extensive hepatic replacement due to liver metastases has ... · extensive hepatic replacement due...

Extensive hepatic replacement due to liver metastases has no effect on 5-fluorouracil pharmacokinetics

Jan Gerard Maring1, Henk Piersma2, Albert van Dalen3, Harry J.M. Groen4, Donald R.A.

Uges5, Elisabeth G.E. De Vries6

1Department of Pharmacy, Diaconessen Hospital Meppel and Bethesda Hospital

Hoogeveen; 2Department of Internal Medicine, Martini Hospital Groningen; 3Department of Radiology, Diaconessen Hospital Meppel; Departments of 4Pulmonary

Diseases, 5Pharmacy and 6Medical Oncology, University Hospital Groningen, The Nether-

lands

Cancer Chemother Pharmacol 2003;51:167-173

Chapter 4.1

70

Influence of liver metastases on 5-fluorouracil pharmacokinetics

71

Erosion at work. Zabriskie point. Death Valley USA 1995.

Chapter 4.1

70


71

Abstract

Aim The influence of liver metastases on the pharmacokinetics of 5-fluorouracil (5-FU)

and its metabolite 5,6-dihydrofluorouracil (DHFU) was studied in patients with liver me-

tastases from gastrointestinal cancer and compared with a control group of patients with

non-metastatic gastrointestinal cancer.

Methods Patients were assigned to two different groups based on the presence of liver

metastases. The percentage of hepatic replacement was determined with CT and ultra-

sonography and classified as <25 %, 25-50 or > 50% from the total liver volume. Chemo-

therapy consisted of leucovorin 20 mg/m2/day plus 5-FU 425 mg/m2/day, both during 5

days. Blood sampling was carried out on the first day of the first chemotherapy cycle. 5-FU

and DHFU were quantified by HPLC in plasma. A four compartment parent drug - metabo-

lite model with non-linear Michaelis-Menten elimination from the central compartment

of the parent drug (5-FU) was applied to describe 5-FU and DHFU pharmacokinetics.

Results No effect of liver metastases on 5-FU clearance was observed between the two

groups. The effect of 18 covariables on pharmacokinetic parameters was also studied in

univariate correlation analyses. Body surface area was positively correlated with the dis-

tribution volume of 5-FU in the central compartment and with Vmax

(r = 0.65 and r = 0.54

respectively).

Conclusions There is no need for dose adjustment of 5-FU as standard procedure in

patients with liver metastases and mild to moderate elevations in liver function tests.

Introduction

Fluorouracil (5-FU) is widely used in chemotherapeutic regimens for the treatment

of breast-, colorectal- and head and neck cancer. The cytotoxic mechanism of 5-FU is

complex, requiring intracellular bioconversion of 5-FU into cytotoxic nucleotides. Inhibi-

tion of thymidylate synthase by the metabolite 5-fluoro-2’-deoxyuridine-5’-monophos-

phate (FdUMP) is thought to be the main mechanism of cytotoxicity [1]. The cytotoxicity

is caused by only a small part of the administered 5-FU dose, as the majority of 5-FU is

rapidly metabolised into inactive metabolites (see figure 1). The initial and rate-limiting

enzyme in the catabolism of 5-FU is dihydropyrimidine dehydrogenase (DPD), catalysing

a reduction of 5-FU into 5,6-dihydrofluorouracil (DHFU). Subsequently DHFU is degraded

into fluoro-β-ureidopropionic acid (FUPA) and fluoro-β-alanine (FBAL) [2]. Several groups

have suggested a major role of DPD in the regulation of 5-FU metabolism and thus in the

amount of 5-FU available for cytotoxicity [3-7]. DPD is present in many tissues, but the

highest activity is found in the liver and, since liver blood flow is relatively high, this organ

is considered as major site for 5-FU degradation [2,8]. In the last decades the role of liver

metastases and liver function impairment on 5-FU pharmacokinetics has been subject to

Chapter 4.1

72


73

debate.

The effects of liver dysfunction on the pharmacokinetics of drugs are generally difficult to

predict. Liver dysfunction is usually diagnosed on the basis of liver function tests, but for

most drugs the correlation between test values and drug metabolism is only weak. Liver

dysfunction due to liver metastases is an even more complicated issue. Liver metastases

displace healthy liver tissue and this may directly affect the liver metabolic capacity due

to tissue loss, but also indirectly due to tissue damage caused by cholestasis as a result of

compression of surrounding structures including bileducts. Furthermore, liver dysfunc-

tion may affect plasma volume, serum albumin, drug protein binding, and hepatic blood

flow, all of which can influence pharmacokinetic parameters in complex ways.

In the last decade, several groups have studied the effects of liver metastases on 5-FU

clearance, in particular during continuous infusion. In an early small study, a reduction of

clearance of continuously infused 5-FU was reported in four patients with gastrointesti-

nal carcinoma and hepatic metastases [9]. Contrary to this, others found no correlation

between drug clearance and hepatic function tests in a larger study with 187 patients

��

��

��

��

��

��

��

Figure 1 Metabolism of 5-FU. 5-Fluoro-2’-deoxyuridine-5’-monophosphate (FdUMP) is the cytotoxic product resulting from a multi-step 5-FU activation route . FdUMP inhibits the enzyme thymidylate synthase (TS), which leads to intracellular accumulation of deoxy-uridine-monophospate (dUMP) and depletion of deoxy-thymidine-monophospate (dTMP). This causes arrest of DNA synthesis. The initial and rate-limiting enzyme in the catabolism of 5-FU is dihydropyrimidine dehydrogenase (DPD), catalys-ing the reduction of 5-FU into 5,6-dihydrofluorouracil (DHFU). Subsequently, DHFU is degraded into fluoro-β-ureidopropionic acid (FUPA) and fluoro-β-alanine (FBAL).

Chapter 4.1

72


73

receiving 5-FU as continuous infusion [10]. More recently, Etienne et al. studied several co-

variables affecting 5-FU clearance during continuous infusion in 104 patients with various

cancers [11]. They found no effect of liver metastases in a subgroup of seven patients,

but did not further specify the involvement of concurrent liver dysfunction. Although

these studies suggest that liver metastases do not affect 5-FU clearance during continu-

ous infusion, it is important to realise that the pharmacokinetic behaviour of 5-FU after

bolus injection or short time infusion differs from that during continuous infusion. After

rapid infusion, 5-FU displays non-linear pharmacokinetic behaviour [12-15], probably due

to saturation of the DPD enzyme at higher plasma levels, although the exact mechanism

is still unclear [16]. Thus, data on 5-FU clearance obtained during continuous infusion do

not necessarily predict the situation after bolus injection. So far, however, only few reports

have been published regarding the effects of liver metastases and/or liver dysfunction on

the pharmacokinetics of bolus injected 5-FU [12,17,18].

Christophidis et al. studied the bioavailability of 5-FU after oral and intravenous drug ad-

ministration in 12 patients with liver metastases, and concluded that the bioavailability

was not related to liver function test abnormalities or metastatic deposits [12]. Nowa-

kowska-Dulawa also found no effect of liver metastases on 5-FU clearance in 20 patients

with colorectal cancer (compared to 8 controls) [17]. Unfortunately, in both studies patient

characteristics and liver function test results were not further specified. More recently,

Terret et al. studied the dose and time dependencies of 5-FU pharmacokinetics in 21

patients and also included some liver function parameters in their analysis [18]. Their data

suggest that 5-FU clearance might increase with the volume of hepatic replacement.

Since in most studies information regarding the extent of liver metastatic involvement is

concise or lacking, we decided to design a protocol to study the effects of this parameter

on the pharmacokinetics of 5-FU and DHFU after bolus injection. This study should provide

more insight in potential interactions between liver function and 5-FU pharmacokinetics.

