Liver International 2005: 25: 367–379 Printed in Denmark. All rights reserved
Copyright r Blackwell Munksgaard 2005
Basic Studies
DOI: 10.1111/j.1478-3231.2005.01033.x
Hepatobiliary transporter expression in human hepatocellular carcinoma
Zollner G, Wagner M, Fickert P, Silbert D, Fuchsbichler A, Zatloukal K, Denk H, Trauner M. Hepatobiliary transporter expression in human hepatocellular carcinoma. Liver International 2005: 25: 367–379. r Blackwell Munksgaard 2005 Abstract: Background/Aims: Treatment of hepatocellular carcinoma (HCC) is hampered by resistance to chemotherapy, which might be mediated by multidrug resistance P-glycoproteins (MDR P-gps) and MDR-associated proteins (MRPs). The effectiveness of cytostatics could be further impeded by reduced hepatocellular drug uptake into HCCs. Therefore, we aimed to determine P-gp, MRP and organic anion transporting protein OATP2 (SLC21A6) expression in HCC. Furthermore, we investigated expression of the major bile salt uptake system Na1/taurocholate cotransporter NTCP (SLC10A1), since bile salt-coupled chemotherapeutics were proposed to increase therapeutic drug enrichment in HCC. Material/Methods: mRNA and protein expression and tissue distribution of P-gps, MRPs, OATP2 and NTCP were assessed in HCC and peritumorous non-neoplastic tissue by reverse transcription polymerase chain reaction, Western blotting and immunohistochemistry, respectively. Results: Expression of P-gps (multidrug export pump MDR1 (ABCB1), phospholipid flippase MDR3 (ABCB4), sister of P-glycoprotein SPGP (ABCB11)) and basolateral MRP homologue MRP3 (ABCC3) showed a trend for decreased levels in HCC but was highly variable among individual tumors. In contrast, canalicular conjugate export pump MRP2 (ABCC2) expression was generally maintained or even showed a trend towards increased levels. NTCP and OATP2 expression was markedly reduced in most HCCs (Po0.05). Expression of the genuine drug transporter, the concentrative nucleoside transporter (CNT1), was highly variable and showed a trend for reduced levels in HCC. Summary/Conclusions: MRP2 seems to be the major candidate transporter involved in chemoresistance and reduced expression of OATP2 may further contribute to low drug accumulation in HCCs. Overexpression of drug exporters is not a general feature of HCC but could account for chemoresistance of individual cases. Since expression of uptake systems is generally reduced in HCC, bile salt-coupled therapeutics may not represent a suitable strategy to overcome insufficient drug enrichment.
Hepatocellular carcinoma (HCC) is the most frequent primary malignancy of the liver, with Abbreviations: BSEP, bile salt export pump, also called sister of P-glycoprotein (SPGP, ABCB11); CNT1, concentrative nucleoside transporter 1 (SLC28A1); GAPDH, glyceraldehyde-3phosphate dehydrogenase; HCC, hepatocellular carcinoma; MDR, multidrug resistance; MDR1, multidrug export pump (ABCB1); MDR3, phospholipid flippase (ABCB4); MRP2, canalicular conjugate export pump (ABCC2); MRP3, basolateral MRP homologue (ABCC3); NTCP, Na1/taurocholate cotransporter (SLC10A1); OATP2, organic anion transporting protein (SLC21A6); P-gp, P-glycoprotein; RT-PCR, reverse transcription polymerase chain reaction.
