CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES

CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES CHAPTER 20 – MRP2, THE APICAL EXPORT PUMP FOR ANIONIC CONJUGATES

I NTRODUCTION

The ATP-dependent unidirectional transport

of anionic conjugates, such as bilirubin

glucu-ronosides and leukotriene C4(LTC4), across the

apical membrane domain of polarized cells

plays an important role in the elimination and

detoxification of endogenous and xenobiotic

substances This process has been

function-ally characterized by measurements of

ATP-dependent transport of labeled conjugates into

inside-out membrane vesicles prepared from

the apical membranes of hepatocytes (Ishikawa

et al., 1990; Kitamura et al., 1990; Kobayashi

et al., 1988) This transport function was

ori-ginally described as a glutathione S-conjugate

transport system (Ishikawa et al., 1990) and as

a canalicular multispecific organic anion

trans-porter (abbreviated cMOAT) (Oude Elferink

et al., 1993) Subsequent cloning, expression and

functional analysis of the recombinant protein

has established that the apical conjugate export

pump is encoded by the MRP2 (ABCC2) gene

(Büchler et al., 1996; Cui et al., 1999; Evers et al.,

1998; Paulusma et al., 1996; Taniguchi et al.,

1996) Antibodies raised against various

epi-topes of MRP2 from several species, including

human, monkey, dog, rabbit, rat and mouse,

served to localize the MRP2 glycoprotein to the

apical membrane of polarized cells, including

hepatocytes (Büchler et al., 1996; Keppler and

Kartenbeck, 1996), kidney proximal tubules

(Schaub et al., 1997, 1999), intestinal epithelia

(Fromm et al., 2000; Mottino et al., 2000;

van Aubel et al., 2000), gallbladder (Rost et al.,

2001) and lung The apical localization of MRP2and its broad substrate specificity for variousconjugates qualify this ATP-binding cassette(ABC) transporter as an important terminal com-ponent in detoxification, subsequent to the phase

I and phase II reactions of xenobiotic olism The latter is comprised predominantly

metab-of cytochrome P450-catalyzed oxidations andconjugation reactions catalyzed by varioustransferases Hepatocytes and kidney proximaltubule epithelia are the major sites for detoxifica-tion and excretion of xenobiotics, and in both celltypes MRP2 contributes to the vectorial trans-port of these substances In addition, in the liver,MRP2 contributes to the bile-salt-independentbile flow, as evidenced by the strong reduction inbile flow in mutant rats lacking the Mrp2 protein

(Jansen et al., 1985; Keppler and König, 1997) In

the intestine, the apical localization of MRP2may counteract the entry of toxic or carcino-genic MRP2 substrates from the intestinal lumeninto the epithelia and into the blood circulation

(Dietrich et al., 2001) Thus, in the intestinal

tract, MRP2 may have a similar protective role asproposed previously for MDR1 P-glycoprotein

(Benet et al., 1999) In addition to this protective

role, MRP2 has been shown directly to conferresistance to several chemotherapeutic agents

including cisplatin (Cui et al., 1999).

A hereditary defect of the hepatobiliary ination of anionic conjugates has long been

elim-ABC Proteins: From Bacteria to Man ISBN 0-12-352551-9

Copyright 2003 Elsevier Science Ltd All rights of reproduction in any form reserved

20

P UMP FOR A NIONIC

known in human Dubin–Johnson syndrome

(Dubin and Johnson, 1954; Sprinz and Nelson,

1954), and in several animal species (for review

see Roy Chowdhury et al., 1994), including

two different mutant rat strains (Jansen et al.,

1985; Takikawa et al., 1991) These mutant rat

strains served as models which facilitated the

localization of the defect in ATP-dependent

conjugate transport to the hepatocyte

canalicu-lar membrane (Ishikawa et al., 1990; Kitamura

et al., 1990) and enabled the cloning of the rat

Mrp2 cDNA (Büchler et al., 1996; Ito et al., 1997;