Patients and Methods

Patients

Patients, aged 18 years and older, scheduled to receive adjuvant or palliative 5-FU treatment

for gastrointestinal cancer, and formerly chemotherapy naive, were included. Patients with

anaemia (Hb < 6 mmol/l), known disorders of hemostasis (e.g. haemophilia), severe renal

failure (GFR < 30 ml/min) or a history of alcohol or drug abuse were excluded. Patients

were assigned to two different groups based on the presence of liver metastases, that were

identified and measured with ultrasound and/or CT imaging. Other pre-treatment meas-

urements were body weight, height, blood cell counts, standard liver function tests (ALT,

AST, LDH, ALP, bilirubin) and markers for the liver synthesis function (pseudocholineste-

rase, albumin and antithrombin-III). Chemotherapy consisted of leucovorin 20 mg/m2/day,

Chapter 4.1

74


75

administered as short time infusion, followed by 5-FU 425 mg/m2/day, administered as

bolus intravenous injection during 2 min. Both drugs were given on 5 consecutive days,

in a 28-day cycle (Mayo regimen). On the day of blood sampling, leucovorin was infused

after the last sample. On the following 4 days, the same 5-FU dose was administered as

short time infusion (10-15 min), after the administration of leucovorin. During the first

chemotherapy cycle, toxicity was scored according to the Common Toxicity Criteria. The

study was approved by the Institutional Medical Ethics Review Board in the participating

hospitals and written informed consent was obtained from all patients.

Quantification of liver metastatic involvement

The diameter of the metastases was measured with standard CT and ultrasound imaging

software. From each lesion the maximum diameter was measured. This diameter was

used to calculate the volume of the metastatic lesions in relation to the total liver volume.

Four levels were defined for semi-quantification of the extent of liver metastatic disease,

according to Hunt et al. [19]. Absence of liver metastases was classified as level 0 (control

group), less than 25 % liver metastatic involvement as level 1, 25-50 % as level 2, and more

than 50% as level 3.

Collection of blood samples

For pharmacokinetic sampling, a canule was placed intravenously in the arm of the

patient contralateral to the side of drug administration. Blood samples of 5 ml were

collected in heparinised tubes just before, and 2, 5, 10, 20, 30, 45, 60, 80, 100, 120, 150 and

180 min after 5-FU injection. Collected samples were immediately placed on ice and sub-

sequently centrifuged at 2,500 g for 10 min at 4°C and stored at –80 °C till analysis. The

plasma samples were analysed for 5-FU and DHFU concentrations by high-performance

liquid chromatography (HPLC).

Chemicals

5-FU and chlorouracil were obtained from Sigma Chemical Co (Zwijndrecht, the

Netherlands). 5,6-dihydro-5-fluorouracil was kindly provided by Roche Laboratories

(Basel, Switzerland). Human heparinised plasma was obtained from the Red Cross Blood

Bank (Groningen, the Netherlands). All other chemicals were of analytical grade.

Reversed phase HPLC analysis

5-FU and DHFU concentrations were measured by HPLC analysis using a modification of

the method described by Ackland et al. [20]. Briefly, 100 µl chlorouracil internal standard

solution (80 mg/l in water) was added to 1 ml plasma sample, and this mixture was

vortexed and subsequently deproteinated with 50 µl of a 50% (w/v) trichloracetic acid

solution. After centrifugation at 8,000 g for 2 min the supernatant was transferred into a

20 ml centrifuge tube and neutralised with 1 ml 1 M sodium acetate solution. Then 5 ml

Chapter 4.1

74


75

ethylacetate was added and the mixture was vortexed during 10 min. After separation of

the organic and aqueous layers by centrifugation at 5,000 g for 5 min, the ethylacetate

layer was transferred into a 10 ml tube and evaporated under a stream of nitrogen at 25 °C.

The residue was dissolved in 100 µl ultrapure water and 20 µl was injected. 5-FU and DHFU

standards ranging from 0.5 to 20 mg/l were prepared in human plasma. The chromato-

graphic system consisted of a Waters 616 pump equipped with a Waters 717+ autosam-

pler. The separation of 5-FU and DHFU was accomplished by gradient elution at ambient

temperature on a Phenomenex Prodigy ODS 3 column (I.D. 250 x 4.6 mm, 5µm) equipped

with a guard column (30 x 4.6 mm) of the same material (both purchased from Bester,

Amstelveen, The Netherlands). Mobile phase A consisted of 1.5 mM K3PO

4 and 1% (v/v)

methanol in water (pH=6.0) and mobile phase B of 1.5 mM K3PO

4 and 5% (v/v) methanol

in water (pH=6.0). The gradient was programmed as follows: 100% A during 2 min; 100%

A → 100% B in 0.5 min; 100% B during 7 min; 100% B → 100% A in 0.5 min; 100% A during

10 min. Drug detection was performed using a Waters 996 Photo Diode Array UV detector

interfaced with a Millenium 2010 Chromatography Manager Workstation. Spectra were

acquired in the 201-300 nm range. 5-FU was monitored at 266 nm and DHFU at 205 nm.

The internal standard chlorouracil was monitored at both wavelengths. The limit of quan-

tification in plasma was 0.1 mg/l for both 5-FU and DHFU.

Pharmacokinetic analysis

The pharmacokinetic analyses were performed in the ADAPT II Maximum Likelihood

Parameter Estimation program (version 4.0; University of Southern California, Los Angeles,

Ca). The pharmacokinetic data of the first 15 patients were tested in 8 different parent drug-

metabolite pharmacokinetic models, characterised by linear or non-linear (Michaelis-Menten)

parent drug (5-FU) elimination from a central compartment and distribution of 5-FU and me-

tabolite (DHFU) over one or two compartments. Variance for the observations was assumed to

be proportional to the measured values and set at 10%. In each model the patient’s data were

fitted individually and for each data set the Akaike Information Criterion (AIC) was calculated

[21]. The model with the lowest summarised AIC value was selected as the better one. The

area under the curve (AUC 0→3h

) of 5-FU and DHFU was calculated using the trapezoidal

rule. The total clearance of 5-FU was calculated by dividing the administered dose by the

AUC.

Statistical analysis

Patient data were analysed as two groups, based on the presence of liver metastases.

Clinical chemistry and pharmacokinetic data in both groups were compared with a two

sided Student’s t-test. In case of unequal variances as indicated by the Kolmogorov-

Smirnov test, the log transformed data were used, or data were tested with the non-para-

metric Mann-Whitney U-test. The study was powered (>80%) to detect a 25% difference in

population means, assuming a standard deviation in pharmacokinetic parameters of 35%.

Chapter 4.1

76


77

Correlations between clinical chemistry, demographic and pharmacokinetic data were

tested by Spearman correlation analysis. Statistical significance was at p<0.05. Analyses

were performed with the SYSTAT 7.0 statistical program (SPSS inc. 1997).

Results

Patients

Patients in this study were treated in the Martini Hospital Groningen, the Bethesda

Hospital Hoogeveen or the Diaconessen Hospital Meppel. Between December 1997 and

January 2001, 18 patients were included in the control group and 16 in the liver metas-

tases group. In one patient belonging to the control group, largely reduced clearance of

5-FU was observed. DNA sequence analysis of the gene encoding DPD revealed that this

patient was heterozygous for a G→A point mutation in the DPYD gene. The resulting

protein from the mutated allele is inactive and total DPD activity in this patient was

low. Data on this patient were published elsewhere [22]. The statistical analyses were

performed both with and without inclusion of this patient in the control group. The liver

metastatic involvement of 9 patients in the metastases group was classified as level 1, that

of 4 patients as level 2 and that of 3 patients as level 3. An overview of the patient char-

acteristics is represented in table 1. An overview of treatment related toxicity, observed

during the first cycle, is shown in table 2.

Pharmacokinetics

The mean pharmacokinetic curves of 5-FU and DHFU as measured in both treatment

groups are represented in figure 2. The model, selected for calculating 5-FU and DHFU phar-

macokinetics of the patients in this study, is a four-compartment parent drug-metabolite

model with Michaelis-Menten elimination from the first towards the third compartment

(see figure 3). The model is described by four differential equations:

��

� ��

�

��

��

�

��

��

��

��

��

��

� ��

�

��

��

�

��

�

��

��

��

Chapter 4.1

76


77

live

r m

etas

tase

sco

ntr

ols

Su

bg

rou

ps

on

per

cen

tag

e o

f li

ver

met

asta

tic

invo

lvem

ent

< 2

5%

25

-50

%>

50

%

par

amet

er(n

orm

al r

ang

e)n

=1

6n

=1

8(n

=9)

(n=

4)(n

=3)

gen

der

M/F

9/7

13/5

4/5

4/0

1/2

age

(yr)

45-

78

(64)

4

5-80

(66)

57-

78(6

4) 4

5-76

(64)

47-

70(6

4)

wei

gh

t(k

g)

60-

94(7

8)

56-

85(7

3) 6

3-90

(78)

60-

94(7

9) 7

5-79

(76)

S-cr

eati

nin

e(4

0-10

0 µm

mo

l/l)

53

-131

(77)

