The first two authors contributed equally to this work.
Cancer Cell Biology
Expression and localization of human multidrug resistance protein (ABCC) family members in pancreatic carcinoma
Article first published online: 1 FEB 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 115, Issue 3, pages 359–367, 20 June 2005
How to Cite
König, J., Hartel, M., Nies, A. T., Martignoni, M. E., Guo, J., Büchler, M. W., Friess, H. and Keppler, D. (2005), Expression and localization of human multidrug resistance protein (ABCC) family members in pancreatic carcinoma. Int. J. Cancer, 115: 359–367. doi: 10.1002/ijc.20831
- Issue published online: 7 APR 2005
- Article first published online: 1 FEB 2005
- Manuscript Accepted: 25 OCT 2004
- Manuscript Received: 21 JUL 2004
- Deutsche Forschungsgemeinschaft. Grant Number: Ko 2120/1-1
- multidrug resistance protein;
- pancreatic carcinoma;
- ABCC family;
Pancreatic ductal adenocarcinoma is among the top 10 causes of death from cancer in industrialized countries. In comparison with other gastrointestinal malignancies, pancreatic cancer is one of the tumors most resistant to chemotherapy. An important mechanism of tumor multidrug resistance is increased drug efflux mediated by several transporters of the ABC superfamily. Especially BCRP (ABCG2), MDR1 P-glycoprotein (ABCB1) and members of the MRP (ABCC) family are important in mediating drug resistance. The MRP family consists of 9 members (MRP1–MRP9) with MRP1–MRP6 being best characterized with respect to protein localization and substrate selectivity. Here, we quantified the mRNA expression of BCRP and of all MRP family members in normal human pancreas and pancreatic carcinoma and analyzed the mRNA level of the transporters most abundantly expressed in pancreatic tissue, BCRP, MRP1, MRP3, MRP4 and MRP5, in 37 tissue samples. In addition, we determined the localization of the 4 MRP proteins in normal human pancreas and in pancreatic carcinoma. The expression of BCRP, MRP1 and MRP4 mRNA did not correlate with tumor stage or grading. On the other hand, the expression of MRP3 mRNA was upregulated in pancreatic carcinoma samples and was correlated with tumor grading. The MRP5 mRNA level was significantly higher in pancreatic carcinoma tissue compared to normal pancreatic tissue. These data suggest that MRP3 and MRP5 are involved in drug resistance of pancreatic tumors and that quantitative analysis of their expression may contribute to predict the benefit of chemotherapy in patients with pancreatic cancer. © 2005 Wiley-Liss, Inc.
Pancreatic cancer is the 4th to 5th leading cause of cancer-related death in most Western industrialized countries1 with an average survival after diagnosis of 3 to 6 months. In Europe, it is the 8th most common cancer with approximately 74,000 newly diagnosed cases per year.2 In spite of impressive advances in the field of diagnostic imaging of the pancreas, the availability of numerous tumor markers and an aggressive therapeutic approach, the prognosis of pancreatic carcinoma continues to be poor, with less than 5% surviving beyond 5 years. Surgical resection is possible in up to 40% of the patients with localized disease, but even in this group of patients, prognosis is relatively poor.3, 4 Most treatment failures are due to local recurrence, hepatic metastases or both and occur within 1 to 2 years after surgery.5, 6 Adjuvant therapy may improve long-term survival7, 8, 9, 10 but its routine use is not universal9 because the results of randomized trials have been inconclusive.8 In case of nonresectable pancreatic carcinomas infiltrating the retroperitoneal plexi or the superior mesenteric artery, chemotherapy might be the option of choice for the treatment. Until now, much impact on survival has not been achieved regarding the different chemotherapies, with maximum median survival times lying between 4 and 9 months and relatively low response rates.11, 12 Therefore, in comparison with other gastrointestinal malignancies, pancreatic cancer seems to be one of the most resistant tumors to chemotherapy. This fact underscores the urgency to find novel therapeutic strategies to understand the mechanism of drug resistance in pancreatic carcinoma in order to develop more effective drugs in the future.
