CANCER TREATMENT REVIEWS (2005) 31, 90–105

www.elsevierhealth.com/journals/ctrv

LABORATORY–CLINICAL INTERFACE

Pharmacology of oxaliplatin and the use of

pharmacogenomics to individualize therapy
D.M. Kweekela,*, H. Gelderblomb, H.-J. Guchelaara

a
Department of Clinical Pharmacy and Toxicology, Leiden University Medical Center, P.O. Box 9600,
2300 RC Leiden, The Netherlands
b
Department of Clinical Oncology, Leiden University Medical Center, Leiden, The Netherlands

KEYWORDS Summary Oxaliplatin is a relatively new platinum analogue that is currently used
SNP; in pharmacotherapy of metastatic colorectal cancer. Its dose-limiting toxicity is
Polymorphism; sensory neuropathy, which can be modulated by infusion of calcium and magnesium.
Pharmacogenetic; Oxaliplatin exerts its anti-tumour effects by platinum-adduct formation, binding to
Oxaliplatin; cellular proteins and possibly interfering with RNA synthesis as well. If they are not
Colorectal removed from DNA, oxaliplatin adducts are lethal. Cellular defense mechanisms
prevent adduct formation (e.g., glutathione-S-transferase) or remove DNA adducts
(e.g., nucleotide excision repair). Depending on the activity of necessary enzymes in
these cellular defense pathways, oxaliplatin induced damage varies from one indi-
vidual to another.
There is growing evidence that polymorphisms in genes coding for DNA repair
enzymes and metabolic inactivation routes contribute to the interindividual differ-
ences in anti-tumour efficacy and toxicity of oxaliplatin. Single nucleotide polymor-
phisms (SNPs) may yield inactive enzymes, or increased gene transcription and
hence increased enzyme production. This review covers findings of recent investiga-
tions on the associations of SNPs and clinical outcome after oxaliplatin chemother-
apy in metastatic colorectal cancer.

c 2004 Elsevier Ltd. All rights reserved.

Introduction

Development of platinum analogues

Oxaliplatin (Eloxatin
) is a relatively new platinum

* Corresponding author. analogue that has been licensed in the European
E-mail address: D.M.Kweekel@lumc.nl (D.M. Kweekel). Union since 1999, and in the United States since

0305-7372/$ - see front matter c 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ctrv.2004.12.006
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 91

2002. The drug is currently applied in new promis- oxaliplatin, have recently been approved. Interest-
ing chemotherapeutic regimens in the treatment of ingly, oxaliplatin shows no inheritant cross resis-
advanced colorectal cancer. tance with both cisplatin and carboplatin. This is
The first platinum analogue cisplatin (Platinol
) especially relevant for the treatment of colorectal
was described in 1844 as Peyrone’s chloride. Its cancer, a disease that is known to be extremely
cytotoxic properties were unrecognized until 1965 insensitive to the earlier platinum analogues. At
when Rosenberg and his colleagues observed inhibi- the same time the toxicity profile of oxaliplatin is
tion of bacterial growth by an electric current. La- favourable, with frequencies for ototoxicity of
ter it appeared that this phenomenon was due to <1% and for renal toxicity <3%, except for unusual
the formation of cytotoxic compounds, e.g., cis- toxicity with regard to peripheral sensory nerves.
platin in an ammonia buffer around the platinum Sensory neuropathy usually arises during infusion,
electrode and not attributable to direct effects of affects hands, feet and the perioral area and is en-
the electric current itself.1 hanced by cold. These effects appear to be cumu-
Cisplatin was investigated in several clinical tri- lative and they generally reverse within 4–6
als in the early 1970s and became available for clin- months after treatment discontinuation.3
ical use in 1978. Cisplatin has little effect on Sensory neuropathy affects about 85–95% of all
colorectal cancer but it changed the prognosis for patients. An explanation for this unusual high inci-
ovarian and especially for testicular cancer dence is proposed by Grolleau et al.4 Chelation of
patients to a great extent. However, cisplatin calcium ions by the oxalate group was suggested
causes severe side effects of which renal toxicity to block the voltage-gated sodium channels of sen-
and peripheral neuropathy are dose-limiting. Renal sory neurons, producing acute toxicity to the neu-
toxicity caused by cisplatin is clinically manifested ron, and in the long term inducing neuropathy.
by elevation of blood urea nitrogen, serum creati- Indeed, infusion with calcium and magnesium salts
nine and electrolyte disturbances (e.g., hypomag- significantly reduced the incidence of sensory
nesemia). Adequate intravenous hydration, slow neurotoxicity in patients with advanced colorectal
cisplatin infusion rates and simultaneous adminis- cancer without affecting tumour response.5 Only
tration of mannitol are applied to circumvent renal five of 69 patients (7%) receiving calcium/magne-
toxicity. Cisplatin has a high emetogenic potential sium infusions experienced distal paresthesias,
but its gastrointestinal side effects such as nausea whereas these adverse events were reported for
and vomiting can be effectively controlled by 17 of 65 patients (26%) in the control group. McK-
administration of dopamine agonists, corticoster- eage suggests an alternative mechanism for oxa-
oids and especially serotonin antagonists, alone liplatin induced neurotoxicity assuming that
or in combination. In contrast, ototoxicity and oxaliplatin inhibits rRNA synthesis in ganglionic
neurotoxicity by cisplatin are difficult to control sensory nerves, causing damage to sensory nerve
and quite often a reason to stop treatment or re- nucleoli.6 Peripheral neuropathy, diarrhoea and
duce dose. Ototoxicity is characterized by tinnitus leucopenia can be modulated by using a dose
(often reversible) and hearing loss (irreversible), schedule based on circadian rhythm (chronophar-
especially in the high-frequency range. Its severity macology or chronomodulation).7 Time-dependent
is usually related to the cumulative dose received dosage is effective in limiting gastrointestinal and
in subsequent therapeutic courses. Neurotoxicity neurological toxicities, and at the same time
involves peripheral neuropathy of the upper and anti-tumour activity is increased compared to con-
lower limbs, including paresthesias, weakness, tinuous infusion of oxaliplatin. Research indicates
tremors and loss of taste.1 that, although scarcely used, chronomodulation is
not a unique feature of oxaliplatin, and that cis-
platin and carboplatin adverse effects can be mod-
Oxaliplatin ulated in the same manner.8,9 However, due to its
complexity and for practical reasons chronophar-
In an attempt to overcome the renal and gastroin- macology of oxaliplatin has only been used by a se-
testinal side effects of cisplatin, less toxic plati- lected number of specialized medical facilities.
num analogues were developed and as a result,
carboplatin has replaced cisplatin in many chemo-
therapeutic regimens. Carboplatin has a different Efficacy and tolerability
spectrum of toxicity, as its primary toxic effects
are haematological.2 Novel platinum compounds The US Food and Drug Administration (FDA) have
are still being tested (e.g., the orally effective now approved oxaliplatin for treatment of
satraplatin) and others, such as nedaplatin and metastatic colorectal cancer in combination with
92 D.M. Kweekel et al.

5-fluorouracil (5-FU) and leucovorin (LV). Patients center trial demonstrated that a regimen consist-
receiving oxaliplatin should have experienced ing of 5-FU/LV and oxaliplatin was more effective
recurrence or progression of metastatic disease and better tolerated than the topo-isomerase I
within 6 months of completion of first-line 5-FU/ inhibitor irinotecan in combination with oxalipla-
LV + irinotecan combined therapy. Although grade tin, or than irinotecan with 5-FU/LV. Survival rate
3 and 4 haematopoietic and gastrointestinal toxic- was significantly lower for the 5-FU/LV/irinotecan
ity is limited for 5-FU monotherapy, combined regimen and response rate was significantly higher
administration with oxaliplatin significantly for 5-FU/LV/oxaliplatin (45%) compared to oxalipl-
increases the incidence of thrombocytopenia, neu- atin/irinotecan (35%, p = 0.03) and 5-FU/LV/irino-
tropenia, diarrhoea and nausea.10 Oxaliplatin com- tecan (31%, p = 0.002) combinations.13
bination therapy (85 mg/m2 every 2 weeks or
130 mg/m2 every 3 weeks) has a twofold higher
response rate compared to 5-FU/LV therapies and Individualization of therapy
also improves progression free survival (PFS) in
chemotherapy-naive patients11 (Table 1). In addi- Although response rates shown in various clinical
tion, patients who underwent curative resection trials vary depending on patient selection and con-
of stage II or III colon tumours had 23% less chance comitant therapy, it is clear that oxaliplatin is a
of relapse within 3 years if they had received 5-FU/ promising new agent in the scarce armamentarium
LV + oxaliplatin compared to 5-FU/LV alone as of drugs available for the treatment of metastatic
adjuvant treatment.12 colorectal cancer. However, the clinical efficacy
The efficacy of oxaliplatin in combination with and toxicity for the individual patient are still lar-
5-FU/LV was compared to 5-FU/LV or oxaliplatin gely unpredictable, and at the start of chemother-
monotherapy in metastatic colorectal cancer pa- apy it is unclear which chemotherapeutic regimen
tients who experienced relapse or progression the individual patient will benefit most from.
within 6 months of first-line therapy. At interim Objective response rates for various chemothera-
analysis, patients with the combination regimen peutic regimens range from 10% to 50%.
had a significantly higher response rate and longer For some other drugs, such as antiepileptics and
median time to progression (9.9%, 4.6 months) antibiotics, patient outcome can be described by
compared to 5-FU/LV (0%, p = 0.0002, 2.7 months) pharmacokinetic and/or pharmacodynamic model-
and oxaliplatin alone (1%, 1.6 months). The most ling, based on measurements of the drug in blood
frequent grade 3 and 4 side effects for patients or other body fluids. For example, phenytoin serum
receiving combination therapy were neutropenia levels are monitored in epileptic patients, as they
(44%), neuropathy (7%), nausea (11%), vomiting are useful for predicting efficacy and toxicity in
(9%) and diarrhoea (11%).10 A randomized multi- the individual patient. Unfortunately, for anti-

