AJP - GI Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Am J Physiol Gastrointest Liver Physiol 292: G1272-G1282, 2007. First published February 1, 2007; doi:10.1152/ajpgi.00474.2006
0193-1857/07 $8.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
292/5/G1272    most recent
00474.2006v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Luo, Q.
Right arrow Articles by Angeletti, R. H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Luo, Q.
Right arrow Articles by Angeletti, R. H.

LIVER AND BILIARY TRACT

Altered protein expression at early-stage rat hepatic neoplasia

Qilie Luo,1,3 Linda Siconolfi-Baez,2,3 Pallavi Annamaneni,4 Mark T. Bielawski,4 Phyllis M. Novikoff,4 and Ruth Hogue Angeletti1,2,3

Departments of 1Developmental and Molecular Biology and 2Biochemistry, 3Laboratory for Macromolecular Analysis, and 4Department of Pathology, Albert Einstein College of Medicine, Bronx, New York

Submitted 12 October 2006 ; accepted in final form 25 January 2007


   ABSTRACT
TOP
 ABSTRACT
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Protein expression patterns were analyzed in a rat model of hepatic neoplasia to detect changes reflecting biological mechanism or potential therapeutic targets. The rat resistant hepatocyte model of carcinogenesis was studied, with a focus on the earliest preneoplastic lesion visible in the liver, the preneoplastic hyperplastic nodule. Expression differences were shown by two-dimensional polyacrylamide gel electrophoresis and image analysis. Polypeptide masses were measured by peptide mass fingerprinting using matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF) and their sequences were obtained by tandem mass spectrometry. Alterations in expression of cytoskeletal and functional proteins were demonstrated, consistent with biological changes known to occur in the preneoplastic cells. Of particular interest was the differential expression of a serine protease inhibitor (serpin) with a role implicated in angiogenesis. Serpin, implicated in the inhibition of angiogenesis, is present in normal liver but has greatly reduced expression at the preneoplastic stage of liver cancer development. Immunofluorescence microscopy with antibodies to this serpin, kallistatin, supports the proteomic identification. Immunofluorescence microscopy with antibodies to the blood vessel marker von Willebrand factor provides evidence for neovascularization in the liver containing multiple preneoplastic nodules. These observations suggest that at an early stage of liver carcinogenesis reduction or loss of angiogenesis inhibitors may contribute to initiation of neoangiogenesis. A number of other identified proteins known to be associated with hepatomas are also present at early-stage neoplasia.

liver; preneoplastic nodules; hepatoma; differentially expressed proteins; angiogenesis

HEPATOCELLULAR CARCINOMA (HCC) is the one of most frequent causes of death by cancer worldwide, with no reliable diagnosis prior to late stages of disease and no cure except surgery (13, 34, 47, 56, 63, 66). The development of HCC is invariably associated with liver damage resulting from chronic hepatitis, extensive alcohol intake, or toxins, sequentially leading to liver cirrhosis, dysplastic lesions, and finally invasive liver carcinoma (13). However, the mechanism of HCC initiation and progression and how specific lesions interact to produce its aggressive characteristic remain poorly understood. Thus there is some urgency to deepen understanding of the process of neoplasia and to elucidate biomarkers for early diagnosis as well as to identify potential targets for early therapeutic intervention. Proteomics approaches are now being employed in a variety of models to achieve these goals (18, 35, 40, 60–62, 78–80).

Rodent models of human cancer provide powerful tools to investigate cancer molecular biology and therapy (14, 18). Here, we used the highly reproducible rat liver cancer model, the resistant hepatocyte (RH) model, to study cancer development in a multistage process, with the successive appearance of distinctive cell populations (3, 14, 15). The RH model shows synchronous appearance of unique cell populations in well-defined stages that progress to increasing malignancy (14, 16, 17). Each cell type has a characteristic phenotype that can be analyzed by a variety of genetic, biochemical, and cytological techniques. Metastasis of the liver cancer to lung and other tissues has also been observed in this model. We initiated proteomic studies of liver containing preneoplastic nodules, which represent the earliest preneoplastic lesion visible on gross examination of the liver without prior immunohistochemical or histochemical staining. After an additional 6- to 8-mo period, a subset of preneoplastic nodules develops into larger neoplastic persistent nodules and then into hepatomas. Those nodules that do not progress into hepatomas redifferentiate into hepatocytes morphologically indistinguishable from normal hepatocytes (12).