62-1

84(7

7)

53-

131

(75)

71-

94(8

8) 6

1-81

(71)

AST

(<48

U/l

)

12-2

01(4

2) §

1

3-24

(20)

12-

60(3

4) §

26

-137

(59)

§

89-2

01(1

11) §

ALT

(<42

U/l

)

8-

234

(56)

§

5-

44(3

1)

8-58

(37)

§

38-2

02(7

4) §

50

-234

(66)

§

LDH

(200

-500

U/l

)

234-

2662

(5

54) §

17

3-39

6(2

72)

2

34-1

035

(449

) §

348-

1165

(689

) §

839-

2662

(235

7) §

ALP

(<12

5 U

/l)

65-1

221

(146

) §

54

-108

(79)

6

5-16

3(1

07) §

121

-266

(1

81) §

34

4-12

21

(454

) §

Bili

rub

in to

tal

(<17

µm

mo

l/l)

3-29

3(1

3)

5-

14(9

)

3-1

3(1

1)

8-59

(13)

31

-293

(40)

§

alb

um

in(3

5-55

g/l

) 2

4-39

(36)

2

5-43

(36)

31-

39(3

7) 3

6-38

(37)

24-

30(2

6) §

Ach

E-as

e(5

.4-1

3.2

x103 U

/l)

1.2

-8.3

(4.4

)

1.9-

6.5

(3.9

) 3

.0-8

.3(4

.9) §

1.2

-6.3

(3.7

) 1

.5-2

.8(2

.1)

AT-3

(80-

120

%)

5

2-13

7(9

2)

82

-126

(96)

83

-137

(96)

81

-100

(90)

52

-119

(57)

§ Sta

tistic

ally

diff

eren

t fro

m co

ntro

ls at

p<0

.05

(Man

n W

hitn

ey U

test

)

Tabl

e 1

Patie

nt C

hara

cter

istic

s

Data

are

pre

sent

ed a

s ran

ge. T

he m

edia

n va

lue

is pl

aced

bet

wee

n br

acke

ts.

Chapter 4.1

78


79

The compartments 1 and 2 represent the central and peripheral compartment for 5-FU

(parent drug) pharmacokinetics, compartments 3 and 4 are the central and peripheral

compartment for DHFU (metabolite) pharmacokinetics. The X values indicate the amount

of drug in each compartment respectively. Data are imported in the model as plasma drug

concentrations, measured in compartment 1 (5-FU) and compartment 3 (DHFU) respec-

tively. The volume of compartment 1 and 3 is calculated by dividing the drug amount

by drug concentration. The k-values represent linear distribution- and elimination rate

constants, and the Vmax

and Km

values represent Michaelis-Menten constants for non-linear

elimination from the first compartment. The Km

value was kept constant during fitting of

patient data. To determine the best fitting value, this parameter was varied between 0.5

and 15 mg/l. Km

=5 mg/l was selected as most optimal value, based upon the lowest sum-

marised AIC. Rinf

represents the infusion rate of 5-FU in mg/h. No differences existed in

model parameters between the 2 treatment groups and no correlations were measured

between individual model parameters and patient characteristics. Therefore, the calcu-

lated model parameters are listed in table 3 as mean values obtained from all 33 patients

in this study. The interindividual variation in 5-FU clearance was considerable, resulting in

metastases (n=15) no metastases (n=18)

CTC grade I II III IV I II III IV

gastrointestinal

nausea 6 0 0 0 5 0 0 0

vomiting 3 0 0 0 2 0 0 0

diarrhea 1 0 0 0 4 2 0 0

mucositis 3 4 0 0 2 3 0 0

Flu-like symptoms

fever 1 1 0 0 1 0 1 0

malaise 1 1 0 0 6 1 0

Others

dermatological 1 0 0 0 2 0 0 0

eyes 1 0 0 0 2 0 0 0

Toxicity of any kindb 10 (66%) 13 (72%)

aThe figures represent the number of patients suffering from a particular type of toxicity (graded ac-cording to the Common Toxicity Criteria version 2.0) bTotal number of patients suffering from toxicity of any kind at any grade.

Table 2 Overview of side effects during the first cyclea.

Chapter 4.1

78


79

� ��

��

��

��

��

��

��

��

��

��

�

��

��

� ��

��

��

�

��

��

��

��

Figure 2 Pharmacokinetics of 5-FU. Shown are 5-FU ( ) and DHFU ( ) plasma levels observed in control patients (n=18) and 5-FU ( ) and DHFU ( ) plasma levels observed in patients with liver metastases (n=16). All values are depicted as mean ± SD.

Figure 3 Four compartiment parent drug - metabolite pharmacokinetic model describing 5-FU and DHFU pharmacokinetics.

Chapter 4.1

80


81

5-FU AUCs ranging from the lowest to the highest value over a factor 3 (not including the

patient with DPD deficiency).

Correlation between patient covariables and pharmacokinetic parameters

We identified 18 patient covariables that were each tested in a univariate Spearman cor-

relation analysis. Tested covariables were sex, age, height, weight, body surface area (BSA),

lean body mass (LBM), creatinine, urea, level of liver metastatic involvement, AST, ALT, ALP,

LDH, bilirubin, albumin, AT-III, pseudocholinesterase and 5-FU dose. Positive correlations

were found between the BSA and the Vmax

and V1 values, regardless the level of liver meta-

static involvement (r = 0.65 and r = 0.54 respectively)

Discussion

The persistent uncertainty regarding the effects of liver metastases and liver dysfunction

on the pharmacokinetics of bolus injected 5-FU- made us to design the current study. We

decided to target the study on the extent of liver metastatic involvement and also decided

to include a detailed characterisation of the liver function in the study design.

We eventually included 34 patients from which 16 had liver metastases. We choose to

classify the extent of liver metastatic involvement in these patients as categorical rather

than continuous variable, according to Hunt et al.[19], since accurate measurement of the

percentage hepatic replacement is difficult to perform. More subtle differences between

patients cannot be detected in this approach, but this was not considered disadvanta-

Model parameter mean (SD)

Vmax (h-1) 1472 (356)

V1 (l) 15.5 (5.3)

K12 (h-1) 7.73 (4.13)

K21 (h-1) 6.24 (2.77)

V3 (l) 97 (42)

K34 (h-1) 6.18 (12.83)

K43 (h-1) 4.91 (7.38)

K30 (h-1) 1.80 (1.71)

AUC 5-FU (mg.h/l) 10.1 (3.7)

Cl 5-FU (ml/min) 1485 (537)

AUC DHFU (mg.h/l) 5.6 (2.0)

Table 3 Pharmacokinetic parameters

Chapter 4.1

80


81

geous, since small differences will probably be clinically irrelevant.

All patients received treatment according to the Mayo Clinics scheme. The 5-FU dose was

administered as short-time infusion (5-10 min), according to common practice in the

participating hospitals, but on the day of blood sampling, 5-FU was given as 2 min bolus

injection to warrant precise and standardised drug administration. The short half-life of

5-FU necessitated this standardisation, since uncertainty in this parameter can hamper

pharmacokinetic calculations.

To describe the non-linear pharmacokinetics, we applied a relatively complex model

with Michaelis-Menten elimination from the central compartment, based on the model

proposed by Collins et al [13]. Our model was selected on the basis of ‘best fit’ (objectified

by the Akaike Information Criterion) from a series of 8 different variants on the Collins

model. The final model that we used is almost identical to the model proposed by Terret et

al. [18] in their analysis of 5-FU pharmacokinetics. We also included the metabolite DHFU

in our model, but the additional value of this seemed limited, since large 95% confidence

intervals were found around the calculated K34

, K43

and K30

values.

The consequences of non-linear Michaelis-Menten pharmacokinetics are most profound

at plasma levels exceeding the Km

value. In the case of 5-FU, such plasma levels are

reached after bolus injection but not during continuous infusion. A reduction of liver DPD

capacity, due to liver metastases and/or liver dysfunction, might result in lower Vmax

values

and, thus, in a lower 5-FU clearance immediately after bolus injection. Our results do not

confirm this, as patients in both treatment groups displayed similar pharmacokinetics.

No significant correlations were found between pharmacokinetic parameters, including

overall 5-FU clearance, and liver function parameters.

The fact that most patients with liver metastases had mild to moderate liver dysfunction

might explain our observations. However, in three patients with extensive (level 3) meta-

static disease, attended by cholestase as indicated by high bilirubin and ALP levels, 5-FU

clearance also was unchanged. These patients further displayed high transaminase levels

(indicating cell damage) and low albumin and AT-III levels (indicating loss of function).