Table 1 Comparison of different chemotherapeutic regimens in metastatic colorectal disease

Previous treatment Study treatment
5-FU/LV 5-FU/LV + OX OX p value
Response (%)
Chemotherapy-naive11 22.3 50.7 – 0.0001
Pretreated (5-FU/LV + IRI)10 0 9.9 1.3 0.0001
Progression-free survival (months)
Chemotherapy-naive11 6.2 9.0 – 0.0003
Time to progression (months)
Pretreated (5-FU/LV + IRI)10 2.7 4.6 1.6 0.0003
Overall survival (months)
Chemotherapy-naive11 14.7 16.2 – 0.12
Pretreated (5-FU/LV + IRI)10 8.8 9.9 8.1 0.09
Overall toxicity (%)
Chemotherapy-naive11 5.5 17.8 – Grade 4
Pretreated (5-FU/LV + IRI)10 41 73 46 Grade 3+4
Response was defined as complete or partial response based on WHO criteria11 or RECIST criteria.10
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 93

tumour therapy there are generally no such simple Oxaliplatin shows similar chemical behaviour
concentration–effect relationships14 and clinical and has a comparable mechanism of action as com-
efficacy depends upon a diversity of factors, pared to the other platinum derivatives. First, the
including inherited and acquired drug resistance pro-drug oxaliplatin is activated by conversion to
of tumour tissue or host body. monochloro, dichloro and diaquo compounds by
In addition to classical therapeutic drug moni- non-enzymatic hydrolysis and displacement of the
toring15 the newly evolving field of pharmacogenet- oxalate group (Fig. 2). The kinetics of hydrolysis
ics advocates drug choice taking into account differ among platinum compounds, being slower
genetic differences among patients and/or for oxaliplatin than for cisplatin.17 The highly reac-
tumours. Pharmacogenetics gives us more insight tive monochloro, dichloro and diaquo intermedi-
into clinically relevant issues of response diversity, ates react with sulphur- and amino groups in
and will therefore most likely change the manage- proteins, RNA and DNA. Its anti-tumour effects
ment of cancer therapy in the future.16 are thought to be related to the formation of Pt–
In this manuscript the pharmacogenetics rele- DNA adducts. Other reactions include irreversible
vant to oxaliplatin chemotherapy are reviewed. binding to biomolecules such as albumin, cysteine
After a brief introduction on the pharmacology of (Cys), methionine (Met) and reduced glutathione
oxaliplatin, genetic polymorphisms in genes coding (GSH), which are in fact the first steps of in vivo
for DNA-repair, biotransformation and colorectal biotransformation and cellular detoxification7
tumour response are discussed and their impact (Fig. 2).
on clinical outcome is reviewed. This information After a 2-h infusion of oxaliplatin 70% of the
gives detailed insight into (experimental) drug is bound to plasma proteins (mostly albumin)
approaches on how to individualize oxaliplatin- thereby losing its anti-tumour potential and 5 days
based chemotherapy in colorectal cancer aiming after a single infusion this fraction increases to
at increasing drug efficacy while minimizing che- about 95%.18 The estimated volume of distribution
motherapy-induced toxicity. is 35L for total oxaliplatin and 440L if only the
unbound fraction is considered.19,20 Plasma
obtained by filtration (PUF) or centrifugation
Pharmacokinetics (PUC) is not totally identical but adequately
reflects the so-called ‘free’ platinum fraction
The oxaliplatin molecular structure consists of a (which represents the active drug). Both PUF
central platinum atom (Pt), surrounded by a 1,2- and PUC give an overestimate of the free plati-
diaminocyclohexane group (DACH) and a bidentate num concentration because they both include
oxalate ligand (Fig. 1). Due to its DACH ligand, ste- inactive oxaliplatin bound to amino acids and
reochemical isomers of the oxalate–Pt–DACH other small biomolecules present in blood. A sig-
complex exist, of which the trans-1-(R,R)–DACH– nificant correlation exists between PUF and total
Pt isomer (oxaliplatin) was shown to be the most platinum, and between the free fraction and total
cytotoxic.7 platinum (Fig. 3). The measurement of platinum

R1 R2 R3 R4

cisplatin H H Cl Cl

H2
H
O
R1 N R3 carboplatin H
O
Pt
R2 N R4 O
H2
O

oxaliplatin
O O

O O

Figure 1 Chemical structure of oxaliplatin and parent platinum derivatives, showing the various ligands and the
trans-1-(R,R)-DACH-Pt isomer configuration for oxaliplatin.
94 D.M. Kweekel et al.

2+
H O H H
N N Cl
O C N OH2
Pt Pt Pt
N O C N Cl N OH2
H H H
O

Oxaliplatin Dichloro intermediate Diaquo intermediate

H COOH
DNA adduct
N S
NH2 formation
Pt
NH 2
N S
H
COOH
Pt-(DACH)(Cys)2

+ 2+
H H
N N H2 COOH
S N
Pt NHGlu Pt
N NH 2 COOH N S
H CH2 C OOH H
CH 3

Pt-(DACH)(GSH) Pt-(DACH)(Met)

2+
H NHGlu
N S
CONHCH2 C OOH
Pt
CONHCH2 C OOH
N S
H
NHGlu
Pt-(DACH)(GSH)2

Figure 2 Non-enzymatic hydrolysis reactions of oxaliplatin in vivo and in vitro. The aquated derivatives of oxaliplatin
are considered to be the biologically active species, capable of adduct formation with various sulphide- and amino
groups. Such groups are abundant in cellular DNA and biomelocules. Cellular detoxification processes competing with
DNA-adduct formation include conjugation of the aquated compound to glutathione (GSH), methionine (Met) and
cysteine (Cys). The conjugation products are subsequently excreted from the cell and eliminated from the body.

in PUF/PUC or total plasma can give us valuable 21 platinum species are found in urine, but the
insight into the pharmacological behaviour of the main constituents are glutathione (GSH)–DACH–
active platinum species, though measured indi- Pt and several creatinine derivatives.21 Only 2% of
rectly. The main compound in PUF after a 2-h the dose is excreted by faeces. Substantial uptake
infusion period was found to be monochloro- of platinum takes place in erythrocytes, where it is
DACH–Pt.7,21 trapped by reaction with small intracellular bio-
Oxaliplatin species are widely distributed among molecules. The maximal Pt concentration in eryth-
various tissue sites and free platinum is eliminated rocytes is reached 3 h after infusion, followed by a
from the body, mainly by renal clearance. Up to slow decline with a mean half-life of 29–50 days,
50% of the dose is eliminated within the first day, close to the half-life of the erythrocytes them-
depending on renal function. In contrast to carbo- selves, indicating that once inside, oxaliplatin is
platin there is no simple relationship between renal trapped in the erythrocyte and does not substan-
function and platinum exposure (AUC).22 At least tially diffuse into plasma.7
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 95

TOTAL atoms modify the three-dimensional DNA struc-

FREE ture, which inhibits the normal DNA synthesis and
BOUND -repair processes. A direct relationship between
500
450
the cytotoxic effect and the number of platinum
400 atoms was shown for nedaplatin, while upregula-
350 tion of DNA repair decreases sensitivity to, e.g.,
cisplatin.17 Platinum trans adducts are 30 times
Pt (µg/L)