Mayer et al., 1995; Paulusma et al., 1996)

The absence of functional MRP2 protein from

human liver has been recognized as the cause

of Dubin–Johnson syndrome (Kartenbeck et al.,

1996; Keppler and Kartenbeck, 1996) and many

naturally occurring mutations in the MRP2

(ABCC2) gene have been identified (Mor-Cohen

et al., 2001; Paulusma et al., 1997; Tsujii et al.,

1999; Wada et al., 1998) Mutations and

poly-morphisms in the human MRP2 gene that affect

MRP2 function may be relevant for adverse drug

reactions because of an impaired hepatobiliary

and renal clearance of anionic drug conjugates

Moreover, such polymorphisms may also affect

the oral bioavailability of drugs that are

sub-strates for intestinal MRP2 or become subsub-strates

following conjugation inside intestinal epithelia

The first cDNA fragment of Mrp2 (Abcc2;

for-merly described as cMrp and cMoat) was

iden-tified in 1995 in a comparative analysis of

normal and transport-deficient GY/TR⫺mutant

rat liver (Mayer et al., 1995) Using degenerate

oligonucleotides complementary to human

MRP1 mRNA, an MRP1-related 347 bp cDNA

fragment was amplified from normal rat liver

but not from RNA from transport-deficient

liver (Mayer et al., 1995) Subsequently, the

full-length cDNA encoding an MRP1-related

protein now known as Mrp2 was cloned and

further analyzed (Büchler et al., 1996; Paulusma

et al., 1996) At present, the full-length MRP2

cDNA sequences and deduced amino acid

sequences from five mammalian species areknown, including the orthologues from human,

rat, rabbit, mouse and dog (Büchler et al., 1996; Conrad et al., 2001; Fritz et al., 2000; Paulusma

et al., 1996; Taniguchi et al., 1996; van Aubel

et al., 1998) These five mammalian MRP2

pro-teins are highly homologous, with amino acididentities ranging from 77% for the identitybetween the MRP2 proteins from rat and dog, to87% identity for the proteins from rat andmouse Furthermore, MRP2-related sequences

from other organisms including Caenorhabditis elegans (Broeks et al., 1996) and the plant Arabidopsis thaliana (Rea et al., 1998) (Chapter 17)

have been described and, in part, functionallycharacterized Within the human MRP (ABCC)subfamily, MRP2 shows the highest degree ofsimilarity to MRP1 with 48% identity (Cole

et al., 1992), followed by MRP3 with 47% tity (Kiuchi et al., 1998), and MRP6 with 38% identity (Kool et al., 1999) The lowest degree

iden-of amino acid identity was found betweenMRP2 and MRP8 (GenBank accession

XM_040766) and CFTR (Riordan et al., 1989)

with 29% and 26% identity, respectively

The identity of human MRP2 with respect to

MDR1 (Ambudkar et al., 1999), a member of

the P-glycoprotein (ABCB) subfamily, is only18%, underlining a major difference between theABCB and the ABCC transporter subfamilies.Differences between the proteins belonging tothe two subfamilies are also apparent based

on studies of the membrane topology of thesetransporters In contrast to the typical organi-zation described for members of the ABCB subfamily with two transmembrane domainsand two ATP-binding domains, MRP2, as well

as MRP1, MRP3, MRP6 and MRP7, contains anadditional NH2-proximal membrane-spanning

domain (Figure 20.1) (Borst et al., 1999; Büchler

et al., 1996; Hipfner et al., 1997; König

et al., 1999a) This additional domain is

repre-sented by an extension of approximately 200amino acids when compared with the length ofthe ABCB subfamily members Another strik-ing feature of MRP2 was found in studies onthe localization of the NH2-terminus Owing to

an odd number of predicted transmembranehelices, the NH2-terminus was predicted to beextracytosolic on the basis of computational