Drug resistance is attributable to several processes taking place in neoplastic cells. One of these processes is the decreased accumulation of drugs within cancer cells because of increased drug efflux. Proteins mediating this drug efflux mostly belong to the large superfamily of ABC transporters. Especially members of the ABCB family including MDR1 P-glycoprotein13 and members of the ABCC family14, 15, 16 have been shown to be responsible for mediating multidrug resistance. In addition, the BCRP (ABCG2) protein, a member of the ABCG family was shown to mediate resistance against several anticancer drugs.17 Substrates for BCRP include mitoxantrone, methotrexate and topoisomerase I inhibitors.17 The human ATP-binding cassette transporter family C (symbol ABCC) consists of 12 members, 9 of which comprise the group of multidrug resistance proteins (MRP1–MRP9; ABCC1–ABCC6 and ABCC10–ABCC12).14, 15, 16 MRPs are integral membrane proteins mediating the ATP-dependent export of organic anions out of cells. So far, the family members MRP1–MRP6 are the best characterized paralogs with respect to their substrate spectrum. MRP1, MRP2, MRP3 and MRP6 transport lipophilic compounds conjugated with glutathione, glucuronate or sulfate.14, 15, 16, 18 Substrates for MRP4 and MRP5 include cyclic nucleotides and nucleotide analogs.19, 20, 21 Furthermore, MRP4 has been identified as a cotransporter for reduced glutathione with bile salts22 and as a transporter for prostaglandins23 and the steroid dehydroepiandrosterone-3-sulfate (DHEAS).24 In addition to endogenous compounds, MRP family members are able to export a variety of organic anions of toxicological relevance and are important in conferring resistance to cytotoxic and antiviral drugs.19, 20, 23 Whereas the expression of MDR1 P-glycoprotein has been analyzed in detail in pancreatic carcinoma,25, 26 the knowledge on the expression and localization of MRP family members and of BCRP is very limited. In pancreas the expression of MRP3 (ABCC3) has been analyzed by Northern blotting and RT-PCR.27, 28, 29, 30 In addition, the MRP3 protein has been detected in normal pancreatic tissue and in pancreatic adenocarcinoma.30 Furthermore, the protein expression of MRP1 (ABCC1) and MRP2 (ABCC2) has been analyzed in rat and human tissue samples with chronic pancreatitis.31
In our study, we analyzed the mRNA expression of all 9 human MRP (ABCC) family members and of BCRP (ABCG2) in normal pancreatic tissue and pancreatic carcinoma. Based on these data, we subsequently quantified the mRNA expression of BCRP (ABCG2) and of the 4 most abundant MRP family members MRP1 (ABCC1), MRP3 (ABCC3), MRP4 (ABCC4) and MRP5 (ABCC5) by semiquantitative real-time RT-PCR in 31 pancreatic carcinoma samples and in 6 samples from healthy tissue donors and correlated these mRNA expression data to clinical parameters. In addition, we studied the localization of the 4 MRP family members by immunofluorescence analysis. These expression and localization studies contribute to the understanding of the role of drug transporters in patients with normal pancreatic tissue and in pancreatic carcinoma thus supporting the understanding of the impact of these transporters in causing intrinsic multidrug resistance.
Material and methods
Patients and tissue collection
Normal human pancreatic tissue was obtained through an organ donor program from 6 individuals who were free of any pancreatic disease. The median age of the organ donors was 44 years (percentiles 35/50). Pancreatic cancer tissue samples were obtained from 31 patients (14 women and 17 men) undergoing a pylorus-preserving Whipple resection for ductal adenocarcinoma of the pancreas. The median age of the patients with pancreatic cancer was 63 years (percentiles 58/71). According to the international classification of the UICC, there were 8 stage I, 5 stage II, 10 stage III and 8 stage IV pancreatic cancers. Tumor grading was well differentiated in 7 cases, moderately differentiated in 17 cases and undifferentiated in 7 cases. The median survival time of all patients together regardless of their tumor stage was 12 months (percentiles 5.8/25.3). None of the patients received chemotherapy before surgery. In pancreatic carcinomas with early tumor stages (I + II) the median survival was 20 months (percentiles 10.5/31.5), in late tumor stages (III + IV) the median survival was only 10 months (percentiles 5.0/21.8).
