Irinotecan-induced Diarrhea: Functional
Significance of the Polymorphic ABCC2
FA de Jong1, TJ Scott-Horton2, DL Kroetz3, HL McLeod2, LE Friberg4, RH Mathijssen1, J Verweij1,
S Marsh2and A Sparreboom1,5
Interindividual pharmacokinetic variability of the anticancer agent irinotecan is high. Life-threatening diarrhea is
observed in up to 25% of patients receiving irinotecan and has been related with irinotecan pharmacokinetics and
UGT1A1 genotype status. Here, we explore the association of ABCC2 (MRP2) polymorphisms and haplotypes with
irinotecan disposition and diarrhea. A cohort of 167 Caucasian cancer patients who were previously assessed for
irinotecan pharmacokinetics (90-min infusion given every 21 days), toxicity, and UGT1A1*28 genotype were genotyped
for polymorphisms in ABCC2 using Pyrosequencing. Fifteen ABCC2 haplotypes were identified in the studied patients.
The haplotype ABCC2*2 was associated with lower irinotecan clearance (28.3 versus 31.6l/h; P¼0.020). In patients who
did not carry a UGT1A1*28 allele, a significant reduction of severe diarrhea was noted in patients with the ABCC2*2
haplotype (10 versus 44%; odds ratio, 0.15; 95% confidence interval, 0.04–0.61; P¼0.005). This effect was not observed
in patients with at least one UGT1A1*28 allele (32 versus 20%; odds ratio, 1.87; 95% confidence interval, 0.49–7.05;
P¼0.354). This study suggests that the presence of the ABCC2*2 haplotype is associated with less irinotecan-related
diarrhea, maybe as a consequence of reduced hepatobiliary secretion of irinotecan. As the association was seen in
patients not genetically predisposed at risk for diarrhea due to UGT1A1*28, confirmatory studies of the relationships of
ABCC2 genotypes and irinotecan disposition and toxicity are warranted.
Irinotecan is a widely used anticancer drug that has been
approved for the treatment of advanced colorectal cancer.1
The mechanism of action of irinotecan is associated with
topoisomerase I inhibition by the active metabolite 7-ethyl-
10-hydroxycamptothecin (SN-38), which results in cytotoxic
effects on rapidly dividing cells. The main dose-limiting
toxicities of irinotecan include severe (grade 3 or 4)
myelosuppression with an incidence of about 15–20%, and
delayed-type severe diarrhea, which is characteristically seen
in 20–25% of patients about 5 days after the start of therapy.2
The occurrence of diarrhea, in particular, has significant
clinical ramifications, as it affects the dose that can be safely
administered, and is occasionally associated with life-
threatening events. Although the mechanism by which
irinotecan induces delayed-type diarrhea has not yet been
elucidated, prior investigations have suggested that this side
effect is associated with interindividual variability in both
systemic and intestinal exposure to SN-38.3Hence, identi-
fication of the environmental and genetic factors affecting the
pharmacokinetic profile of SN-38 following irinotecan
treatment could aid in predicting or adapting appropriate,
individualized doses of this drug.
The primary pathway of elimination of SN-38 is a Phase II
glucuronic acid conjugation reaction that results in the
formation of SN-38-glucuronide (SN-38G), and that is
mediated by UDP glucuronosyltransferase isoforms – in
decreasing order of importance – UGT1A1, UGT1A6,
UGT1A7, UGT1A9, and UGT1A10.4In particular, UGT1A1
contains many genetic variants influencing the expression
and functional properties of its encoded protein. Polymorph-
nature publishing group
Received 14 June 2006; accepted 2 October 2006. doi:10.1038/sj.clpt.6100019
Previously presented at the 2006 Annual Meeting of the American Society for Clinical Pharmacology and Therapeutics, Baltimore, MD, 10 March 2006.
1Department of Medical Oncology, Erasmus University Medical Center –Daniel den Hoed Cancer Center, Rotterdam, The Netherlands;2Department of Medicine,
Washington University School of Medicine, St Louis, Missouri, USA;3Department of Biopharmaceutical Sciences, University of California, San Francisco, California,
USA;4Division of Pharmacokinetics and Drug Therapy, Department of Pharmaceutical Biosciences, Uppsala University, Uppsala, Sweden;5Clinical Pharmacology
Research Core, National Cancer Institute, Bethesda, Maryland, USA. Correspondence: A Sparreboom (email@example.com)
42VOLUME 81 NUMBER 1 | JANUARY 2007 | www.nature.com/cpt
isms resulting in absent or very low UGT1A1 activity have
been associated with three heritable unconjugated hyper-
bilirubinemia syndromes: Crigler–Najjar syndrome type 1
and 2, and Gilbert’s syndrome.5–8Gilbert’s syndrome is
frequently seen in Caucasians and has been associated with
the presence of an additional, seventh, dinucleotide (TA)
insertion (UGT1A1*28) in the TATA-box of the UGT1A1
promoter region,7,8leading to a considerable reduced enzyme
expression of about 30–80%.7,9–12The UGT1A1 activity is
inversely related to the number of TA-repeats, varying from
5 to 8.9Although the frequency distribution of the number of
TA-repeats differs among ethnic populations,13the presence
of five or eight alleles is rare compared to the six ((TA)6TAA;
UGT1A1*1) and seven ((TA)7TAA; UGT1A1*28) alleles.