analysis by the TMAP program (Büchler et al., 1996) (see section on mutations in the MRP2

gene) This was recently directly established byimmunofluorescence microscopy studies using

an antibody directed against the NH2-terminus

of MRP2 (Cui et al., 1999) The extracellular

localization of the NH2-terminus of MRP1 was

also demonstrated by glycosylation site

muta-tional studies and epitope insertion experiments

(Hipfner et al., 1997; Kast and Gros, 1998) and,

based on sequence similarities, it is expected

that MRP3 and MRP6 will be the same

The human MRP2 gene has been localized

to chromosome 10q23–q24 (Taniguchi et al.,

1996) It spans approximately 65 kbp and

con-tains 32 exons with a high proportion of class 0

introns (Tsujii et al., 1999) The size of coding

exons ranges from 56 bp (exon 6) to 255 bp

(exon 10), and each nucleotide-binding domain

is encoded by three exons (Toh et al., 1999;

Tsujii et al., 1999) Comparison of the genomic

organization of the human MRP2 and MRP1

genes shows that they display remarkable

sim-ilarities as indicated by size and number of

exons (Grant et al., 1997) Furthermore, human

MRP1, MRP2 and MRP3 (GenBank accession

AC004590) have 21 identical splice sites based

on an amino acid alignment of the three

cog-nate proteins (Tsujii et al., 1999) Despite the

fact that these three human MRP family

mem-bers share a relatively moderate degree of amino

acid identity, their similar genomic tion suggests a close evolutionary relationship,possibly originating from gene duplication

rat (Büchler et al., 1996) and human

hepato-cytes (Keppler and Kartenbeck, 1996; Paulusma

et al., 1997), and was, therefore, termed cular MRP (cMRP) (Büchler et al., 1996), or

canali-canalicular multispecific organic anion

trans-porter (cMOAT) (Paulusma et al., 1996, 1997;

Taniguchi et al., 1996) Soon after the cloning

Figure 20.1 A predicted membrane topology model for human MRP2 Amino acids in the

nucleotide-(ATP)-binding domains are indicated, with the Walker A and B motifs in red and the ABC

transporter family signature in blue Mutated amino acids in patients with Dubin–Johnson syndrome

are indicated as white stars on the polypeptide chain, whereas splice site mutations are indicated as

pentagonal stars near the polypeptide chain ( See Chapter 3 for detailed discussion on topologies).

of MRP2 from rat and human liver, MRP2

mRNA was also detected in kidney, duodenum

and peripheral nerve (Kool et al., 1997; Schaub

et al., 1997; van Aubel et al., 1998) Kidney

proximal tubule epithelia are a particularly

good example for the extrahepatic apical

local-ization of MRP2 (Schaub et al., 1997, 1999).

Table 20.1summarizes the tissues from

differ-ent mammalian species in which MRP2 mRNA

and/or protein has been detected so far

Localization of MRP2 in the apical membrane of

hepatocytes, kidney proximal tubules, and

epithelial cells of gallbladder, small intestine,

colon and lung was confirmed by

immunofluo-rescence microscopy or immunohistochemistry

(Table 20.1 and Figure 20.2) In the placenta,

MRP2 is localized to the apical

syncytiotro-phoblast membrane (St-Pierre et al., 2000) In

addition to the apically localized MRP2,

polar-ized cells express other MRP homologues

in the basolateral membrane (Keppler et al.,

2001) For example, MRP6 is highly expressed

in rat liver (Hirohashi et al., 1998; Madon et al.,

2000), and in human liver and kidney (Kool

et al., 1999), and is localized to the basolateral

membrane of human hepatocytes and kidney

proximal tubule epithelia (Figure 20.2).