This observation suggests that the influence of extensive liver metastatic disease,

including liver damage, on 5-FU pharmacokinetics is at leas not dramatic. Recently, Terret

et al. performed a NONMEM analysis to identify covariables that affect 5-FU model pa-

rameters and they observed that their Vmax

values tended to increase with the volume of

liver metastatic involvement [18]. In our study this correlation was only weak (r = 0.21, see

figure 4) and at least clinically irrelevant. These results suggest that the metabolism of 5-

FU in metastatic tumour tissue at least equals that in healthy liver tissue. Extensive 5-FU

uptake in metastatic tissue indeed has been demonstrated during 18F-fluorouracil labelled

positron emission tomography in patients with liver metastases from colorectal cancer

[23]. Although the DPD activity in liver metastases from colon cancer seems to be lower

than in adjacent normal liver tissue [24], hepatic arterial blood flow is generally increased

in liver metastatic disease [19,25]. Since the elimination rate of drugs with a very large

Chapter 4.1

82


83

extraction ratio is strongly depending on the hepatic blood flow, an increase of liver blood

flow in metastatic disease might compensate for a reduced DPD activity in parts of the

liver replaced by metastases.

More or less expected on the basis of the almost identical pharmacokinetic profiles,

treatment related toxicity was comparable in both treatment groups. The incidence of

diarrhoea and malaise was somewhat higher in the control group, but this is probably

accidental. One patient in the control group experienced excessive toxicity as a result of

decreased 5-FU clearance due to DPD deficiency. A full description of the pharmacoki-

netics of 5-FU in this patient was published elsewhere [22]. Interestingly, we found that

irrespective of metastatic status 5-FU clearance was significantly lower in patients experi-

encing mucositis (1696 ± 640 vs. 1209 ± 290 ml/min, p<0.05, see figure 5). Such a trend was

not observed for nausea. This finding might be a coincidence, but it might also be possible

that late toxic effects such as mucositis are more related to pharmacokinetics than direct

toxic effects such as nausea. It has been shown that the extent of salivary excretion is a

predictor for the development of mucositis and salivary excretion is generally higher in

patients with low drug clearance [26].

Based on current data we conclude that there is no need for dose adjustment of 5-FU in

patients with liver metastases and mild to moderate elevations in liver function tests. Both

the extent of liver metastatic involvement and liver function parameters were included

�

��

��

��

��

��

� � � �

��

��

Figure 4 Correlation between level of hepatic replacement and calculated Vmax (h-1) value.

Chapter 4.1

82


83

as parameters in this study to this allowed a more comprehensive evaluation of liver

metastatic disease on 5-FU pharmacokinetics. We believe that there is no reason to expect

increased 5-FU toxicity in patients with liver metastases due to a reduced 5-FU clearance.

Therefore we do not recommend 5-FU dose reduction as standard procedure in these

patients.

Acknowledgement

We would like to thank Roche Basel for providing dihydrofluorouracil chemical standard, Dr.

Hans Proost (Department of Pharmacokinetics and Drug Delivery, University Groningen,

NL) for his advice on pharmacokinetic modelling, Barbara Bong and Dr. Robert de Jong

��

�

��

��

��

��

��

��

��

��

��

��

��

�

��

�

��

��

��

��

��

��

��

��

��

��

��

�

��

��

��

��

Figure 5 Correlation between 5-FU clearance (ml/min) and the occurrence of mucositis (upper panel) and nausea (lower panel).

Chapter 4.1

84


85

(Martini Hospital Groningen, NL), Henk de Korte (Diaconessen Hospital Meppel, NL) and

Janny Haasjes (Bethesda Hospital Hoogeveen, NL) for patient inclusions, the departments

of radiology of the participating hospitals for help on calculations regarding the extent of

liver metastatic involvement, and last but not least all nurses of the oncology wards of the

participating hospitals for their assistance during blood sampling.

Chapter 4.1

84


85

References

1. Pinedo HM, Peters GF. Fluorouracil: biochemistry and pharmacology. J Clin Oncol 1988;6:1653-

1664

2. Heggie GD, Sommadosi JP, Cross DS, Huster WJ, Diasio RB. Clinical pharmacokinetics of 5-fluoro-

uracil and its metabolites in plasma, urine and bile. Cancer Res 1987;47:2203-2206

3. Chazal M, Etienne MC, Renee N, Bourgeon A, Richelme H, Milano G. Link between dihydropy-

rimidine dehydrogenase activity in peripheral blood mononuclear cells and liver. Clin Cancer Res

1996;2: 506-510

4. Etienne MC, Lagrange JL, Dassonville O, Fleming R, Thyss A, Renee N, Schneider M, Demard F,

Milano G. Population study of dihydropyrimidine dehydrogenase in cancer patients. J Clin Oncol

1994;12:2248-2253

5. Fleming RA, Milano G, Thyss A, Etienne MC, Renee N, Schneider M, Demard F. Correlation between

dihydropyrimidine dehydrogenase activity in peripheral mononuclear cells and systemic clear-

ance of fluorouracil in cancer patients. Cancer Res 1992;52: 2899-2902

6. Harris BE, Song R, Soong SJ, Diasio RB. Relationship between dihydropyrimidine dehydrogenase

activity and plasma 5-fluorouracil levels with evidence for circadian variation of enzyme activity

and plasma drug levels in cancer patients receiving 5-fluorouracil by protracted continuous infu-

sion. Cancer Res 1991;50:197-201

7. Lu Z, Zhang R, Diasio RB. Population characteristics of hepatic dihydropyrimidine dehydrogenase

activity, a key metabolic enzyme in 5-fluorouracil chemotherapy. Clin Pharmacol Ther 1995;58:

512-522

8. Ho DH, Townsend L, Luna MA, Bodey GP. Distribution and inhibition of dihydrouracil dehydroge-

nase activities in human tissues using 5-fluorouracil as substrate. Anticancer Res 1986;6:781-784

9. Floyd RA, Hornbeck CL, Byfield JE, Griffiths JC, Frankel SS. Clearance of continuously infused 5-flu-

orouracil in adults having lung or gastrointestinal carcinoma with or without hepatic metastases.

Drug Intell Clin Pharm 1982;16: 665-667

10. Fleming RA, Milano GA, Etienne MC, Renee N, Thyss A, Schneider M, Demard F. No effect of dose,

hepatic function, or nutritional status on 5-FU clearance following continuous (5-day) 5FU infu-

sion. Br J Cancer 1992;6: 668-672

11. Etienne MC, Chatelut E, Pivot X, Lavit M, Pujol A, Canal P, Milano G. Co-variables influencing 5-fluo-

rouracil clearance during continuous venous infusion. A NONMEM analysis. Eur J Cancer 1998;34:

92-97

12. Christophidis N, Vajda FJE, Lucas I, Drummer O, Moon WJ, Louis WJ. Fluorouracil therapy in pa-

tients with carcinoma in the large bowel: a pharmacokinetic comparison of various rates and

routes of administration. Clin Pharmacokinet 1978;3:330-336

13. Collins JM, Dedrick RL, King FG, Speyer JL, Myers CE. Nonlinear pharmacokinetic models for 5-flu-

orouracil in man: intravenous and intraperitoneal routes. Clin Pharmacol Ther 1980;28:235-246

14. McDermot BJ, van den Berg HW, Murphy RF.Nonlinear pharmacokinetics for the elimination of

5-fluorouracil after intravenous administration in cancer patients. Cancer Chemother Pharmacol

Chapter 4.1

86

1982;9:173-178

15. Van Groeningen CJ, Pinedo HM, Heddes J, Kok RM, De Jong AP, Wattel E, Peters GJ, Lankelma J.

Pharmacokinetics of 5-fluorouracil assessed with a sensitive mass spectrometric method in pa-

tients on a dose escalation schedule. Cancer Res 1988;48:6956-6961

16. Gamelin E, Boisdron-Celle M. Dose monitoring of 5-fluorouracil in patients with colorectal or

head and neck cancer - status of the art. Crit Rev Oncol Hematol 1999;30: 71-79

17. Nowakowska-Dulawa E. Circadian rhythm of 5-fluorouracil pharmacokinetics and tolerance.

Chronobiologica 1990;17:27-35

18. Terret C, Erdociain E, Guimbaud R, Boisdron-Celle M, McLeod HL, Fety-Deporte R, Lafond T, Game-

lin E, Bugat R, Canal P, Chatelut E. Dose and time dependencies of 5-fluorouracil pharmacokinet-

ics. Clin Pharmacol Ther 2000;68:270-279

19. Hunt TM, Flowerdew ADS, Britten AJ, Fleming JS, Karran SJ, Taylor I. An association between

parameters of liver blood flow and percentage hepatic replacement with tumour. Br J Cancer