Gernot Zollner1, Martin Wagner1, Peter Fickert1, Dagmar Silbert1, Andrea Fuchsbichler2, Kurt Zatloukal2, Helmut Denk2, Michael Trauner1 1 Division of Gastroenterology and Hepatology, Department of Medicine, and 2Institute of Pathology, Medical University, Graz, Austria
Key words: bile salt-coupled drugs – chemoresistance – multidrug resistance Michael Trauner, MD, Laboratory of Experimental and Molecular Hepatology, Division of Gastroenterology and Hepatology, Department of Internal Medicine, Medical University, Auenbruggerplatz 15, A-8036 Graz, Austria. Tel: 143-316-385-2863 Fax: 143-316-385-3062 e-mail:
[email protected] Received 10 December 2003, accepted 13 July 2004
increasing incidence accounting for as many as one million deaths annually worldwide, (1–3). Orthotopic liver transplantation (OLT) or liver resection represents the best treatments for HCC (4). However, most patients cannot be subjected to potential curative OLT or resection because of extensive tumor involvement of the liver, metastasis, invasion of the hepatic or portal vein or advanced underlying hepatocellular disease at the time of diagnosis (5, 6). Systemic chemotherapy is not considered for these patients since this is associated with low response rates (1). Chemoembolization represents a novel treatment option 367
Zollner et al. with the possible advantage of delivering high concentrations of chemotherapeutics to the tumor with less systemic side effects (1). However, response rates and patient benefits of chemoembolization and systemic chemotherapy are not satisfactory (1). Drug therapy of cancer in general is hampered by multidrug resistance (MDR) whereby tumor cells possess cross-resistance to diverse chemotherapeutic agents (7). Molecular investigations of MDR resulted in the characterization of genes coding for several cellular drug efflux transporters such as MDR P-glycoproteins (Pgps) and MDR-associated proteins (MRPs) (7). Expression of P-gps and MRPs was detected in various hematological and solid malignancies and expression levels correlated with clinical resistance to chemotherapy (8–19). Overexpression of P-gps and MRPs might also be involved in mediating chemoresistance of HCC (20–22). Members of the P-gp (multidrug export pump MDR1 (ABCB1), phospholipid flippase MDR3 (ABCB4) and bile salt export pump, also called sister of P-glycoprotein (SPGP/BSEP/ABCB11) and MRP (canalicular conjugate export pump MRP2 (ABCC2), basolateral MRP homologue MRP3 (ABCC3)) transporter family are expressed in hepatocytes already under physiologic conditions and play an important role in bile formation and excretion of various toxic substances, including drugs and their metabolites (23). As such, MDR1 mediates excretion of various hydrophobic cationic substances including anticancer drugs (Fig. 1) (23). MDR3 acts as a phospholipid flippase but may also be involved in drug transport (Fig. 1) (24, 25). Another member of the P-gp family, the sister of P-gp (SPGP), represents the main canalicular bile salt export pump (BSEP), and is also capable of transporting drugs (Fig. 1) (24, 26). Most prominent members of the MRP family are the canalicular multispecific organic anion transporter MRP2 and the basolateral MRP homologue MRP3. In vitro studies demonstrated that both can confer resistance to various chemotherapeutics (Fig. 1) (27, 28). Since these multidrug transporters are already expressed in the liver under physiologic conditions, one might expect that maintenance or even up-regulation of transporter gene expression could mediate chemoresistance of HCC. However, analysis of P-gp expression in HCC in vivo and in carcinoma or hepatoblastoma cell lines in vitro revealed divergent results and it still remains controversial as to whether P-gps are involved in drug resistance (20, 21, 29, 30). MRPs may be the more likely candidates in mediating chemoresistance of HCCs (22). 368