300
250 less toxic than cis adducts due to more effective
200 repair.23 Despite the fact that oxaliplatin only
150
forms trans adducts, it induces more efficient or
100
50
different damage since a 10-fold lower adduct ratio
0 was found relative to cisplatin at an equitoxic dose
0 24 48 72 96 120 of oxaliplatin.17 In addition, more DNA strand
time (hrs) breaks are formed with oxaliplatin compared to
Figure 3 Platinum concentrations during a 120-h oxa- cisplatin. There is some evidence suggesting that
liplatin constant rate infusion (20 mg/m2/day for 5 days) the formation of Pt adducts with oxaliplatin is not
with combined continuous administration of fluorouracil the only mechanism explaining the drug’s anti-
and leucovorin. A correlation between platinum in tumour activity but interference with RNA and pro-
plasma and ultrafiltrate (free) is evident at day 1. During teins may add to this effect. It was found that in
days 3–5, the ultrafilterable (‘free’) platinum concen- mouse leukaemia cell cultures oxaliplatin inter-
tration remains relatively stable whereas the amount of feres with RNA synthesis whereas cisplatin did
bound oxaliplatin (attached to biomolecules) varies not.3 There is some evidence that oxaliplatin con-
considerably during the day [adapted from Levi (7)]. jugates with sulfhydryl groups in hydrophobic pock-
ets of different cellular proteins that may be poorly
reactive with the hydrophilic cisplatin. This conju-
gation will result in inactive proteins that will
Pharmacodynamics interfere with cell function, increasing cellular
toxicity.24
The cytotoxic activity of oxaliplatin is initiated by Compared to the amino ligands of both cisplatin
formation of a DNA adduct between the aquated and carboplatin, the DACH ligand of oxaliplatin is
oxaliplatin derivative and a DNA base. Initially, remarkably bulkier and more hydrophobic. In con-
only monoadducts are formed but eventually oxa- trast to other DACH complexes however, oxalipla-
liplatin attaches simultaneously to two different tin shows relatively good water solubility.
nucleotide bases resulting in DNA cross-links. Com- Together, these properties result in a greater
pared to cisplatin, this conversion takes more time, deformation of tumour DNA by steric hindrance of
but in vitro the two-step process is generally com- adduct formation and this may explain the more
pleted in about 15 min.3 effective inhibition of DNA synthesis by oxaliplatin
The adducts are formed with the N-7 positions of as compared to, e.g., cisplatin.17
guanine and adenine preferentially and in most Interestingly, cellular DNA repair mechanisms
cases these reactions result in intrastrand cross- seem to differ in their response to Pt or Pt–DACH
links. In the cell approximately one of every complexes. The mismatch repair complex (MMR),
100,000 bases can be cross-linked by a platinum e.g., is unable to bind to a DACH–Pt adduct site
atom, resulting in 10,000 platinum atoms per because of its great steric distortion of the DNA
cell.17 About 60% of intrastrand platinum adducts structure.3 After DNA-adduct formation by oxa-
are formed between two guanine bases and 30% liplatin, (tumour) cells will activate cellular
are formed between an adenine and a guanine repair mechanisms. In general, DNA repair is car-
base. Only 1% of all platinum adducts is of the ried out by enzymes that consist of several ami-
interstrand type or the result of a reaction binding no- and sulphur groups. Therefore, oxaliplatin
platinum to a biomolecule and a DNA base (DNA– can be covalently bound to these repair enzymes
protein cross-link). The precise cytotoxic efficacy as well, impairing their function. In order to ob-
of the different types of DNA adducts are as yet tain new repair proteins DNA transcription is acti-
unknown.7 vated, but this transcription may be hindered
In general, the cytotoxic efficacy of platinum because of the existing DNA damage.24 If substan-
compounds in cancer cells can be related to inhibi- tial DNA damage persists this may ultimately lead
tion of DNA synthesis or to saturation of the cellu- to the activation of apoptotic pathways and cell
lar capacity to repair Pt–DNA adducts. Platinum death.25
96 D.M. Kweekel et al.

Pharmacogenetics

Several mechanisms are described that confer de-

creased sensitivity or resistance to oxaliplatin,
including diminished cellular drug accumulation,
increased intracellular drug detoxification and in-
creased Pt–DNA adduct repair (Table 2).
The uptake of platinum by cells is not com-
pletely understood but there is evidence that
decreased accumulation is the most common
mechanism of resistance to cisplatin. Comparable
observations are described for oxaliplatin resistant
cell lines.26 Platinum uptake by cells is an energy Figure 4 Schematic view of cellular defence mecha-
requiring process, but it is not saturable and possi- nisms involved in oxaliplatin resistance. Cellular uptake
bly involves transport by a yet unidentified efflux and efflux determine the level of oxaliplatin in intracel-
pump.27 Once inside the cell, conjugation to gluta- lular fluid. In plasma, extracellular conjugation of oxa-
thione (catalyzed by the enzyme glutathione-S- liplatin to plasma proteins (mainly albumin) results in
renal excretion of inactive drug species. Once inside the
transferase, GST) effectively inactivates platinum
cell, the oxaliplatin prodrug is hydrolyzed to mono-
compounds before DNA damage is induced. This chloro-, dichloro and diaquo active species which form
conjugation reaction is followed by cellular excre- DNA adducts. Intracellular conjugation to glutathione
tion and is therefore related to cellular drug resis- effectively inactivates these highly reactive oxaliplatin-
tance as well. A number of studies indicate an species before DNA damage occurs, followed by cellular
important role of GST in oxaliplatin resistance. excretion into plasma. DNA damage is repaired by
Platinum-induced DNA adducts can be repaired nucleotide excision repair (NER), base excision repair
by several enzyme systems: base excision repair (BER) and replicative bypass.
(BER), nucleotide excision repair (NER), post-repli-
cation repair and mismatch repair (MMR). Induction An overview of several pharmacogenomic stud-
of the enzymes involved in these systems results in ies on the effects of polymorphisms in DNA repair
increased DNA repair activity, more efficient ad- and detoxification of oxaliplatin will be given in
duct removal and hence decreased sensitivity to the following sections. Polymorphic distributions
platinum drugs. of these genes in colorectal cancer patients as well
However, the overall sensitivity of a cell is mul- as therapeutic consequences are summarized in
tifactorial and the relative importance of each pro- Table 3.
cess on ultimate drug sensitivity is difficult to
predict (Fig. 4).
There is growing evidence that common gene
variants (polymorphisms) affect the activity of cel- DNA repair polymorphisms
lular DNA repair and platinum conjugation. Such
polymorphisms in genes coding for enzymes in- Since the primary anti-tumour mechanism of oxa-
volved in oxaliplatin accumulation, detoxicification liplatin is the formation of Pt–DNA adducts (ulti-
and Pt-DNA repair may influence cellular response mately leading to cell cycle arrest and apoptosis),
to oxaliplatin. polymorphisms in genes involving the repair of

Table 2 Mechanisms involved in loss of sensitivity to oxaliplatin (inherited or acquired resistance)

Basic mechanism Cause of decreased sensitivity Reference
Decreased accumulation Unidentified uptake mechanism Unclear Unknown
Unidentified efflux pump Probable Unknown
Increased detoxification Conjugation to Met, Cys Not described n.a.
Glutathione-S-transferase (GSTP) Induction of enzyme 71
Decreased DNA repair NER-system Induction of enzymes 61,63
BER-system Induction of enzymes 34
MMR-system Not involved 29
Post-replication bypass Unclear 31
NER: nucleotide excision repair; BER: base excision repair; MMR: mismatch repair; n.a.: not applicable.
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 97

Table 3 Polymorphisms associated with clinical outcome in oxaliplatin combination therapies of advanced
colorectal cancer
Nucleotide Rs-Numbers wt/wt% wt/mt% mt/mt% Ref Population Effects
change
ERCC1 C496T 11615 34 38 28 62 mcrc-A Not studied
Asn118Asn 20-43- 45-36-0-55 36-21-0-9 73 3mcrc-B wt › survival
100-36
ERCC2 A2251C 13181 30 56 14 42 2/3mcrc-C wt/wt › response,
›survival
Lys751Gln – – 14 80 1mcrc-A No association
found
25-57- 61-29-33-9 14-14-0-0 73 3mcrc-B wt/wt › survival
67-91
ERCC2 G965A 1799793 47 41 12 42 2/3mcrc-C No association
found
Asp321Asn
ERCC2 C499A 238406 32 55 13 42 2/3mcrc-C wt/wt › response,
mt fl response
Arg156Arg 26-38- 58-54-17-73 16-8-0-9 73 3mcrc-B No association
83-18 found
XRCC1 G1301A 25487 41 – – 37 mcrc-C wt/wt › response
Arg399Gln 39-36- 52-57-33-27 9-7-20-9 73 3mcrc-B No association
50-64 found
25 27 5 81 Gastrintes-D Not studied
GSTP1 A342G 1695 49 42 9 71 2mcrc-C wt/wt fl survival,
risk of death
Ile105Val – – 11 80 1mcrc-A No association
45-50- 46-43-67-27 9-7-17-9 73 3mcrc-B wt/wt fl time to
17-64 progression,
wt/wt survival
EGFR 9–23 11568315 22\ 41\ 37\ 62 mcrc-A wt/wt fl survival,
repeats › progression
No. of CA
Notation: amino acid substitution (first column; wt-codon-mt), nucleotide change (second column; wt-nucleotide position-mt
allele). Rs-numbers and nucleotide position according to NCBI database. Symbols: wt = wildtype, mt = mutant, gastrintest = solid
tumours, 75% gastrointestinal, mcrc = metastatic colorectal cancer (undefined), 1mcrc = first line, 2mcrc = second-line etc., \ for
EGFR wt/wt% is both alleles 16 CA repeats,wt/mt% one 16 CA and one 18 CA repeat allele, mt/mt% is one 16 CA and one 20 CA
allele. Race distribution of populations; (A) not described, probably predominantly Caucasian, (B) Caucasian-Hispanic-blacks-Asian
(respectively), (C) 75% Caucasian, P13% Hispanic, (D) 97% Caucasian.81

these adducts, such as nucleotide excision repair hMSH2, hMSH3 and hMSH6 genes. The gene prod-
(NER), base excision repair (BER), mismatch repair ucts form complexes that exert their effects during
(MMR) and other post-replicative repair pathways DNA repair and are designated as hMutSa
(Table 2), may affect oxaliplatin efficacy. (hMSH2 + hMSH6) and hMutSb (hMSH2 + hMSH3).
Together, these complexes recruit hMutLa
(hMLHl + hPMS2).
Mismatch repair Decreased sensitivity to cisplatin has been cor-
related with defects in one or more of these com-
Mismatch repair (MMR) is a DNA repair pathway plexes, of which hMLH1 appears to be the most
that corrects base mispairs and small strand loops important.3 After recognition of a damaged DNA
that occur during replication. Loss of MMR function site, replicative bypass is initiated as an attempt
results in an increased spontaneous mutation rate. to resynthesize undamaged DNA. However, once
The MMR system consists of six different proteins, replicative bypass occurs in a normal cell, the
originating from the hMLH1, hMLH2, hPMS2, hMutSa/hMutLa complex will bind to the Pt–DNA
98 D.M. Kweekel et al.

