Lung resistance-related protein (LRP) plays an important role in chemoresistance of tumor cells probably by altering nuclear-cytoplasmic transport processes. We analyzed the association between LRP expression and hepatocarcinogenesis in humans and rats by RT-PCR, immunoblotting, and immunohistochemistry. LRP was found in hepatocytes and bile epithelia of normal human and rat liver showing distinct interindividual variations. In human tissues, the LRP expression levels of dysplastic liver nodules, hepatocellular adenomas, and carcinomas were highly variable, including decreased but also distinctly increased staining intensities. Mean expression levels, however, were comparable to the surrounding tissue. Considerable levels of LRP mRNA and protein were also found in human hepatoma cell lines. To study LRP expression from the beginning of hepatocarcinogenesis onward, rats were subjected to a tumor initiation/promotion protocol leading to preneoplastic hepatocytes present as single cells or multicellular clones, followed by adenoma and carcinoma. All of the (pre)neoplastic rat liver lesions expressed, comparable to the surrounding tissue, considerable amounts of LRP. We conclude that LRP might be one mechanism involved in the intrinsically high but variable chemoresistance of normal and (pre)neoplastic hepatocytes.
- vault particle
- human liver
- hepatocellular carcinoma
liver cancer is among the seven leading causes of cancer deaths worldwide. Prognosis is usually poor, and no effective chemotherapeutic treatment is presently available (4, 15). One explanation may be that the major function of the liver is to process and detoxify numerous structurally diverse compounds (40). Drug metabolites are extruded from the hepatocytes by specific membrane proteins such as the multispecific organic anion transporters that comprise members of the multidrug resistance protein family (MRP), the bile acid transporter, ion-motive ATPases, and P-glycoprotein (21, 25). These transporters contribute to the broad chemoresistance of hepatocytes known as intrinsic multidrug resistance (MDR). The liver, like several other tissues, preserves protection mechanisms during malignant progression resulting in intrinsically drug-resistant tumors (12).
Additionally, to a protective function, MDR might also play a role in hepatocarcinogenesis. (Pre)malignant hepatocytes have been shown to acquire further resistance to cytotoxic agents by activation ofP-glycoprotein and phase II detoxifying enzymes, including the placentar glutathione-S-transferase (GSTp) (10, 30,35, 38). This might provide a selection advantage to the (pre)neoplastic hepatocytes in a toxic environment and thus drive the further development to malignant tumors (11, 32, 35, 37,41).
A newly identified protein, called lung resistance-related protein (LRP) due to its discovery in a drug-resistant lung cancer cell line, has been associated with cellular chemoresistance (36). LRP, also known as major vault protein, is the main structural component of vaults, the largest ribonucleoprotein particles known so far. Besides multiple copies of LRP (100 kDa), the vault complex is composed of two proteins (p240 and p193) and an untranslated RNA (vRNA) molecule. The two larger protein components were identified as telomerase-associated protein-1 and as a new poly-ADP-ribose-polymerase (vault PARP) (22, 23). Vaults have been detected originally in preparations of coated vesicles from rat liver (20). The importance of LRP in detoxification processes is substantiated by the fact that it is highly expressed in epithelia potentially exposed to toxins (17). Scheffer et al. (36) and Berger et al. (2, 3) have shown a distinct correlation of LRP expression with chemoresistance against diverse antineoplastic drugs including platinum agents in various cancer cell lines derived from lung cancer, gliomas, and other tumors. Furthermore, a distinct LRP expression was found also in highly chemoresistant soft tissue leiomyosarcomas and malignant gastrointestinal stromal tumors (33). In several malignancies, LRP expression has predictive value for poor response to chemotherapy (36). LRP induction in colon cancer cells by sodium butyrate reduced nuclear accumulation of the anticancer drug doxorubicin (24). These results corroborate the hypothesis that vaults are involved in nucleocytoplasm transport (7).
The possible role of LRP for intrinsic or acquired MDR of hepatocytes and their (pre)malignant counterparts has not been extensively investigated so far. Two studies gave conflicting results about the expression and possible function of LRP in normal hepatocytes (17, 39). We therefore studied the extent of LRP expression in human and rat hepatocarcinogenesis. In both species we observed considerable levels of LRP in unaltered liver and all steps of liver (pre)neoplasia. Thus LRP vaults might be one factor contributing to the high intrinsic chemoresistance of liver tumor.