In contrast to MDR because of increased drug export via P-gps and MRPs, insufficient drug uptake may also contribute to chemoresistance of HCC. Organic anion transporting proteins (OATPs) such as OATP2/SCL21A6 (new nomenclature OATP1B1/SLCO1B1) play an important role in drug absorption and drug disposition (31). Reduced OATP2/SLC21A6 would thus lead to decreased drug uptake into neoplastic cells. Furthermore, hepatic uptake systems may also represent attractive targets for targeting of chemotherapeutics to HCC. Cytostatic drugs covalently bound to bile acids have been proposed as a novel approach to provide specific drug targeting to the liver, thus minimizing systemic side effects, and to overcome insufficient drug enrichment in HCCs (32–36). Uptake of drugs linked to bile acids as carrier molecules would strictly depend on preserved expression of the Na1/taurocholate cotransporter (NTCP/SLC10A1) but also OATP2/SLC21A6 (33, 34, 37, 38). Residual bile acid uptake was detectable in hepatoblastoma cells but at lower levels compared with normal liver (39). This would favor the distribution of compound drugs to normal rather than malignant hepatocytes. However, there is only limited information about expression of NTCP or OATP2/SLC21A6 in HCCs (37, 40, 41). The present study therefore aimed (i) to investigate the expression of P-gps, MRPs and OATP2 as potential candidates mediating MDR of HCCs and (ii) to determine whether expression of NTCP is preserved in HCCs. Expression of NTCP but also OATP2 is of particular interest since chemotherapeutics linked to bile acids could overcome insufficient accumulation of cytostatic drugs within tumor cells. Material and methods
Liver specimens and patient characteristics
HCC and peritumorous non-neoplastic liver tissue were derived from 10 patients during liver resection in the period 1996–1999. The limited number of studied cases is because of the two major recent developments that interfere with sufficient tissue sampling in HCC: (i) Liver biopsy is no longer recommended in HCC fullfilling standard biochemical and radiological criteria (42) and (ii) liver transplantation with preceding chemoembolization rather than liver resection is performed for curative treatment of HCC (1). Tissue was immediately snap-frozen and stored in liquid nitrogen until RNA extraction, Western blot analysis and immunohistochemistry. HCCs were classified according to histological grade
Transporter expression in HCC MDR1 BS , OA , OC MTX , BS -coupled drugs
OATP2
X
OC Taxol Anthracyclines Vinca alkaloids
MDR3
Phospholipids Vinblastine
NTCP Na
OA Cisplatin Anthracyclines Vincristine Vinblastine MTX
BS BS -coupled drugs
MRP3
OA Etoposide MTX
BS Vinblastine
MRP2
SPGP ( BSEP)
Fig. 1. Hepatobiliary transport systems investigated in hepatocellular carcinoma (HCC) in this study. Schematic representation showing a hepatocellular carcinoma cell with the transport systems investigated. Members of the multidrug resistance P-glycoprotein (MDR P-gp) and multidrug resistance-associated protein (MRP) family are already expressed in normal hepatocytes where they play an important role for biliary excretory function; moreover, these efflux pumps may mediate chemoresistance of various malignancies: MDR1 (ABCB1) excretes various organic cations and drugs including cytostatic agents (in italic). MDR3 (ABCB4) is a phospholipid flippase and might also be involved in drug transport. Another member of the P-gp family possibly contributing to chemoresistance is the canalicular bile salt export pump, also known as sister of P-gp (SPGP/BSEP, ABCB11). MRP2/ABCC2 and MRP3/ABCC3 can also confer resistance to chemotherapeutics as they are able to transport various drugs including chemotherapeutics. In contrast to these export pumps, the Na1/taurocholate cotransporter (NTCP/SLC10A1) mediates uptake of bile salts while a family of Na1-independent organic anion transporting polypeptides (OATPs, with its most prominent member OATP2/SLC21A6 new nomenclature OATP1B1/SLC1B1) also transports (non-bile salt) organic anions, cations and drugs in exchange with anions. Bile-salt coupled chemotherapeutics were proposed in order to increase drug enrichment within HCCs, which would require specific uptake via these transporters. X , anions; BS , bile salts; MTX, methotrexate; OA , organic anions; OC1, organic cations.
