefi-tinib resistance both in vitro [4,7,8] and in vivo [9]. Moreover, considering that ABCG2 is highly expressed in the gastrointestinal tract where it plays a role in the regulation of the uptake of several xeno biotics [10], and since gefitinib is an orally active compound, one might predict an important role for ABCG2 in the elimination of this drug [3].

Several naturally occurring SNPs in the ABCG2 gene have been described that might have an important impact on ABCG2 protein

expression, localization and function. In partic-ular, the nonsynonymous SNP ABCG2 421C/A, which results in a glutamine to lysine amino acid substitution at position 141 (Q141K), has been associated with markedly decreased levels of ABCG2 protein expression [11–13] and/or activity [14–16]. Of note, Mizuarai and colleagues reported a decreased ATPase activ-ity for the ABCG2 421C/A variant compared with the wild-type ABCG2 [15], while Morisaki et al. demonstrated both a reduced ATPase activity and an altered cellular localization of the ABCG2 421C/A [14]. This SNP was asso-ciated with increased oral bioavailability of several drugs, including the TKI gefitinib and erlotinib [6], the statins simvastatin and fluvas-tatin [17], as well as the topoisomerase 1 inhibi-tor dif lomotecan [18]. Interestingly, Cusatis et al. reported a strong association between the ABCG2 421C/A polymorphism and diarrhea in gefitinib-treated non-small-cell lung carcinoma (NSCLC) patients [19]. They specifically demon-strated that seven (44%) out of 16 patients het-erozygous for ABCG2 421C/A developed diar-rhea, versus only 13 (12%) out of 108 patients homozygous for the wild-type genotype. The authors suggested that the reduced protein lev-els and altered ATPase activity of the ABCG2 421C/A variant in the intestine might affect the oral absorption and/or elimination pathways of

Aims: The current study investigates whether or not functional polymorphisms in the ATP-binding cassette transporter gene ABCG2 might affect gefitinib activity and/or toxicity in non-small-cell lung cancer (NSCLC) patients. Materials & methods: Towards this end, ABCG2 polymorphisms and expression were assessed in DNA and tumors from 94 NSCLC patients treated with gefitinib, whereas their associations with toxicity/response and time-to-progression/overall survival were evaluated using Pearson-c2 and log-rank-test, respectively. Results: Patients carrying an ABCG2 -15622T/T genotype or harboring at least one TT copy in the ABCG2 (1143C/T, -15622C/T) haplotype developed significantly more grade 2/3 diarrhea (p < 0.01). No associations were found between polymorphisms and outcome. Consistently, ABCG2 protein levels in tumors were not significantly different between patients harboring different ABCG2 variants. Conclusion: The ABCG2 -15622C/T polymorphism and ABCG2 (1143C/T, -15622C/T) haplotype resulted in a gefitinib-dependent, moderate-to-severe diarrhea suggesting that these pharmacogenetic markers should be considered to optimize NSCLC treatment.

Original submitted 12th August 2010; Revision submitted 19th October 2010.

KEYWORDS: ABCG2 n gefitinib n lung cancer n pharmacogenetics n toxicity Clara Lemos*, Elisa Giovannetti†*,Paolo A Zucali, Yehuda G Assaraf, George L Scheffer, Tahar van der Straaten, Armida D’Incecco, Alfredo Falcone, Henk-Jan Guchelaar, Romano Danesi, Armando Santoro, Giuseppe Giaccone, Carmelo Tibaldi & Godefridus J Peters†Author for correspondence:Department of Medical Oncology, VU University Medical Center, De Boelelaan 1117, 1081HV Amsterdam, The Netherlands andDepartment of Internal Medicine, University of Pisa, via Roma 55, 56100 Pisa, Italy Tel.: +31 204 442 633 Fax: +31 204 443 844 [email protected] *These authors contributed equally to this study For a full list of affiliations please see the back page

For reprint orders, please contact: [email protected]

ReseaRch aRticle Lemos, Giovannetti, Zucali et al.

Pharmacogenomics (2011) 12(2)160 future science group

gefitinib, thereby presumably increasing the steady-state gefitinib plasma concentrations hence resulting in diarrhea.

Two new ABCG2 polymorphisms, the ABCG2 1143C/T and ABCG2 -15622C/T, were recently identified [20,21]. These poly morphisms have also been associated with decreased levels of ABCG2 expression in different tissues [21]. Furthermore, the ABCG2 -15622C/T polymorphism was marginally associated with diarrhea in erlo-tinib-treated patients [20], whereas the ABCG2 (-15622C/T, 1143C/T) haplotype was signifi-cantly associated with toxicity in sunitinib-treated patients [22]. However, no data are available on the impact of these ABCG2 polymorphisms on gefitinib-induced toxicity. Furthermore, consider-ing that ABCG2 is also expressed in lung tumor tissue [23], ABCG2 polymorphisms might also play a role in drug response and clinical outcome of gefitinib-treated patients.

Gefitinib has been recently registered in Europe as a first-line treatment for patients with EGFR-activating mutations, but additional factors could help in adapting individualized therapy, especially for patients with a low fre-quency of somatic mutations such as Caucasians. Germline polymorphisms might affect the genetic susceptibility to both tumor response and skin/gastrointestinal toxicity because they occur in both normal and tumor tissues. Therefore, the aim of the current study was to retrospectively evaluate the correlation between candidate ABCG2 polymorphisms and gefitinib-induced toxicity, as well as clinical outcome in gefitinib-treated NSCLC patients.

Materials & methods n Patients

A total of 49 and 45 NSCLC patients were treated within the expanded access program of gefitinib at Humanitas Clinical Institute (Rozzano, Milan, Italy) and Livorno Civil Hospital (Livorno, Italy), respectively. Patients were enrolled between November 2005 and January 2008. The selection was based on diag-The selection was based on diag-nosis of histologically confirmed NSCLC with measurable, locally advanced or metastatic dis-ease, progressing/relapsing after chemotherapy, or with contraindications for chemotherapy. The clinical trial and the study on patients’ speci-mens was approved by the ethics committees (ClinicalTrials.gov ID-NCT00831454 [101]) and conducted in accordance with principles stated in the Declaration of Helsinki. Patients received oral gefitinib at a dose of 250 mg/day and were evaluated for response after 2 months according

to the Response Evaluation Criteria In Solid Tumors (RECIST) criteria. Tumor response was assessed by CT scan, with a confirmatory evalu-ation repeated in patients with complete/partial response and stable disease at least 4 weeks after the initial determination of response. Toxicities were assessed using the NCI common termino-logy criteria (Common toxicity criteria manual version 3.0, available from [102]).

n DNA isolation DNA was isolated from blood or paraffin-embedded tumor samples using the microDNA-kit (Qiagen, Hilden, Germany). DNA yields and integrity were checked at 260–280 nm with NanoDrop® 1000 Detector (NanoDrop Technologies, NC, USA).