adduct, actually increasing its cytotoxicity. This cer since colon tumour cells are frequently MMR
apparent contradiction is explained by ‘futile cy- deficient as a result of promoter hypermethylation.
cling’ originally proposed by Goldmacher in 1986 About 5–15% of sporadic colorectal tumours is MMR
and supported by recent findings. defective and this defect is frequently associated
Following the futile cycling model, MMR proteins with loss of hMLH1 and hMSH2 expression.24 To
bind to the platinum cross-link during replication date, no polymorphisms in the MMR pathway genes
bypass of the adduct. The MMR proteins select for are known that influence the anti-tumour effects
the parental DNA strand and remove the newly syn- of oxaliplatin.
thesized DNA (Fig. 5). The Pt–DNA adduct is thus
retained and a new attempt by replication bypass
enzymes fails likewise. Replicative bypass
Moreover, binding of MMR proteins shields the
DNA adducts from repair enzymes, prolonging their Post-replication repair is a replicative bypass of
lifetime and retaining DNA damage. The continuous damaged DNA-strands without the introduction of
operations of these ‘futile cycles’ would lead to cell gaps or discontinuities into the newly synthesized
death caused by gaps and strand breaks.28 As a re- strand. DNA repair enzymes involved in replicative
sult of these observations, MLH1 functions as a dam- bypass are able to carry on DNA synthesis despite
age recognition unit like high mobility group protein the presence of adducts like Pt on the leading
(HMG1), which is in concordance with its observed strand. Adducts therefore do not represent an
function in cell cycle regulation and apoptosis.13 absolute block of DNA replication, but the helical
In vitro studies showed that MMR is not involved distortions associated with adducts do affect en-
in oxaliplatin induced DNA-damage repair, whereas zyme action and accuracy. This type of repair is
it serves as an important mechanism in cisplatin essential since the persistent presence of gaps or
and carboplatin adduct repair. The conformational faults can be lethal to cells. Normally, replicative
distortion of the oxaliplatin DNA complex is differ- bypass takes place primarily during cell replication,
ent from the cisplatin and carboplatin adduct and but in cisplatin-resistant cell lines this process ap-
this, together with the less polar properties of pears to be active as well. Some investigators
the DACH–ligand, contributes to a recognition fail- distinguish between replicative bypass and transle-
ure of MMR proteins to detect oxaliplatin adducts. sion synthesis, according to the type of enzyme
Indeed, the Escherichia coli derived MMR enzyme that is responsible for replication with the latter
MutS binds cisplatin adducts with a twofold greater involving only DNA polymerases, and the first
affinity compared to oxaliplatin. This difference is involving any replication enzyme. Replicative by-
even greater after addition of adenosine diphos- pass is either error free or error prone, depending
phate (ADP).29 These observations favour the use on the type of polymerase and replication mecha-
of oxaliplatin over cisplatin in MMR deficient tu- nism involved. DNA repair involving template
mours. Cisplatin adducts are preserved in MMR pro- switching is generally seen as error free replica-
ficient cells by shielding or ‘futile cycling’, whereas tion, whereas direct translesion synthesis is consid-
they are effectively removed in MMR-deficient ered error prone.28 Replicative bypass of DNA
cells. This fundamental difference is particularly lesions may determine tolerance to platinum
of importance for the treatment of colorectal can- adducts.3

Figure 5 The futile cycling model proposed by Goldmacher. A nucleotide adduct in a DNA strand results in helical
distortion, which is recognized by the hMutLa/hMutSa MMR complex (1). The strands are separated (2) and mismatch
repair is attempted (3), but the newly synthesized strand is removed (4A). The continued action of these ‘futile cycles’
results in the persistence of adducts or induction of strand breaks, causing cell death. In MMR-deficient cells, however,
the newly synthesized strand is retained, discarding the damaged nucleotide (4B). In the short term, MMR deficiency
therefore protects against apoptosis caused by DNA-adduct formation.
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 99

DNA polymerases have a number of similar func- sitivity to ionizing radiation.36 In individuals with
tional domains: a palm domain containing three the mutant Arg399 fi Gln codon increased DNA
carboxylate groups that is involved in the nucleo- damage marker levels are found due to inadequate
tide transferase reaction, a finger region and a repair or increased damage tolerance.34 Patients
thumb region which are important for positioning with at least one of the mutant alleles have a more
on the DNA strand. Four polymerase families are than fivefold risk of combined oxaliplatin/5-FU
distinguished, based on sequential differences.28 chemotherapy failure compared to patients with
Functional absence of polymerase g is associated two wild type alleles.37 Other common non-conser-
with reduced ability to repair UV-damaged DNA vative polymorphisms in the coding region of
and a high incidence of tumours.30 DNA polyme- the XRCC1 gene include Arg194 fi Trp and
rases b, c and g are known to discriminate in vitro Arg280 fi His.35 However, these polymorphisms ap-
between Pt adducts with different carrier ligands pear to be non-functional and do not show correla-
such as cisplatin and oxaliplatin. This distinction tion with increased levels of DNA damage in vivo.38
is probably related to difference in MMR and Polymorphisms of other enzymes of the BER path-
HMG1 recognition of both adducts.3 It is still un- way include XRCC3 Thr241Met and a XRCC5 A > G
clear which polymerases are involved in translesion substitution in the 30 UTR (untranslated region);
synthesis in vivo, and if resistance to platinum expression of the latter was shown to be correlated
drugs can be caused by induction of polymerase with ERCC2 expression (next section) in head/neck
activity. However, a HCT-8 human colon tumour cancer patients.39,40
cell line was recently reported to express high lev-
els of polymerase b in concordance with cellular
resistance to oxaliplatin.31 Polymorphisms in poly- Nucleotide excision repair
merase enzymes that influence replicative bypass
near oxaliplatin adducts are yet unidentified. Nucleotide excision repair (NER) is a pathway in-
volved in the recognition and repair of damaged
or inappropriate nucleotides. A wide variety of
Base excision repair DNA-damage is repaired by NER, including UV-
induced photoproducts, helix-distorting monoad-
Single-strand breaks resulting from exposure to ducts, cross-links and endogenous oxidative dam-
endogenously produced active oxygen, ionizing age.41 At least six proteins are essential for
radiation or alkylating agents are repaired by the damage recognition and removal by this repair
base excision repair system (BER). X-ray repair pathway. The first step in this process is recogni-
cross-complementing group 1 enzyme (XRCC1) con- tion of a damaged or inappropriate base by XPA
tains a domain which functions as a protein–pro- (xeroderma pigmentosum complementation group
tein interface that interacts with poly(ADP-ribose) A protein) and RPA (replication protein A). The
polymerase (PARP). PARP is a zinc finger-contain- adhesion of XPA and RPA to a DNA strand attracts
ing enzyme that detects strand breaks subse- other repair factors to the site followed by enzy-
quently removes proteins from the DNA helix, matic unwinding of the helix lesion area by XPD.
which in turn become more accessible for repair The XPD gene, also known as ERCC2 (excision repair
enzymes.32,33 Besides PARP, XRCC1 interacts with cross complementing group 2), encodes an ATP-
DNA ligase III and polymerase b. Although BER dif- dependent helicase that is a component of tran-
fers significantly from nucleotide excision repair scription factor TFIIH. Moreover, XPD has functions
(NER, following section), the XPG protein has been in the basal transcription process and is required
shown to have overlapping functions in both for transcriptional activity by RNA polymerase
systems.34 II.41 After partial unwinding of the helix, the XPA.R-
Shen et al.35 identified three polymorphisms in PA complex recruits endonucleases (XPG and the
the XRCC1 gene. One of these, located in exon 10 XPF.ERCC1 complex) by which a double incision is
of this gene, causes the amino acid change made, respectively, in the 30 and 50 sites flanking
Arg399 fi Gln resulting in an amino acid substitu- the lesion. A portion of 22–32 nucleotides is ex-
tion in the PARP binding domain (amino acids cised from the strand and this is followed by DNA
301–402). The polymorphic enzyme is supposed resynthesis and ligation of the new strand to fill
to be less capable of initiating DNA repair due to al- in the resulting gap.3,42–44
tered binding characteristics. Chinese hamster Several distinct genetic disorders are caused by
ovary (CHO) cells with non-conservative XRCC1 defects in the enzymes of the NER pathway such
polymorphisms in this particular domain show re- as xeroderma pigmentosum (XP), Cockayne syn-
duced repair of single strand breaks and hypersen- drome (CS) and thrichothiodystrophy (TTD). These
100 D.M. Kweekel et al.