1989;59:410-414

20. Ackland SP, Garg MB, Dunstan RH. Simultaneous determination of dihydrofluorouracil and 5-

fluorouracil in plasma by high-performance liquid chromatography. Anal Biochem 1997;246:

79-85

21. Bourne DWA. Evaluation of program output. In: Bourne DWA (ed) Mathematical modelling of

pharmacokinetic data. Technomic publishing Co inc. 1995 Lancaster Basel p.107-109

22. Maring JG, Van Kuilenburg ABP, Piersma H, Groen HJM, Uges, DRA, De Vries EGE. Reduced 5-FU

clearance in a patient with low DPD activity due to heterozygosity for a mutant allele of the

DPYD gene. Br J Cancer 2002;86;1028-1033

23. Moehler M, Dimitrakopoulou-Strauss A, Gutzler F, Raeth U, Strauss LG, Stremmel W.18F-labeled

fluorouracil positron emission tomography and the prognoses of colorectal carcinoma patients

with metastases to the liver treated with 5-fluorouracil. Cancer 1998;83:245-253

24. Johnston SJ, Ridge SA, Cassidy J, McLeod HL. Regulation of dihydropyrimidine dehydrogenase in

colorectal cancer. Clin Cancer Res 1999;5:2566-2570

25. Leen E, Goldberg JA, Robertson J, Angerson WJ, Sutherland GR, Cooke TG, McArdle CS. Early de-

tection of occult colorectal hepatic metastases using duplex colour Doppler sonography. Br J

Surg 1993; 80:1249-1251

26. Joulia JM, Pinguet F, Ychou M, Duffour J, Astra C, Bresolle F. Plasma and salivary pharmacokinetics

of 5-fluorouracil (5-FU) in patients with metastatic colorectal cancer receiving 5-FU bolus plus

continuous infusion with high-dose folinic acid. Eur J Cancer 1999;35:296-301

Reduced 5-FU clearance in a patient with low DPD activity due to heterozygosity for a mutant allele of the DPYD gene

Jan Gerard Maring1, André B.P. van Kuilenburg2, Janet Haasjes2, Henk Piersma3, Harry J.M.

Groen4, Donald R.A. Uges5, Albert H. van Gennip2, Elisabeth G.E. de Vries6

1Department of Pharmacy, Diaconessen Hospital Meppel and Bethesda Hospital

Hoogeveen; 2Department of Clinical Chemistry, Academic Medical Center Amsterdam; 3Department of Internal Medicine, Martini Hospital Groningen; Departments of 4Pulmonary Diseases, 5Pharmacy and 6Medical Oncology, University Hospital Groningen,

The Netherlands

Br J Cancer 2002;86;1028-1033

Chapter 4.2

88

Influence of DPD deficiency on 5-fluorouracil pharmacokinetics

89

One out of billions. Lago di Garda. Italy 2000.

Chapter 4.2

88


89

Abstract

Aim 5-FU pharmacokinetics, DPD-activity and DNA sequence analysis were compared

between a patient with extreme 5-FU induced toxicity and six control patients with

normal 5-FU related symptoms.

Methods Patients were treated for colorectal cancer and received chemotherapy consist-

ing of folinic acid 20 mg/m2 plus 5-FU 425 mg/m2 . Blood sampling was carried out on day

1 of the first cycle.

Results The 5-FU AUC in the index patient was 24.1 mg.h/l compared to 9.8 ± 3.6 (range

5.4-15.3) mg.h/l in control patients. The 5-FU clearance was 520 ml/min versus 1293 ± 302

(range 980-1780) ml/min in controls. The activity of DPD in mononuclear cells was lower

in the index patient (5.5 nmol/mg/h) compared to the 6 controls (10.3 ± 1.6, range 8.0-

11.7 nmol/mg/h). Sequence analysis of the DPD gene revealed that the index patient was

heterozygous for a IVS14+1G>A point mutation.

Conclusions Our results indicate that the inactivation of one DPYD allele can result in a

strong reduction in 5-FU clearance, causing severe 5-FU induced toxicity.

Introduction

Fluorouracil (5-FU) is widely used in chemotherapeutic regimens for the treatment

of breast-, colorectal- and head- and neck cancer. The cytotoxic mechanism of 5-FU is

complex, requiring intracellular bioconversion of 5-FU into cytotoxic nucleotides (see

figure 1). Inhibition of thymidylate synthase by the metabolite 5-fluoro-2’-deoxyuridine-

5’-monophosphate is thought to be the main mechanism of cytotoxicity [1]. The cytotox-

icity is caused by only a small part of the administered 5-FU dose, as the majority of 5-FU is

rapidly metabolised into inactive metabolites. The initial and rate-limiting enzyme in the

catabolism of 5-FU is dihydropyrimidine dehydrogenase (DPD), catalysing the reduction

of 5-FU into 5,6-dihydrofluorouracil (DHFU). Several groups have suggested a major role

of DPD in the regulation of 5-FU metabolism and thus in the amount of 5-FU available for

cytotoxicity [2-5]. Indeed, in patients with DPD enzyme deficiency, 5-FU chemotherapy

is associated with severe, life-threatening toxicity [6]. Moreover, a markedly prolonged

elimination half-life of 5-FU has been observed in a patient with complete deficiency

of DPD enzyme activity [7]. Several mutations in the dihydropyrimidine dehydrogenase

gene (DPYD), which encodes for the DPD enzyme have recently been identified [6,8].

Furthermore, the frequency of DPD deficiency has been estimated to be as high as 2-3%

[5,9,10]. To date, a direct correlation between DPD gene mutation and decreased 5-FU

clearance has only been suggested but never been proven. In this study, we provide the

first detailed analysis of 5-FU pharmacokinetics in a patient with low DPD-activity due to

heterozygosity for a mutant allele of the gene encoding DPD.

Chapter 4.2

90


91

Methods

Chemicals

5-FU was obtained from Sigma Chemical Co. (Zwijndrecht, the Netherlands). 5,6-dihydro-

5-fluorouracil was kindly provided by Roche Laboratories (Basel, Switzerland). AmpliTaq

Taq polymerase and BigDye-Terminator-Cycle-Sequencing-Ready-Reaction kit were

supplied by Perkin Elmer (San Jose, CA, USA). A Quaquik Gel Extraction kit was obtained

from Qiagen (Hilden, Germany). Human heparinised plasma was obtained from the Red

Cross Blood Bank (Groningen, the Netherlands). [4-14C ]Thymine (1.85-2.22 GBq/mmol) was

obtained from Moravek Biochemicals (CA, USA) and Lymphoprep (spec.gravity 1.077 g/ml,

280 mOsm) was from Nycomed Pharma AS (Oslo, Norway). Leucosep tubes were supplied

by Greiner (Frickenhausen, Germany). All other chemicals were of analytical grade.

Patient and controls

All patients were treated for colorectal cancer and participated in a protocol that had

been designed to study 5-FU and DHFU pharmacokinetics. The protocol was approved

��

��

��

��

��

��

��

Figure 1 Metabolism of 5-FU. 5-Fluoro-2’-deoxyuridine-5’-monophosphate (FdUMP) is the cytotoxic product resulting from a multi-step 5-FU activation route . FdUMP inhibits the enzyme thymidylate synthase (TS), which leads to intracellular accumulation of deoxy-uridine-monophospate (dUTP) and depletion of deoxy-thymidine-monophospate (dTMP). This causes arrest of DNA synthesis. The initial and rate-limiting enzyme in the catabolism of 5-FU is dihydropyrimidine dehydrogenase (DPD), catalys-ing the reduction of 5-FU into 5,6-dihydrofluorouracil (DHFU). Subsequently, DHFU is degraded into fluoro-β-ureidopropionic acid (FUPA) and fluoro-β-alanine (FBAL).