Freshly removed tissue samples were cut in the operating room and randomly divided for histologic analysis (immediately fixed in paraformaldehyde solution for 12–24 hr and paraffin-embedded for immunohistochemistry) or were snap-frozen in liquid nitrogen and maintained at −80°C until further analysis. Studies were approved by the Human Ethics Committee of the Universities of Bern and Heidelberg.
The polyclonal antisera EAG5,32 FDS,28 SNG22 and AMF19 were raised in rabbits against the carboxy-terminal sequences of human MRP2, MRP3, MRP4 and MRP5, respectively. The EAG5, SNG and AMF antisera were affinity-purified as described33 using MRP2-,34 MRP4-22 and MRP5-expressing19 cells. The monoclonal mouse antibodies QCRL1 against MRP1, M2III-6 against MRP2 and M3II-9 against MRP3 were from Alexis Biochemicals (Günzburg, Germany). The monoclonal rat M6II-31 antibody against MRP6 was a kind gift of Dr. G. Scheffer (Free University Medical Center, Amsterdam, The Netherlands) and is now commercially available from Alexis. Alexa Fluor488-conjugated goat anti-rabbit, anti-mouse or anti-rat IgG were from Molecular Probes (Eugene, OR).
Cryosections (4–5 μm) were prepared with a cryotome (Leica, Bensheim, Germany), air-dried for at least 2 hr and fixed in precooled acetone (−20°C for 10 min). Immunofluorescence staining was carried out as described35 with antibodies or antisera diluted in phosphate-buffered saline (PBS). Antisera FDS and AMF were diluted 1:50 and the fluorochrome-conjugated antibodies 1:300. The monoclonal antibodies and the affinity-purified antibodies were used at the following final concentrations: QCRL1 1–1.5 μg/ml, M2III-6 5 μg/ml, M3II-9 10–13 μg/ml, M6II-31 0.2 mg/ml, affinity-purified EAG5 0.5–1 mg/ml, affinity-purified SNG 0.3–0.5 mg/ml and affinity-purified AMF 0.5–1 mg/ml. Pictures were taken on a confocal laser scanning microscope (LSM510, Carl Zeiss, Jena, Germany). Some cryosections were stained with hematoxylin/eosin and photographed with a digital video camera (Hamamatsu, Hamamatsu, Japan) on an Axiovert S100TV (Carl Zeiss). Pictures were analyzed using Openlab imaging software (Improvision, Coventry, UK).
Real-time semiquantitative RT-PCR
Total RNA was isolated by the single-step guanidinium method. Total RNA (1 μg) was reverse-transcribed using 2.5 μg oligo(dT)18 primer as described.28 Synthesized single-stranded DNAs were purified using Microcon 100 (Millipore, Billerica, MA) filter units. BCRP (ABCG2) and MRP1–MRP9 (ABCC1–ABCC6 and ABCC10–ABCC12) mRNA levels in relation to β-actin mRNA levels were determined using the LightCycler™ System and the FastStart DNA Master SYBR Green I kit (both from Roche, Mannheim, Germany). PCRs were performed according to the manufacturer's instructions with 0.5 μM of the respective sense and 0.5 μM of the respective antisense primers, 4 mM MgCl2, and 1-fold LightCycler-FastStart DNA Master SYBR Green I mix in a total volume of 20 μl including 1 μl of the synthesized sscDNA. Cycling conditions were as follows: 10 min denaturation at 94°C, followed by 45 cycles of 10 sec denaturation at 94°C, 15 sec primer annealing at 64°C and 30 sec of fragment elongation at 72°C. Sequence and position of MRP-specific primers used in this analysis and the amplified fragment lengths are summarized in Table I. For β-actin amplification, the sense primer oActin.for (5′-TGACGGGGTCACCCACACTGTGCCCATCTA-3′) and the antisense primer oActin.rev (5′-CTAGAAGCATTTGCGGTGGAC-GATGGAGGG-3′) was used. The amount of β-actin and MRP1 to MRP9 single-stranded cDNAs was determined as described35 using a serial plasmid dilution (human MRP5 cDNA in pcDNA3.1/Hygro19); from 1 × 106 to 1 × 103 fg as amplification standard. The β-actin mRNA concentration, calculated in relation to the standard curve, was set to 100% and the respective BCRP and MRP mRNA value is given as a percentage of β-actin amplification.