MATERIALS AND METHODS
Animals and treatment.
Male SPF Wistar rats, 3 wk old, were obtained from the Forschungsinstitut für Versuchstierzucht und Versuchstierhaltung (Himberg, Austria). Animals were kept under standardized conditions and were treated as described (26). Briefly,N-nitrosomorpholine (NNM; Sigma, St. Louis, MO), dissolved in PBS (pH 7.4), was given as a single dose of 250 mg NNM/10 ml solution/kg body wt by gavage. After a recovery period of 4 days, animals were treated with phenobarbital (Fluka, Buchs, Switzerland), which was admixed to the diet (Altromin 1321N; Altromin, Lage, Germany) at a daily dose of 50 mg/kg body wt. Controls received basal diet ad libitum. Animals were killed by decapitation under CO2asphyxia. Time points of death were 91 and 266 days post-NNM treatment (for details see Ref. 26). The experiment was performed according to Austrian guidelines for animal care and protection.
Human liver samples.
Patients suffering from cirrhosis (n = 13) or cholangiocellular carcinoma (n = 2) were subjected to liver transplantation. Patients with dysplastic liver nodules (n = 13), hepatocellular adenoma (n = 5), or carcinoma (n = 20) underwent liver surgery. Chemotherapy had not been applied before surgery. For causative factors for the development of the lesions and diagnoses see Table1. The classification of benign and malignant liver lesions followed the International Working Party (16) as well as Edmondson and Steiner (9), respectively. Tissue samples were immediately fixed in 10% buffered formaldehyde; additional samples were stored in liquid nitrogen.
Human hepatoma cell lines HepG2 (HB-8065), Hep2B2.1–7 (HB-8064), Chang liver (CCl-13) and WRL 68 (CL-48), obtained from the American Type Culture Collection (Manassas, VA), were maintained in vitro as described (13). The small-cell lung cancer cell line GLC-4 and its drug-resistant LRP-overexpressing subline GLC-4/ADR (42) were a gift of Dr. E. G. E. de Vries, University of Groningen, The Netherlands.
RNA was extracted from cell lines and RT-PCR was performed as described previously (3). Two oligonucleotide primer sets were used: set I resulted in the amplification of a 284 bp PCR product specific for human LRP (GenBank accession no. X79882); primer set II specific for GAPDH (358 bp) was used as a housekeeping gene (3). Amplification products were separated by acrylamide gel electrophoresis and were stained with ethidium bromide. Dynamics of PCR amplification, as measured by ethidium bromide staining, was evaluated for both genes and revealed an optimum of 20 cycles for GAPDH and 28 cycles for LRP.
Tissue samples or hepatoma cells were suspended in lysis puffer (50 mM Tris, 300 mM NaCl, 0.5% Triton X at pH 7.6, containing several protease inhibitors and PMSF) for 10 min at 4°C, frozen and thawed three times, and centrifuged at 5,000 rpm for 15 min at 4°C. Supernatant was separated by SDS-PAGE and transferred, detected, and quantified using the Gel-Doc System from Bio-Rad (Munich, Germany) as described (3). Primary antibodies (anti-LRP mouse monoclonal antibody from Transduction Laboratories, Lexington, KY; anti-β-actin, clone AC-15 from Sigma) were applied at a working dilution of 1:1,000.
Specimens of human liver, fixed in 4% buffered formaldehyde and rat liver tissue, fixed in Carnoy's solution, were processed as described (14). Serial sections were cut and stained with hematoxylin and eosin and for LRP. For rat samples, an additional section was stained for GSTp.
Immunostaining for LRP and GSTp.