Besides being associated with decreased activity and SN-38
glucuronidation in humans, the presence of UGT1A1*28 is a
well-known risk factor for the occurrence of severe
toxicity.13–16Indeed, there is fairly strong evidence that
individuals with low capacity to glucuronidate SN-38 tend to
have a higher prevalence of irinotecan-induced neutro-
penia.11However, the association between UGT1A1*28
genotype and diarrhea is far less clear. For example, the
presence of one or two UGT1A1*28 alleles explains only less
than half of all cases of severe diarrhea following treatment
with irinotecan.17In addition, other investigations do not
show a relationship between the (homozygous) presence of
UGT1A1*28 and diarrhea as well,18,19indicating the multi-
factorial origin and the clinical complexity of irinotecan-
induced diarrhea. Furthermore, there is evidence beginning
to emerge that polymorphisms in other UGT1A genes may
influence irinotecan-related toxicity. In particular, variations
in the gene encoding the extrahepatically expressed UGT1A7
resulting in lower enzyme activity and less transcriptional
activity have recently been associated with lower gastro-
intestinal irinotecan-induced toxicity and higher antitumor
In addition to metabolism, SN-38 and irinotecan are also
sensitive to direct hepatobiliary secretion mediated by a
number of highly polymorphic members of the ATP-binding
cassette transporters, including ABCC2 (cMOAT; MRP2),20
and, to a lesser extent, ABCC1 (MRP1),21ABCC4 (MRP4),22
and ABCG2 (ABCP; MXR;
BCRP).24In this study, we performed an exploratory analysis
to evaluate the association of single-nucleotide polymorph-
isms (SNPs) in the ABCC2 gene with the pharmacokinetics
and toxicity of irinotecan therapy.
A total of 167 cancer patients (Table 1) diagnosed with a
histologically confirmed malignant solid tumor were treated
with irinotecan. The most common tumor types were of
gastrointestinal and pulmonary origin. All patients (86 male,
81 female) were of European Caucasian descent, were
between 34 and 75 years old (median, 55 years), and were
treated between January 1997 and August 2004.
The typical clearance of irinotecan in the 150 patients who
were sampled for pharmacokinetic purposes was estimated to
be 29.3l/h, with an interindividual variability (CV) of 32%.
The metabolic clearances (uncorrected for fraction metabo-
lized) of SN-38 and SN-38G were 465l/h (49% coefficient of
variation (CV) and 31.9l/h (62% CV), respectively, which is
in line with earlier findings.25,26The means, medians, and
ranges of the individual clearance values and the metabolic
conversion ratios are presented in Table 2.
Side effects of irinotecan therapy were recorded in a subset
of 136 patients receiving the recommended dose (350mg/m2
or 600mg). Grade 3 and 4 diarrhea was observed in 26 of 136
patients (19.1%), and grade 3 and 4 neutropenia was seen in
40 patients (28.8%). Neutropenia, defined as the lowest point
during follow-up (nadir), correlated negatively with both
systemic exposure to irinotecan (r¼?0.276; P¼0.001) and
SN-38 (r¼?0.549; Po0.001). Patients experiencing grade 3
and 4 diarrhea had a higher systemic exposure to SN-38
compared with those experiencing less severe or no diarrhea
(median, 521 versus 615ng?h/ml; Po0.008).
Population frequency of variant ABCC2 and UGT1A1
Six variants in the ABCC2 gene were studied in the current
population (Tables 3 and 4). Depending on the specific
Table 1 Patient demographicsa
Total entered 167
Age (years) 55 (34–75)
Body-surface area (m2)1.86 (1.29–2.38)
Height (m) 1.71 (1.51–1.92)
Weight (kg) 74.4 (38.6–115)
World Health Organization performance status1 (0–2)
Pre-therapy clinical chemistry
Aspartate aminotransferase (U/l) 30 (6–185)
Alanine aminotransferase (U/l)21 (4–225)
Total bilirubin (mM) 8 (3–26)
Serum creatinine (mM)76 (45–151)
Serum albumin (g/l) 40 (30–51)
Total serum protein (g/l)76 (62–88)
White blood cell count (?109/l)
Absolute neutrophil count (?109/l)
aContinuous data are given as median with range in parentheses; categorical data as
the number of patients.
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 81 NUMBER 1 | JANUARY 2007 43
variant, these SNPs were successfully analyzed in 154 to 163
patients. The observed frequency of the variant alleles ranged
between 0.04 and 0.56, which is in line with relative
frequencies reported in Caucasians and Japanese subjects
previously.27–32In particular, relative frequencies of the
variant alleles of ABCC2 ?24C4T, ABCC2 1249 G4A,
and ABCC2 3972 C4T were comparable, although different
distributions across different ethic groups cannot be excluded
given the relatively small number of patients studied. For
example, the association found by Itoda et al.32between
ABCC2 ?24C4T and ABCC2 3972C4T could not be
established in our population. No patients homozygous for
the ABCC2 ?24C4T and ABCC2 IVS26 ?34T4C poly-
morphisms were identified. The frequency distributions of all
SNPs are in Hardy–Weinberg equilibrium. A total of 15
haplotypes were constructed in the 139 patients who had
been successfully genotyped for all six variants, although only
six haplotypes had a frequency greater than 0.02 (Table 4).
Frequencies of the most prevalent haplotypes were compar-
able to those estimated earlier in a group of Caucasian cancer
The number of TA-repeats in the TATA-box of the
promoter region of the UGT1A1 gene was successfully
determined in 147 patients. The frequency of the UGT1A1*28
allele was found to be 28%. The genotype distribution
showed 74 patients (50%) with the *1/*1 genotype, 63
patients (43%) with the *1/*28 genotype, and 10 (7%) with
the *28/*28 genotype, which is in line with findings in other
None of the investigated ABCC2 SNPs was found to be
significantly associated with the investigated irinotecan
pharmacokinetic parameters. However, the ABCC2*2 haplo-
type was associated with lower irinotecan clearance (28.3
versus 31.6l/h; P¼0.020), but no effect on the apparent
clearance of SN-38 or SN-38G or any other parameter
estimates could be demonstrated (Table 5). As predicted
previously,11patients carrying the UGT1A1 *1/*1 genotype
had significantly higher clearance of SN-38 compared with
patients with the *1/*28 or *28/*28 genotypes (median, 387
versus 316 versus 266l/h, respectively; Po0.001), whereas the
clearance of irinotecan was not affected by UGT1A1*28
genotype status. Likewise, the area under the curve (AUC)
ratio of SN-38 and irinotecan was affected by UGT1A1*28
genotype status, with median values of 3.00, 3.64, and 4.49%
(P¼0.001) in the three groups, respectively. In addition, the
AUC ratio of SN-38G and SN-38 (i.e., relative extent of
glucuronidation), and the biliary index were significantly
UGT1A1*28 genotype (see Table 5).