The distribution of MRP2 within an organmay change during different pathophysiolog-ical conditions For example, Mrp2 is homo-geneously distributed throughout a lobule innormal rat liver; however, cholestasis causesMrp2 to concentrate near the central (perive-

nous) area of the liver lobule (Paulusma et al.,

2000) On the subcellular level, selectiveretrieval of Mrp2 from the canalicular mem-brane to pericanalicular vesicles of rat hepa-tocytes has been observed as an early event

of cholestasis by immunofluorescence

micro-scopy (Dombrowski et al., 2000; Kubitz et al., 1999; Rost et al., 1999; Trauner et al., 1997) and immunogold electron microscopy (Beuers et al., 2001; Dombrowski et al., 2000).

Several cell lines have been used for studies

of MRP2 function in intact cells Rat and humanhepatoma cells express endogenous MRP2 inthe apical membrane surrounding apical vac-uoles or bile canaliculus-like structures between

adjacent cells (Cantz et al., 2000; Nies et al.,

1998) Secretion of fluorescent MRP2 substratesinto these apical vacuoles can be observed by

fluorescence microscopy (Figure 20.3) Caco-2

cells, derived from a human colon carcinoma,also express endogenous MRP2 in the apical

TABLE20.1 EXPRESSION AND LOCALIZATION OFMRP2 IN NORMAL TISSUES

membrane (Bock et al., 2000; Walgren et al.,

2001) Transfection of rat or human MRP2

cDNA into Madine–Darby canine kidney

(MDCKII) cells also results in apical localization

of the respective recombinant MRP2 protein

(Cui et al., 1999; Evers et al., 1998) Because

Caco-2 cells and MRP2-expressing MDCKII

cells grow in a polarized fashion on special

membrane filters, these cells are often used as a

model system for studies on the uptake,

trans-cellular transport, and MRP2-mediated export

of substances by intact cells (Cui et al., 2001).

The clinical relevance of MRP2 as an dependent export pump for chemotherapeutic

ATP-agents has not yet been extensively investigated

but has been supported by the localization ofMRP2 in the plasma membrane of renal clear-cell

(Schaub et al., 1999), ovarian (Arts et al., 1999), colorectal (Hinoshita et al., 2000), and hepatocel- lular (Nies et al., 2001) carcinoma cells MRP2

expression was also detected by reverse scriptase (RT)-PCR and immunoblotting in celllines from lung, gastric, uterine and colorectal

tran-cancers (Kool et al., 1997; Minemura et al., 1999;

Narasaki et al., 1997; Young et al., 1999, 2001).

Thus, MRP2 is expressed in several malignanttumor types and may contribute to their resist-ance to a wide variety of antitumor drugs, as

demonstrated in vitro in MRP2-transfected cells (Cui et al., 1999).

Figure 20.2 Localization of MRP2 (red in A–D) and MRP6 (green in A and B) by immunofluorescence in

different human tissues Double-label (A, B) and single-label (C, D) immunofluorescence microscopy of

frozen tissue sections (5 ␮m thickness) was performed as described (König et al., 1999b) Pictures were taken

by confocal laser scanning microscopy (A, B, D) or by conventional fluorescence microscopy (C) MRP2 is

localized to the apical membrane of human hepatocytes (A), proximal tubule epithelia in the kidney (B),

epithelia of the colon (C), and of bronchial epithelia (D) as detected either with the monoclonal antibody

M 2 III-6 (Paulusma et al., 1997) in A, B, D or with the polyclonal antiserum EAG5 in C (Cui et al., 1999;

Schaub et al., 1999) The isoform MRP6 is localized to the basolateral membrane of hepatocytes (A) and

proximal tubule epithelia (B) as detected with the antiserum AQL (König et al., 1999b) Lu, lumen Bars in

A–D, 50 ␮m.