A great deal of pioneering work has been done in proteomic analysis of hepatocellular carcinoma (HCC) (9, 35, 37–39, 41, 43, 55, 61, 79, 80). However, most work has focused on late-stage HCC (18) rather than early-stage HCC. To reduce the mortality due to liver tumor and metastasis, it is important to develop methods for early diagnosis and prognosis of HCC. The application of the RH resistant model will possibly contribute to identifying biomarkers for detection of liver cancer at the early stage.

The goal of the present study was to determine the feasibility of the RH model for providing insights into the development of protein expression differences that are correlated with specific stages of carcinogenesis and/or specific cell types. In this report, we show that differences in protein expression can be identified in the RH model at the preneoplastic nodule stage in the neoplastic process that reflect proteins associated with functions related to tumor development as well as to the extensive reorganization of the cytoskeleton known to occur during carcinogenesis.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Materials. Immobilized pH gradient (IPG) buffer and IPG strips of pH 3–10, 4–5, 5–6, and 6–9 were obtained from Amersham Pharmacia Biotech (Uppsala, Sweden). The 12% SDS-PAGE gel was purchased from Genomic Solutions (Ann Arbor, MI). Acetic acid, acetonitrile, ammonium hydroxide, and methanol were from J. T. Baker (Philipsburg, NJ). Diethynitrosamine (DEN), acethyaminofluorene, DTT, iodoacetamide, 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, silver nitrate, formaldehyde, and {alpha}-cyano- 4-hydroxycinnamic acid were from Sigma (St. Louis, MO). Sequencing grade-modified trypsin was from Promega (Madison, WI). The packed reversed-phase column (15 cm x 0.32 mm ID) was from Micro-Tech Scientific (Sunnyvale, CA). Distilled water was deionized (18 M{Omega}) by use of the Milli-Q system from Millipore (Bedford, MA). Other reagents were from Fisher Scientific (Pittsburgh, PA).

RH model. The RH model of liver cancer was reproduced as previously described (12, 14, 67). Figure 1A outlines the protocol to induce preneoplastic nodules and hepatomas. The model was reproduced in male Fischer 344 rats (Charles River Laboratories, Wilmington, MA) as follows: initiating carcinogen, DEN, at a dose of 200 mg/kg was injected intraperitoneally into rats; 2 wk after the injection, carcinogen, acetyaminofluorene was administered in a time-release pellet designed to release a total of 35 mg for 2 wk (Innovative Research of America, Miami, FL) that was inserted under the skin of the rat's neck; 1 wk after insertion of the pellet, two-thirds surgical partial hepatectomy (PH) was performed. Nine days after PH (i.e., 5 wk and 2 days after initiation), rats were killed and liver tissue was removed for proteomic and morphological analysis. At this time point, multiple preneoplastic hyperplastic nodules are present throughout all the lobes of the liver. They are easily distinguishable from the surrounding normal liver because of their white color in contrast to the brown color of the normal liver, their spherical shape, and their essentially uniform size (~1 µm in diameter). Figure 1B illustrates pryonin-positive spherical preneoplastic nodules. Distinct multiple nodules in a section of the liver are evident especially when methyl-green pryonin (MGP) (48) staining is employed to highlight the nodules. Different profiles of nodule size are evident in the section because the nodules are sectioned at different levels. For protein extraction analysis, sections serial to this section were used for proteomic analysis. We estimate that about 70% of the hepatic cells in a section analyzed for proteins are derived from nodular cells. The nodules also express the glutathione S-transferase (GST)-Pi, a human hepatoma marker (59, 70) (Fig. 1C). The rat nodules are histologically similar to early dysplastic lesions that appear in humans before HCC (36). Bile ductule/oval cells, which also proliferate in this model, are evident (Fig. 1B).


Figure 1
View larger version (82K):
[in this window]
[in a new window]

  		Fig. 1. A: timeline of carcinogen protocol for induction of preneoplastic nodules and hepatomas in a resistant hepatocyte (RH) rat model. Five weeks after initiation of protocol, preneoplastic nodules develop which progress into metastatic hepatomas by 9 mo. This study investigates the liver at stage 4, i.e., liver with multiple preneoplastic hyperplastic nodules. DEN, diethylnitrosamine, initiating carcinogen; AAF, aminoacetylfluorene, a mitosis-inhibiting carcinogen; PH, partial hepatectomy. B: methyl-green pryonin. Multiple preneoplastic nodules are present in the liver and show strong pryonin staining (pink) in contrast to the surrounding hepatocytes. Oval/bile ductule proliferation is also evident (blue stain). C: immunolocalization of glutathione S-transferase (GST)-Pi. A preneoplastic nodule is positive for GST-Pi; no GST-Pi is present in hepatocytes.