Primary antisera of mouse (LRP-56) or rat (LMR5) monoclonal antibodies raised against LRP (both Alexis Biochemicals, San Diego, CA) and rabbit polyclonal IgGs raised against rat Yp-subunit of GSTp (Biotrin International, Dublin, Ireland). Pretreatment and staining of tissue was described by Grasl-Kraupp et al. (14). Primary antibodies, diluted in BSA-TBS (2.5% BSA in 0.05 M Tris, 0.3 M NaCl, pH 7.6) (LRP-56 and LMR5: 1:10; Yp: 1:5,000), were applied overnight at 4°C. Secondary antibodies, diluted in BSA-TBS (biotinylated goat-anti-mouse IgG 1:200; biotinylated rabbit-anti-rat IgG 1:300; biotinylated goat-anti-rabbit 1:600; all from Dakopatts, Glostrup, Denmark) were used for 90 min at RT. Incubation with peroxidase-labeled streptavidin (1:300 in TBS, 45 min RT; Dakopatts) and diaminobenzidine (Sigma) was used for color development. Specificity of immunohistochemistry was confirmed by omitting the primary antibodies.
Quantitative evaluation of GSTp-positive foci.
Preneoplastic rat liver foci were first identified in the GSTp-stained serial section and were then evaluated in the LRP-stained serial section (see Semiquantitative evaluation of immunoreactions) using two microscopes linked by a bridge for overprojection (Zeiss, Germany). Wherever indicated, we also determined the number of component cells per cross section of individual GSTp+ foci.
Semiquantitative evaluation of immunoreactions.
LRP staining of each lesion was compared with the surrounding tissue showing mostly uniform staining. At least 1,000 cells in 5 different areas of lesions and their surroundings were evaluated. Intensity of staining within a lesion was categorized on an arbitrary scale of 0 to 2, where 0 = negative staining; 0.3 and 0.7 = weaker than surrounding tissue; 1 = equal to surrounding tissue; and 1.3, 1.7, and 2 = stronger than surrounding tissue. The extent of staining was estimated by the percentage of cells of a certain staining intensity. Intensity and extent of staining served to calculate the degree of LRP expression, e.g., 30% of the cells were grouped into staining category 0.3 and 70% of the cells into category 1 (0.3 × 0.3 = 0.09; 0.7 × 1 = 0.7; 0.09 + 0.7 = 0.79); the degree of staining was calculated to be 0.79, i.e., 79%, of the surrounding tissue.
Densitometrical analysis of immunoreactions.
Tumor and surrounding tissue were analyzed within the same section. The microscope image (×40 objective; Olympus BH series) was detected by a JVC color-video camera (type 3CCD) and transferred to the image processor Lucia G/4.00 (Lucia, Prague; details:http://www.lim.cz/). The digitized picture, corrected for nonhomogeneous illumination, was transformed to an image of 768 × 512 pixels with a 16-bit gray resolution and possible gray values of 0–255 for each pixel. Transmission refers to the gray value spectrum and expresses the light intensity measured. The optical density (OD) is defined as the negative logarithm (base 10) of the transmission. For each measurement, two points for OD-calibration were chosen (the lightest and darkest point in the sample), whereby gray value 0 means OD 10 and a gray value of 255 means OD 0. OD is evaluated according to the following formula (α-coefficient of the camera of 0.45 is included): OD = −log (pixelgray value + 0.5)/62.5.
The cytoplasm of 30 randomly chosen hepatocytes within the tumor and within the surrounding tissue was measured. For each of the 60 measurements, the sum of ODs of all pixels was normalized to the area evaluated. The ratio between data of the tumor and of the surrounding tissue was calculated (for details see Refs. 27 and 28).
The significance of differences of means was calculated by Kruskal-Wallis or the Wilcoxon tests.
LRP expression in nontumorous liver of humans and rats.
In immunostained sections of human and rat liver, LRP was found exclusively in the cytoplasm of hepatocytes and bile epithelia, corresponding to the location of LRP in the cytoplasmic vesicle fraction (Fig. 1,A–C). The staining intensity was most prominent in the perivenous zone of the liver lobule and decreased toward the periportal zone (Fig. 1 A). Occasionally, single hepatocytes with strong LRP expression were scattered throughout the lobule (Fig. 1 B). Immunostaining with two different antibodies against LRP gave identical results as demonstrated in the case of rat liver carcinoma (Fig. 1, H andI).