F UNCTIONAL A NALYSIS

Elucidation of the physiological function of

MRP1, the first identified member of the MRP

family (see Chapter 19), was closely linked to

the characterization of the membrane proteins

mediating the ATP-dependent transport of

the endogenous glutathione S-conjugate, LTC4

(Keppler, 1992; Leier et al., 1994b) The search

for the molecular identity of the

ATP-depend-ent LTC4 transporter localized to the

hepato-cyte canalicular membrane (Ishikawa et al.,

1990) led subsequently to the identification of

MRP2 (see above) (Büchler et al., 1996; Paulusma et al., 1996; Taniguchi et al., 1996).

Previously, ATP-dependent transport ments using inside-out hepatocyte canalicularmembrane vesicles from normal and Mrp2-deficient GY/TR⫺ rats (Ishikawa et al., 1990)

measure-or EHBR rats (Fernandez-Checa et al., 1992; Takenaka et al., 1995) have been invaluable in

elucidating the substrate specificity of Mrp2

(Keppler and Kartenbeck, 1996; König et al.,

1999a) Since cell lines stably expressing binant rat or human MRP2 became available

recom-(Chen et al., 1999; Cui et al., 1999; Evers et al., 1998; Ito et al., 1998), the substrate specificity of MRP2

has been studied under more defined conditionsusing inside-out membrane vesicles preparedfrom these transfected cells Human MRP2 has

Figure 20.3 Fluo-3 as a fluorescent substrate for MRP2 A, Structure of the fluorescent organic anion

Fluo-3 (Kao et al., 1989; Minta et al., 1989) B, ATP-dependent transport of Fluo-3 into Mrp2-containing canalicular membrane vesicles isolated from rat hepatocytes Vesicle-associated fluorescence was

determined fluorometrically as described; a Km value of 3.7 ␮M was obtained for Fluo-3 as the substrate

for rat Mrp2 (Nies et al., 1998) C, Vectorial transport of Fluo-3 into apical vacuoles of human HepG2 cells Polarized HepG2 cells, derived from hepatocellular carcinoma, were incubated with the non-fluorescent acetoxymethyl ester of Fluo-3, which was taken up by the cells and hydrolyzed to the fluorescent Fluo-3 Fluo-3-filled apical vacuoles (arrowheads pointing to green fluorescence) were observed by fluorescence microscopy as described (Cantz et al., 2000) D, Single Fluo-3-filled vacuole (green) is shown after

immunostaining of MRP2 (red) Merging both fluorescences demonstrates secretion of Fluo-3 into the

apical vacuole of polarized HepG2 cells Bar in C, 25 ␮m; in D, 5 ␮m.

also been purified to homogeneity, and shown to

exhibit substrate-stimulated ATPase and

trans-port activity when reconstituted in

proteolipo-somes (Hagmann et al., 1999, 2002).

Various labeled substrates for MRP2 as well

as for MRP1 are useful in assessing the

trans-port function of both proteins [3H]LTC4 has

become the preferred substrate for transport

measurements because of its high affinity for

both MRP1 and MRP2 (Cui et al., 1999; Leier

et al., 1994a), and its commercial availability.

ATP-dependent transport of [3H]LTC4 into

membrane vesicles is measured by incubating

the labeled substrate with the inside-out

mem-brane vesicles for the desired time period and

subsequently separating membrane vesicles

from extravesicular labeled substrate by rapid

filtration through nitrocellulose filter

mem-branes (Keppler et al., 1998) With substrates

more hydrophobic than [3H]LTC4, which bind

strongly to the filters and to the membrane

vesicles, small Sephadex G-50 columns (Böhme

et al., 1993) or glass filters (Loe et al., 1996) have

been used for the separation of vesicles and

labeled substrate

The apical export pump MRP2 shares a verysimilar substrate spectrum with MRP1 High-

affinity substrates for MRP2 include amphiphilic

anions, particularly those conjugated with

glu-tathione and glucuronate, such as LTC4,

bili-rubin glucuronosides, and 17␤-glucuronosyl

estradiol (Table 20.2) The comparison of both

recombinant proteins shows that MRP1 has a

10-fold higher affinity for LTC4 and a 5-fold

higher affinity for 17␤-glucuronosyl estradiol

than MRP2 (Cui et al., 1999), whereas

mono-and bisglucuronosyl bilirubin are preferred

substrates for MRP2 (Kamisako et al., 1999).