Relationships between genotype and toxicity were analyzed in
the patients receiving 350mg/m2or an equivalent absolute
dose of 600mg.33In 124 patients who had available both
UGT1A1*28 and diarrhea toxicity scores, the presence of at
least one UGT1A1*28 allele was associated with a 1.3-fold
higher (95% confidence interval, 0.53–3.13) occurrence of
severe diarrhea, but this was not statistically significant
(P¼0.587). Likewise, no relationship between UGT1A1*28
genotype and the occurrence of neutropenia was detected
(P¼0.411). For ABCC2*2, similar findings were found. The
presence of the ABCC2*2 allele was not significantly related
with less severe diarrhea (odds ratio, 0.55; 95% confidence
interval, 0.22–1.34; P¼0.185), or neutropenia (P¼0.96).
As shown in Table 6, in 102 patients who had available
both ABCC2 haplotype and UGT1A1*28 genotype and
Table 2 Summary of pharmacokinetic data
ParameterMedian Mean Range
Dose (mg) 600 620 260–875
Infusion duration (h)1.50 1.540.75–2.50
CL irinotecan (l/h)a
REC (%)3.36 3.751.08–24.5
Biliary index9.22 10.220.97–34.3
Biliary index, ratio of molar AUC100irinotecan and REG; CL, clearance (i.e., dose
divided by area under the plasma concentration versus time curve); CLM, metabolic
clearance (i.e., dose divided by area under the plasma concentration versus
time curve); REC, relative extent of conversion (molar AUC100ratio of SN-38 to
irinotecan in %); REG, relative extent of glucuronidation (molar AUC100ratio of
SN-38G to SN-38).
aReported data represent the empirical Bayes estimates.
Table 3 Functional consequences of investigated ABCC2 polymorphisms
ABCC2 genotype PositionEffecta
50-Flanking—Unknownrs1885301 Innocenti et al.28
50-Flanking—Unknown rs2804402Innocenti et al.28
50-UTR—Decreasedrs717620Ito et al.,31Itoda et al.,32Haenisch et al.39
Exon 10V417IUnknownrs2273697Ito et al.,31Itoda et al.,32Kroetz et al.38
ABCC2 IVS26 –34T4CIntron 26—Unknownrs8187698Kroetz et al.38
ABCC2 3972C4TExon 28I324IUnknownrs3740066Ito et al.,31Itoda et al.32
NCBI ID, National Center for Biotechnology Information identification number; 50-UTR, five prime untranslated region.
aNumber represents amino-acid codon.
bProposed functional activity of variant protein relative to reference protein.
44VOLUME 81 NUMBER 1 | JANUARY 2007 | www.nature.com/cpt
received the recommended dose (350mg/m2or 600mg), a
significant reduction of severe diarrhea was noted in patients
with the ABCC2*2 haplotype (10 versus 44%; odds ratio,
0.15; 95% confidence interval, 0.04–0.61; P¼0.005) and in
patients who did not carry a single UGT1A1*28 allele. This
effect was not observed in patients with at least one
UGT1A1*28 allele (32% versus 20%; odds ratio, 1.87; 95%
confidence interval, 0.49–7.05; P¼0.354).
This study provides preliminary evidence for a genetic
predisposition to the pharmacokinetic profile of the anti-
cancer drug irinotecan and suggests that patients with
impaired ABCC2 activity due to an inherited genetic defect
are at an increased risk for irinotecan-induced diarrhea. The
data complement previous knowledge on the clinical
pharmacology of irinotecan and may have important
practical implications for its optimal use.
Numerous polymorphic proteins are involved in irinote-
can elimination. However, altered functionality caused by
inherited variability in a single gene is frequently obscured by
the (compensatory) activity of other enzymes and transpor-
ters and environmental factors.34,35In the past, genotyping
efforts in the context of irinotecan chemotherapy have
mainly focused on the functional significance of Phase II
conjugating pathways involved in the detoxification of the
active irinotecan metabolite, SN-38. For example, numerous
investigations have determined that the UGT1A1*28 poly-
morphism affects irinotecan metabolism and toxicity in
Caucasians. Moreover, the US Food and Drug Administra-
tion has recently incorporated this information in the
package label insert,13in particular regarding its effects on
neutropenia. However, guidelines for clinical investigators
and practicing oncologists on how to screen patients for this
polymorphism are lacking, nor is it clear how dosages should
be adjusted for patients carrying this polymorphism. Most
importantly, a priori determination of this particular
polymorphism has rather poor predictive ability. In fact, it
has been estimated that screening for UGT1A1*28 identifies
only about half of individuals who will eventually experience
(severe) diarrhea following treatment with irinotecan.17In
addition, recently Toffoli et al.18related UGT1A1*28 to
toxicity in 250 patients and concluded that UGT1A1*28 is of
limited relevance to toxicity following treatment with
irinotecan. Likewise, in this study, UGT1A1*28 was not
found to be a clinically relevant predictor for the occurrence
of grade 3 and 4 diarrhea and neutropenia. The incentive for
the current investigation was based on previous data
Table 4 Genotype and haplotype frequencies for investigated polymorphisms
ABCC2 genotypeN Wild type HeterozygousVariantpq
16352 (31.9) 90 (55.2)21 (12.9)0.600.40
15427 (17.5)80 (51.9)47 (30.5) 0.440.56
156 109 (69.9) 47 (30.1) 0 (0.0)0.85 0.15
162 99 (61.1) 54 (33.3)9 (5.6)0.78 0.22
ABCC2 IVS26 ?34T4C
161148 (91.9)13 (8.1) 0 (0.0)0.96 0.04
16166 (41.0) 79 (49.1)16 (9.9) 0.660.34
ABCC2 haplotypeN HomozygousHeterozygous Absentp
GGCGTC (ABCC2*2) 139 11 (7.9)85 (53.2)43 (38.9) 0.34
GGCATC (ABCC2*3) 139 8 (5.8)50 (30.2) 81 (64.0)0.21
AACGTT (ABCC2*4)139 3 (2.2)38 (25.2) 98 (72.6)0.15
AATGTT (ABCC2*5)139 0 (0.0) 39 (28.1)100 (71.9) 0.14
AACGTC (ABCC2*6)139 0 (0.0)17 (12.2) 122 (87.8)0.06
AACGCC (ABCC2*7) 139 0 (0.0)11 (7.9) 128 (92.1)0.04
UGT1A1 genotypeN Wild typeHeterozygousVariantpq
14774 (50.3) 63 (42.9)10 (6.8) 0.720.28
Haplotype, ABCC2 ?1549G4A, ?1019A4G, ?24C4T, 1249G4A, IVS26 ?34T4C, and 3972C4T; homozygous, homozygous wild-type frequency; heterozygous,
heterozygous frequency; absent, frequency of patients in which the haplotype was not detected; p and q, standard Hardy–Weinberg nomenclature for allele frequencies;
N, number of patients studied; variant, homozygous variant frequency.