and growth pattern (tubular, trabecular/pseudoglandular) (43). Five HCCs were grade I and other five were grade II with a trabecular (five HCCs) or a mixed trabecular and tubular (five HCCs) growth pattern. Patients did not receive any chemotherapy prior to resection, nor was transarterial embolization, chemoembolization, radiofrequency ablation or ethanol injection performed. All patients had given their informed consent for the study, and the experimental protocol was approved by the local Ethics Committee in accordance with the ethical guidelines of the 1975 Declaration of Helsinki. RNA extraction, determination of mRNA copy numbers and reverse transcription polymerase chain reaction (RT-PCR)
Total RNA was extracted according to a protocol by Krieg et al. (44). RNA standards with mutated restriction sites were used as internal controls for complementary DNA synthesis and subsequent amplification by polymerase chain reaction (PCR) as described previously (45). In brief,
reverse transcription of wild-type and standard RNA was performed simultaneously for MDR1, MDR3, SPGP (BSEP), MRP2, MRP3, NTCP and OATP2/SLC21A6 using primers published previously (45). For amplification of hepatocyte nuclear factor 1a (HNF1a) the following primers were used: forward 2367–2386, reverse 2920–2901 according to GenBank accession number M57732. After amplification by PCR, wildtype and standard amplicons were distinguished by restriction digestion and separation by agarose gel electrophoresis. Band intensities were detemined by video-densitometry (Intas DigitStores, Goettingen, Germany; RFLP-Scans or ZeroD-Scans software, Scanalytics, Billerica, MA) and transporter mRNA copy numbers were calculated by the ratio of standard to wildtype amplicon intensities and normalized to those of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as described previously (45). For quantification of concentrative nucleoside transporter 1 (CNT1/SLC28A1) mRNA levels, real-time PCR was performed as previously described (46). Reactions were performed in 369
Zollner et al. duplicate and repeated twice. PCR products were checked by sequencing and by gel electrophoresis for correct size and quality. For quantification the standard curve method was used. All data were normalized to GAPDH. The TaqMans oligonucleotides for CNT1 and GAPDH were as follows: CNT1: forward (fwd) primer 5 0 CGTGTTCGTCGCTCTCCTC-3 0 , reverse (rvs) primer 5 0 -CTCGAACGCAATGAATCCTGG3 0 , fluorogenic probe 5 0 -FAM-AGTGTCCTGG AGGGCCGTGTCTTG-3 0 -TAMRA corresponding to GenBank accession number NM_004213; GAPDH: fwd primer 5 0 -CCACATCGCTCA GACACCAT-3 0 , rvs primer 5 0 -ACCAGGCGCC CAATACG-3 0 , flourogenic probe 5 0 -FAM-CAA ATCCGTTGACTCCGACCTTCA-TAMRA-3 0 corresponding to GenBank accession number NM_002046. Preparation of liver membranes
Liver membranes were prepared as described previously with a minor modification (single 100 000 g centrifugation step of gauze-filtered liver homogenates) (47). Protein concentrations were determined using a Bradford kit (BioRad, Richmond, CA). Analysis of transporter protein expression by Western blotting
Similar amounts of protein (20 mg) were loaded onto a 7.5% (MDR1, MRP2, MRP3) or 10% (NTCP, OATP2/SLC21A6) sodium dodecyl sulfate–polyacrylamide gel and subsequently subjected to electrophoresis (48, 49). Western blotting was performed in four surgical liver specimens with sufficient amounts of tissue. Equal protein loading was confirmed by Coomassie blue staining of gels and Ponceau S staining of membranes after transfer (not shown). After electrotransfer to nitrocellulose membranes (BioRad), blots were blocked with Tris-buffered saline containing 0.1% Tween and 5% dry milk for 1 h at room temperature and incubated overnight at 4 1C with polyclonal antibodies against MDR1 (dilution, 1:250; Kamiya Biomedical Company, Seattle, WA), BSEP (dilution, 1:250; Kamiya Biomedical Company), MRP2 (dilution, 1:2000; Alexis Corporation, Lausen, Switzerland), MRP3 (dilution, 1:2000; Alexis Corporation), NTCP (dilution, 1:500; kindly provided by Dr. Bruno Stieger, Zurich, Switzerland) and OATP2/SLC21A6 (dilution, 1:2000; provided by Dr. Bruno Stieger). Blots were reprobed with an anti--actin antibody (dilution, 1:5000, Sigma, Steinheim, Germany) to confirm the specificity of changes in transporter protein levels. Immune370