Chapter 4.2

90


91

by the Medical Ethics Review Committee of the Martini Hospital Groningen and written

informed consent was obtained from all patients. All patients who entered this protocol

were chemotherapy naive. Chemotherapy consisted of folinic acid 20 mg/m2 combined

with 5-FU 425 mg/m2 , both on 5 successive days, in a 28-day cycle. Blood sampling was

carried out on the first day of the first chemotherapy cycle immediately following the 5-FU

dose, administered as bolus intravenous injection over 2 min. Folinic acid was infused after

the end of blood sampling. On the following 4 days, the same 5-FU dose was administered

as short time infusion, after folinic acid administration.

One patient experienced severe toxicity during the first chemotherapy cycle and,

therefore, a screening on DPD deficiency was initiated. Data from seventeen patients who

participated in the same study protocol were analysed for reference pharmacokinetics.

Six patients, who showed no signs of severe toxicity, were randomly selected for reference

DPYD genotyping and DPD enzyme activity. These patients served as controls.

Collection of blood samples

For pharmacokinetic sampling, a canule was placed in the arm of the patient contralateral

from drug administration. Blood samples (5 ml) were collected in heparinised tubes just

before, and 2, 5, 10, 20, 30, 45, 60, 80, 100, 120, 150 and 180 min postinjection. Collected

samples were immediately placed on ice and subsequently centrifuged at 2,500 g for 10

min. The plasma samples were analysed for 5-FU and DHFU concentrations by high-per-

formance liquid chromatography (HPLC) on the day of collection.

Blood samples for DPD analysis were collected 5 to 23 months after blood sampling for

5-FU pharmacokinetics, which corresponds to intervals ranging from 2 to 17 months after

the last 5-FU dose. None of the patients received chemotherapy at that moment.

Reversed phase HPLC analysis

5-FU and DHFU concentrations were measured by HPLC analysis using a modification of

the method described by Ackland et al [11]. Briefly, 100 µl chlorouracil internal standard

solution (80 mg/l in water) was added to 1 ml plasma sample, and this mixture was

vortexed and subsequently deproteinated with 50 µl of a 50% (w/v) trichloracetic acid

solution. After centrifugation at 8,000 g for 2 min the supernatant was transferred into a

20 ml centrifuge tube and neutralised with 1 ml 1M sodium acetate solution. Then 5 ml

ethylacetate was added and the mixture was vortexed during 10 min. After separation of

the organic and aqueous layers by centrifugation at 5,000 g for 5 min, the ethylacetate

layer was transferred into a 10 ml tube and evaporated under a stream of nitrogen at 25

°C. The residue was dissolved in 100 µl ultrapure water and 20 µl was injected. 5-FU and

DHFU standards ranging from 0.5 to 20 mg/l were prepared in human plasma.

The chromatographic system consisted of a Waters 616 pump equipped with a Waters

717+ autosampler. The separation of 5-FU and DHFU was accomplished by gradient

elution at ambient temperature on a Phenomenex Prodigy ODS 3 column (I.D. 250x4.6

Chapter 4.2

92


93

mm, 5µm) equipped with a guard column (30x4.6 mm) of the same material. Mobile phase

A consisted of 1.5 mM K3PO

4 and 1% (v/v) methanol (pH=6.0) and mobile phase B of 1.5

mM K3PO

4 and 5% methanol (pH=6.0).

The gradient was programmed as follows: 100% A during 2 min; 100% A → 100% B in 0.5

min; 100% B during 7 min; 100% B → 100% A in 0.5 min; 100% A during 10 min. Detection

was performed using a Waters 996 Photo Diode Array UV detector interfaced with a

Millenium 2010 Chromatography Manager Workstation. Spectra were acquired in the 201-

300 nm range. 5-FU was monitored at 266 nm and DHFU at 205 nm. The internal standard

chlorouracil was monitored at both wavelengths.


The pharmacokinetic analyses were performed in the ADAPT II computer program (version

4.0; USC Los Angeles). The pharmacokinetic data of both the index patient and seventeen

reference patients (among which the six controls) were tested in 8 different models. In each

model the patient’s data were fitted individually and for each data set the Akaike Information

Criterion (AIC) was calculated. The model with the lowest summarised AIC value was selected

as the better one (data not shown). The model used for calculating 5-FU pharmacokinetics is a

two-compartment model with Michaelis-Menten elimination from the first compartment

and is described by two differential equations:

��

� ��

�

��

��

�

��

��

��

��

��

X1

and X2

indicate the amount of drug in each compartment, respectively. The k-values

represent linear distribution- and elimination rate constants, and the Vmax

and Km

values

represent Michaelis-Menten constants for non-linear elimination from the first compartment.

Rinf

represents the infusion rate of 5-FU.

The area under the curve of 5-FU and DHFU was calculated using the trapezoid rule. The

average systemic clearance of 5-FU was calculated by dividing the administered dose by the

area under the curve (AUC).

Determination of dihydropyrimidine dehydrogenase activity

To investigate whether the 5-FU toxicity might have been caused by a partial deficiency of

DPD, we determined the activity of DPD in peripheral blood mononuclear (PBM) cells.

Therefore PBM cells were isolated from 15 ml EDTA anticoagulated blood and the activity

of DPD was determined according previously decribed methods [12]. In brief, the sample

was incubated in a reaction mixture containing 35mM potassium phosphate pH 7.4, 1mM

dithiothreitol, 2.5mM magnesium chloride, 250µM NADPH and 25 µM [4-14C] thymine.

Chapter 4.2

92


93

After an appropriate incubation time, the reaction catalysed by DPD was terminated by

adding 10 % (v/v) perchloric acid. The reaction mixture was centrifuged at 11,000 g for 5

min to remove protein. The separation of radiolabelled thymine and the reaction products

was performed by reversed phase HPLC. Protein concentrations were determined with a

copper-reduction method using bicinchoninic acid, as decribed by Smith et al. [13].

PCR amplification of coding exons

The DNA from the index and control patients was isolated from PBM cells as previously

described [14]. PCR amplification of exon 14 and flanking intronic regions was carried out

according to Van Kuilenburg et al. [6]. PCR products were separated on 1% agarose gels,

visualised with ethidium bromide and purified using a Qiaquick Gel Extraction kit and

used for direct sequencing.

Sequence analysis

Sequence analysis was carried out on a Applied Biosystems model 377 automated DNA

sequencer using Dye-Terminator method for the DPD cDNA and genomic fragments.

Statistical analysis

Each value, measured in the index patient, was compared to the mean ± 2 SD range of the

corresponding parameter in the control group . Values outside this range were considered

abnormal (p<0.05). We did not match our control patients for age and gender.

Results

Clinical evaluation

Patient characteristics from the index and control patients, as measured before 5-FU ad-

ministration on the first day of the first chemotherapy cycle, are listed in table 1. The index

patient is a 60-year-old white female who received adjuvant chemotherapy for Dukes C

colon carcinoma. She was known with a chronic moderate renal function impairment as

a result of a double sided nephrolithotomy at age 40. The first two injections with a total

dose of 800 mg 5-FU/day were tolerated well by the patient without complications. On the

third day of chemotherapy she experienced nausea and cold shivers. The nausea was suc-

cessfully treated with metoclopramide. The cold shivers remained on days 4 and 5. Twelve

days after administration of the first 5-FU injection, leukopenia (1.5x109 leukocytes/l) and

thrombocytopenia (26x109 platelets/l) developed along with nausea, diarrhoea, stomati-

tis, fever and hair loss. The next day leukocytes and platelets decreased to 0.5x109/l and

12x109/l (both nadir values respectively). During this period the patient developed leuko-

penic fever (40 °C) for which antibiotics were administered. Until day 20 the leucocytes

and platelets remained low (1x109/l and 13x109/l respectively). During the subsequent

Chapter 4.2

94


95

��

�

��

��

� ��

��

��

��

��

��

��

��

Figure 2 Pharmacokinetics of 5-FU. Shown are 5-FU plasma levels observed in a patient with a IVS14+1G→A mutation in the DPYD gene ( ) and the 5-FU plasma levels resulting from simulation of a normal renal function in the same patient. 5-FU plasma levels from control patients are depicted as mean ± SD ( ; n=6).

Index patient Controls (n=6)

mean ± 2 SD

Gender M/F F 4/2

Age (yr) 60 64 ± 12

Weight (kg) 73 77 ± 12

Serum creatinine (µmmol/L) 184 § 81 ± 30

Aspartate aminotransferase (U/L) 23 18 ± 8

Alanine aminotransferase (U/L) 41 29 ± 26

LDH (U/L) 379 302 ± 152

Alkaline phosphatase (U/L) 95 72 ± 24

Bilirubin total (µmmol/L) 7 10 ± 6

Albumin (g/L) 32 37 ± 8

§ outside 95% control range, p<0.05

Table 1 Patient Characteristics

Chapter 4.2

94


95

week the clinical picture and hematological parameters gradually improved and normal-

ised. On day 34 the patient was discharged from the hospital.