n Analysis of polymorphisms The ABCG2 421C/A (dbSNP ID: rs2231142), 1143C/T (dbSNP ID: rs2622604) and -15622C/T polymorphisms were studied with Taqman® (Applied Biosystems, CA, USA) probes-based assays using the Applied Biosystems PRISM® 7500HT instrument equipped with the Sequence Detection System version 2.0 software (Applied Biosystems). Specific primers and probe (Applied Biosystems SNP Genotyping Assays products) for the ABCG2 421C/A polymorphism were obtained from Applied Biosystems (C__15854163_70, TaqMan Drug Metabolism Genotyping Assays), while primers and probes for ABCG2 1143C/T and -15622C/T were designed using the FileBuilder 2.0 software (Applied Biosystems), on the basis of the gene sequence from GenBank: 1143C/T (foward): 5́ -CTC TGA AAG CAC TGT TTT GTA AAG AAT ATCA A-3´; (reverse): 5´- GAG GAT CTT ACA GCT CTC TTA TCA TTT GAA-3´; (probe): 5´-ACT ACT CTG G[T/G] GTA TTT-3´ and -15622 C/T (for-ward): 5 -́GAA ACC CTG TCT GTC TCT ACT AAA ATT ACA-3´; (reverse): 5´-CCT CCA AAG GAC ATG AAC TCA TTC TTT-3´; (probe): 5´-CAG ACA TGC[A/G] CCA CCA-3 ,́ for the FAM/VIC-fluorescent reporter, respectively. The PCR reactions were performed using 20 ng of genomic DNA diluted in 5.94 µl DNAse-RNAse free water, with 6.25 µl of TaqMan Universal PCR Master Mix and 0.31 µl of the assay mix, including the primers and probes in concentrations optimized in prelimi-nary reactions, in a total volume of 12.5 µl. After thermal cycling, the 7500HT instrument deter-mined the allelic content of each sample in the plate by reading the generated fluorescence [8].

ABCG2 polymorphisms & the clinical outcome & toxicity of gefitinib ReseaRch aRticle

www.futuremedicine.com 161future science group

n Tissue microarray constructionA tissue microarray was constructed as previously described [24]. Briefly, paraffin-embedded tumor specimens of 49 patients (Rozzano, Milan, Italy) were collected, and a pathologist selected areas for 1-mm2 punches from tumor cores, tumor borders and normal tissue to include in recipi-ent tissue array block using a specific instrument (Beecher Instruments, WI, USA).

n ImmunohistochemistryTissue microarray sections were deparaf-finized using xylene and rehydrated in alcohol. Endogenous peroxidase activity was inhib-ited by incubating the tissue slides in metha-nol/0.3% H

2O

2 for 30 min. Antigen retrieval

was carried out by boiling (microwave) the sec-tions in 0.01 M of citrate buffer (pH 6.0) for 15 min. After 15 min cooling at room tempera-ture, nonspecific binding was blocked for 30 min at room temperature with phosphate buff-ered saline containing 5% normal goat serum (DakoCytomation, Glostrup, Denmark). Subsequently, the sections were incubated with a mouse anti-ABCG2 monoclonal anti-body BXP-21 [25], diluted 1:50 in phosphate buffered saline with 1% BSA (final concentra-tion: 5 µg/ml), for 60 min at room temperature. Sections were developed using Powervision plus visualization system (ImmunoLogic, Duiven, The Netherlands) and AEC (Invitrogen, CA, USA), and counterstained with hematoxy-lin. Control for nonspecific staining was the replacement of BXP-21 with a nonspecif ic antibody from the same class, mouse IgG2a (DakoCytomation).

n Immunohistochemistry scoringProtein expression determined by immuno-histochemistry (IHC) was evaluated using an Olympus BX50F bright field microscope (Olympus Optical Co. Ltd, Tokyo, Japan) with a 20× objective. Scoring was performed by two independent observers (C Lemos and GL Scheffer). Both the intensity (negative: 0, weak positive: 1, moderately positive: 2, strong positive: 3) and the percentages of positively stained cells were scored. Two tumor cores per patient were evaluated and the average score was calculated. Some cases suffered tissue loss or lack of tumor cell representation to an extent that precluded the evaluation of protein expres-sion. The staining intensity value was multi-plied by the percentage of positive cells [26], yielding a final expression score ranging from 0 to 300.

n Statistical analysisDemographic and clinical information were compared across genotype, using the Pearson-c2 two-sided test. Patients achieving complete (CR) or partial response (PR) were defi ned as �respond-were defined as �respond-ers’, while patients with stable disease (SD) or progressive disease (PD) as �nonresponders’. Additional analyses were performed by group-ing patients with CR/PR and SD (�patients with clinical benefit’) versus PD [27].

Time to progression (TTP) was calculated from the date of the first dose of gefitinib to the date of clinical and/or radiological evidence of progression/death, whichever occurred first, while overall survival (OS) was calculated from the day the treatment was initiated to the end point (death or censoring). The Kaplan–Meier method was used to plot TTP and OS, and the log-rank test to compare curves.

The univariate analysis included several clinical characteristics. Baseline demographic characteristics included sex, performance sta-tus (0–1 vs 2–3), age (<median vs ≥median), clinical stage (IIIB vs IV), histology (adenocarci-noma vs other histology, and bronchioloalveolar carcinoma [BAC] vs other histology), smoking history (never, i.e., <100 cigarettes/lifetime vs former [quit smoking more than 12 months before starting gefitinib] and current smokers) and prior chemotherapy (yes vs no). Clinical treatment effects included toxicity (grade 0–1 vs grade 2+), and response to therapy, categorized as described earlier. Regarding polymorphisms, statistical analyses were performed by grouping patients according to different allelic activity, as previously described, for example, ABCG2 421C/C versus C/A-A/A, ABCG2 1143C/C-C/T versus T/T and ABCG2 -15622C/C-C/T versus T/T [19,20,22].

The nonparametric Kruskal–Wallis test was used when analyzing the correlation of ABCG2 polymorphisms with ABCG2 protein expression determined by IHC.

All the analyses of the samples were carried out in a blinded fashion relative to clinical out-come. Data were analyzed using SPSS-17 soft-ware (SPSS, Inc., IL, USA). p-values between 0.05 and 0.1 were regarded as marginally sug-gestive of an association (�trend toward a sig-nificant correlation’), but only p < 0.05 were considered statistically significant. However, in the analyses for clinical outcome and toxicity according to the three studied ABCG2 poly-morphisms and ABCG2 haplotype, a Bonferroni correction required a p < 0.05/4 = 0.0125 for statistical significance.