MRP2 is also able to transport non-conjugated

compounds such as the penta-anionic

fluo-rescent dye Fluo-3, the model compound for

hepatic transport studies sulfobromophthalein,

the anionic anticancer drug methotrexate, the

HMG-CoA reductase inhibitor pravastatin, and

the angiotensin-converting enzyme inhibitor

temocaprilat (Table 20.2).

Fluorescent substrates, such as the Ca2⫹indicator Fluo-3, are useful for studies of

-MRP2 function in intact cells (Cantz et al., 2000;

Nies et al., 1998) The fluorescent glutathione

S-conjugates glutathione bimane (Oude Elferink

et al., 1993; Roelofsen et al., 1995) and

glu-tathione methylfluorescein (Roelofsen et al.,

1998) are also likely substrates for Mrp2 because

they are not transported from hepatocytes

into bile of Mrp2-deficient mutant rats The

ATP-dependent transport of Fluo-3 by Mrp2was demonstrated using inside-out membranevesicles from normal and Mrp2-deficient rat

hepatocytes (Nies et al., 1998) (Figure 20.3) and

shown to represent the predominant exportpump mediating extrusion of Fluo-3 across theapical membrane of polarized cells Polarized

rat (Ihrke et al., 1993) and human (Sormunen

et al., 1993) hepatoma cells form apical

vac-uoles between adjacent cells and express MRP2

in the apical membrane (Cantz et al., 2000; Nies

et al., 1998) Secretion of Fluo-3 and other

fluo-rescent anions into these apical vacuoles is

read-ily observed by fluorescence microscopy (Figure 20.3) MRP2-mediated Fluo-3 secretion into

apical vacuoles is inhibited by cyclosporin Abut not by the selective MDR1 P-glycoproteininhibitor LY335979 Recently, probenecid-sensi-tive efflux of carboxyfluorescein has been usedfor estimation of MRP2 transport activity in

intact cells (Mor-Cohen et al., 2001).

api-tine, colon and bronchia (Table 20.1, Figure 20.2)

suggests that a physiological function of MRP2

is to excrete endogenous metabolites and biotics, and to prevent toxic compounds fromentering the body It has been reported that theMrp2-deficient GY/TR⫺rats have a much lowerexcretion rate for the food-derived carcinogen2-amino-1-methyl-6-phenylimidazo[4,5-b]

xeno-pyridine (PhIP) and its glucuronate conjugatecompared to wild-type Wistar rats (Dietrich

et al., 2001) MRP2 has also been shown to act synergistically with the glutathione S-transferase

GST␲1-1 in the detoxification of the cytotoxicand genotoxic agent 4-nitroquinoline 1-oxide

(4-NQO) (Morrow et al., 2001).

Because of the similar substrate spectrum ofMRP2 and MRP1, which has been shown toconfer resistance to different anticancer drugswhen overexpressed in mammalian cells (see Chapter 19)(Cole et al., 1994; Grant et al., 1994),

it has been proposed that MRP2 may also conferdrug resistance by pumping drug conjugates ordrug–glutathione complexes out of the cell

Northern blot analyses and RNase protection

TABLE20.2 SUBSTRATE SPECIFICITY OF HUMAN AND RATMRP2

MRP2 (human recombinant)

S-Glutathionyl 2,4-dinitrobenzene 6.5 Evers et al., 1998

S-Glutathionyl ethacrynic acid Evers et al., 1998

Bilirubin

17␤-Glucuronosyl estradiol 7.2 Cui et al., 1999

p-Aminohippurate 880 Leier et al., 2000

Mrp2 (rat recombinant)