The toxicity observed in the six control patients was limited to mild nausea (n=4), vomiting

(n=2) and CTC grade 1 stomatitis (n=1).


The clearance of 5-FU was considerable slower in the index patient than in the six control

patients. In all control patients the plasma level at t=90 min was below 0.1 mg/l, whereas

in the index patient the plasma level was still 3.8 mg/l at this time point (see figure 2).

The AUC in the patient suffering from toxicity was 24.1 mg.h/l compared to 15.3 mg.h/l

as highest AUC value in control patients. We calculated an average systemic clearance of

only 520 ml/min versus 980-1780 ml/min in controls. The Vmax value, calculated by phar-

macokinetic modelling was 548 mg/h, while the Vmax values of control patients ranged

from 984 to 1772 mg/h (see table 2). The pharmacokinetic data of the six control patients

did statistically not differ from data of the reference group. The effect of the impaired

renal function of the index patient on 5-FU clearance was studied by pharmacokinetic

modelling.

The excretion of 5-FU in urine was measured in five patients of the reference group

(including two control patients) and kurine

was estimated 0.5 ± 0.08 h-1. This kurine

value, in-

dividually normalised on calculated GFR, was used during subsequent modelling of other

patient data. A normal renal function was simulated in the index patient by replacing the

GFR related kurine

by kurine

=0.6. Renal function impairment appeared to have only a slight

effect on 5-FU clearance. In the index patient, we estimated an additional 18% increase of

Index patient Index patient

kurine

=0.6

simulated data

Controls

(n=6)

Controls

Kurine

=0

simulated data

Dose (mg/m2) 447 431 ± 64

AUC 5-FU (mg.h/L) 24.1 § 20.4 § 10.0 ± 6.6 § 11.8 ± 8.6 §

Cl 5-FU (mL/min) 553 § 637 § 1493 ± 840 § 1306 ± 886

Vmax (1/h) 548 § 1329 ± 680 §

V1 /kg (L) 0.25 0.23 ± 0.06

K12 (1/h) 9.0 5.1 ± 7.6

K21 (1/h) 7.5 5.2 ± 6.6

§ data of index patient outside 95% control range, p<0.05

Table 2 Overview of pharmacokinetic parameters. Data are presented as single observation or as mean ± 2 SD

Chapter 4.2

96


97

the AUC due to the renal insufficiency upon a 108% higher AUC due to partial DPD defi-

ciency. Simulation of anuria in the control group (Kurine

= 0), revealed a 16 ± 4% increase of

the 5-FU AUC. In the reference group, the estimated increase was 15 ± 5 %.

Activity of DPD in PBM cells

The activity of DPD in PBM cells was lower in the patient experiencing severe toxicity (5.5

nmol/mg/h) compared to the 6 control patients (8.0-11.7 nmol/mg/h; mean 9.6).

Genomic sequence analysis

Sequence analysis of the DPD gene showed that the patient was heterozygous for a G→A

point mutation in the invariant GT splice donor site (IVS14+1G>A), leading to the skipping

of exon 14 directly upstream of the mutated splice donor site during DPD pre-mRNA

splicing. Sequence analysis of the DPD gene of the six control patients revealed no gene

mutations.

Discussion

5-FU remains the major drug in the treatment of advanced colorectal cancer. DPD is the

key metabolic enzyme in 5-FU degradation and since more than 80% of the dose is meta-

bolised by this enzyme, DPD activity is one of the main factors determining drug exposure

[2-5]. It is generally accepted that DPD activity in the liver is responsible for the majority of

5-FU catabolism [15,16], but PBM cells are often used as a surrogate for liver DPD activity,

since these cells are better accessible [2-5]. Several groups have suggested that markedly

diminished DPD activity in PBM cells is strongly related to the risk of developing severe

5-FU toxicity due to reduced 5-FU clearance [4-6]. Although total DPD deficiency is rare in

adults, about 2-3% of the population has a low PBM-DPD enzyme level and, thus, is at risk

to develop severe toxicity when treated with 5-FU [5,9,10]. In only few reports however,

the effect of DPD-deficiency on 5-FU clearance has been objectively quantified. Diasio

et al. administered a test dose of 25 mg/m2 5-FU to a patient with non-detectable DPD-

activity in PBM cells and found a very low 5-FU clearance rate [7]. This patient was probably

homozygous for a mutant DPD allele, although the genetic cause was never elucidated.

Stephan et al. reported severe toxicity in a female patient after treatment comprising

folinic acid 500 mg/m2 as 2 h intravenous infusion plus 125 mg orally followed by 5-FU

2 g/m2 as a 24 h continuous infusion [17]. They found a 5-FU plasma level of 0.3 mg/l on

day 15 after administration, which implies a dramatic overexposure to 5-FU. This patient

could not have been homozygous deficient because the DPD activity in lymphocytes was

within the normal range. The role of PBM-DPD activity as an indicator for 5-FU clearance

is, however, questionable. Lu et al. found a correlation between PBM-DPD activity and

DPD activity measured in the liver [9], but others reported a weak, non-significant rela-

Chapter 4.2

96


97

tionship [10]. Both Harris et al. and Fleming et al. initially reported a correlation between

PBM-DPD activity and 5-FU clearance but re-examination of this relationship in a larger

set of patients revealed a markedly weakened correlation [2,3,5]. These data indicate that

PBM-DPD activity is not a strong and reliable indicator of 5-FU clearance. To some extent,

the variability in DPD activity might have been caused by the composition of the isolated

PBM cells [12]. Another important factor is the timing of cell sampling in relation to 5-FU

administration, since McLeod et al. showed that 5-FU is able to inhibit DPD-activity [18].

In this paper, we report a combined pharmacokinetic and genetic analysis of the DPD

gene and demonstrate that a single G→A point mutation in the invariant splice donor site

IVS14+1 of the DPD gene has profound impact on the clearance of 5-FU. We also found a

lower PBM-DPD activity in the index patient compared to 6 controls. The DPD activity was

comparable to the activity observed in obligate heterozygotes (5.5 ± 2.1 nmol/mg/h, n=8)

in previous work, but the measured value also fits within the range for normal controls

(10.0 ± 3.4 nmol/mg/h, range 3.4-18 nmol/mg/h, n=22) [6]. This might indicate that not

only obligate heterozygotes, but also low normal homozygotes are at risk for developing

severe toxicity when treated with 5-FU. It is however as yet unclear whether the pharma-

cokinetic profile of 5-FU is identical in both groups. Unfortunately, we did not measure

DPD activity and 5-FU pharmacokinetics on the same day. DPD activity was measured at

least two months after completing chemotherapy and might have been changed since

the first chemotherapy cycle as a result of chemotherapy itself or disease state. Although

5-FU has a direct inhibitory effect on DPD activity, as was shown by McLeod et al., we

believe that it is not likely that this effect will continue until two months after the last

dose [18]. Furthermore, it has been shown that DPD activity is lower in (breast-) cancer

patients compared to healthy persons, suggesting an effect of disease state [19]. We tried

to minimise this effect by matching index patient and controls for this variable (all Dukes

C, with no signs of disease progression at time of blood sampling for DPD). Finally, timing

of blood sampling is an important parameter in relation to DPD activity, since the activity

of DPD exhibits a circadian pattern, with lowest activity around 1 p.m. [2]. All our patients

received chemotherapy between 10 a.m. and 3 p.m. and blood samples for DPD analysis

were collected between 11 a.m. and 2 p.m. to minimise the effect of circadian variation.

Thus, unless unfortunate late DPD sample collection, we believe our results to be repre-

sentative for DPD activity during pharmacokinetic sampling.

The structural organization of the DPD gene has recently been described. It is 150 kb

in length and consists of 23 exons ranging in size from 69 to 1404 bp [20]. The G→A

mutation changes an invariant GT splice donor site into AT which leads to skipping of a

165 bp exon immediately upstream of the mutated spice donor site during the splicing of

DPD pre-mRNA. As a consequence, a 165 bp fragment encoding the amino acid residues

581-635 of the primary sequence of the DPD protein is lacking in the mature DPD mRNA,

which results in an enzyme without catalytic activity [14,21]. Analysis of the prevalence of

the various mutations among cancer patients with partial DPD deficiency showed that the

Chapter 4.2

98


99

G→A mutation in the invariant splice donor site is the most common one (43%) [6].