complexes were detected using horseradish-conjugated goat anti-rabbit or sheep anti-mouse IgG F(ab 0 )2 fragments (dilution, 1:1000; Dako, Glostrup, Denmark) according to the ECL Western blotting detection system (Amersham, Buckinghamshire, UK) (47). After exposing the blots to Trimax XDA Plus (3 M Imation, Rochester, NY), band intensities were determined with an Intas Digit-Stores system and RFLP-Scans or ZeroD-Scans Software (Scanalytics). Immunohistochemistry and immunofluorescence microscopy
Immunohistochemistry was performed on frozen liver sections (4 mm thickness) fixed for 10 min with acetone and for 10 min with chloroform at room temperature using MDR1 (dilution, 1:50), C219 (dilution, 1:30; Signet, Dedham, MA, USA), BSEP (dilution, 1:100; provided by Dr. Bruno Stieger), and MRP2 (dilution, 1:25) and MRP3 antibodies (dilution, 1:50). Specific reactions were detected using the Dako ChemMates detection kit (Dako), alkaline phosphatase conjugated, using Fast Red (Boehringer Ingelheim, Heidelberg, Germany) as a indicator substrate. As immunohistochemical staining of NTCP did not render satisfactory results, immunofluorescence microscopy was performed using a NTCP antibody (dilution, 1:25; supplied by Dr. Bruno Stieger). Fluoresceine isothiocyanate-conjugated swine anti-rabbit Ig (Dako) was used as secondary antibody. Fluorescent staining was visualized using an MRC 600s (BioRad) scanning confocal device attached to a Zeiss Axiophot as described (45). Negative controls were performed by omitting the primary antibodies. Only samples that were prepared in parallel in all steps were compared. Statistical analysis
Data are reported as arithmetic means SEM unless stated otherwise. Differences in overall steady-state mRNA levels between HCCs and peritumorous non-neoplastic tissues were analyzed by Student’s t-test using the SigmaStats statistic program (Jandel Scientific, San Rafael, CA, USA). A P value of o0.05 was considered significant. Results
MDR P-glycoprotein expression is variable in HCC
MDR1 mRNA levels were decreased in 8 and increased in two HCCs compared with corresponding non-tumorous tissue, and overall MDR1 mRNA expression showed a trend
Transporter expression in HCC B MDR1 mRNA / GAPDH mRNA
A 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00
NT T NT T NT T NT T MDR1 9.6 1.0
0.2 1.0 0.5
1.0
7.0
1.0
1.5
0.8 1.0
1.2 1.0
1.5
β-actin
NT
C
1.0
1.0
T
D
E
Fig. 2. MDR1 expression in hepatocellular carcinoma (HCC). Total RNA and liver membrane fractions were isolated from HCC and perineoplastic tissue and analyzed by competitive reverse transcription polymerase chain reaction (A) and Western blotting (B) as described in Material and methods. In addition, MDR1 tissue distribution was analyzed by immunohistochemistry (C–E). (A) MDR1 steady-state mRNA expression in unaffected perineoplastic tissue (left part, non-tumorous, NT) and in the corresponding HCC (right part) are represented by the same symbol and are connected by lines. The horizontal bar represents mean values of overall mRNA levels of each group. (B) Representative Western blots of non-tumorous (NT) liver tissue and HCC (T). Symbols depicting the liver samples correspond to symbols used in (A). Data are expressed as the fold change compared to respective non-tumorous sample. (C) Regular canalicular immunohistochemical staining pattern for MDR1 in peritumorous non-neoplastic tissue. (D) Staining is preserved in HCC with a tubular (pseudoglandular) growth pattern and is positive on the luminal membranes of pseudoglandular structures. (E) In contrast, MDR1 staining is almost absent on tumor cells arranged in a trabecular pattern. Original magnification 20.