S-Glutathionyl 2,4-dinitrobenzene 0.2 Ito et al., 1998

Bilirubin

17 ␤-Glucuronosyl estradiol 6.9 Cui et al., 1999

Sulfatolithocholyl taurine 3.9 Akita et al., 2001

Mrp2 (rat; normal/mutant BCM)a

N-Acetyl LTE4 5.2 Ishikawa et al., 1990

S-Glutathionyl 2,4-dinitrobenzene Ishikawa et al., 1990

S-Glutathionyl sulfobromophthalein Ishikawa et al., 1990

Bilirubin

Jedlitschky et al., 1997

Glucuronosyl nafenopin Jedlitschky et al., 1994

Glucuronosyl grepafloxacin 7.2 Sasabe et al., 1998

Glucuronosyl SN38 carboxylatea Chu et al., 1997

Glucuronosyl SN38 lactonea Chu et al., 1997

Sulfatolithocholyl taurine 1.5 Akita et al., 2001

Sulfatochenodeoxycholyl taurine 8.8 Akita et al., 2001

Compounds listed have been identified as substrates by measurement of their ATP-dependent transport into inside-out membrane vesicles from cells expressing the recombinant MRP2/Mrp2 in comparison with membrane vesicles from control vector-expressing cells In addition, measurement of ATP-dependent transport into hepatocyte canalicular membranes vesicles from Mrp2-deficient mutant rats (GY/TR⫺and EHBR) compared with those from normal rats are presented (summarized by

König et al., 1999a).

aAbbreviations: BCM, bile (hepatocyte) canalicular membranes; E3040, 2-methylamino-4-(3-pyridylmethyl)benzothiazole; Fluo-3, 1-[2-amino-5-(2,7-dichloro-6-hydroxy- 3-oxo-3H-xanthen-9-yl)]-2-(2⬘-amino-5⬘-methyl-phenoxy)-ethane-N,N,N⬘,N⬘,-tetraacetic acid penta ammonium salt; SN38, de-esterified metabolite of CPT11 (7-ethyl-10-hydroxycamptothecin).

6-hydroxy-5,7-dimethyl-assays indicate that a correlation between MRP2

expression and multidrug resistance may exist

(Kool et al., 1997; Taniguchi et al., 1996)

MRP2-related drug resistance was also suggested

by its cloning from the cisplatin-resistant

human cancer cell lines KB-CDP4 and P-CDP5

(Taniguchi et al., 1996) In both cisplatin-resistant

cell lines, MRP2 mRNA was overexpressed

relative to non-resistant parental cell lines A

correlation between cisplatin resistance and

MRP2 expression was also demonstrated for

several additional cell lines by RNase

protec-tion assays and Northern blot analyses (Kool

et al., 1997; Minemura et al., 1999) Finally, drug

resistance conferred by MRP2 has also been

demonstrated by the use of antisense cDNA,

studied in the human hepatoma cell line HepG2

(Koike et al., 1997) The amount of MRP2

mRNA was reduced, leading to elevated

intra-cellular glutathione levels and enhanced

sensi-tivity to anticancer drugs including cisplatin,

vincristine, doxorubicin and the camptothecin

derivatives CPT11 and SN38

Direct evidence for MRP2-mediated drug resistance was obtained by transfection

multi-studies with MRP2 cDNA (Cui et al., 1999;

Hooijberg et al., 1999) In MRP2-transfected

MDCKII cells and HEK293 cells, MRP2 was

localized to the plasma membrane, which is a

prerequisite for the measurement of

MRP2-mediated resistance (Cui et al., 1999) Expression

of both recombinant rat and human MRP2 inMDCKII and HEK293 cells leads to significantresistance to cisplatin, etoposide, vincristine,

doxorubicin and epirubicin (Table 20.3) The

ability of MRP2 to confer resistance to cisplatinhas also been shown in MRP2-transfected LLC-

PK1 cells (Chen et al., 1999) Moreover, MRP2

confers resistance to the antifolate ate in transfected human ovarian carcinoma

methotrex-2008 cells (Hooijberg et al., 1999) The possible

clinical relevance of MRP2-mediated drugresistance was further suggested by the detec-tion of MRP2 in various carcinoma samples

Naturally occurring mutations in the MRP2 gene

have been discovered in humans (Mor-Cohen

et al., 2001; Paulusma et al., 1997; Toh et al., 1999; Tsujii et al., 1999; Wada et al., 1998) and rat (Ito et al., 1997; Paulusma et al., 1996).