syndromes are clinical distinct disorders that are It remains uncertain whether the XPD A751C
characterized by defects in excision repair of UV- polymorphism causes an increase or a decrease in
induced DNA damage. Most of the XP cases are NER activity since, depending on the assay used
caused by ERCC2 (XPD) defects. XP group D pa- to test DNA repair capacity, conflicting results are
tients are extremely sensitive to UV light and have reported on this issue.42 However, the enzyme
a 1000-fold risk to develop cancer.41,45 A number of seems to play an important role in colorectal can-
causative mutations in the ERCC2 gene have been cer since it was found that ERCC2 (and ERCC1)
identified (Fig. 6),46–48 nearly all involving alter- expression in tumour tissue were high compared
ation of seven highly conserved protein elements to adjacent normal mucosa.50 Besides the possibil-
necessary for unwinding of the DNA helix. In the ity of modulating enzyme activity, the amino acid
majority of patients (72%) described by Taylor change Lys > Gln may alter intracellular ERCC2
et al., parts of the gene coding for helicase domain levels by affecting post-transcriptional and/or
V or VI were altered. post-translational stability. However, Dabholkar
Apart from the deleterious mutations in XP pa- et al.51 could not demonstrate a relationship be-
tients, a number of common non-conservative tween ERCC2 protein levels and tumour response
polymorphisms are described in healthy humans to platinum analogues in ovarian cancer patients.
at nucleotide positions 156, 312 and 751 of the Therefore, it seems more likely that the amino acid
ERCC2 gene and having a linked prevalence.42 A sig- change influences the interaction of the ERCC2 pro-
nificant relationship with clinical response to plat- tein with other NER enzymes, or that the A751C
inum-based chemotherapy was found for the polymorphism is linked to nucleotide changes in
Lys751 fi Gln polymorphism only. This SNP causes nearby-located genes, like ERCC1 and XRCC1.42
an amino acid change in exon 23 about 50 base It was shown by the host cell reactivation assay
pairs upstream from the poly(A) signal, apparently that both the A751C and the G312A genotypes are
affecting protein function but not resulting in an associated with less optimal DNA repair capacity.41
alteration of any of the seven helicase domains. However, both the G312A and the conservative
In one study, patients with two mutated alleles C156A single nucleotide polymorphisms were not
were found 6–12 times more likely to have pro- found to be individually associated with response
gression of advanced colorectal cancer compared or survival after 5-FU/oxaliplatin chemotherapy.
to the other genotypes.42 Metastatic colorectal The C156A polymorphism transforms a high usage
cancer patients treated with oxaliplatin/5-FU codon (CGC) into a low usage codon (CGA) and
showed different tumour response for the various might have yet unknown consequences for cellular
genotypes; 24% responders in the Lys/Lys group, processes. The interrelationship between the three
versus 10% in the Lys/Gln and 10% in the Gln/Gln ERCC2 polymorphisms and the net effect on tumour
groups, respectively. Similar findings have recently response or survival is as yet unclear and remains
been reported for lung cancer patients receiving to be determined.42
various platinum agents: Gln/Gln individuals had Although ERCC2 is indisputably indispensable
significantly more risk of death during follow-up for NER function, it was found to be less crucial
(OR 4.5) compared to the other genotypes.49 Con- for the repair of interstrand cross-links compared
versely, a significant association was found for to ERCC1 or ERCC4.52,53 This might be reflected by
the presence of two Gln alleles and low NER repair the fact that a number of patients with ERCC2 de-
capacity for benzo(a)pyrene adducts in a cohort of fects are known, whereas to date no case reports
lung cancer patients.41 have been published identifying live human

Figure 6 Mutations of the XPD gene found in XP patients. Most mutations involve alteration of helicase domains I–VI.
Upper part: mutations described by Taylor et al. (46); lower part: described by Kobayashi (47) and Broughton (48).
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 101

subjects suffering from an ERCC1 disorder. Most that the Asn118 fi Asn (C fi T), however silent,
information considering the importance of the may be related to outcome and survival by chang-
ERCC1 enzyme is therefore based on observations ing ERCC1 expression.
in cell lines equipped with artificial NER defects Since ERCC1 combines with ERCC4 (XPF) to form
and mouse knockout models. For example, in hu- the endonuclease complex XPF-ERCC1, it may not
man ovarian cancer cells, the C200 variant was be surprising that cells lacking ERCC4 show similar
found to be cisplatin resistant due to enhanced sensitivity to the cross-linking agent mitomycin
NER activity and ERCC1 expression.54 Further- compared to ERCC1 deficient cells.53 ERCC4 upreg-
more, cisplatin cross-link removal by UV-20 CHO ulation in human KCP-4 carcinoma cells was associ-
cells lacking ERCC1 was small compared to its ated with cisplatin and oxaliplatin resistance.64 In
wild-type (AA8) counterpart.55 Since the overall XP group F patients numerous non-synonymous
rate and efficacy of the NER process is compara- polymorphisms can be found, predominantly in
ble for cisplatin and oxaliplatin3, the lack of exons 8 and 11.65 Polymorphisms in exon 11 are
ERCC1 is considered to influence oxaliplatin sensi- known to interfere with ERCC1 complex formation,
tivity as well. Indeed, decreased sensitivity to which may result in rapid degradation of ERCC1.66
oxaliplatin was associated with high ERCC1 The ERCC4 Arg415Gln substitution, located near
expression levels in HCT116 colorectal cancer some XPF disease related mutations in exon 8,
cells.56 Similar observations were made by Ar- was found to increase breast cancer risk.67 This
nould et al. in a panel of colon cancer cell lines; mutation is found in a highly conserved gene region
ERCCl and XPA expression were found to be pre- and protein function may therefore be affected as
dictive of oxaliplatin sensitivity.31 ERCC1 null predicted by the sorting intolerant from tolerant
mice are extremely small at birth and die within substitutions (SIFT) database tool.68,69
3 weeks of age. Death is related to liver failure,
most likely by elevated levels of oxidative DNA
damage, but the brain and kidneys are affected
as well.57,58 Other polymorphisms
The ERCC1 enzyme consists of 297 amino acids
and appears to harbour a nuclear localization signal Glutathione-S-transferase
and a domain characteristic of a DNA-binding pro-
tein. In normal individuals, a common polymor- Polymorphisms associated with cellular platinum
phism at exon 4 can be found at codon 118, drug clearance affect oxaliplatin efficacy by lower-
changing AAC into the less used AAT codon. This ing intracellular concentrations of the drug. This
is a conservative nucleotide change, as both codons so-called cellular detoxification is the result of
result in the same amino acid (Asn) being incorpo- conjugation of Pt drug to biomolecules such as
rated into the protein. However silent, this poly- methionine, cysteine and glutathione (Fig. 2).
morphism was found to be associated with The latter conjugation is catalyzed by the glutathi-
decreased ERCC1 expression in ovarian cancer one-S-transferase enzyme (GST). The conjugation
cells59 but, contrasting, with increased intratumor of toxic and carcinogenic electrophilic molecules
levels of ERCC1 in 32 patients with metastatic colo- with glutathione by GST is followed by cellular
rectal cancer treated with 5-FU/oxaliplatin chemo- excretion of the conjugate thereby protecting
therapy.60 Shirota et al. reported a significant DNA and other macromolecules from damage. Sev-
association between intratumoural ERCC1 mRNA eral subclasses of the GST enzyme are distin-
levels and survival after oxaliplatin based chemo- guished: the a, p, l, h, and f subclasses (alpha,
therapy in 5-FU resistant colorectal tumors.61 Pa- pi, mu, theta and zeta).
tients with intratumoural mRNA P 4.9 · 103 A single nucleotide polymorphism (SNP) in exon
(relative to the expression of the house-keeping 5 at position 313 (A fi G) in the GSTP1 (p) gene re-
gene b-actin) had a median survival of 1.9 months, sults in a valine being incorporated into this en-
compared to 10.2 months for patients with mRNA zyme at site 105 instead of the usual isoleucine
expression below 4.9 · 103. When 7.4 · 103 was (Ile105 fi Val). The mutant GSTP1 (p) enzyme is
chosen as the expression cutoff value for non- less potent in detoxification of carcinogens70 and
response, median survival for the high expression individuals with two mutant alleles have shown a
group was 150 days compared to 292 days in the significant survival benefit from combined oxalipla-
low expression group.62 An inverse association be- tin/5-FU treatment with a median survival of 24.9
tween ERCC1 expression and survival of gastric months compared to only 7.9 months for
cancer patients treated with platinum agents was metastatic colorectal cancer patients with two
found in another study.63 These findings indicate wild-type alleles.71 A similar result was found in a
102 D.M. Kweekel et al.