towards reduced levels in HCC (59 21% of non-tumorous control tissues) without reaching statistical significance (n.s.) (Fig. 2A). MDR3 mRNA levels were reduced in seven and increased in three HCCs compared with non-tumorous tissue, but overall mRNA expression did not change significantly in HCC (78 14% of non-tumorous tissues, n.s.) (not shown). Total liver membranes were extracted from four HCCs where sufficient amounts of tissue were available. Western blot analysis using a specific anti MDR1 antibody revealed decreased P-gp protein levels in two HCCs (23% and 54% of corresponding nontumorous tissue, respectively) and increased protein levels in the other two HCCs (960% and 690% of corresponding non-tumorous tissue, respectively) (Fig. 2B). The observed highly variable P-gp expression further confirms the heterogeneity of HCCs, which has also been observed in previous studies (21, 30). Unaltered b-actin levels supported the specificity of the observed changes (to avoid redundancy, Western blots for b-actin are only shown for Fig. 2). Immunohistochemistry for MDR1 revealed a distinct canalicular
staining in controls (Fig. 2C) and a positive staining of luminal membranes of tumor cells forming pseudoglands (tubules) in HCCs (Fig. 2D). In contrast, no signal was observed in areas with trabecular arrangements of tumor cells in HCC (Fig. 2E). SPGP/BSEP mRNA levels were increased in five HCCs and decreased in the other five compared with corresponding control tissue, and average mRNA levels in HCC were not significantly different (94 22% of non-tumorous tissue, n.s.) (Fig. 3A). SPGP/BSEP protein levels were increased in one tumor (418% of corresponding non-tumorous tissue), decreased in two HCCs (79% and 28% of corresponding non-tumorous tissue, respectively) and virtually absent in one HCC (2% of corresponding nontumorous tissue) (Fig. 3B). Immunohistochemistry for BSEP was positive on luminal membranes of tumor cells forming pseudoglands (Fig. 3D) and decorated a canalicular pattern in trabecular structures in HCC (Fig. 3E). Immunohistochemistry for P-gps using the C219 antibody recognizing MDR1, MDR3 as 371
Zollner et al. B BSEP mRNA / GAPDH mRNA
A 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
T
NT
1.0
0.8 1.0
T
NT
T
NT
T
1.0
0.3
BSEP
NT
C
NT
4.2 1.0 0.1
T
D
E
Fig. 3. Bile salt export pump, also called sister of P-glycoprotein (SPGP/BSEP) expression in hepatocellular carcinoma HCC. Total RNA and liver membrane fractions were isolated from HCC and perineoplastic tissue and analyzed by competitive reverse transcription polymerase chain reaction (A) and Western blotting (B) as described in Material and methods. In addition, SPGP/BSEP tissue distribution was analyzed by immunohistochemistry (C–E). (A) SPGP/BSEP steady-state mRNA expression in unaffected perineoplastic tissue (left part, non-tumorous, NT) and in the corresponding HCC (right part) are represented by the same symbol and are connected by lines. The horizontal bar represents mean values of overall mRNA levels of each group. (B) Representative Western blots of non-tumorous (NT) liver tissue and HCC (T). Symbols depicting the liver samples correspond to symbols used in (A). Data are expressed as the fold change compared to respective non-tumorous sample. (C) Regular canalicular immunohistochemical staining pattern for SPGP/BSEP in peritumorous non-neoplastic tissue. (D) Staining is preserved on luminal membranes of pseudoglandular structures in HCC and (E) on canalicular membranes of tumor cells with a trabecular growth pattern. Original magnification 20.