We believe that in our patient the IVS14+1G>A mutation explains the dramatic reduction

in 5-FU clearance compared to controls. This is in line with the observation that at least

80% of the 5-FU dose is catabolised by DPD. Although our patient had a moderately

impaired renal function, we analysed that the effect of renal function on 5-FU clearance is

only limited. This was expected, as only about 10% of the 5-FU dose is excreted in urine.

It is important, however, to realise that in patients with reduced DPD capacity, the con-

tribution of renal excretion to total clearance is relatively increased. As a consequence,

the impact of renal insufficiency on the AUC is larger in DPD deficient than in DPD pro-

ficient patients. Thus, in the index patient, the impaired renal function might have been

contributed to development of more severe toxicity, additionally to that caused by DPD

deficiency.

Development of rapid assays to detect mutations in the DPYD gene makes it possible to

carry out a genetic screening prior to the start of chemotherapy containing 5-FU. However

it is important to identify those mutations that result in a defect DPD protein.

So far, 19 molecular defects in the DPYD gene such as point mutations and deletions due

to exon skipping have been reported, but not all mutations result in a DPD enzyme de-

ficiency [6]. Incomplete correlation between DPD phenotype and genotype is clinically

important and suggests that DPD polymorphisms are likely to be complex [8].

Our results indicate that low DPD activity, due to the inactivation of one DPYD allele

results in a strong reduction in 5-FU clearance, measured on day 1 of chemotherapy. In-

hibition of the yet reduced DPD activity by 5-FU itself during subsequent days may lead

to further reduction of 5-FU clearance and this may further add to the development of

severe toxicity. In order to identify those mutations that result in reduced 5-FU clearance,

monitoring of 5-FU plasma levels using a limited sampling strategy can be helpful in

patient selection. This requires, however, rapid plasma level analysis, because results from

the first 5-FU infusion must be available before the second dose is administered. Unfor-

tunately, no rapid 5-FU (immuno-)assay is available yet, and therefore in most hospitals

therapeutic drug monitoring of 5-FU is probably not feasible.

Alternatively, the use of DPD inhibitors such as eniluracil in combination with a reduced

5-FU dose might eliminate the problem of DPD-related variable clearance of 5-FU. Clinical

trials are in progress to evaluate such combinations as an alternative for standard 5-FU

therapy.

Acknowledgements

We thank Dr. J.H. Proost (Department of Pharmacokinetics and Drug Delivery, University

Groningen) for advice on pharmacokinetic modelling. We thank Roche Basel for providing

dihydrofluorouracil chemical standard.

Chapter 4.2

98


99

References

1. Pinedo HM and Peters GF. Fluorouracil: biochemistry and pharmacology. J Clin Oncol 1988;6:

1653-1664

2. Harris BE, Song R, Soong SJ and Diasio RB. Relationship between dihydropyrimidine dehydroge-

nase activity and plasma 5-fluorouracil levels with evidence for circadian variation of enzyme

activity and plasma drug levels in cancer patients receiving 5-fluorouracil by protracted continu-

ous infusion. Cancer Res 1990;50:197-201

3. Fleming RA, Milano G, Thyss A, Etienne MC, Renee N, Schneider M and Demard F. Correlation be-

tween dihydropyrimidine dehydrogenase activity in peripheral mononuclear cells and systemic

clearance of fluorouracil in cancer patients. Cancer Res 1992;52:2899-2902

4. Lu Z, Zhang R and Diasio RB. Dihydropyrimine dehydrogenase activity in human peripheral

blood mononuclear cells and liver: population characteristics, newly identified patients, and

clinical implication in 5-fluorouracil chemotherapy. Cancer Res 1993;53: 5433-5438

5. Etienne MC, Lagrange JL, Dassonville O, Fleming R, Thyss A, Renee N, Schneider M, Demard F and

Milano G. Population study of dihydropyrimidine dehydrogenase in cancer patients. J Clin Oncol

1994;12: 2248-2253

6. Van Kuilenburg ABP, Haasjes J, Richel DJ, Zoetekouw L, Van Lenthe H, Waterham HR, De Abreu RA,

Maring JG, Vreken P and Van Gennip AH. Clinical implications of dihydropyrimidine dehydroge-

nase (DPD) deficiency in patients with severe 5-fluorouracil associated toxicity. Identification of

new mutations in the DPD gene. Clin Cancer Res 2000;6:4705-4712

7. Diasio RB, Beavers TL and Carpenter JT. Familial deficiency of dihydropyrimine dehydrogenase.

Biochemical basis for familial pyrimidinemia and severe 5-fluorouracil induced toxicity. J Clin

Invest 1988;81:47-51

8. Collie-Duguid ES, Etienne MC, Milano G and McLeod HL. Known variant DPYD alleles do not ex-

plain DPD deficiency in cancer patients. Pharmacogenetics 2000;10:217-223

9. Lu Z, Zhang R and Diasio RB. Population characteristics of hepatic dihydropyrimidine dehydro-

genase activity, a key metabolic enzyme in 5-fluorouracil chemotherapy. Clin Pharmacol Ther

1995;58:512-522

10. Chazal M, Etienne MC, Renee N, Bourgeon A, Richelme H, Milano G Link between dihydropy-

rimidine dehydrogenase activity in peripheral blood mononuclear cells and liver. Clin Cancer Res

1996;2:506-510

11. Ackland SP, Garg MB, Dunstan RH. Simultaneous determination of dihydrofluorouracil and 5-fluo-

rouracil in plasma by high-performance liquid chromatography. Anal Biochem 1997; 246:79-85

12. Van Kuilenburg ABP, Van Lenthe H, Tromp A, Veldman PC, Van Gennip AH. Pitfalls in the diagnosis

of patients with partial dihydropyrimidine dehydrogenase deficiency. Clin Chem 2000:46:9-17.

13. Smith PK, Krohn RI, Hermanson GT, Mallia AK, Gartner FH, Provenzano MD, Fujimoto EK, Goeke NM,

Olson BJ, Klenk DC. Measurement of protein using bicinchoninic acid. Anal Biochem 1985;150:76-

85

14. Van Kuilenburg ABP, Vreken P, Beex LV, Meinsma R, Van Lenthe H, De Abreu RA and Van Gennip

Chapter 4.2

100

AH. Heterozygosity for a point mutation in an invariant splice donor site of dihydropyrimine de-

hydrogenase and severe 5-fluorouracil related toxicity. Eur J Cancer 1997;33:2258-2264

15. Fleming RA, Milano GA, Etienne MC, Renee N, Thyss A, Schneider M, Demard F. No effect of dose,

hepatic function, or nutrional status on 5-FU clearance following continuous (5-day) 5-FU infu-

sion. Br J Cancer 1992;66:668-672

16. Ho DH, Townsend L, Luna M and Bodey GP. Distribution and inhibition of dihydrouracil dehy-

drogenase activities in human tissues using 5-fluorouracil as substrate. Anticancer Res 1986;6:

781-784

17. Stephan F, Etienne MC, Wallays C, Milano G and Clergue F. Depressed hepatic dihydropyrimidine

dehydrogenase activity and fluorouracil related toxicities. Am J Med 1995;99:685-688

18. McLeod HL, Sludden J, Hardy SC, Lock RE, Hawksworth GM and Cassidy J. Autoregulation of 5-

fluorouracil metabolism. Eur J Cancer 1998;34:623-1627

19. Lu Z, Zhang R, Carpenter JT and Diasio RB. Decreased dihydropyrimine dehydrogenase activity in

a population of patients with breast cancer: implication for 5-fluorouracil-based chemotherapy.

Clin Cancer Res 1998;4:325-329

20. Johnson MR, Wang K, Tillmanns S, Albin N and Diasio RB. Structural organization of the human

dihydropyrimidine dehydrogenase gene. Cancer Res 1997;57:1660-1663

21. Wei X, McLeod HL, McMurrough J, Gonzalez FJ and Fernandez-Salguero P. Molecular basis of

the human dihydropyrimidine dehydrogenase deficiency and 5-fluorouracil toxicity. J Clin Invest

1996;98:610-615

extensive hepatic replacement due to liver metastases has ... · extensive hepatic replacement due...

Documents