retrospective study among 240 breast cancer pa- has been demonstrated that tumour EGFR geno-
tients treated with radiotherapy or various alkylat- type is non-prognostic of survival in head/neck
ing agents (inducing DNA adducts). Individuals with cancer patients, whereas germline genotype could
the 105Val allele had superior survival (hazard ratios predict survival adequately. In 33% of cases, tu-
for death were 0.8 and 0.3 for Ile/Val and Val/Val mour DNA had a smaller number of EGFR dinucle-
genotypes, respectively) compared to the 105Ile al- otide repeats compared to normal tissue, probably
lele. Apparently, the 105Val allele is an independent due to microsatellite instability. This favours the
prognostic factor in breast cancer because the ob- use of peripheral blood specimens for genotyping
served survival benefit was shown to be irrespec- EGFR.75
tive of the treatment used.72 Other common The exact mechanism explaining how EGFR lev-
polymorphisms in the GSTT1 (h) and GSTM1 (l) els relate to oxaliplatin/fluorouracil efficacy is
genes include deletions that result in complete loss not known. Probably the EGFR genotype is related
of enzyme activity in homozygous individuals. How- to basic mechanisms for cell growth or cell death
ever, no association with altered survival or clinical making it prognostic for the response to chemo-
response in patients with advanced colorectal can- therapy and patient outcome in general. This cor-
cer treated with oxaliplatin/5-FU was observed for relation was shown for a number of malignancies
the GSTT1 and GSTM1 genotypes.71,73 It seems including tumours of bladder, breast and lungs.76
therefore likely that the h and l subclass enzymes When oxaliplatin is combined with the tyrosine ki-
play a less important role in colorectal tissue (can- nase inhibitor ZD1839 (Iressa
), a synergistic effect
cer) cells as compared to the p subclass. Recent is observed in colon cancer cell lines. Platinum ad-
findings confirm that p subclass enzymes are over duct removal is reduced, with 90% of oxaliplatin in-
expressed in colorectal cancer tissues relative to duced Pt–DNA adducts remaining unrepaired, and
normal mucosa.50 apoptosis is prolonged compared to oxaliplatin
administration alone.77 Combination of oxaliplatin
with EGFR antibodies or tyrosine kinase inhibitors
Epidermal growth factor receptor seems a promising new direction in colorectal can-
cer therapy.
The epidermal growth factor receptor (EGFR) is a
170 kD transmembrane protein that is composed
of an extracellular ligand binding domain, a trans-
membrane lipophilic region and an intracellular Discussion
protein tyrosine kinase site. The binding of a li-
gand results in the dimerization (followed by Pharmacogenetics sheds new light on the classical
internalization) of EGFR molecules or heterodi- pharmacological question and understanding why
merization with other closely related receptors individuals respond differently to various drug
such as HER2/neu. This in turn activates the tyro- treatments. Current anti-cancer drug treatments
sine kinase domain via autophosphorylation, are effective in only a minority of patients and it
resulting in signal transduction and regulating cel- is not yet possible to reliably predict which patient
lular growth. Endogenous ligands to this receptor benefits from a certain chemotherapeutic drug
include epidermal growth factor and transforming treatment or which patient will experience (life-
growth factor a (TGFa). The EGFR is essential for threatening) drug toxicity. Pharmacogenetics is
normal cellular function and increased levels of aiming at individuating drug therapy thereby
EGFR mRNA are associated with metastasis and increasing the potential efficacy of novel antican-
more aggressive tumour growth.74 EGFR, alterna- cer drugs such as oxaliplatin.
tively called HER1, is over expressed in many Several polymorphisms influencing drug sensiti-
types of cancers, including colorectal cancer. vity, e.g., affecting repair of Pt–DNA adducts or
The EGFR gene transcription and enzyme expres- cellular detoxicification, are described but not
sion are presumably inversely related to the num- yet routinely applied in the pharmacotherapeutic
ber of CA repeats in intron 1. It was shown that an treatment of colorectal cancer. One reason for
individual with two alleles with 16 repeats (16/16) this is that it is still unclear how the different
has worse response to second-line oxaliplatin/flu- polymorphisms interrelate. Moreover, patients
orouracil treatment compared to the 16/18 and are frequently treated by multiple drugs at one
16/20 genotypes. Median survival rates for pa- time, and the efficacy of these drugs will be
tients with metastasized colorectal cancer were influenced by several concurrent interindividual
63, 133 and 805 days for the 16/16, 16/18 and genetic differences. Since germline SNPs are
16/20 genotypes, respectively.62 Conversely, it polymorphisms that originate from the ovary and
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 103

sperm parental DNA, these SNPs are theoretically peutic treatment, thereby ‘tailoring’ anti-neoplas-
found both in normal and in tumour tissue. In con- tic drug therapy in a rational manner.
trast, tumour tissue may have additional non-
germline mutations that initially caused neoplastic
growth, for example by overexpressing oncogenic
proteins. However, analysis on gene mRNA expres- References
sion profiles has shown that only 0.2–10% of all
1. Holland Frei, editor. Cancer Medicine, vol. 1. Hamilton
gene transcripts are differentially expressed be-
(Ont.): BC Becker Inc; 2003.
tween colorectal tumours and non-malignant tis- 2. van Glabbeke M, Renard J, Pinedo HM, Cavalli F, Vermorken
sue. Most of the transcripts with abnormal high C, Sessa C, et al. Iproplatin and carboplatin induced
expression are related to malignancy, and many toxicities: overview of phase II clinical trial conducted by
others have comparable expression to normal the EORTC Early Clinical Trials Cooperative Group (ECTG).
Eur J Cancer Clin Oncol 1988;24:255–62.
cells.78 Germline SNPs therefore give a good rep-
3. Di Francesco AM, Ruggiero A, Riccardi R. Cellular and
resentation of the main bulk of tumour DNA, apart molecular aspects of drugs of the future: oxaliplatin. Cell
from oncogenic SNPs. For this reason, it is rational Mol Life Sci 2002;59:1914–27.
and preferable to develop pharmacogenetic tests 4. Grolleau F, Gamelin L, Boisdron-Celle M, Lapied B, Pelhate
based on germline SNPs in, e.g., lymphocytes, E, Gamelin E. A possible explanation for a neurotoxic effect
of the anticancer agent oxaliplatin on neuronal voltage-
which represent a more accessible tissue com-
gated sodium channels. J Neurophysiol 2001;85:2293–7.
pared to tumour material. 5. Gamelin L, Boisdron-Celle M, Delva R, Guerin-Meyer V, Ifrah
This strategy seems to be very promising as sug- A, Morel A, et al. Prevention of oxaliplatin-related neuro-
gested by recently published studies.79 Stoehl- toxicity by calcium and magnesium infusions: a retrospec-
macher et al. studied survival and time to tive study of 161 patients receiving oxaliplatin combined
with 5-Fluorouracil and leucovorin for advanced colorectal
progression in a group of refractory colorectal can-
cancer. Clin Cancer Res 2004;10:4055–61.
cer patients treated with 5-FU/oxaliplatin combi- 6. McKeage MJ, Hsu T, Screnci D, Haddad G, Baguley BC.
nation therapy. Of the 10 polymorphisms studied Nucleolar damage correlates with neurotoxicity induced by
in peripheral blood samples (GSTP1, GSTM1, GSTT1, different platinum drugs. Br J Cancer 2001;85:1219–25.
XRCC1, TS 28 base pair repeat and 6 base pair dele- 7. Levi F, Metzger G, Massari C, Milano G, Levi F, Mishima M,
et al. Oxaliplatin: pharmacokinetics and chronopharmaco-
tion, ERCC2 Lys751Gln and Arg156Arg, XPA and
logical aspects. Clin Pharmacokinet 2000;Jan(5):1–21PMID.
ERCC1 Asn118Asn), only ERCC2 Lys751Gln, GSTP1 8. Kobayashi M, Wood PA, Hrushesky WJ. Circadian chemo-
Ile105Val, ERCC1 Asn118Asn and TS-30 UTR (a 6 base therapy for gynecological and genitourinary cancers. Chro-
pair deletion likely associated with 5-FU responsive- nobiol Int 2002;19:237–51.
ness) were associated with overall survival. How- 9. Boughattas NA, Levi F, Fournier C, Hecquet B, Lemaigre G,
Roulon A, et al. Stable circadian mechanisms of toxicity of
ever, combining these polymorphisms increased
two platinum analogs (cisplatin and carboplatin) despite
predictability of survival; patients homozygote for repeated dosages in mice. J Pharmacol Exp Ther 1990;
two or more favourable SNPs had a median survival 255:672–9.
time of 17.4 months, compared to 10.2 months for 10. Rothenberg ML, Oza AM, Bigelow RH, Berlin JD, Marshall JL,
one or 5.4 months for none of the favourable allele Ramanathan RK, et al. Superiority of oxaliplatin and fluo-
rouracil–leucovorin compared with either therapy alone in
combinations (p < 0.001).73 In this study no associa-
patients with progressive colorectal cancer after irinotecan
tion was found between anti-tumour response and and fluorouracil–leucovorin: interim results of a phase III
genotype, probably due to the low response rate trial. J Clin Oncol 2003;21:2059–69.
of <10% observed in the heavily pretreated pa- 11. de Gramont A, Figer A, Seymour M, Homerin M, Hmissi A,
tients. Different results were reported by McLeod Cassidy J, et al. Leucovorin and fluorouracil with or without
oxaliplatin as first-line treatment in advanced colorectal
and colleagues, who studied the association be-
cancer. J Clin Oncol 2000;18:2938–47.
tween 28 germline SNPs and the response to first- 12. Andre T, Boni C, Mounedji-Boudiaf L, Navarro M, Tabernero
line treatment (5-FU/irinotecan, 5-FU/oxaliplatin T, Hickish T, et al. Oxaliplatin, fluorouracil, and leucovorin
and oxaliplatin/irinotecan) of advanced colorectal as adjuvant treatment for colon cancer. N Engl J Med
cancer. Preliminary results of this prospective study 2004;350:2343–51.
13. Goldberg RM, Sargent DJ, Morton RF, Fuchs CS, Ramanathan
indicate that GSTP1, Ile105Val and ERCC2
SK, Williamson SK, et al. A randomized controlled trial of
Lys751Gln polymorphisms are not related to oxali- fluorouracil plus leucovorin, irinotecan, and oxaliplatin
platin responsiveness or toxicity.80 In conclusion, combinations in patients with previously untreated meta-
(combined) SNP-analysis may help us predict static colorectal cancer. J Clin Oncol 2004;22:23–30.
responsiveness and survival benefit in colorectal 14. van Warmerdam LJ. Tailor-made chemotherapy for cancer
patients. Neth J Med 1997;51:30–5.
cancer patients treated with oxaliplatin containing
15. Van Gelder T, Hesselink DA, Van Hest RM, Mathot RA, Van
regimens. Hopefully, this type of genetic informa- Schaik R. Pharmacogenetics in immunosuppressive ther-
tion may help us to select individual patients that apy: the best thing since TDM?. Ther Drug Monit
have a high potential for benefit from chemothera- 2004;26:343–6.
104 D.M. Kweekel et al.