|Sense primer||Position||Antisense primer||Position||Fragment length (bp)|
|MRP1 (ABCC1)||5′-CTGACAAGCTAGACCATGAATGT-3′||bp 4244 – bp 4269||5′-TCACACCAAGCCGGCGTCTTT-3′||bp 4596 – bp 4576||353|
|MRP2 (ABCC2)||5′-CTTCGGAAATCCAAGATCCTGG-3′||bp 4354 – bp 4375||5′-TAGAATTTTGTGCTGTTCACATTCT-3′||bp 4637 – bp 4613||284|
|MRP3 (ABCC3)||5′-GGACCCTGCGCATGAACCTG-3′||bp 4133 – bp 4152||5′-AGGCAAGTCCAGCATCTCTGG-3′||bp 4582 – bp 4562||450|
|MRP4 (ABCC4)||5′-GGATCCAAGAACTGATGAGTTAAT-3′||bp 3621 – bp 3644||5′-TCACAGTGCTGTCTCGAAAATAG-3′||bp 3978 – bp 3956||358|
|MRP5 (ABCC5)||5′-GCTGTTCAGTGGCACTGTCAG-3′||bp 3834 – bp 3854||5′-TCAGCCCTTGACAGCGACCTT-3′||bp 4314 – bp 4294||481|
|MRP6 (ABCC6)||5′-CACTGCGCTCCAGGATCAGC-3′||bp 4010 – bp 4029||5′-CAGACCAGGCCTGACTCCTG-3′||bp 4511 – bp 4492||502|
|MRP7 (ABCC10)||5′-AGGACAGGGCCTTGTGGCAG-3′||bp 4043 – bp 4062||5′-TCAGGGACCTCCGAGTGAGG-3′||bp 4479 – bp 4460||437|
|MRP8 (ABCC11)||5′-GAAGTCCTCCTTGGGCATGGC-3′||bp 3540 – bp 3560||5′-TTATCTCAGTGAAGAAGTGGCTGT-3′||bp 4149 – bp 4126||610|
|MRP9 (ABCC12)||5′-AGAGACACAATAATGAAACTCCCA-3′||bp 3706 – bp 3729||5′-CTACAATCTGACTTCTGCTGCTA-3′||bp 4080 – bp 4058||375|
|BCRP (ABCG2)||5′-CAAAGGCAGATGCCTTCTTCG-3′||bp 1502 – bp 1522||5′-CATACTGAATTAAGGGGAAATTTAA-3′||bp 1989 – bp 1965||488|
The age and survival of all patients are reported as median together with the percentiles. Furthermore, we analyzed the median of MRP3 and MRP5 mRNA levels according to the 4 tumor stages (UICC), and according to the different tumor grade (I, II and III).
Differences between groups were analyzed using the Kruskal-Wallis and Mann-Whitney tests for non-parametric data with p < 0.05 considered statistically significant. For survival analysis related to MRP3 and MRP5 expression the log-rank test was used. The correlation of MRP3 and MRP5 mRNA levels with tumor grading and tumor stage was analyzed using the Pearson test.