aNumber represents number of patients, with percentage in parentheses.
bData are given as relative frequency.
cThe presence of an additional, seventh TA-repeat in the promoter region of UGT1A1 is defined as variant (UGT1A1*28), compared to the presence of six TA-repeats (wild type;
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 81 NUMBER 1 | JANUARY 200745
indicating that ABCC2 appears to be the principal transpor-
ter involved in hepatobiliary secretion of irinotecan, SN-38,
as well as SN-38G,20and that multiple functional poly-
morphic variants of ABCC2 have been described.29,31,32,36
ABCC2 is a member of the ABC transmembrane proteins
that bind ATP and use its energy to drive the transport of
various substrates across cell membranes. The ABCC2 gene is
located on chromosome 10q24 and encodes a 1,545-amino-
acid polypeptide, and, like several other transporters of the
ABC superfamily, is found on the apical membrane of
polarized cells in the liver, the kidneys, and the intestines, and
is endogenously expressed at highest levels in the canalicular
membrane of the hepatocyte.37Significant variability of the
relative expression of ABCC2 has been demonstrated.38It is
conceivable that the decreased transport of irinotecan into
the bile by ABCC2 leads to increased hepatic metabolism,
increased transport of irinotecan into the bile by ABCB1 or
ABCG2, or increased transport back into the circulation by
Innocenti et al.28reported the first evidence for a
functional variant of ABCC2 and its influence on inter-
individual irinotecan pharmacokinetic variability. ABCC2
3972C4Twas found to potentially affect ABCC2 activity and
it was suggested that ABCC2 may have a significant impact
on irinotecan clearance. It has been reported that the variant
ABCC2 ?24C4T allele was related with lower mRNA levels
in normal tissue and with an almost 20% reduced activity in
vitro,39which is in line with the suggestion that patients wild
type for this polymorphism may clear irinotecan faster than
patients carrying at least one variant ABCC2 ?24C4T
allele.40In contrast, Zamboni et al.41did not find any effect of
ABCC2 ?24C4Ton the disposition of 9-nitrocamptothecin
Table 6 Functional significance of the ABCC2*2 haplotype
UGT1A1*28 DiarrheaAbsent (high activity) Present (low activity) Odds ratioa(P-value)
Absent (high activity)Grade 0–2
Present (low activity) Grade 0–2
aOdds ratio (with 95% confidence interval in parentheses) reflects protective effect of presence of ABCC2*2 haplotype on occurrence of diarrhea.
Table 5 Pharmacokinetic parameters as a function of genotypea
Present (N=48)Absent (N=75)P-valuec
CL irinotecan (l/h)d
28.3 (13.1–50.4)31.6 (14.0–48.4)0.020
349 (150–1,683)324 (143–1,087)0.996
52.0 (9.91–214) 62.4 (2.24–145) 0.460
REC (%) 3.11 (1.08–7.27)3.55 (1.98–24.5)0.071
REG4.49 (1.41–14.6) 3.91 (1.45–36.3)0.245
Biliary index 8.49 (2.52–27.1) 7.98 (0.97–32.2)0.841
*1/*1 (N=68)*1/*28 (N=58) *28/*28 (N=8)P-valuef
CL irinotecan (l/h)d
29.5 (13.1–50.41)29.7 (11.7–50.4)29.2 (17.3–41.1) 0.919
387 (199–1,683) 316 (107–833) 266 (68.8–495)
57.6 (23.7–214) 53.8 (2.24–163)69.6 (34.3–125)0.326
REC (%)3.00 (1.08–7.35)3.64 (1.88–24.5)4.49 (2.40–8.10)0.001
REG 4.41 (1.26–12.2) 4.02 (1.84–36.3)2.24 (1.38–3.97)0.005
Biliary index7.95 (1.92–27.1)9.04 (0.97–24.5) 18.8 (7.10–34.3)0.014
Biliary index, ratio of molar AUC100irinotecan and REG; CL, clearance (i.e., dose divided by area under the plasma concentration versus time curve); CLM, metabolic clearance
(i.e., dose divided by area under the plasma concentration versus time curve); N, number of patients; REG, relative extent of glucuronidation (molar AUC100ratio of SN-38G to
SN-38); uncorrected for fraction metabolized REC, relative extent of conversion (molar AUC100ratio of SN-38 to irinotecan in %).
aData are reported as median, with range in parentheses.
bPresent is defined as the presence of one or two ABCC2*2, whereas absent is defined as the absence of the ABCC2*2 allele.
cP-values from the non-parametric Mann–Whitney U-test.
dReported data represent the empirical Bayes estimates.
eThe presence of an additional, seventh TA-repeat in the promoter region of UGT1A1 is defined as variant (UGT1A1*28), compared to the presence of six TA-repeats (wild-type;
fP-values from the non-parametric Kruskal–Wallis test, unless stated otherwise.