LRP immunostaining was confirmed by immunoblot analyses of cell extracts obtained from 15 nonmalignant human liver samples (Fig.2). A single band of about 100 kDa was detectable, which agrees with the molecular size of human LRP. Extracts derived from GLC-4 and GLC-4/ADR cells were used as negative and positive controls, respectively. Of the 15 livers, 13 expressed considerable amounts of LRP. Generally, distinct interindividual variations in LRP expression were evident. Among the two almost negative cases, one patient suffered from alcoholic liver disease (Fig.2, lane 1), the other from primary biliary cirrhosis due to Byler's disease (lane 15). The highest LRP expression was found in a noncirrhotic, nonmalignant liver tissue derived from a patient suffering from cholangiocellular carcinoma (lane 13).
Expression of LRP-mRNA in human hepatoma cell lines.
By RT-PCR, mRNA of LRP was detectable in all human hepatoma cells tested showing marginal differences among the cell lines (Fig.3 A). Also, LRP protein was present in all cell lines. The protein concentration was highest in CCL-13 cells and lowest in Hep-G2 cells (Fig. 3 B).
Expression of LRP in human benign and malignant hepatocellular tumors.
All cases of benign and malignant hepatocellular tumors studied displayed a positive immunoreaction for LRP in pre(malignant) cells. Like in normal liver, LRP was found exclusively in the cytoplasm (Fig.1, F and G). LRP-stained tissue sections were evaluated by both semiquantitative scoring and densitometry. Data obtained with the two methods correlated significantly (Fig.4; Pearson's test:r 2 = 0.7127; P < 0.0001). In Fig. 5 data from the densitometric evaluation are shown. Generally, adenomas and grade 1 carcinomas tended to an enhanced, low grade dysplastic nodules and grade 2 carcinomas to a decreased LRP expression compared with the surrounding tissue. High grade dysplastic nodules and carcinomas grade 3 tended to decreased but also intensely increase immunoreactions (Figs. 1 G and 5). The mean expression levels of both premalignant and malignant tissues did, however, not differ significantly from the surrounding tissue.
Expression of LRP in GSTp+ preneoplasia, hepatocellular adenoma, and carcinoma of the rat.
Details on the model system used for studying hepatocarcinogenesis in the rat were published elsewhere (26). GSTp staining was used to detect all stages of hepatocarcinogenesis ranging from initiated single cells, multicellular foci (Fig. 1, D and E), hepatocellular adenoma, and finally, to carcinoma (Fig. 1, H and I). Similar to the human situation, LRP was expressed in all (pre)malignant stages of the rat liver analyzed. Immunostaining with two different antibodies against LRP gave identical results. As one example, serial sections of a rat liver carcinoma stained with the two anti-LRP antibodies are shown (Fig. 1, H and I). Expression levels did not vary significantly between (pre)malignant lesions and the adjacent tissue (Fig.6).
Expression of LRP, the major vault protein, was studied in unaltered and (pre)malignant hepatocytes, which are known for their intrinsic and acquired chemoresistance toward a great variety of cytotoxic agents. We found considerable expression of LRP in hepatocytes and bile duct epithelia in almost all liver samples studied. To this end, two papers reported somewhat conflicting results on the occurrence of LRP in human liver. Whereas a low expression of LRP in hepatocytes and a higher one in bile epithelia was reported (17), others (39) failed to detect LRP in the liver. Vaults were originally discovered in extracts from rat liver (20). Primary cultures of human hepatocytes treated with the hepatomitogens epidermal growth factor or hepatocyte growth factor stably expressed significant amounts of LRP mRNA. At the protein level, LRP-expression even tended to increase within 33 days of culture (34). These data suggest an inherent and pronounced expression of vaults in human hepatocytes, which may relate to the specific function of this organ for activation, metabolism, and excretion of xenobiotics and endogenous toxins.
In human and rat liver, we observed the highest LRP-expression in hepatocytes in the perivenous zone of the liver acinus. This contrasts to the intrahepatic distribution of P-glycoprotein being highest in the periportal and lowest in the perivenous zone and suggests different functions of these two protection proteins in normal liver (38). Periportal and perivenous hepatocytes differ in their biochemical capacity, reflecting different functions in normal liver physiology. Metabolic activities like utilization of glucose and amino acids as well as the synthesis of urea, cholesterol, and bile localize preferentially to the periportal area, whereas synthesis of glycogen, fatty acids, and glutamine, and xenobiotic metabolism predominate in the perivenous area (18). Thus vaults may fulfill a defined physiological function in the liver mainly executed by the hepatocytes within the perivenous area.