16. McLeod HL, Yu J. Cancer pharmacogenomics: SNPs, chips, 37. Stoehlmacher J, Ghaderi V, Iobal S, Groshen S, Tsao-Wei D,
and the individual patient. Cancer Invest 2003;21:630–40. Park D, et al. A polymorphism of the XRCC1 gene predicts
17. Desoize B, Madoulet C, Graham MA, Lockwood GF, Greens- for response to platinum based treatment in advanced
lade D, Brienza S, et al. Particular aspects of platinum colorectal cancer. Anticancer Res 2001;21:3075–9.
compounds used at present in cancer treatment. Crit Rev 38. Lunn RM, Langlois RG, Hsieh LL, Thompson CL, Bell DA.
Oncol Hematol 2002;1077:317–325PMID. XRCC1 polymorphisms: effects on aflatoxiri Bl-DNA adducts
18. Culy CR, Clemett D, Wiseman LR. Oxaliplatin. A review of its and glycophorin A variant frequency. Cancer Res
pharmacological properties and clinical efficacy in meta- 1999;59:2557–61.
static colorectal cancer and its potential in other malignan- 39. Charles J, Monzo M, Artells R, Amat R, Foro P, Fuster D, et
cies. Drugs 2000;60:895–924. al. XPC, XPD/ERCC2, UMPS, Cyp1b1 and XRCC5 single
19. Sanofi Synthelabo the Netherlands. Prod Info Eloxatin; 2004. nucleotide polymorphisms in head and neck cancer. In:
20. Liu J, Kraut E, Bender J, Brooks R, Balcerzak S, Grever M, Asco proceedings 2004 general poster session, 5588; 2004.
et al. Pharmacokinetics of oxaliplatin (NSC 266046) alone 40. Sarries C, Alberola V, De Las Penas R, Camps C, Massuti B,
and in combination with paclitaxel in cancer patients. Garcia-Gomez R, et al. Combined DNA repair gene single
Cancer Chemother Pharmacol 2002;49:367–74. nucleotide polymorphisms in gemcitabine/cisplatin-treated
21. Graham MA, Lockwood GF, Greenslade D, Brienza S, Bayssas non-small-cell lung cancer patients. In: Asco proceedings
E, Gamelin E, et al. Clinical pharmacokinetics of oxalipla- 2004 poster discussion, 7031; 2004.
tin: a critical review. Clin Cancer Res 2000;1066: 41. Spitz MR, Wu X, Wang Y, Wang LE, Shete S, Amos CI, et al.
205–218PMID. Modulation of nucleotide excision repair capacity by XPD
22. Calvert AH, Egorin MJ. Carboplatin dosing formulae: gender polymorphisms in lung cancer patients. Cancer Res
bias and the use of creatinine-based methodologies. Eur J 2001;61:1354–7.
Cancer 2002;38:11–6. 42. Park DJ, Stoehlmacher J, Zhang W, Tsao-Wei DD, Groshen S,
23. Mello JA, Lippard SJ, Essigmann JM. DNA adducts of cis- Lenz HJ. A xeroderma pigmentosum group D gene polymor-
diamminedichloroplatinum(II) and its trans isomer inhibit phism predicts clinical outcome to platinum-based chemo-
RNA polymerase II differentially in vivo. Biochemistry therapy in patients with advanced colorectal cancer. Cancer
1995;34:14783–91. Res 2001(Dec. 15):8654–8.
24. Raymond E, Faivre S, Chaney S, Woynarowski J, Cvitkovic E. 43. Bessho T, Mu D, Sancar A. Initiation of DNA interstrand cross-
Cellular and molecular pharmacology of oxaliplatin. Mol link repair in humans: the nucleotide excision repair system
Cancer Ther 2002:227–35. makes dual incisions 50 to the cross-linked base and removes
25. Guchelaar HJ, Vermes A, Vermes I, Haanen C. Apoptosis: a 22- to 28-nucleotide-long damage-free strand. Mol Cell
molecular mechanisms and implications for cancer chemo- Biol 1997;17:6822–30.
therapy. Pharm World Sci 1997;19:119–25. 44. Bessho T, Sancar A, Thompson LH, Thelen MP. Reconstitu-
26. Mishima M, Samimi G, Kondo A, Lin X, Howell SB, Hector S, tion of human excision nuclease with recombinant XPF–
et al. The cellular pharmacology of oxaliplatin resistance. ERCC1 complex. J Biol Chem 1997;272:3833–7.
Eur J Cancer 2002;1176:1405–1412PMID. 45. Reardon JT, Vaisman A, Chaney SG, Sancar A. Efficient
27. Akiyama S, Chen ZS, Sumizawa T, Furukawa T. Resistance to nucleotide excision repair of cisplatin, oxaliplatin, and bis-
cisplatin. Anticancer Drug Des 1999;14:143–51. aceto-ammine-dichloro-cyclohexylamine-platinum(IV) (IV)
28. Chaney SG, Vaisman A. Specificity of platinum–DNA adduct (JM216) platinum intrastrand DNA diadducts. Cancer Res
repair. J Inorg Biochem 1999;77:71–81. 1999(Aug. 15):3968–71.
29. Zdraveski ZZ, Mello JA, Farinelli CK, Essigmann JM, Marinus 46. Taylor EM, Broughton BC, Botta E, Stefanini M, Sarasin A,
MG. MutS preferentially recognizes cisplatin- over oxalipla- Jaspers NG, et al. Xeroderma pigmentosum and trichothi-
tin-modified DNA. J Biol Chem 2002;277:1255–60. odystrophy are associated with different mutations in the
30. Broughton BC, Cordonnier A, Kleijer WJ, Jaspers NG, XPD (ERCC2) repair/transcription gene. Proc Natl Acad Sci
Fawcett H, Raams A, et al. Molecular analysis of mutations USA 1997;94:8658–63.
in DNA polymerase eta in xeroderma pigmentosum-variant 47. Kobayashi T, Kuraoka I, Saijo M, Nakatsu Y, Tanaka A,
patients. Proc Natl Acad Sci USA 2002;99:815–20. Someda Y, et al. Mutations in the XPD gene leading to
31. Arnould S, Hennebelle I, Canal P, Bugat R, Guichard S. xeroderma pigmentosum symptoms. Hum Mutat 1997;9:
Cellular determinants of oxaliplatin sensitivity in colon 322–31.
cancer cell lines. Eur J Cancer 2003:112–9. 48. Broughton BC, Thompson AF, Harcourt SA, Vermeulen W,
32. Hartwig A, Blessing H, Schwerdtle T, Walter I. Modulation of Hoeijmakers JH, Botta E, et al. Molecular and cellular
DNA repair processes by arsenic and selenium compounds. analysis of the DNA repair defect in a patient in xeroderma
Toxicology 2003;193:161–9. pigmentosum complementation group D who has the clinical
33. Shall S, de Murcia G. Poly(ADP-ribose) polymerase-1: what features of xeroderma pigmentosum and Cockayne syn-
have we learned from the deficient mouse model?. Mutat drome. Am J Hum Genet 1995;56:167–74.
Res 2000;460:1–15. 49. Ricardi U, Regazzoni S, Ghio P, Giachino D, De Marchi M,
34. Duell EJ, Wiencke JK, Cheng TJ, Varkonyi A, Zuo ZF, Ashok Novello S, et al. The K751Q polymorphism of XPD gene in
TD, et al. Polymorphisms in the DNA repair genes XRCC1 and lung cancer patients receiving platinum-based chemother-
ERCC2 and biomarkers of DNA damage in human blood apy. In: Asco proceedings 2004 integrated education session,
mononuclear cells. Carcinogenesis 2000:965–71. 2500; 2004.
35. Shen MR, Jones IM, Mohrenweiser H. Nonconservative amino 50. Glasgow SC, Yu J, Lineham DC, Watson MA, McLeod HL.
acid substitution variants exist at polymorphic frequency in Modeling the response of colorectal cancer to common
DNA repair genes in healthy humans. Cancer Res 1998;1: chemotherapy using multivariate gene expression. In: Asco
604–8. proceedings 2004 oral presentation, 2008; 2004.
36. Shen MR, Zdzienicka MZ, Mohrenweiser H, Thompson LH, 51. Dabholkar M, Bostick-Bruton F, Weber C, Bohr VA, Egwuagu
Thelen MP. Mutations in hamster single-strand break repair E, Reed E. ERCC1 and ERCC2 expression in malignant tissues
gene XRCC1 causing defective DNA repair. Nucleic Acids Res from ovarian cancer patients. J Natl Cancer Inst 1992;84:
1998;26:1032–7. 1512–7.
Individualization of Oxaliplatin pharmacotherapy in colorectal cancer 105