46VOLUME 81 NUMBER 1 | JANUARY 2007 | www.nature.com/cpt
and its 9-aminocamptothecin metabolite, illustrating the
difficulties associated with in vivo pharmacokinetic–pharma-
cogenetic analyses. Additionally, Colombo et al.29reported
that both ABCC2 ?24C4T and ABCC2 1249G4A did not
affect cellular exposure to nelfinavir in vivo. Although
Hirouchi et al.42suggest that the latter substitution may
not affect the in vivo functioning of ABCC2, recently, this
ABCC2 1249G4A polymorphism was found to be associated
with altered expression of the ABCC2 gene in the liver.38In
particular, compared with haplotypes containing the variant
ABCC2 1249G4A allele, haplotypes containing the wild-type
allele, like the ABCC2*2 haplotype as identified in this study,
had significant lower mRNA levels.38
It is interesting to note that individuals predicted to have
high capacity to glucuronidate SN-38 and low capacity to
transport via ABCC2 are relatively protected from severe
diarrhea, but those with low capacity to glucuronidate and
low capacity to transport are relatively more likely to
experience severe diarrhea. This might be speaking to the
relative affinities of the transporter to flux SN-38, irinotecan,
or SN-38G, and, if verified, could be a critical step forward in
our understanding of the complex pharmacology or irino-
tecan. Assuming that the ABCC2*2 haplotype encodes a
decreased function ABCC2, our finding that the presence of
the ABCC2*2 haplotype results in significantly increased
systemic irinotecan exposure suggests that other transporters
did not compensate completely for the loss of apical
transport capacity, and that the overall hepatobiliary secre-
tion is impaired.
Reduced hepatobiliary secretion of irinotecan may explain
the observed association in this study between the ABCC2*2
haplotype and the lower frequency of irinotecan-induced
diarrhea. Previous work has demonstrated that carboxyles-
terases are highly expressed in enterocytes and mediate local
activation of irinotecan into SN-38.43In this scenario, less
biliary excretion of irinotecan and hence less local activation
within intestinal enterocytes may play a crucial role in the
pathogenesis of late-onset diarrhea observed with irinotecan
chemotherapy. It is theoretically plausible that the presence of
a variant UGT1A1*28 allele may indirectly override the
protective effect of the ABCC2*2 haplotype on diarrhea,
because of reduced glucuronidation capacity to detoxify
enterically formed SN-38 that results in increased exposure to
this toxic metabolite in the enterocytes. All of these scenarios
assume that carriers of the ABCC2*2 haplotype have
decreased biliary transport function.
In addition, these data might suggest as well that SN-38G
is highly dependent upon ABCC2 for transport to the gut. In
this scenario, low capacity to transport SN-38G that is
hydrolyzed to SN-38 in the gut by microbial b-glucuroni-
dases where it is locally toxic,44–46and high capacity to form
SN-38G might protect one from severe diarrhea. With low
capacity to form SN-38G and low capacity to transport via
ABCC2, the scale is perhaps tipped toward the transport of
SN-38 to the gut via other transporters (or via systemic
circulation) to the gut where it is toxic. Alternatively, if the
ABCC2*2 haplotype represents normal transport function,
then a high glucuronidation capacity (UGT1A1*1) coupled
with a high ABCC2 transport activity might be consistent
with biliary elimination of SN-38G and decreased enterocyte
exposure to the toxic SN-38. Understanding the functional
significance of the ABCC2 variants and haplotypes will help
decipher this complex multigene pathway that mediates
irinotecan pharmacokinetics and toxicity.
The hypothesis that decreased local SN-38 detoxification
in patients carrying the UGT1A1*28 allele has a key role in
the pathogenesis of diarrhea, as opposed to increased
hepatobiliary SN-38 secretion, may explain why prophylactic
interventions (e.g., administration of the antibiotic neomy-
cin before treatment) aimed at reducing intraluminal SN-38
exposure are relatively ineffective in reducing the risk for
severe diarrhea.17,47This is perhaps with the exception of the
use of active charcoal administered orally, which might be
effective for all patients through the adsorption of both
intraluminal irinotecan and SN-38.48This supposition also
explains why patients excreting more SN-38G, as observed
in patients not carrying the UGT1A1*28 allele, and which
metabolite is almost completely de-glucuronidated by
microbial b-glucuronidases to form SN-38,49are in general
subject to experience less severe diarrhea.50In addition,
irinotecan and, in particular, SN-38 are supposed to
undergo enterohepatic (re)cycling.51,52Although not clearly
established yet, this recycling is likely highly influenced by
the intestinal expression of metabolizing enzymes and apical
and basolateral functional expression of ABC transporters
and other drug transporters.53In clinical practice, other
factors cooperate with the ABCC2*2 haplotype and/or
UGT1A1*28 polymorphism in the development of irinote-
can-related diarrhea as well, such as other functional
polymorphisms and haplotypes in these genes or in the
genes of other enzymes (like CYP3A4, CYP3A5, UGT1A7,
and UGT1A9) and drug transporters (like OATP1B1,
ABCG2, ABCC1, and ABCB1) involved in irinotecan
elimination, used co-medication, use of complementary
alternative medicines, spread of the disease, and the presence
of microorganisms producing b-glucuronidases. As late-
onset diarrhea remains a severe side effect of irinotecan
therapy that occurs rather frequently, additional research to
define predictive measures of its occurrence in the individual
patient is warranted.
In conclusion, this study suggests that the polymorphic
ABCC2 transporter may have a crucial role in the occurrence
of irinotecan-related diarrhea, particularly in individuals who
UGT1A1-mediated glucuronidation of the active metabolite,
SN-38. It is of most importance to consider these distinctions
because of the implications they have on Food and Drug
Administration labeling and on clinical practice. It is possible
that in the context of specific genetic backgrounds, the
UGT1A1*28 allele is not always associated with more toxicity
and perhaps may even be protective. Confirmation of the
current observation is warranted.
because of impaired
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 81 NUMBER 1 | JANUARY 200747
MATERIALS AND METHODS
Patient treatment. Patients with histological evidence of a solid
tumor that was potentially sensitive to treatment with single-agent
irinotecan were considered eligible for this study. The inclusion and
exclusion criteria, pre-medication schedules, and protocols for
treatment of drug-induced side effects have been summarized
before.34Irinotecan was given once every 3 weeks as a 90-min
continuous intravenous infusion. Other concurrent chemotherapy
or other drugs, food supplements, and/or herbal preparations
known or suspected to interfere with the pharmacokinetics of
irinotecan were not allowed. Clinical protocols, including blood
sampling for the purpose of pharmacokinetic and pharmacogenetic
analyses, were approved by the local ethics board, and all patients
provided written informed consent.