Expression of BCRP and MRP family member mRNAs in human pancreas and pancreatic carcinoma
The first aim was to investigate the mRNA expression of BCRP and of all 9 human MRP family members in normal human pancreas and pancreatic carcinoma. This expression analysis was performed using the LightCycler™ system and primer pairs directed against BCRP and each of the MRP paralogs. These primer pairs (Table I) were tested for their specificity and affinity against the respective cDNA template to ensure that this experimental setup can be used not only for detecting a specific BCRP or MRP mRNA but also for quantifying the amount of the respective mRNA in a given sample. Using these primer pairs and RNA from normal pancreatic tissue and 4 different pancreatic carcinoma samples cDNA fragments of the expected sizes were amplified for BCRP, for MRP1 to MRP5 and MRP7 (Fig. 1). For MRP6 mRNA, only very weak amplification products of the expected size were detected (Fig. 1). The MRP7 mRNA level was low in all samples investigated (<<0.1% β-actin mRNA), whereas no amplification product was observed for MRP8 mRNA. MRP9 mRNA was detected only in normal human pancreas as indicated by a weak amplification product (Fig. 1) with a mRNA level below 0.1% of β-actin mRNA. Because the MRP family members MRP1 to MRP6 are best characterized with respect to protein localization and transport properties, with MRP1 to MRP5 being expressed in human pancreas, and because MRP7 and MRP9 were only weakly amplified, we further quantified the mRNA levels of the MRP family members MRP1 to MRP5 in 37 tissue samples from normal pancreas and pancreatic carcinoma.
For BCRP, the mRNA values varied between lower than 0.1% and 18.7% of the respective β-actin value, with most values being lower than 1% of the β-actin mRNA (Table II). The values of MRP1 mRNA levels were between 0.1% and 3.0% of the quantity of the β-actin mRNA levels of the respective sample (Table II). Amplified MRP2 cDNA fragments were detected in 32 of 37 samples but in all cases the amount of the MRP2 mRNA in relation to β-actin was below 0.1% confirming the weak MRP2 amplification products in the first expression analysis (Fig. 1). Therefore, we did not include the data on MRP2 mRNA quantification into Table II. The quantity of MRP3 mRNA was moderate to high in all samples and varied between 0.3% and 44.0% of the respective β-actin mRNA value. The quantities of MRP4 mRNA varied between 0.2% and 1.6% of the quantity of β-actin mRNA with 1 sample being below 0.1% β-actin mRNA level. MRP5 mRNA was detectable in all 37 samples, with values between 0.2% and 5% of the respective β-actin mRNA level, only in 2 pancreatic carcinoma samples the values in relation to β actin were below 0.1%.
|Patient number||Gender||Age||G||Stage(UICC)||PT||PN||PM||Survival(months)||mRNA expression (%)|
Immunolocalization of MRP family members in normal pancreas and pancreatic carcinoma
Immunofluorescence signals were detected for MRP1, MRP3, MRP4 and MRP5 in normal pancreas and pancreatic carcinoma tissue (Figs. 2, 3). However, no MRP2-positive staining was observed in normal pancreas and pancreatic carcinoma when using the affinity-purified EAG5 antibody (Fig. 2a) or the M2II-6 antibody (not shown), which is in agreement with a study by Sandusky et al.36 Normal pancreas and pancreatic carcinoma were also negative when stained with the M6II-31 antibody, whereas this antibody readily detected MRP6 on liver sections being used as positive control. Similar negative staining results for MRP6 in normal human pancreas were obtained by Scheffer et al.37 In contrast, MRP3 was readily detectable with the FDS antibody (Figs. 2, 3) or the M3II-9 antibody (not shown). MRP3 was predominantly localized in the basolateral membrane of epithelial cells of ducts in normal pancreas and in the cells of ductal pancreatic carcinoma (Fig. 2b, 3b), which is in accordance with a study by Scheffer et al.30 In addition, MRP3 was also present in the plasma membrane of acinar cells in normal pancreas. Similar localizations in duct cells, acinar cells and pancreatic cancer cells were observed for MRP4 and MRP5 (Figs. 2c,d, 3c,d). MRP1 mRNA was detectable in normal pancreas and pancreatic carcinoma (Table II), the source of the MRP1 mRNA probably being fibroblasts of the connective tissue, which showed strong staining (Fig. 3f), rather than acinar cells or pancreatic carcinoma cells, which lacked MRP1 staining (Fig. 3e,f). Because the affinities of the antibodies towards their respective antigen differ from one antibody to the other, the staining intensities do not reflect the exact relative amounts of the MRP proteins to each other. The stainings rather give a qualitative assessment of the localization of different MRP isoforms. However, when normal pancreatic tissue and pancreatic carcinoma samples with high MRP3 mRNA expression levels were analyzed exemplarily with identical settings on a confocal laser scanning microscope (Fig. 3g,h), staining for MRP3 with the FDS antibody was more intense in the carcinoma sample than in normal human pancreas.