Results n Patients’ characteristics

& clinical outcomePatient characteristics and their association with clinical outcome are summarized in Table 1. Response and TTP data were available from all patients, while OS was available from 92 patients. There were 87 deaths (event rate 94.6%) while two patients were alive without PD at last contact (January 2010). A total of 17 out of the 94 patients evaluated (18.1%) had PR, while 36 (38.3%) and 41 (43.6%) had SD and PD, respectively. Response was signifi-cantly higher among females and never smokers, whereas other patient characteristics were not associated with response. Median TTP and OS were 3.2 and 8.0 months, respectively. Negative smoker history and BAC histology were asso-ciated with significantly longer TTP and OS. Older patients also had a significantly longer TTP, whereas the female gender was associated with a significantly longer OS. Finally, both response and clinical benefit were associated with significantly longer TTP and OS.

n ABCG2 polymorphismsPrevious studies demonstrated no differences in polymorphisms analyzed in tumors and normal tissues [28–30]. However, we performed prelimi-nary studies of ABCG2 421C/A polymorphism in 45 paired DNA samples isolated from germ-line and lung cancer, hence displaying identical interindividual genotypes between normal and malignant tissues. Therefore, for the 94 patients enrolled in the present study the genotyping was performed with DNA extracted from the available 49 tumors and 45 blood samples. The wild-type ABCG2 421C/A genotype (CC) had a frequency of 88.3%, whereas the CA and AA genotypes were found in 10.6 and 1.1% of the patients, respectively. For the ABCG2 1143C/T polymorphism, the frequencies of the CC, CT and TT genotypes were 59.3, 37.4 and 3.3%, respectively. Regarding the ABCG2 -15622C/T promoter polymorphism, the CC, CT and TT variants were observed in 52.8, 39.3 and 7.9% of the cases, respectively. For the ABCG2 haplo type we grouped the patients according to the com-bination of their ABCG2 1143C/T and ABCG2 -15622C/T polymorphisms, as previously described [22]. A total of seven (8%) patients had at least one copy of TT in the ABCG2 haplotype, while 81 (92%) individuals had no TT copies in this haplotype (Table 2). All polymorphisms fol-lowed Hardy–Weinberg’s equilibrium (Table 2), as calculated with the SNP analyzer software [103].

Furthermore, the allelic frequencies for these ABCG2 polymorphisms were comparable with those reported in previous studies in Caucasian NSCLC patients [19,20]. Linkage disequilib-rium between each pair of SNPs was calcu-lated and plotted using the Innate Immunity PGA, NHLBI Program LDPlotter tool (see SupplemenTary Figure 1; www.futuremedicine.com/doi/suppl/10.2217/pgs.10.172). No significant correlations were observed between ABCG2 polymorphisms and patients’ characteristics, including age, gender, smoking history, clinical stage, performance status and histology.

n Polymorphisms & clinical outcomeConsidering that ABCG2 is a drug efflux trans-porter with the ability to extrude gefinitib out of the cells, we hypothesized that the ABCG2 polymorphisms, with potential implications on ABCG2 protein expression and transport func-tion, might affect the drug response rate and clinical outcome. However, the ABCG2 poly-morphisms evaluated were not significantly associated with response. By grouping patients with clinical benefit versus those without clini-cal benefit, we also did not find an association with the studied polymorphisms. In addition, no association was observed between ABCG2 poly-morphisms and TTP and OS (Table 2). However, a trend toward a significant correlation was observed between the ABCG2 haplotype and TTP, with patients harboring at least one TT copy having shorter TTP (p = 0.06), as well as between the ABCG2 -15622C/T polymorphism and TTP (p = 0.07).

n Polymorphisms & toxicityAmong the 87 patients evaluable for skin toxic-ity, 41 (47.1%) had no toxicity while 24 (27.6%) developed a grade 1 skin rash and 22 (25.3%) developed a grade 2 or 3 skin rash. Regarding the 85 patients evaluable for gastrointestinal toxicity, 27 patients (31.8%) developed grade 1, while 6 patients (7.0%) had grade 2 or 3 diarrhea, while 52 (61.2%) exhibited no toxicity. None of our patients experienced lung toxicity. According to previous analyses [20,28], we grouped our popu-lation in patients with toxicity (grade 1+) ver-sus patients without toxicity (grade 0) and in patients with �low’ (grade 0–1) versus patients with �high’ (grade 2–3) toxicity.

No significant correlations were observed between skin or gastrointestinal toxicities and TTP or OS. Younger patients (<63.5 years), females and patients who had received previous treatment more frequently developed a higher



grade (>1) skin rash (p = 0.03, p = 0.05 and p = 0.02, respectively), while stage IV patients experienced significantly more skin toxicity (p = 0.02).

The associations between ABCG2 poly-morphisms and toxicity are described in Table 3; the ABCG2 -15622C/T polymorphism was significantly associated with gastrointestinal

toxicity, in particular with moderate to severe diarrhea. A total of 50% of the patients develop-ing grade 2 or 3 gastrointestinal toxicity carried the TT genotype. By contrast, only three out of 74 (4.1%) patients with the CC-CT vari-ants developed grade 2/3 diarrhea. Regarding the ABCG2 haplotype, we also found a strong association with higher grade diarrhea: 42.9%

Table 1. Clinical outcome according to clinical characteristics.

Characteristic Patients, n (%)

Response, n (%)†

p-value‡ TTP, months (95% CI)

p-value§ OS, months (95% CI)

p-value§

Patients (n) 94 17 (18.1) 3.2 (2.6–3.8) 8.0 (5.4–10.6)

Baseline demographic characteristics

Age, median years

– <63.5 47 (50.0) 8 (17.0) 0.79 3.0 (2.6–3.4) 0.05 7.9 (6.3–9.5) 0.40

– ≥63.5 47 (50.0) 9 (19.1) 5.7 (2.6–8.8) 10.4 (6.8–14.0)

Sex

– Male 53 (56.4) 3 (5.7) <0.01 3.0 (2.8–3.2) 0.06 7.0 (4.4–9.6) <0.01

– Female 41 (43.6) 14 (34.1) 6.0 (3.6–8.4) 13.0 (8.0–18.0)

Smoking history

– Smokers 64 (68.8) 4 (6.3) <0.01 3.0 (2.7–3.3) <0.01 8.0 (5.4–10.6) 0.02

– Never smokers¶ 29 (31.2) 13 (44.8) 5.8 (4.0–7.6) 8.0 (2.7–13.3)

Clinical stage

– IIIB 9 (9.6) 0 (0.0) 0.14 3.0 (2.8–3.2) 0.07 6.0 (0.0–12.2) 0.11

– IV 85 (90.4) 17 (20.0) 3.3 (1.9–4.7) 8.0 (5.5–10.5)

ECOG PS

– 0–1 75 (79.8) 12 (16.0) 0.30 3.3 (1.6–5.0) 0.78 9.6 (7.2–12.0) 0.93

– 2–3 19 (20.2) 5 (26.3) 2.0 (1.7–2.3) 3.0 (0.4–5.6)