Pharmacokinetic analysis. Blood samples of 5–7ml each for
pharmacokinetic studies were collected at up to 20 serial time
points between time 0 and 3 weeks after the first drug administra-
tion as documented in detail elsewhere.17,25,33,54–57The samples were
centrifuged to obtain plasma, which was stored frozen until the time
of analysis. Concentrations of irinotecan, SN-38, and SN-38G were
determined using a validated method based on liquid chromato-
graphy with fluorescence detection, as described elsewhere.58
Previously developed population models for irinotecan and its
metabolites formed the basis for the pharmacokinetic modeling;
however, here, a three-compartment model for SN-38 was
supported by the data.25,26The analysis was performed using the
software package NONMEM version V (GloboMax, Hanover, MD).
Individual parameters were derived as empirical Bayes estimates
using the POSTHOC option. Metabolic conversion ratios, based on
molar AUCs that were simulated up to 100h after infusion,
including the AUC ratio of SN-38 to irinotecan (relative extent of
conversion in %), the AUC ratio of SN-38G to SN-38 (relative
extent of glucuronidation), and the biliary index (i.e., the ratio of
the AUC of irinotecan and relative extent of glucuronidation) were
also calculated as described previously.34
Pharmacodynamic analysis. Toxicity was scored on a five-point
ordinal scale (grades 0, 1, 2, 3, or 4) using the National Cancer
Institute – Common Toxicity Criteria, version 2.0. Grade 3 and 4
diarrhea scores were grouped and categorized as severe, relative to
grade 0–2. A full blood count, including neutrophils, was obtained
on each individual once every week for at least 3 weeks after drug
administration. Besides the absolute nadir values, the percent
decrease at nadir from baseline was used for the classification of
severity of neutropenia and leukopenia.
Genotype analysis. Whole blood or plasma was used to isolate
genomic DNA using the PureGene Blood Kit (Gentra, Minneapolis,
MN) and the DNA Blood midi kit (Qiagen, Valencia, CA),
respectively, according to the manufacturer’s instructions. Variations
in ABCC2 at positions ?1549G4A, ?1019A4G, ?24C4T,
1249G4A, IVS26 ?34T4C, and 3972C4T were analyzed by
polymerase chain reaction and Pyrosequencing, as described
previously,59using the Pyrosequencing AB PSQ hs96A instrument
and software (Uppsala, Sweden).
The number of TA-repeats in the TATA-box of the promoter
region of the UGT1A1 gene was determined by sizing of polymerase
chain reaction products obtained with UGT1A1 specific primers as
described previously.60Genotypes were assigned as *1/*1, *1/*28,
and *28/*28, where *1 represents the reference allele containing six
TA-repeats (UGT1A1*1), whereas *28 represents the variant allele
containing seven TA-repeats (UGT1A1*28), respectively. The
genotype was called variant if it differed from the Refseq consensus
sequence for the SNP position (http://www.ncbi.nlm.nih.gov/
LocusLink/refseq.html). Genotype–frequency analysis of Hardy–-
Weinberg equilibrium was carried out using hwsim (http://
krunch.med.yale.edu/hwsim/). Haplotypes were determined using
the software package PHASE version 0.9.61
Statistical considerations. All data are presented as median values
with range in parenthesis, unless indicated otherwise. Linkage
disequilibrium between different pairs of genetic variants was
determined in terms of the classical statistic D0. The absolute value
for D0of 1 denotes complete linkage disequilibrium, whereas a value
of 0 denotes complete linkage equilibrium. The association of the
variant genotypes and haplotypes with the pertinent pharmaco-
kinetic and pharmacodynamic parameters was evaluated using the
non-parametric Mann–Whitney U-test or the Kruskal–Wallis one-
way analysis of variance on ranks. To relate continuous variables,
Spearman’s rho correlation test performed, whereas a chi-squared
test was used to detect associations between dichotomous variables.
Because this study was mainly exploratory in intent, no adjustments
were performed to evaluate the significance of the multiple
comparisons. Two-tailed P-values of less than 0.05 were considered
to be statistically significant. All statistical calculations were
performed in the software package SPSS version 10.0.7 (SPSS,
We acknowledge the seminal information provided on the web
by Dr Mark J Ratain (University of Chicago, Chicago, IL, USA;
http://www.pharmgkb.org/) to the effect that ABCC2 polymorphisms
may play a role in the pharmacokinetics of irinotecan. This research
was supported in part by NIH grants GM61390 (DLK) and U01 GM63340
(TJS-H, HLM, and SM).
CONFLICTS OF INTEREST
The authors declared no conflict of interest.
& 2007 American Society for Clinical Pharmacology and Therapeutics
1.Chau, I. & Cunningham, D. Adjuvant therapy in colon cancer – what,
when and how? Ann. Oncol. 17, 1347–1359 (2006).
Kawahara, M. Irinotecan in the treatment of small cell lung cancer: a
review of patient safety considerations. Expert Opin. Drug Saf. 5,
Araki, E. et al. Relationship between development of diarrhea and the
concentration of SN-38, an active metabolite of CPT-11, in the intestine
and the blood plasma of athymic mice following intraperitoneal
administration of CPT-11. Jpn. J. Cancer Res. 84, 697–702 (1993).
Ciotti, M. et al. Glucuronidation of 7-ethyl-10-hydroxycamptothecin
(SN-38) by the human UDP-glucuronosyltransferases encoded at the
UGT1 locus. Biochem. Biophys. Res. Commun. 260, 199–202 (1999).
Bosma, P.J. et al. Sequence of exons and the flanking regions of human
bilirubin-UDP-glucuronosyltransferase gene complex and identification
of a genetic mutation in a patient with Crigler–Najjar syndrome, type I.
Hepatology 15, 941–947 (1992).