Relationship of MRP3 and MRP5 mRNA expression levels to histopathologic and clinical parameters
The statistical analysis of MRP1, MRP4 and BCRP mRNA in relation to histopathologic and clinical parameters demonstrated that there was no correlation of the mRNA expression level to carcinoma stages or grades. Therefore, these 3 transporters were not included into the statistical analysis. The MRP3 mRNA expression level in pancreatic carcinoma tissue samples was, irrespective of the tumor stage (tumor stage classified by UICC with UICC I = T1-2N0M0, UICC II = T3N0M0 and T1-3N1M0, UICC III = T4N0-1M0, and UICC IV = T1-4N0-1M1 with T = tumor, N = lymph node, and M = metastasis) significantly higher than the MRP3 mRNA level in normal pancreatic tissue (Mann-Whitney test p = 0.02 for UICC stage I and stage II and p = 0.03 for UICC stage II and IV). With regard to the correlation between the MRP3 mRNA expression in lower tumor stages (I+II) with higher stages (III+IV), the mRNA expression did not significantly increase in the higher tumor stages (p = 0.7). The comparison of the histopathological data (tumor grade of differentiation with grade I = highly differentiated, grade II = less-differentiated and grade III = low-differentiated) and the MRP3 mRNA expression data revealed that tumor differentiation was related to MRP3 mRNA expression levels. The Mann-Whitney test showed significant differences of MRP3 mRNA expression levels between grade 1 vs. 2 (p = 0.001), between grade 1 vs. 3 (p = 0.017) and between grade 2 vs. 3 (p = 0.028). Between normal pancreatic tissue and tissue samples with tumor grade 1, there was no significant relationship (p = 1). However, significant differences in MRP3 mRNA expression levels between normal pancreatic tissue and pancreatic carcinoma grade 2 and grade 3 were observed (Mann-Whitney test p = 0.01 or p = 0.035) (Fig. 4). The Kaplan-Meier graph (Fig. 5) shows a difference in survival between patients (n = 8) with a high MRP3 mRNA expression level (>10% β-actin mRNA expression level) and patients (n = 23) with moderate to low MRP3 mRNA expression levels (<10% β-actin mRNA expression level; p = 0.0005). A lower MRP3 mRNA expression was associated with a better survival prognosis.