A distinct interindividual heterogeneity in the LRP level in liver tissues was detectable. This may be due to an inherent, genetically determined LRP expression pattern, as shown for many genes involved in chemoresistance and metabolism of xenobiotics and that often underlies the different interindividual sensitivities of humans against carcinogens and toxins (29). On the other side, many of the metabolic functions of the liver may be readily induced, e.g., on functional load, drug metabolism increases its capacity dramatically (40). Likewise, LRP expression in lung tumor cell lines (3) and 1 of 3 human hepatoblastoma xenocrafts (1) was upregulated by exposure to cytotoxic drugs and to benzo-a-pyrene, a genotoxic carcinogen occurring in tobacco smoke and smoked food products (6). Accordingly, the heterogeneity observed in our mainly cirrhotic, nonmalignant liver samples may also be based on exogeneous or endogeneous toxins, such as ethanol or accumulation of inflammatory intermediates. However, it appears to be independent of cirrhosis, because the two samples almost lacking LRP were also derived from patients suffering from cirrhosis due to ethanol abuse and Byler's disease, respectively. It will be interesting to investigate whether the interindividual differences in the LRP expression level in human liver are more or less constant or are changeable by the administration of various cytotoxic and noncytotoxic xenobiotics.
Intrinsic and acquired forms of MDR may underlie the limited success obtained by chemotherapeutic treatments against different types of malignant tumors including liver cancer (15, 33). Accordingly, in experimental hepatocarcinogenesis, MDR plays a key role in the development of tumors. In addition to the intrinsic chemoresistance of normal hepatocytes, (pre) malignant liver cells acquire further resistance to the cytotoxic/cytocidal effects of various hepatotoxins and hepatocarcinogens, e.g., acetylaminofluorene, nitrosamines, and aflatoxin B1 (11, 37). This resistance provides a selection advantage in a toxic environment, i.e., proliferation of normal hepatocytes is suppressed, whereas resistant cells in (pre)neoplasia multiply. As a result, tumor promotion and finally malignant progression may be accelerated. The biochemical basis for this phenomenon is an increase of metabolic pathways favoring inactivation and extrusion of toxic compounds, e.g., conjugation reactions by the GSTp and other GSH-transferases, or membrane transport by P-glycoprotein and MRP-family members (5, 8, 10,11, 19, 31, 32, 37). Also, in humans, resistance to cytotoxins seems to be important for hepatocarcinogenesis. These toxins may stem from exogenous (ethanol, aflatoxin B1) and/or endogenous sources (enterotoxins, excessive storage of metabolic by-products, inflammatory by-products) (4). Our data suggest that LRP is not involved in the acquired forms of chemoresistance in hepatocarcinogenesis, because LRP expression was not generally upregulated in liver (pre)neoplasia. The variable expression of LRP in liver carcinomas may reflect the genetic instability within the tumor, rather than activation of a distinct program. Nevertheless, the level of LRP was high in most of the tumors, which is the first evidence that LRP may be involved in MDR of liver cancer, as observed in the clinical situation. Because LRP was inducible in one of three adriamycin-treated human hepatoblastoma xenocrafts (1) and in lung cancer cell lines (2), it appears likely that LRP is induced in (pre)malignant liver tissue by the administration of chemotherapeutic drugs. Data on the regulation of LRP in liver tissue might clarify the still enigmatic role of vaults in normal and tumor cell physiology.
In conclusion, our data derived from human and rat tissues demonstrate that LRP is expressed in normal and (pre)malignant hepatocytes. This suggests a possible function of LRP in the basal chemoprotection of benign and malignant liver tissue.
We thank Helga Koudelka and Juliana Krejsa for excellent technical assistance.
The present study was supported by the Herzfelder'sche Familienstiftung and by the Jubiläumsfonds derÖsterreichischen Nationalbank, Project No. 8817.
Address for reprint requests and other correspondence: W. Berger, Institute of Cancer Research, Div. of Applied and Exp. Oncology, Borschkegasse 8a, A-1090 Vienna, Austria (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
July 25, 2002;10.1152/ajpgi.00195.2002
- Copyright © 2002 the American Physiological Society