Histology

– Adenocarcinoma 52 (55.3) 10 (19.2) 0.75 4.5 (2.8–6.2) 0.68 9.6 (6.5–12.7) 0.27

– Other histology 42 (44.7) 7 (16.7) 3.0 (2.8–3.2) 7.7 (7.0–8.4)

– BAC 13 (13.8) 4 (30.8) 0.20 8.0 (4.5–11.5) 0.04 34.0 (22.7–45.3) <0.01

– Other histology 81 (86.2) 13 (16.0) 3.0 (2.7–3.3) 7.6 (6.7–8.6)

Previous treament

– No 14 (14.9) 2 (14.3) 0.69 5.7 (2.3–9.1) 0.59 16.0 (0.0–35.9) 0.40

– Yes 80 (85.1) 15 (18.8) 3.1 (2.8–3.4) 8.0 (5.9–10.1)

Treatment effects

Response

– PR 17 (18.1) ND ND 12.0 (0.0–25.4) <0.01 32.0 (23.6–40.4) <0.01

– SD + PD 77 (81.9) ND ND 3.0 (2.8–3.2) 7.4 (6.4–8.5)

Clinical benefit

– PR + SD 53 (56.4) ND ND 7.6 (6.2–9.1) <0.01 14.0 (8.5–19.5) <0.01

– PD 41 (43.6) ND ND 2.0 (1.9–2.1) 4.6 (3.4–5.8)

Skin rash (grade)

– 0–1 65 (74.7) 11 (16.9) 0.54 3.1 (2.8–3.4) 0.24 7.7 (6.5–8.9) 0.20

– 2–3 22 (25.3) 5 (22.7) 4.9 (1.8–8.0) 11.1 (8.8–13.4)

Diarrhea (grade)

– 0–1 79 (92.9) 13 (16.5) 0.30 3.2 (2.5–3.9) 0.45 8.0 (5.4–10.6) 0.21

– 2–3 6 (7.1) 2 (33.3) 1.9 (0.7–3.1) 7.4 (0.0–18.0)†Percentages were calculated with respect to n of the correspondent characteristic. ‡Pearson-c2 two-sided test.§Log-rank test.¶Never (<100 cigarettes/lifetime) vs former (quit smoking more than 12 months before starting gefitinib) and current smokers.BAC: Bronchioloalveolar carcinoma; ECOG: Eastern Cooperative Oncology Group; ND: Not determined; OS: Overall survival; PD: Progressive disease; PR: Partial response; PS: Performance status; SD: Stable disease; TTP: Time to progression.



(three out of seven) patients carrying at least one TT copy in the ABCG2 haplotype had grade 2/3 diarrhea in comparison to 4.1% (three out of 73) patients without any copy of TT.

In order to keep the probability of find-ing a statistically significant difference purely by chance to a minimum, the usual nominal level (p = 0.05) has been lowered to 0.0125 by Bonferroni correction for multiple comparisons. After this adjustment, the ABCG2 -15622T/T polymorphism and ABCG2 haplotype were still significantly correlated to grade 2/3 diarrhea.

We did not observe significant association between ABCG2 polymorphisms and skin toxicity, although the ABCG2 1143C/T poly-morphism showed a trend toward a significant association with skin rash. In addition, we did not observe a significant association between the ABCG2 421C/A polymorphism and any kind of toxicity.

n Polymorphisms & ABCG2 protein expressionABCG2 protein expression was evaluated by IHC in 49 patients using a tissue microarray. Owing to tissue loss or lack of tumor cell representation in both tumor cores of two patients, ABCG2 protein expression was evaluable in a total of 47 patients. Figure 1 shows representative images of immuno-histochemical staining intensities for ABCG2. A total of 24 (51.1%) patients had no ABCG2 pro-tein expression in their lung tumor tissue, while 23 (48.9%) patients displayed variable expres-sion of the ABCG2 protein. We compared the ABCG2 protein expression across the different genotypes, but we did not find any association between the ABCG2 polymorphisms and the lev-els of protein expression in the lung tumor tissue of this group of patients (Figure 2). As previously mentioned, ABCG2 expression might affect gefi-tinib resistance phenotypes [9]. Therefore, we also studied the correlation between ABCG2 expres-sion in the lung tissue with response and clinical outcome. However, we did not find any associa-tion between the protein expression of ABCG2 and response, clinical benefit, TTP and OS (data not shown).

DiscussionIn the present study, we evaluated the associa-tions between ABCG2 polymorphisms and clin-ical outcome and toxicity in advanced NSCLC patients treated with gefitinib. To the best of our knowledge, this is the first report to identify a correlation between the ABCG2 -15622C/T poly morphism and the ABCG2 (1143C/T, Ta

ble

2. C

linic

al o

utc

om

e ac

cord

ing

to

po

lym

orp

his

ms.

Poly

mo

rph

ism

Pati

ents

, n

(%

)H

WE

p-v

alu

eA

llelic

fr

equ

enci

esG

eno

typ

e g

rou

ps

Pati

ents

in

gen

oty

pe

gro

up

, n (

%)

Res

po

nse

, n

(%

)p

-val

ue†

TTP

mo

nth

s(9

5% C

I)p

-val

ue‡

OS

mo

nth

s(9

5% C

I)p

-val

ue‡

AB

CG

2 4

21C

/A

CC

83

(88

.3)

0.55

C 0

.94

CC

83

(88

.3)

16 (1

9.3

)0.

413.

1 (2

.3–3

.9)

0.95

8.0

(5.

3–1

0.7

)0.

38

CA

10 (1

0.6

)A

0.0

6C

A–A

A11

(11.

7)

1 (9

.1)

3.3

(1.7

–4.9

)8

.3 (1

.4–1

5.2)

AA

1 (1

.1)

AB

CG

2 11

43C

/T

CC

54

(59.

3)

0.67

C 0

.78

CC

–CT

88

(96

.7)

17 (1

9.3

)0.

40

3.2

(1.8

–4.6

)0.

36

8.0

(5.

5–1

0.5

)0.

94

CT

34

(37.

4)

T 0.

22TT

3 (3

.3)

0 (0

.0)

3.0

(1.2

–4.8

)7.

7 (5

.0–1

0.4

)

TT3

(3.3

)

AB

CG

2 -1

5622

C/T

CC

47 (

52.8

)0.

99

C 0

.72

CC

–CT

82 (

92.1

)16

(19.

5)

0.74

4.0

(2.7

–5.3

)0.

078

.3 (

5.8

–10.

8)

0.91

CT

35 (3

9.3

)T

0.28

TT7

(7.9

)1

(14

.3)

2.8

(0.5

–5.1

)7.

7 (7

.0–8

.4)

TT7

(7.9

)

AB

CG

2 h

aplo

typ

e§

TT–T

T +

TT–

othe

r7

(8.0

)TT

–TT

+

TT–o

ther

7 (8

.0)

1 (1

4.3

)0.

732.

8 (0

.5–5

.1)

0.0

67.