In contrast to the MRP3 mRNA levels, MRP5 mRNA expression levels were not related to tumor grade (Fig. 4) or tumor stage. A correlation in the sense of rising expression of MRP5 mRNA level together with the rising tumor stage could not be shown like for MRP3 mRNA expression levels (Mann-Whitney test p = 0.52). MRP5 mRNA expression levels in pancreatic carcinoma tissue samples were significantly higher than in normal pancreatic tissue samples for the tumor stages I, II and III (UICC stage I: p = 0.005, UICC stage II: p = 0.009, UICC stage III: 0.02). In UICC stage IV, however, there was no significant difference of the MRP5 mRNA expression level in comparison to normal pancreatic tissue (p = 0.1). MRP5 mRNA expression levels in tumors with grading 1 and 3 were similar, and the mRNA level in grade 2 tumors was not significantly different (Fig. 4). Furthermore, the Mann-Whitney test did not show significant differences of MRP5 mRNA expression levels between grade 1 vs. 2 (p = 0.087), between grade 1 vs. 3 (p = 0.8) or between 2 vs. 3 (p = 0.187). In contrast, MRP5 mRNA expression levels in normal pancreatic tissue samples were significantly lower in comparison to pancreatic carcinoma samples.
Increased export of cytotoxic drugs from cells is one major mechanism of drug resistance. Proteins involved in this increased drug efflux mostly belong to the large family of ABC transporters. In our study, we investigated the expression of BCRP, a member of the ABCG family and the expression and localization of members of the MRP (ABCC) family of export pumps in normal human pancreatic tissue and pancreatic carcinoma. Semiquantitative real-time RT-PCR demonstrated the expression of BCRP (ABCG2), MRP1 (ABCC1), MRP3 (ABCC3), MRP4 (ABCC4) and MRP5 (ABCC5) mRNA in normal pancreatic tissue and in pancreatic carcinoma samples with different tumor gradings. In addition, cDNA fragments of the expected length were amplified for MRP2 (ABCC2), MRP6 (ABCC6) and MRP7 (ABCC10), but for these the quantification of the mRNA levels relative to the respective β-actin mRNA levels demonstrated mRNA levels below 0.1% (for MRP6 below 0.01%) and, therefore, may reflect amplification products originating from nonpancreatic cells within the tissue samples used for RNA isolation. MRP9 (ABCC12) was amplified only in normal pancreatic tissue, whereas MRP8 (ABCC11) was neither amplified in normal nor in the pancreatic carcinoma samples (Fig. 1). Based on the semiquantitative RT-PCR data, we subsequently analyzed the expression levels of BCRP, MRP1, MRP3, MRP4 and MRP5 mRNA in 31 different pancreatic carcinoma tissue samples and in 6 samples from normal human pancreas. In addition, the localization of the 4 MRP family members was analyzed.
The quantification of the BCRP and the MRP family member mRNA levels demonstrated that BCRP, MRP1 and MRP4 mRNA levels did not change significantly between normal pancreatic tissue and pancreatic carcinoma samples with different tumor gradings. Differences between samples may thus reflect interindividual variations rather than adaptive mechanisms during tumor development. Furthermore, MRP1 protein was localized only in fibroblasts and not in acinar cells or pancreatic carcinoma cells. Expression of MRP1 protein in human38 and mouse39 fibroblasts has been demonstrated earlier and supports the importance of this export pump in protecting cells of the connective tissue. In contrast, MRP3 mRNA expression was significantly lower in normal pancreatic tissue compared to pancreatic cancer tissue (Fig. 4). The MRP3 protein mediates the efflux of bile salts and several other organic anions out of cells,40, 41, 42 and confers resistance to several anticancer drugs including vincristine, etoposide, teniposide and methotrexate.29, 43, 44, 45 Several of these substances have been used in clinical studies for the treatment of pancreatic cancer.46, 47, 48 Interestingly, in most cases, the beneficial effect was very low. Especially methotrexate treatment had a marginal antitumor activity against pancreatic carcinoma,46, 48 suggesting that upregulation of MRP3 in pancreatic carcinoma may play a role in the resistance of these tumors. The upregulation of MRP3 in pancreatic carcinoma without prior treatment of patients with anticancer drugs could be explained by the fact that MRP3 was found to be an inducible transporter. Studies on the regulation of MRP3 protein and mRNA expression have demonstrated that MRP3 is upregulated in obstructive cholestasis30, 49 and in regenerating liver.50 Liver regeneration is accompanied with dedifferentiation of hepatocytes and a change in cellular polarity. The fact that MRP3 mRNA expression correlated with tumor grading in pancreatic carcinoma may demonstrate this inducible expression during dedifferentiation. Furthermore, a correlation of the MRP3 mRNA level with survival of patients after resection has been observed. Patients with a MRP3 mRNA level below 10% β-actin mRNA expression level have a longer survival period than patients with a higher MRP3 mRNA expression level. Although this correlation between MRP3 mRNA level and survival of patients is demonstrated only in a small group of patients, MRP3 expression may be an important factor for prognosis.