Some of these mutations were shown to beassociated with the absence of the MRP2 protein from the hepatocyte canalicular mem-

brane (Büchler et al., 1996; Kartenbeck et al.,

1996) In humans, the Dubin–Johnson drome, originally described in 1954, is an autoso-mal recessively inherited disorder characterized

syn-TABLE20.3 MRP2-MEDIATED RESISTANCE TO ANTICANCER DRUGS IN TRANSFECTED CELLS

I Stably transfected MDCKII cells

II Stably transfected HEK293 cells

by conjugated hyperbilirubinemia (Dubin and

Johnson, 1954; Sprinz and Nelson, 1954) The

liver of patients with Dubin–Johnson

syn-drome appears dark blue or black because of

deposition of a dark pigment in the

pericanalic-ular area (Roy Chowdhury et al., 1994) The

deficient transport of anionic conjugates,

includ-ing monoglucuronosyl bilirubin and

bisglu-curonosyl bilirubin, from hepatocytes into bile

is caused by the absence of the MRP2 protein

from the canalicular membrane (Kartenbeck

et al., 1996; Keppler and Kartenbeck, 1996;

Paulusma et al., 1997; Tsujii et al., 1999)

Estab-lished mutations identified in patients with

Dubin–Johnson syndrome include splice site

mutations leading to exon deletions with

sub-sequent premature termination codons (Kajihara

et al., 1998; Toh et al., 1999; Wada et al., 1998),

missense mutations (Mor-Cohen et al., 2001; Toh

et al., 1999), a nonsense mutation leading to a

premature termination codon (Paulusma et al.,

1997; Tsujii et al., 1999), and a deletion

muta-tion leading to the loss of two amino acids

in the second nucleotide-binding domain,

(Tsujii et al., 1999) (Figure 20.4 and Table 20.4).

Interestingly, all mutations identified so far are

located in the COOH-proximal half of theMRP2 protein and only two of them are located

in a predicted extracellular loop (Figures 20.1 and 20.4).

The MRP2 membrane topology has been predicted using several different algorithmsincluding TMpred (http://www.ch.embnet.org/software/TMPRED_form.html), TopPred2(http://bioweb.pasteur.fr/seqanal/interfaces/toppred.html), and SOSUI (http://sosui.pro-teome.bio.tuat.ac.jp/sosuiframe0.html) Thesealgorithms predict four, or at most five, trans-membrane helices for the region between thefirst and the second ATP-binding domain ofhuman MRP2 None of the programs predictsix transmembrane helices for this domain ofMRP2 Because an even number of transmem-brane helices is required between the twocytosolic ATP-binding domains, we have usedthe four-transmembrane-helix topology for theprediction of the location of mutations andpolymorphisms in the COOH-proximal por-

tion of the protein (Figures 20.1 and 20.4).

In addition to the known mutations in Dubin–Johnson syndrome, seven base pair changes, of

which five are in the coding region of the MRP2

Figure 20.4 Schematic membrane topology of human MRP2 with the locations of intron–exon boundaries

of the MRP2 gene indicated in yellow Yellow numbers indicate the number of the exon encoding the

NH 2 -proximal amino acid sequence Thus, the COOH-terminal amino acid sequence is encoded by exon 32 Mutations causing Dubin–Johnson syndrome or polymorphisms are indicated in red as detailed in Table 20.4.