well as SPGP/BSEP (50) revealed distinct canalicular staining in controls, whereas immunostaining was completely absent in three out of nine HCCs. Interestingly, C219 immunostaining was absent in HCCs with a trabecular growth pattern, while areas with tubular (pseudoglandular) cell arrangements showed positive staining of luminal hepatocyte membranes forming pseudorosettes. Expression of MRP2 and MRP3 is generally maintained or even increased in individual HCCs
MRP2 mRNA levels were increased in seven out of 10 HCCs compared with respective surrounding non-tumorous tissue, and overall mRNA levels showed a trend towards increased expression without reaching statistical significance (144 15% of non-tumorous tissue, n.s.) (Fig. 4A). MRP2 protein expression was increased in all four HCCs investigated by Western blot analysis (153%, 209%, 131% and 132% of corresponding non-tumorous tissues, respectively) 372
(Fig. 4B). However, average protein levels were not statistically different from non-tumorous tissue. Canalicular immunohistochemical staining for MRP2 was present in all but one HCC. Staining was more prominent on apical membranes in pseudoglands (Fig. 4D) than on canalicular membranes of tumor cells arranged in trabecular structures in HCC (Fig. 4E). Positive MRP2 staining of HCCs with trabecular growth pattern indicates a partially preserved hepatocellular architecture with distinct canalicular membranes in high- and medium-grade differentiated HCCs investigated in this study. Therefore, absent P-gp staining cannot be attributed to frank loss of hepatocyte membrane polarization but rather might reflect loss of P-gp expression. MRP3 mRNA expression was decreased in seven of 10 HCCs and overall mRNA levels in HCCs were slightly, but not statistically significantly, reduced (61 13% of non-tumorous tissue, n.s.) (Fig. 5A). Western blotting revealed an increased protein expression in two HCCs (201%
Transporter expression in HCC B MRP2 mRNA / GAPDH mRNA
A 1.2 1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0
NT
NT
T
NT
T
3.1 1.0 1.3
1.0
1.3
T
NT
MRP2 1.0
NT
C
T
1.9 1.0
T
D
E
Fig. 4. Canalicular conjugate export pump MRP2 expression in hepatocellular carcinoma (HCC). Total RNA and liver membrane fractions were isolated from HCC and perineoplastic tissue and analyzed by competitive reverse transcription polymerase chain reaction (A) and Western blotting (B) as described in Material and methods. In addition, MRP2 tissue distribution was analyzed by immunohistochemistry (C–E). (A) MRP2 steady-state mRNA expression in unaffected perineoplastic tissue (left part, non-tumorous, NT) and in the corresponding HCC (right part, tumor, T) are represented by the same symbol and are connected by lines. The horizontal bar represents mean values of overall mRNA levels of each group. (B) Representative Western blots of non-tumorous (NT) liver tissue and HCC (T). Symbols depicting the liver samples correspond to symbols used in (A). Data are expressed as the fold change compared to respective non-tumorous sample. (C) Regular canalicular immunohistochemical staining pattern for MRP2 in peritumorous non-neoplastic tissue. (D) Staining is preserved or even enhanced on luminal membranes of pseudoglandular structures in HCC. (E) Canalicular membranes of tumor cells with a trabecular growth pattern also stain positive. Original magnification 20.
and 162% of corresponding non-tumorous tissues, respectively) and was unaltered in two HCCs (93% and 102% of corresponding nontumorous tissues, respectively); however, overall MRP3 protein levels were not significantly different from non-tumorous tissue (Fig. 5A). Immunostaining for MRP3 revealed a weak staining of basolateral membranes of hepatocytes in nontumorous tissue predominantly in hepatocytes adjacent to portal fields (Fig. 5C). No specific basolateral signal was observed in three out of four HCCs (Fig. 5D, E). However, hepatocytes showed a diffuse cytoplasmatic reaction with the MRP3 antibody in one case, which could be interpreted as disturbed targeting of MRP3 towards the basolateral membrane (Fig. 5D). Expression of basolateral uptake systems OATP2/ SLC21A6, NTCP and CNT1 is decreased in HCC
OATP2/SLC21A6 steady-state mRNA levels were reduced in seven of 10 HCCs compared