The missing correlation of MRP3 mRNA levels with the tumor stage could be explained by the fact that chemoresistance is supposed to correlate less with tumor extension, reflected by the tumor stage, than with the differentiation of the tumor. Therefore, the inducible nature of MRP3 together with its ability to confer drug resistance suggests that MRP3 is important in mediating resistance and could be a potential marker for chemoresistance in pancreatic adenocarcinoma.
MRP5 mRNA expression levels in pancreatic carcinoma were neither related to tumor grade nor to tumor stage. Therefore, determination of MRP5 mRNA expression may be less predictive for prognosis and chemoresistance. Higher MRP5 mRNA expression levels in pancreatic carcinoma samples compared to normal pancreatic tissue may indicate that MRP5 could account, at least in part, for the intrinsic drug resistance of ductal pancreatic carcinoma. Many chemotherapeutic agents have been tested in the treatment of pancreatic cancer, but only 5-fluorouracil and mitomycin C have been shown to have beneficial effects.51, 52 Therefore, several other cytotoxic agents have been actively investigated in the treatment of pancreatic carcinoma and from these, gemcitabine has become increasingly the chemotherapeutic drug of choice.8, 53, 54 Gemcitabine is a deoxycytidine analog that competes for the incorporation into DNA.55 Nucleoside analogs are an important class of drugs used in the treatment of cancer and viral infections. Several studies have shown that resistance against these drugs can be mediated by a change in uptake transporters and metabolizing enzymes.56, 57 However recent studies suggest that also efflux pumps are important modulators of resistance against nucleoside-based analogs. Thus, it has been shown that nucleotide analogues and cyclic nucleotides are substrates for MRP423, 58, 59 and MRP5.19, 21, 59, 60 Overexpression of one of these export pumps may cause resistance against these anticancer and antiviral agents. Although MRP5-mediated transport of gemcitabine phosphates using inside out-oriented membrane vesicles has not yet been described and MRP5-mediated resistance against gemcitabine is controversally discussed,21, 61 one study demonstrated that HEK293 cells overexpressing the human MRP5 protein are 2-fold more resistant to gemcitabine compared to vector control-transfected cells.61 Therefore, increased MRP5 expression as observed in pancreatic carcinoma samples may influence the chemotherapy following curative resection by mediating resistance against this cytotoxic agent.
In conclusion, we have shown that the ABCG2 family member BCRP and the MRP family members MRP1, MRP3, MRP4 and MRP5 are expressed in human pancreas and pancreatic carcinoma and that the MRP3, MRP4 and MRP5 proteins were localized to the plasma membranes of ductular and acinar cells. The quantitative mRNA expression analysis demonstrated that MRP3 and MRP5 mRNA levels change during tumor development, whereas BCRP, MRP1 and MRP4 mRNA levels remained stable, when comparing the mRNA levels in normal human pancreatic tissue and pancreatic carcinoma. In addition, MRP3 mRNA levels significantly correlated with tumor grade and prognosis of the patients.
We thank E. Böhm, M. Brom, J. Longin and M. Meinhardt for their excellent technical help during our study, H. Spring for his expert help in obtaining confocal laser scanning micrographs and I. Esposito for her support in the analysis of the histopathological samples.
- 61Human multi-drug resistance protein 5 (MRP5) confers resistance to gemcitabine. Proc Am Assoc Cancer Res 2002; 43: 3868., , , , , , , .