7 (7

.0–8

.4)

0.93

Oth

er–o

ther

81 (

92.0

)O

ther

–oth

er81

(92

.0)

16 (1

9.8

)4

.0 (2

.7–5

.3)

9.6

(7.3

–11.

9)

A t

otal

of

94

, 91,

89

and

88

sam

ple

s o

ut o

f th

e 9

4 p

atie

nts

wer

e ev

alua

ble

fo

r A

BC

G2

421C

/A, A

BC

G2

114

3C/T

, AB

CG

2 -1

5622

C/T

and

AB

CG

2 ha

plo

typ

e an

alys

is, r

esp

ecti

vely

, whi

le in

thr

ee a

nd fi

ve s

ubje

cts

the

AB

CG

2 11

43C

/T a

nd A

BC

G2

-156

22C

/T g

enot

ypes

wer

e no

t d

etec

tab

le, r

esp

ecti

vely

. † P

ears

on

-c2

two

-sid

ed t

est.

‡Lo

g-r

ank

test

.§A

BC

G2

114

3C/T

and

-15

622C

/T.

HW

E: H

ard

y–W

einb

erg

equi

libri

um; O

S: O

vera

ll su

rviv

al; T

TP: T

ime

to p

rog

ress

ion.



-15622C/T) haplotype with moderate-to-severe (grade 2/3) diarrhea in gefitinib-treated patients. The current findings should prompt further prospective trials to validate the role of these ABCG2 polymorphisms for treatment selection and improve patient quality of life and compliance. Indeed, the toxicity profile of gefi-tinib is mild, but diarrhea occurred in 49% of patients in the NSCLC trials and could affect the quality of life. The patients may have as many as five to seven episodes of diarrhea a day. Notably, for more than seven stools a day (NCI Common Toxicity Criteria Grade 3 or 4), gefi-tinib should be stopped for up to 14 days until diarrhea has resolved. When gefitinib treat-ment is resumed, it should be at a lower dose, while patients should be instructed on diarrhea management, including diet and antidiarrheal regimens. Conversely, patients harboring geno-types more resistant to drug toxic effects could be treated with higher doses.

Several previous studies have been devoted to identify molecular markers that can predict response and toxicity to TKIs in NSCLC. An important group of potential markers includes germline polymorphisms in relevant genes such as the direct targets of these drugs (i.e., EGFR and VEGFR), metabolizing enzymes (CYP enzymes) and efflux transporters (ABCB1 and ABCG2) [20,22,28]. In this respect, we have previ-ously reported an association of the AKT1-SNP4 polymorphism with TTP and OS in advanced NSCLC patients treated with gefitinib, as well

as a correlation between the EGFR -191C/A, -216G/T and R497K polymorphisms and gefitinib-induced gastrointestinal toxicity [28]. EGFR-activating mutations were also predic-tive for response, progression and OS, but not for toxicity, as reported previously [29]. Indeed, while polymorphisms might increase the genetic susceptibility to tumor response and skin/gas-trointestinal toxicity because they occur in both normal and tumor tissues, activating muta-tions in EGFR should not play a role in toxicity because they are harbored solely by tumor tissues. Moreover, polymorphisms can be assessed with a simple blood test, and their analysis is easier to adopt in the routine clinical setting than tumor gene-expression arrays, which require core needle biopsies of patient’s tumors with laser microdis-section and sub sequent sophisticated infrastruc-ture. In particular, assessing germline polymor-phisms is extremely appealing in the advanced cancer setting, when diagnosis is usually per-formed from small needle biopsy samples and the handling of tumor material can be problematic.

Our results corroborate the findings of sev-eral published studies, demonstrating that our population is representative of Caucasian populations treated with gefitinib. Therefore, in the current study we explored the impact of ABCG2 polymorphisms in the same population of NSCLC patients.

ABCG2 is a member of the G subfamily of ABC transporters. This integral transmembrane pump protein is expressed in apical membranes

Table 3. Toxicity according to polymorphisms.

Polymorphism Skin rash (grade 0 vs 1+)

p-value† Skin rash (grade 0–1 vs 2–3)

p-value† Diarrhea (grade 0 vs 1+)

p-value† Diarrhea(grade 0–1 vs 2–3)

p-value†

ABCG2 421C/A

CC 36 vs 42 0.59 57 vs 21 0.30 45 vs 31 0.28 70 vs 6 0.38

CA–AA 5 vs 4 8 vs 1 7 vs 2 9 vs 0

ABCG2 1143C/T

CC–CT 36 vs 46 0.06 60 vs 22 0.30 47 vs 33 0.15 74 vs 6 0.62

TT 3 vs 0 3 vs 0 3 vs 0 3 vs 0

ABCG2 -15622C/T

CC–CT 35 vs 41 0.57 58 vs 18 0.26 47 vs 27 0.74 71 vs 3 <0.01

TT 4 vs 3 4 vs 3 4 vs 3 4 vs 3

ABCG2 haplotype‡

TT–TT + TT–other 4 vs 3 0.55 4 vs 3 0.27 4 vs 3 0.76 4 vs 3 <0.01

Other–other 34 vs 41 57 vs 18 46 vs 27 70 vs 3The statistical analyses for skin toxicity were performed in a total of 87 (ABCG2 421C/A), 85 (ABCG2 1143C/T), 83 (ABCG2 -15622C/T) and 82 (ABCG2 haplotype) patients and the statistical analyses for gastrointestinal toxicity were performed in a total number of 85 (ABCG2 421C/A), 83 (ABCG2 1143C/T), 81 (ABCG2 -15622C/T) and 80 (ABCG2 haplotype) patients. The worst toxicity grade for each patient was reported. †Pearson-c2 two-sided test.‡ABCG2 1143C/T and -15622C/T.



of several organs, including the liver, kidney, brain and intestine [25]. The apical membrane localization of ABCG2 in epithelial cells along the gastrointestinal tract suggests a major role for this transporter in the first line of defense against xenobiotics [30]. Therefore, it is reason-able to assume that polymorphisms that impact on the expression and/or efflux function of this pump protein, might markedly alter the absorp-tion, bioavailability and elimination of orally administered drugs, as is the case with gefitinib, thereby affecting the drug-induced toxicity or efficacy [31–33].