Aono, S. et al. Identification of defect in the genes for bilirubin
UDP-glucuronosyl-transferase in a patient with Crigler–Najjar syndrome
type II. Biochem. Biophys. Res. Commun. 197, 1239–1244 (1993).
Bosma, P.J. et al. The genetic basis of the reduced expression of bilirubin
UDP-glucuronosyltransferase 1 in Gilbert’s syndrome. N. Engl. J. Med.
333, 1171–1175 (1995).
Monaghan, G. et al. Genetic variation in bilirubin
UPD-glucuronosyltransferase gene promoter and Gilbert’s syndrome.
Lancet 347, 578–581 (1996).
Beutler, E., Gelbart, T. & Demina, A. Racial variability in the
UDP-glucuronosyltransferase 1 (UGT1A1) promoter: a balanced
polymorphism for regulation of bilirubin metabolism? Proc. Natl. Acad.
Sci. USA 95, 8170–8174 (1998).
10. Desai, A.A., Innocenti, F. & Ratain, M.J. Pharmacogenomics: road to
anticancer therapeutics nirvana? Oncogene 22, 6621–6628 (2003).
11. Innocenti, F. et al. Genetic variants in the UDP-glucuronosyltransferase
1A1 gene predict the risk of severe neutropenia of irinotecan. J. Clin.
Oncol. 22, 1382–1388 (2004).
48VOLUME 81 NUMBER 1 | JANUARY 2007 | www.nature.com/cpt
12. Burchell, B. & Hume, R. Molecular genetic basis of Gilbert’s syndrome. Download full-text
J. Gastroenterol. Hepatol. 14, 960–966 (1999).
13. de Jong, F.A. et al. Role of pharmacogenetics in irinotecan therapy.
Cancer Lett. 234, 90–106 (2006).
14. McLeod, H.L. & Watters, J.W. Irinotecan pharmacogenetics: is it time to
intervene? J. Clin. Oncol. 22, 1356–1359 (2004).
15. McLeod, H.L., King, C.R. & Marsh, S. Application of pharmacogenomics in
the individualization of chemotherapy for gastrointestinal malignancies.
Clin. Colorectal Cancer 4(suppl. 1): S43–S47 (2004).
16. Marsh, S. & McLeod, H.L. Pharmacogenetics of irinotecan toxicity.
Pharmacogenomics 5, 835–843 (2004).
17. de Jong, F.A. et al. Prophylaxis of irinotecan-induced diarrhea with
neomycin and potential role for UGT1A1*28 genotype screening: a
double-blind, randomized, placebo-controlled study. Oncologist 11,
18. Toffoli, G. et al. The role of UGT1A1*28 polymorphism in the
pharmacodynamics and pharmacokinetics of irinotecan in patients with
metastatic colorectal cancer. J. Clin. Oncol. 24, 3061–3068 (2006).
19. Carlini, L.E. et al. UGT1A7 and UGT1A9 polymorphisms predict response
and toxicity in colorectal cancer patients treated with capecitabine/
irinotecan. Clin. Cancer Res. 11, 1226–1236 (2005).
20. Sugiyama, Y., Kato, Y. & Chu, X. Multiplicity of biliary excretion
mechanisms for the camptothecin derivative irinotecan (CPT-11), its
metabolite SN-38, and its glucuronide: role of canalicular multispecific
organic anion transporter and P-glycoprotein. Cancer Chemother.
Pharmacol. 42(suppl.): S44–S49 (1998).
21. Luo, F.R. et al. Intestinal transport of irinotecan in Caco-2 cells and
MDCK II cells overexpressing efflux transporters Pgp, cMOAT, and MRP1.
Drug Metab. Dispos. 30, 763–770 (2002).
22. Norris, M.D. et al. Expression of multidrug transporter MRP4/ABCC4 is a
marker of poor prognosis in neuroblastoma and confers resistance to
irinotecan in vitro. Mol. Cancer Ther. 4, 547–553 (2005).
23. Jansen, W.J. et al. CPT-11 sensitivity in relation to the expression of
P170-glycoprotein and multidrug resistance-associated protein. Br. J.
Cancer 77, 359–365 (1998).
24. Kawabata, S. et al. Breast cancer resistance protein directly confers
SN-38 resistance of lung cancer cells. Biochem. Biophys. Res. Commun.
280, 1216–1223 (2001).
25. Mathijssen, R.H. et al. Prediction of irinotecan pharmacokinetics by use
of cytochrome P450 3A4 phenotyping probes. J. Natl. Cancer Inst. 96,
26. Xie, R. et al. Clinical pharmacokinetics of irinotecan and its metabolites:
a population analysis. J. Clin. Oncol. 20, 3293–3301 (2002).
27. Kitagawa, C. et al. Genetic polymorphisms of the multidrug resistance-
associated protein 2 gene (ABCC2) and irinotecan toxicity. J. Clin. Oncol.,
2004 Annual Meeting Proceedings (Post-Meeting Edition) 22, 2009 (2004).
28. Innocenti, F. et al. Pharmacogenetic analysis of interindividual irinotecan
(CPT-11) pharmacokinetic (PK) variability: evidence for a functional
variant of ABCC2. J. Clin. Oncol., 2004 Annual Meeting Proceedings
(Post-Meeting Edition) 22, 2010 (2004).
29. Colombo, S. et al. Influence of ABCB1, ABCC1, ABCC2, and ABCG2
haplotypes on the cellular exposure of nelfinavir in vivo. Pharmacogenet.
Genomics 15, 599–608 (2005).
30. Innocenti, F. et al. Irinotecan (CPT-11) pharmacokinetics (PK) and
neutropenia: interaction among UGT1A1 and transporter genes. J. Clin.
Oncol., 2005 ASCO Annual Meeting Proceedings 23, 2006 (2005).
31. Ito, S. et al. Polymorphism of the ABC transporter genes, MDR1, MRP1
and MRP2/cMOAT, in healthy Japanese subjects. Pharmacogenetics 11,
32. Itoda, M. et al. Polymorphisms in the ABCC2 (cMOAT/MRP2) gene found
in 72 established cell lines derived from Japanese individuals: an associa-
tion between single nucleotide polymorphisms in the 50-untranslated
region and exon 28. Drug Metab. Dispos. 30, 363–364 (2002).