with the respective non-tumorous tissue; overall mRNA expression in HCC was reduced to 58 21% of non-tumorous tissue (Po0.01) (Fig. 6A). Protein expression was significantly reduced in all four HCCs (38%, 11%, 9% and 66% of non-tumorous tissues; overall reduction to 30 23, respectively, Po0.01) (Fig. 6B). mRNA expression of basolateral NTCP was reduced in all but one HCC; overall expression in HCC was reduced to 41 27% of non-tumorous tissue (Po0.05) (Fig. 7A). Protein levels were significantly reduced in all four HCCs as revealed by Western blot analysis (49%, 55%, 17% and 39% of corresponding non-tumorous tissues, respectively; overall reduction to 35 14%, Po0.05). Immunostaining for NTCP revealed a distinct basolateral pattern in controls (Fig. 7C), whereas no specific signals were detected in seven out of nine HCCs (Fig. 7D, E). Only two HCCs showed positive immunoreactivity irrespective of growth pattern or grade. 373
Zollner et al. B MRP3 mRNA / GAPDH mRNA
A 0.16 0.14
T
NT
T
NT T
NT
T
1.0
0.9 1.0 4.8 1.0 1.2
1.0
2.7
MRP3
0.10 0.08 0.06 0.04 0.02 0.00
NT
C
NT
0.12
T
D
E
Fig. 5. MRP3 expression in hepatocellular carcinoma (HCC). Total RNA and liver membrane fractions were isolated from HCC and perineoplastic tissue and analyzed by competitive reverse transcription polymerase chain reaction (A) and Western blotting (B) as described in Material and methods. In addition, MRP3 tissue distribution was analyzed by immunohistochemistry (C–E). (A) MRP3 steady-state mRNA expression in unaffected perineoplastic tissue (left part, non-tumorous, NT) and in the corresponding HCC (right part, tumor, T) are represented by the same symbol and are connected by lines. The horizontal bar represents mean values of overall mRNA levels of each group. (B) Representative Western blots of non-tumorous (NT) liver tissue and HCC (T). Symbols depicting the liver samples correspond to symbols used in (A). Data are expressed as the fold change compared to respective non-tumorous sample. (C) Weak staining of basolateral membranes of hepatocytes and more pronounced staining of basolateral membranes of cholangiocytes (arrows) in peritumorous non-neoplastic tissue. (D) Basolateral staining is completely absent on tumor cells arranged in a tubular growth pattern, whereas the cytoplasm shows a granular reaction with MRP3. (E) MRP3 staining reveals a weak basolateral signal in an HCC with a trabecular growth pattern. Original magnification 20.
We hypothesized that down-regulation of OATP2/SLC21A6 might be the consequence of reduced levels of HNF1a, which is an essential transactivator of OATP2/SLC21A6 (51, 52). Since the limited amount of tissue material precluded isolation and extraction of nuclei for Western and gelshift analysis, only steady-state mRNA levels of HNF1a were determined. HNF1a steady-state mRNA levels were reduced in eight of 10 HCCs compared with the respective non-tumorous tissue, but overall mRNA expression in HCC showed only a trend for reduced levels (64 18% of non-tumorous tissue, n.s.) (data not shown). However, OATP2/SLC21A6 mRNA levels did not correlate with HNF1a mRNA levels. In addition to the uptake systems NTCP and OATP2/SLC21A6, we investigated expression of one ‘‘true’’ drug transporter: the concentrative nucleoside transporter CNT1. CNT1 mediates active transport of nucleosides and nucleoside 374
drugs such as cytarabine or gemcitabine (53). The latter has also been used for treatment of HCCs, but only showed marginal antitumor effects (54, 55). CNT1 mRNA levels were reduced in eight of 10 HCCs compared with the respective non-tumorous tissue and overall mRNA expression in HCC was reduced to 58 29% of nontumorous tissue, which failed to reach statistical significance (data not shown). Discussion
HCC with contraindications for transplantation or resection represents a malignancy with poor prognosis because of limited therapeutic options and resistance to chemotherapeutic drugs (1, 3). The MDR phenotype of many other malignancies is partially attributed to overexpression of transport proteins belonging to the P-gp and MRP families (7). Evidence exists that chemoresistance of HCCs might also be mediated by
Transporter expression in HCC OATP2 mRNA / GAPDH mRNA
A
B 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0
NT
T
NT
T
NT
T
NT
T
0.4 1.0 0.1
1.0
0.1
1.0
0.7
OATP2 1.0 P