The ABCG2 1143T/T and -15622T/T poly-morphic variants were recently associated with decreased ABCG2 mRNA expression in various tissues, including the intestine [21]. Although lower ABCG2 mRNA levels are not necessarily associ-ated with lower protein levels and/or lower activ-ity, they can be indicative of a potential impact of these polymorphisms on ABCG2 expression and function. Consistently, we observed that indi-viduals carrying the ABCG2 -15622T/T geno-type had significantly more grade 2–3 diarrhea, as well as patients harboring one copy of TT in the ABCG2 haplotype. Given the small number (n = 7) of patients harboring the risk genotype,

in order to determine whether or not other fac-tors could potentially explain the occurrence of diarrhea, we carefully examined their baseline demographic and biological characteristics, which were similar to the average of the studied popula-tion. Furthermore, we used the Bonferroni cor-rection to reduce false-positive findings, demon-strating that the ABCG2 -15622T/T poly morphic variant and the ABCG2 TT haplotype were still significantly associated with grade 2/3 diar-rhea. Our results are in line with the previously reported association of the ABCG2 TT haplotype with increased toxicity (any toxicity > grade 2) in sunitinib-treated patients [22]. Moreover, this haplotype has also been associated with increased erlotinib exposure [20]. Altogether, these results suggest that the TT genotype is associated with a lower expression and/or decreased activity of ABCG2, which thereby affects the elimination of these TKIs and increases the drug-induced toxicity. These polymorphisms should be used together with EGFR -191C/A, -216G/T and R497K in order to identify optimal candidates for gefitinib treatment, thereby reducing the risk of gastrointestinal toxicity that markedly affects patient’s quality of life, provided that their role is firmly established in prospective trials.

Figure 1. ABCG2 protein expression in non-small-cell lung cancer tissue specimens determined by immunohistochemical analysis. Representative staining intensities: (A) negative: 0; (B) weak positive: 1; (C) moderately positive: 2; and (D) strong positive: 3 (arrows).



In the present study, no associations were detected with skin toxicity. These data are in agreement with the study by Rudin et al., which skin toxicity was correlated with erlotinib exposure levels, but not with ABCG2 421C/A, 1143C/T or -15622C/T [20]. Furthermore, the 250 mg/day dose of gefitinib used in our patients is well below the maximum tolerated dose and the incidence of skin toxicity was significantly lower than with 500 mg/day gefitinib [34].

Cusatis et al. reported a strong association between the ABCG2 421C/A polymorphism and diarrhea in gefitinib-treated NSCLC patients [19]. Furthermore, they suggested that their results reported for gefitinib would be representative for other orally ingested drugs that are trans-ported by ABCG2 such as erlotinib. However, Rudin et al. found no correlation between the ABCG2 421C/A polymorphism and diarrhea or skin rash in erlotinib-treated patients [20]. Also, Akasaka et al. found no association between this polymorphism and diarrhea in gefitinib-treated Japanese NSCLC patients [35]. Consistently, we did not find any association between the ABCG2 421C/A polymorphism and gefitinib-induced toxicity in our population. Of note, several in vitro studies demonstrate that the ABCG2 421C/A polymorphism is associated with lower ABCG2 protein expression [11,12] and transport function [16]. Moreover, Kobayashi et al. assessed the ABCG2 protein expression in human pla-centas in relation to the ABCG2 421C/A poly-morphism and found that the protein level was significantly lower in patients homozygous for the A421 allele than in those homozygous for the C421 allele, while hetero zygotes had

an intermediate expression [13]. By contrast, Urquhart et al. showed that the ABCG2 protein and mRNA expression in the human intestine did not significantly differ between patients car-rying different variants of the ABCG2 421C/A polymorphism, which might explain the lack of effect that this polymorphism had on gefitinib-induced diarrhea in our population of NSCLC patients [36].

ABCG2 is also expressed in the lungs [23], where it might play a role in the efficacy of gefitinib. It is well described that ABCG2 over expression is commonly associated with multidrug resistance, as this transporter has the ability to extrude its drug substrates out of the cells, thereby decreasing their intracellular accumulation [37]. Therefore, we hypothesized that different ABCG2 polymorphisms could have an impact on the treatment response, as well as on the TTP and OS. However, we did not find any significant association between the evaluated ABCG2 polymorphims and response, clinical benefit, TTP and OS. We did, however, observe a trend towards a significant correlation between the ABCG2 haplotype and TTP, with patients carrying at least one TT copy having shorter TTP. It is not likely that these results are directly linked with gefitinib resistance, since this haplotype was not associated with response. In order to further understand these results, we analyzed the ABCG2 protein expres-sion in the lung tumor tissue of 49 patients. In this patient population, we did not find a cor-relation between the different variants of each polymorphism and the levels of protein expres-sion, although IHC is only semiquantitative.

p = 0.2596 p = 0.1602 p = 0.8427

AB

CG

2 p

rote

in e

xpre

ssio

n

(IH

C g

rad

ing

)

250

200

150

100

50

0

CC CA AA CC CT TT CC CT TT

ABCG2 421C/A ABCG2 1143C/T ABCG2 -15622C/T

Figure 2. Association between ABCG2 421C/A, 1143C/T and -15622C/T genotypes with ABCG2 protein expression in non-small-cell lung cancer patients. For each patient, two tumor cores were scored based both on the intensity and percentage of positively stained cells. The average of the two cores was calculated and plotted on the graph. Bars represent the median protein expression of each group. Differences in protein expression among the genotypes were analyzed with the Kruskal–Wallis test. IHC: Immunohistochemistry.



Furthermore, we failed to find a correlation between ABCG2 protein expression and response, clinical benefit, TTP and OS, which might be a result of the small patient group available for this analysis. To our knowledge, this is the first study evaluating ABCG2 protein expression in relation to these three ABCG2 polymorphisms in lung tissue. As mentioned earlier, the ABCG2 1143C/T and -15622C/T polymorphisms were associated with decreased levels of ABCG2 expression [21]. However, Poonkuzhali et al. evaluated ABCG2 mRNA levels, while we determined the levels of protein expression, which might not be correlated. In addition, the tissues analyzed by Poonkuzhali et al. did not include the lung [21], suggesting that the impact of these poly morphims on pro-tein expression might be tissue-specific. The same might apply to the ABCG2 421C/A poly-morphism, which was associated with decreased levels of protein expression in human placen-tas [13], but not in human intestine [36] and lung (present study).

ConclusionOur findings demonstrate a strong associa-tion between the ABCG2 -15622C/T poly-morphism and the ABCG2 haplotype with grade 2/3 diarrhea induced by gefitinib. Since these polymorphisms can be assessed with a simple blood test, the ABCG2 haplotype has also been previously associated with sunitinib-induced toxicity and erlotinib accumulation [20,22], and novel EGFR TKIs, such as the

potent, irreversible, dual EGFR and HER2-TKI BIBW 2992, induce a higher degree of severe diarrhea [37,38], ABCG2 genotyping might be an easy test to adopt in the routine clinical setting to enhance the selection of patients for different TKI that are currently available. In particular, these germline poly-morphisms, together with EGFR -191C/A, -216G/T and R497K [28], might improve the stratification of patients for TKI treatments by identifying genetically high-risk subgroups for gastrointestinal toxicity. Therefore, larger pro-spective trials are warranted to validate these findings, which might be applied to the future practice of cancer treatment.

Financial & competing interests disclosureThe authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investi gations involving human subjects, informed consent has been obtained from the participants involved.

Executive summary

� Novel targeted drugs, including the EGFR tyrosine kinase inhibitor (TKI) gefitinib, are transport substrates of the ATP-binding cassette transporter, ABCG2.