33. de Jong, F.A. et al. Flat-fixed dosing of irinotecan: influence on
pharmacokinetic and pharmacodynamic variability. Clin. Cancer Res. 10,
34. Mathijssen, R.H. et al. Irinotecan pathway genotype analysis to predict
pharmacokinetics. Clin. Cancer Res. 9, 3246–3253 (2003).
35. de Jong, F.A. et al. ABCG2 pharmacogenetics: ethnic differences in allele
frequency and assessment of influence on irinotecan disposition. Clin.
Cancer Res. 10, 5889–5894 (2004).
36. Wada, M. Single nucleotide polymorphisms in ABCC2 and ABCB1 genes
and their clinical impact in physiology and drug response. Cancer Lett.
234, 40–50 (2006).
37. Ito, K. et al. Apical/basolateral surface expression of drug transporters
and its role in vectorial drug transport. Pharmaceut. Res. 22, 1559–1577
38. Kroetz, D.L. et al. 1249G4A polymorphism of ABCC2 (MRP2) is
associated with altered gene expression in human liver. J. Clin. Oncol.,
2006 Annual Meeting Proceedings Part I. 24, 13072 (2006).
39. Haenisch, S. et al. Influence of polymorphisms of ABCB1 and ABCC2 on
mRNA and protein expression in normal and cancerous kidney cortex.
Pharmacogenom. J. [E-pub ahead of print, 2006].
40. Zhou, Q. et al. Pharmacogenetic profiling across the irinotecan
pathway in Asian patients with cancer. Br. J. Clin. Pharmacol. 59,
41. Zamboni, W.C. et al. Disposition of 9-nitrocamptothecin and its
9-aminocamptothecin metabolite in relation to ABC transporter
genotypes. Invest. New Drugs 24, 393–401 (2006).
42. Hirouchi, M. et al. Characterization of the cellular localization, expression
level, and function of SNP variants of MRP2/ABCC2. Pharmaceut. Res. 21,
43. Wadkins, R.M. et al. Discovery of novel selective inhibitors of human
intestinal carboxylesterase for the amelioration of irinotecan-induced
diarrhea: synthesis, quantitative structure–activity relationship analysis,
and biological activity. Mol. Pharmacol. 65, 1336–1343 (2004).
44. Takasuna, K. et al. Protective effects of kampo medicines and baicalin
against intestinal toxicity of a new anticancer camptothecin derivative,
irinotecan hydrochloride (CPT-11), in rats. Jpn. J. Cancer Res. 86, 978–984
45. Takasuna, K. et al. Involvement of beta-glucuronidase in intestinal
microflora in the intestinal toxicity of the antitumor camptothecin
derivative irinotecan hydrochloride (CPT-11) in rats. Cancer Res. 56,
46. Takasuna, K. et al. Inhibition of intestinal microflora beta-glucuronidase
modifies the distribution of the active metabolite of the antitumor
agent, irinotecan hydrochloride (CPT-11) in rats. Cancer Chemother.
Pharmacol. 42, 280–286 (1998).
47. Sharma, R., Tobin, P. & Clarke, S.J. Management of chemotherapy-
induced nausea, vomiting, oral mucositis, and diarrhoea. Lancet Oncol.
6, 93–102 (2005).
48. Michael, M. et al. Phase II study of activated charcoal to prevent
irinotecan-induced diarrhea. J. Clin. Oncol. 22, 4410–4417 (2004).
49. Kehrer, D.F. et al. Modulation of irinotecan-induced diarrhea by
cotreatment with neomycin in cancer patients. Clin. Cancer Res. 7,
50. Xie, R. et al. Clinical pharmacokinetics of irinotecan and its
metabolites in relation with diarrhea. Clin. Pharmacol. Ther. 72,
51. Atsumi, R., Suzuki, W. & Hakusui, H. Identification of the metabolites of
irinotecan, a new derivative of camptothecin, in rat bile and its biliary
excretion. Xenobiotica 21, 1159–1169 (1991).
52. Chabot, G.G. et al. Population pharmacokinetics and
pharmacodynamics of irinotecan (CPT-11) and active metabolite SN-38
during phase I trials. Ann. Oncol. 6, 141–151 (1995).
53. de Jong, F.A. et al. Hepatic transport, metabolism and biliary excretion
of irinotecan in a cancer patient with an external bile drain. Cancer Biol.
Ther. 5, 1105–1110 (2006).
54. de Jonge, M.J. et al. Pharmacokinetic, metabolic, and pharmacodynamic
profiles in a dose-escalating study of irinotecan and cisplatin. J. Clin.
Oncol. 18, 195–203 (2000).
55. Kehrer, D.F. et al. Modulation of irinotecan metabolism by ketoconazole.
J. Clin. Oncol. 20, 3122–3129 (2002).
56. Mathijssen, R.H. et al. Effects of St. John’s wort on irinotecan
metabolism. J. Natl. Cancer Inst. 94, 1247–1249 (2002).
57. Sparreboom, A. et al. Phase I and pharmacokinetic study of irinotecan
in combination with R115777, a farnesyl protein transferase inhibitor.
Br. J. Cancer 90, 1508–1515 (2004).
58. Sparreboom, A. et al. Liquid chromatographic determination of
irinotecan and three major metabolites in human plasma, urine and
feces. J. Chromatogr. B 712, 225–235 (1998).
59. Marsh, S. et al. Pyrosequencing of clinically relevant polymorphisms.
Methods Mol. Biol. 311, 97–114 (2005).
60. Saeki, M. et al. Comprehensive UGT1A1 genotyping in a Japanese
population by pyrosequencing. Clin. Chem. 49, 1182–1185 (2003).
61. Schaid, D.J. et al. Caution on pedigree haplotype inference with
software that assumes linkage equilibrium. Am. J. Hum. Genet. 71,
CLINICAL PHARMACOLOGY & THERAPEUTICS | VOLUME 81 NUMBER 1 | JANUARY 200749