52. Murray D, Rosenberg E. The importance of the ERCC1/ breast cancer risk. Cancer Epidemiol. Biomarkers Prev
ERCC4[XPF] complex for hypoxic-cell radioresistance does 2003;12:1200–4.
not appear to derive from its participation in the nucleotide 68. Ng PC, Henikoff S. SIFT: predicting amino acid changes that
excision repair pathway. Mutat Res 1996;364:217–26. affect protein function. Nucleic Acids Res 2003;31:3812–4.
53. Andersson BS, Sadeghi T, Siciliano MJ, Legerski R, Murray D. 69. Ng PC, Henikoff S. Accounting for human polymorphisms
Nucleotide excision repair genes as determinants of cellular predicted to affect protein function. Genome Res
sensitivity to cyclophospharnide analogs. Cancer Chemother 2002;12:436–46.
Pharmacol 1996;38:406–16. 70. Watson MA, Stewart RK, Smith GB, Massey TE, Bell DA.
54. Ferry KV, Hamilton TC, Johnson SW. Increased nucleotide Human glutathione S-transferase polymorphisms: relation-
excision repair in cisplatin-resistant ovarian cancer cells: ship to lung tissue enzyme activity and population frequency
role of ERCC1-XPF. Biochem Pharmacol 2000;60:1305–13. distribution. Carcinogenesis 1998;19:275–80.
55. Meyn RE, Jenkins SF, Thompson LH. Defective removal of 71. Stoehlmacher J, Park DJ, Zhang W, Groshen S, Tsao-Wei DD,
DNA cross-links in a repair-deficient mutant of Chinese Yu MC, et al. Association between glutathione S-transferase
hamster cells. Cancer Res 1982;42:3106–10. P1, T1, and M1 genetic polymorphism and survival of
56. Boyer J, McLean EG, Aroori S, Wilson P, McCulla A, Carey PD, patients with metastatic colorectal cancer. J Natl Cancer
et al. Characterization of p53 wild-type and null isogenic Inst 2002;94:936–42.
colorectal cancer cell lines resistant to 5-fluorouracil, 72. Sweeney C, McClure GY, Fares MY, Stone A, Coles BF,
oxaliplatin, and irinotecan. Clin Cancer Res 2004;10: Thompson PA, et al. Association between survival after
2158–67. treatment for breast cancer and glutathione S-transferase
57. Chipchase MD, O’Neill M, Melton DW. Characterization of P1 Ile105Val polymorphism. Cancer Res 2000;60:5621–4.
premature liver polyploidy in DNA repair (Ercc1)-deficient 73. Stoehlmacher J, Park DJ, Zhang W, Yang D, Groshen S,
mice. Hepatology 2003;38:958–66. Zahedy S, et al. A multivariate analysis of genomic poly-
58. Me Whir J, Selfridge J, Harrison DJ, Squires S, Melton DW. morphisms: prediction of clinical outcome to 5-FU/oxalipl-
Mice with DNA repair gene (ERCC-1) deficiency have atin combination chemotherapy in refractory colorectal
elevated levels of p53, liver nuclear abnormalities and die cancer. Br J Cancer 2004;91:344–54.
before weaning. Nat Genet 1993;5:217–24. 74. Kirn ES, Khuri FR, Herbst RS. Epidermal growth factor
59. Yu JJ, Lee KB, Mu C, Li Q, Abernathy TV, Bostick-Bruton F, receptor biology (IMC-C225). Curr Opin Oncol 2001;13:
et al. Comparison of two human ovarian carcinoma cell 506–13.
lines (A2780/CP70 and MCAS) that are equally resistant to 75. Magné N, Pereira S, Etienne MC, Formento JL, Francoual M,
platinum, but differ at codon 118 of the ERCC1 gene. Int J Demard F, et al. Analysis of the dinucleotide repeat
Oncol 2000;16:555–60. polymorphism in the epidermal growth factor receptor
60. Park D, Stoehlmacher J, Lenz HJ. ERCC1 polymorphism (EGFR) in tumor and normal tissue of head and neck cancer
associated with ERCC1 mRNA expression. Proc Am Assoc patients. In: Asco proceedings 2004 poster session, 5518;
Cancer Res 2002;43:1591 [abstr.]. 2004.
61. Shirota Y, Stoehlmacher J, Brabender J, Xiong YP, Uetake H, 76. Mendelsohn J, Baselga J. Status of epidermal growth factor
Danenberg KD, et al. ERCC1 and thymidylate synthase receptor antagonists in the biology and treatment of cancer.
mRNA levels predict survival for colorectal cancer patients J Clin Oncol 2003;21:2787–99.
receiving combination oxaliplatin and fluorouracil chemo- 77. Xu JM, Azzariti A, Severino M, Lu B, Colucci G, Paradiso A.
therapy. J Clin Oncol 2001;19:4298–304. Characterization of sequence-dependent synergy between
62. Lenz HJ. Pharmacogenomics in colorectal cancer. Semin ZD1839 (‘‘Iressa’’) and oxaliplatin. Biochem Pharmacol
Oncol 2003;30:47–53. 2003;66:551–63.
63. Metzger R, Leichman CG, Danenberg KD, Danenberg PV, 78. Notterman DA, Alon U, Sierk AJ, Levine AJ. Transcriptional
Lenz HJ, Hayashi K, et al. ERCC1 mRNA levels complement gene expression profiles of colorectal adenoma, adenocar-
thymidylate synthase mRNA levels in predicting response cinoma, and normal tissue examined by oligonucleotide
and survival for gastric cancer patients receiving combina- arrays. Cancer Res 2001;61:3124–30.
tion cisplatin and fluorouracil chemotherapy. J Clin Oncol 79. Gurubhagavatula S, Zhou W, Park S, Neuberg DS, Lynch TJ,
1998;16:309–16. Wain JC, et al. Nucelotide excisionrepair gene haplotypes
64. Mukai M, Kanzaki A, Chen ZS, Miyashita H, Sumizawa T, and advanced non-small cell lung cancer prognosis. In: Asco
Furukawa T, et al. Enhanced nucleotide excision repair in proceedings 2004, 3040; 2004.
cisplatin resistant human KB carcinoma cells. Oncol Rep 80. McLeod HL, Sargent DJ, Marsh S, Fuchs CS, Ramanathan
2002;9:839–44. SK, Williamson SK, et al. Pharmacogenetic analysis of
65. Fan F, Liu C, Tavare S, Arnheim N. Polymorphisms in the systemic toxicity and response after 5-fluorouracil (5FU)/
human DNA repair gene XPF. Mutat Res 1999;406:115–20. CPT-11, 5FU/oxaliplatin (oxal), or CPT-11/oxal therapy for
66. Matsumura Y, Nishigori C, Yagi T, Imamura S, Takebe H. advanced colorectal cancer (CRC): Results from an inter-
Characterization of molecular defects in xeroderma pig- group trial. Proc Am Soc Clin Oncol 2003;22:253 [abstr.
mentosum group F in relation to its clinically mild symp- 1013].
toms. Hum Mol Genet 1998;7:969–74. 81. Mathijssen RH, Marsh S, Karlsson MO, Xie R, Baker SD,
67. Smith TR, Levine EA, Perrier ND, Miller MS, Freimanis RI, Verweij J, et al. Irinotecan pathway genotype analysis to
Lohman K, et al. DNA-repair genetic polymorphisms and predict pharmacokinetics. Clin Cancer Res 2003;9:3246–53.