� The SNP ABCG2 421C/A, 1143C/T and -15622C/T might have an important impact on ABCG2 protein expression, function and localization, and subsequently, on gefitinib activity.

� The aim of the current study was to retrospectively evaluate the correlation between these candidate ABCG2 polymorphisms and gefitinib-induced toxicity, as well as clinical outcome in gefitinib-treated non-small-cell lung carcinoma (NSCLC) patients.

Pharmacogenetics of ABCG2 in gefitinib-treated NSCLC patients � Patients carrying ABCG2 -15622T/T genotype or harbouring at least one TT copy in ABCG2 (1143C/T, -15622C/T) haplotype developed

significantly more grade 2/3 diarrhea (p < 0.01). � No associations were found between polymorphisms and response, time to progression or overall survival. � Consistently, ABCG2 protein levels in tumors were not significantly different between patients carrying different ABCG2 variants.

Conclusion & future perspective � The results of the present study demonstrated a strong association between the ABCG2 -15622C/T polymorphism and the ABCG2

haplotype with gefitinib-dependent, moderate-to-severe diarrhea, suggesting that these pharmacogenetic markers should be considered to optimize NSCLC treatment.

� Since the ABCG2 haplotype has been previously associated with sunitinib-induced toxicity and erlotinib accumulation, and novel EGFR-TKIs such as the irreversible EGFR/HER2-TKI BIBW 2992 induce a higher degree of severe diarrhea, it may prove a very helpful tool for treatment selection with different TKIs in cancer patients.

� Hopefully, these biomarkers will be further validated in prospective studies and used to select cancer patients to be treated with TKIs in the near future.



BibliographyPapers of special note have been highlighted as:n of interestnn of considerable interest

1 Assaraf YG: The role of multidrug resistance efflux transporters in antifolate resistance and folate homeostasis. Drug Resist. Updat. 9, 227–246 (2006).

2 Robey RW, Polgar O, Deeken J, To KW, Bates SE: ABCG2: determining its relevance in clinical drug resistance. Cancer Metastasis Rev. 26, 39–57 (2007).

3 Lemos C, Jansen G, Peters GJ: Drug transporters: recent advances concerning BCRP and tyrosine kinase inhibitors. Br. J. Cancer 98, 857–862 (2008).

4 Elkind NB, Szentpétery Z, Apáti A et al.: Multidrug transporter ABCG2 prevents tumor cell death induced by the epidermal growth factor receptor inhibitor Iressa (ZD1839, gefitinib). Cancer Res. 65, 1770–1777 (2005).

5 Nakamura Y, Oka M, Soda H et al.: Gefitinib (�Iressa’, ZD1839), an epidermal growth factor receptor tyrosine kinase inhibitor, reverses breast cancer resistance protein/ABCG2-mediated drug resistance. Cancer Res. 65, 1541–1546 (2005).

6 Li J, Cusatis G, Brahmer J et al.: Association of variant ABCG2 and the pharmacokinetics of epidermal growth factor receptor tyrosine kinase inhibitors in cancer patients. Cancer Biol. Ther. 6, 432–438 (2007).

7 Yanase K, Tsukahara S, Asada S, Ishikawa E, Imai Y, Sugimoto Y: Gefitinib reverses breast cancer resistance protein-mediated drug resistance. Mol. Cancer Ther. 3, 1119–1125 (2004).

8 Lemos C, Kathmann I, Giovannetti E, Calhau C, Jansen G, Peters GJ: Impact of cellular folate status and epidermal growth factor receptor expression on BCRP/ABCG2-mediated resistance to gefitinib and erlotinib. Br. J. Cancer 100, 1120–1127 (2009).

9 Usuda J, Ohira T, Suga Y et al.: Breast cancer resistance protein (BCRP) affected acquired resistance to gefitinib in a �never-smoked’ female patient with advanced non-small cell lung cancer. Lung Cancer 58, 296–299 (2007).

10 Robey RW, To KK, Polgar O et al.: ABCG2: a perspective. Adv. Drug Deliv. Rev. 61, 3–13 (2009).

11 Imai Y, Nakane M, Kage K et al.: C421A polymorphism in the human breast cancer resistance protein gene is associated with low expression of Q141K protein and low-level drug resistance. Mol. Cancer Ther. 1, 611–616 (2002).

12 Kondo C, Suzuki H, Itoda M et al.: Functional analysis of SNPs variants of BCRP/ABCG2. Pharm. Res. 21, 1895–1903 (2004).

13 Kobayashi D, Ieiri I, Hirota T et al.: Functional assessment of ABCG2 (BCRP) gene polymorphisms to protein expression in human placenta. Drug Metab. Dispos. 33, 94–101 (2005).

14 Morisaki K, Robey RW, Ozvegy-Laczka C et al.: Single nucleotide polymorphisms modify the transporter activity of ABCG2. Cancer Chemother. Pharmacol. 56, 161–172 (2005).

15 Mizuarai S, Aozasa N, Kotani H: Single nucleotide polymorphisms result in impaired membrane localization and reduced ATPase activity in multidrug transporter ABCG2. Int. J. Cancer 109, 238–246 (2004).

16 Bram EE, Ifergan I, Grimberg M, Lemke K, Skladanowski A, Assaraf YG: C421 allele-specific ABCG2 gene amplification confers resistance to the antitumor triazoloacridone C-1305 in human lung cancer cells. Biochem. Pharmacol. 74, 41–53 (2007).

17 Keskitalo JE, Pasanen MK, Neuvonen PJ, Niemi M: Different effects of the ABCG2 c.421C>A SNP on the pharmacokinetics of fluvastatin, pravastatin and simvastatin. Pharmacogenomics 10, 1617–1624 (2009).

18 Sparreboom A, Gelderblom H, Marsh S et al.: Diflomotecan pharmacokinetics in relation to ABCG2 421C>A genotype. Clin. Pharmacol. Ther. 76, 38–44 (2004).

19 Cusatis G, Gregorc V, Li J et al.: Pharmacogenetics of ABCG2 and adverse reactions to gefitinib. J. Natl Cancer Inst. 98, 1739–1742 (2006).

nn First study to demonstrate the importance of ABCG2 polymorphisms in the toxicity of gefitinib.

20 Rudin CM, Liu W, Desai A et al.: Pharmacogenomic and pharmacokinetic determinants of erlotinib toxicity. J. Clin. Oncol. 26, 1119–1127 (2008).

nn First study to demonstrate the importance of ABCG2 polymorphisms in the pharmacokinetics of erlotinib.

21 Poonkuzhali B, Lamba J, Strom S et al.: Association of breast cancer resistance protein/ABCG2 phenotypes and novel promoter and intron 1 single nucleotide polymorphisms. Drug Metab. Dispos. 36, 780–795 (2008).

22 van Erp NP, Eechoute K, van der Veldt AA et al.: Pharmacogenetic pathway analysis for determination of sunitinib-induced toxicity. J. Clin. Oncol. 27, 4406–4412 (2009).

nn First study to demonstrate the importance of ABCG2 1143C/T and -15622C/T polymorphisms in the toxicity of sunitinib.

23 Diestra JE, Scheffer GL, Català I et al.: Frequent expression of the multi-drug resistance-associated protein BCRP/MXR/ABCP/ABCG2 in human tumours detected by the BXP-21 monoclonal antibody in paraffin-embedded material. J. Pathol. 198, 213–219 (2002).

24 Zucali PA, Ruiz MG, Giovannetti E et al.: Role of cMET expression in non-small-cell lung cancer patients treated with EGFR tyrosine kinase inhibitors. Ann. Oncol. 19, 1605–1612 (2008).

25 Maliepaard M, Scheffer GL, Faneyte IF et al.: Subcellular localization and distribution of the breast cancer resistance protein transporter in normal human tissues. Cancer Res. 61, 3458–3464 (2001).

26 Lagendijk JH, Mullink H, van Diest PJ, Meijer GA, Meijer CJ: Tracing the origin of adenocarcinomas with unknown primary using immunohistochemistry: differential diagnosis between colonic and ovarian carcinomas as primary sites. Hum. Pathol. 29, 491–497 (1998).

27 Tibaldi C, Giovannetti E, Vasile E et al.: Correlation of CDA, ERCC1, and XPD polymorphisms with response and survival in gemcitabine/cisplatin-treated advanced non-small cell lung cancer patients. Clin. Cancer Res. 14, 1797–1803 (2008).

28 Giovannetti E, Zucali PA, Peters GJ et al.: Association of polymorphisms in AKT1 and EGFR with clinical outcome and toxicity in non-small cell lung cancer patients treated with gefitinib. Mol. Cancer Ther. 9, 581–593 (2010).

29 Marsh S, Mallon MA, Goodfellow P, McLeod HL: Concordance of pharmacogenetic markers in germline and colorectal tumor DNA. Pharmacogenomics 6, 873–877 (2005).

30 McWhinney SR, McLeod HL. Using germline genotype in cancer pharmacogenetic studies. Pharmacogenomics 10, 489–493 (2009).

31 Lynch TJ, Bell DW, Sordella R et al.: Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N. Engl. J. Med. 350, 2129–2139 (2004).

n Important study to demonstrate the importance of EGFR-activating mutations in the anticancer effect of gefitinib in non-small-cell lung cancer.

32 Sarkadi B, Homolya L, Szakacs G, Varadi A: Human multidrug resistance ABCB and ABCG transporters: participation in a chemoimmunity defense system. Physiol Rev. 86, 1179–1236 (2006).



33 Cusatis G, Sparreboom A: Pharmacogenomic importance of ABCG2. Pharmacogenomics 9, 1005–1009 (2008).

34 Perez-Soler R, Saltz L: Cutaneous adverse effects with HER1/EGFR-targeted agents: is there a silver lining? J. Clin. Oncol. 23, 5235–5246 (2005).

35 Akasaka K, Kaburagi T, Yasuda S et al.: Impact of functional ABCG2 polymorphisms on the adverse effects of gefitinib in Japanese patients with non-small-cell lung cancer. Cancer Chemother. Pharmacol. 66, 691–698 (2010).

36 Urquhart BL, Ware JA, Tirona RG et al.: Breast cancer resistance protein (ABCG2) and drug disposition: intestinal expression, polymorphisms and sulfasalazine as an in vivo probe. Pharmacogenet. Genomics 18, 439–448 (2008).

37 Doyle LA, Ross DD: Multidrug resistance mediated by the breast cancer resistance protein BCRP (ABCG2). Oncogene 22, 7340–7358 (2003).

38 Eskens FA, Mom CH, Planting AS et al.: A Phase I dose escalation study of BIBW 2992, an irreversible dual inhibitor of epidermal growth factor receptor 1 (EGFR) and 2 (HER2) tyrosine kinase in a 2-week on, 2-week off schedule in patients with advanced solid tumours. Br. J. Cancer 98, 80–85 (2008).

n Websites101 Identification of EGFR-TKIs sensitivity or

resistance markers in the tumor may help in optimal patient selection http://clinicaltrials.gov/ct2/show/NCT00831454

102 CTEP protocol development http://ctep.cancer.gov/reporting/ctc.html

103 ISTECH http://snp.istech.info/istech/board/login_form.jsp

Affiliations � Clara Lemos

Department of Medical Oncology, VU University Medical Center, De Boelelaan 1117, 1081HV, Amsterdam, The Netherlands

� Elisa GiovannettiDepartment of Medical Oncology, VU University Medical Center, De Boelelaan 1117, 1081HV, Amsterdam, The Netherlands and Department of Internal Medicine, University of Pisa, via Roma 55, 56100 Pisa, Italy

� Paolo A ZucaliDepartment of Medical Oncology & Hematology, Istituto Clinico Humanitas, Rozzano, Milan, Italy

� Yehuda G AssarafThe Fred Wyszkowski Cancer Research Laboratory, Department of Biology, Technion-Israel Institute of Technology, Haifa 32000, Israel

� George L SchefferDepartment of Pathology, VU University Medical Center, De Boelelaan 1117, 1081HV Amsterdam, The Netherlands

� Tahar van der StraatenDepartment of Clinical Pharmacy & Toxicology, Leiden University Medical Center, Albinusdreef 2, NL-2333ZA Leiden, The Netherlands

� Armida D’InceccoDepartment of Oncology, Azienda USL-6 of Livorno, via Alfieri 36, 57100 Livorno, Italy

� Alfredo FalconeDepartment of Oncology, Azienda USL-6 of Livorno, via Alfieri 36, 57100 Livorno, Italy and University of Pisa, via Roma 55, 56100 Pisa, Italy

� Henk-Jan GuchelaarDepartment of Clinical Pharmacy & Toxicology, Leiden University Medical Center, Albinusdreef 2, NL-2333ZA Leiden, The Netherlands

� Romano DanesiDepartment of Internal Medicine, University of Pisa, via Roma 55, 56100 Pisa, Italy

� Armando SantoroDepartment of Medical Oncology & Hematology, Istituto Clinico Humanitas, Rozzano, Milan, Italy

� Giuseppe GiacconeMedical Oncology Branch, National Cancer Institute, Bethesda, MD 20892, USA

� Carmelo TibaldiDepartment of Oncology, Azienda USL-6 of Livorno, via Alfieri 36, 57100 Livorno, Italy

� Godefridus J PetersDepartment of Medical Oncology, VU University Medical Center, Amsterdam, De Boelelaan 1117, 1081HV Amsterdam, The Netherlands

Impact of ABCG2 polymorphisms on the clinical outcome and toxicity of gefitinib in non-small-cell lung cancer patients

Documents