 
Livers from eight male Fischer rats of the same age consisting of two different groups were used for this study. One experimental group of four rats was used for developing the RH model of liver cancer to stage 4 preneoplasia, and the second group consisted of untreated sham-operated control rats. All were killed on the same day. The RH protocol produces successive stages of liver cancer progression at predictable times as defined in the literature (12, 14, 67). If any step is omitted, the sequential process of cellular changes does not occur (49, 51, 53, 67). All animal procedures were approved by the Institutional Animal Care and Use Committee of the Institute for Animal Studies at the Albert Einstein College of Medicine.

Liver samples from RH model and control sham rat livers. The right lobe of liver was removed from RH rats 9 days after PH, when early-stage nodules appear. The nodule-enriched livers were quickly frozen in liquid nitrogen for proteomic analysis (14, 67). For histological and immunocytochemical analysis, liver tissue was either frozen in cold methyl-butane surrounded by dry ice or immersed in aldehyde fixatives (49, 51, 53). All liver samples were obtained at 9 days after PH (i.e., 5 wk and 2 days after DEN initiation) and were designated RH stage 4 nodules according to a staging naming system defined in our laboratory and based on the published literature (14, 17, 67). Liver samples from the same lobes were analyzed from four RH-model and four control rats.

Protein extract samples of controls and RH liver. The frozen liver tissue was cut into thin slices (30-µm thickness and ~20-mm diameter) under –20°C using Cryostat (Belair Instrument, Fanwood, NY). The slices were suspended in sample buffer containing protease inhibitor cocktail (Sigma), 7 M urea, 2 M thiourea, 4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 5% non-detergent subphobetaines 256, 1% DTT, 0.5–1% IPG buffer, and 0.006% bromophenol. The liver tissue was completely dissolved in sample buffer through occasional vortexing and sonication during 2 h. The suspension was centrifuged in a Beckman TL-100 tabletop ultracentrifuge at 31K x1000 for 45 min at 15°C. The supernatant was taken and stored in aliquots at –70°C or directly used for two-dimensional polyacrylamide gel electrophoresis (2D-PAGE). The amount of protein loaded was ~130 µg for analytical gels (pH 3–10 and pH 6–9 IPG strips) and ~650 µg for preparative gels (pH 4–5 and pH 5–6 IPG strips).

2D-PAGE and protein visualization. Isoelectric focusing (IEF) was completed in IPGphor (Amersham Pharmacia Biotech). IPG strips were applied according to the manufacturer's instructions. Proteins were focused at 20°C for 1 h at 200 V, 1 h at 500 V, 1 h at 1,000 V, and 9 h at 8,000 V after rehydration for 12 h at 30 V, a total voltage hour 75 kVh. After the IEF step, IPG gel strips were saved in tubes and stored at –70°C for future use. The IEF gel strips were equilibrated in a buffer containing 50 mM Tris, 6 M urea, 2% SDS, 30% glycerol, and 1% DTT for 15 min, followed by 4% iodacetamide in the same buffer for another 15 min. The second-dimension electrophoresis was carried out until the bromophenol blue dye marker reaches the bottom of the gel. Silver stain method was modified from Ref. 65. The procedures are the same as described previously (45).

Image and statistical analyses. The gels were scanned as 16-bit gray TIFF images with a ScanMaker III (Microtek Lab, Ontario, CA). The image analysis expression data generation was performed by applying the software package Investigator HT Analyzer Version 2.1 (Genomic Solutions) according to the manufacturer's instructions. Average volumes of the differentially expressed protein spots were calculated after normalization using data of three gels obtained by densitometric analysis. The maximal level of the protein spot volume was defined as 100%. In the test of significance, the data were subjected to stringent statistical analysis, and the t-test (19) was applied to select the significant differences of spots. Statistical significance (P < 0.05) between nodule and normal (n = 3–5) was determined by the software.

Protein identification. Protein spots were excised from the gel before the destaining procedures (24) were carried out. The whole protocol of protein in-gel digestion is the same as performed before (45) except that two steps, protein reduction and alkylation, were excluded in this experiment. The matrix-assisted laser desorption ionization (MALDI) mass spectrometric analysis of supernatants from in-gel digestion, and database searches were performed as before (45). Some identifications were checked with mass spectrometry (MS)/MS to screen the case, more than one protein in one spot. A reversed-phase capillary HPLC column (15 cm x 320 µm ID) was attached to the Eldex gradient system (Eldex Laboratories, Napa, CA). A 50-µl sample loop was applied to load supernatant from an in-gel tryptic digest of protein spot into the capillary column. The column was then coupled to a LCQ MS (Finnigan Mat, San Jose, CA). A gradient for column elution was developed over 40 min at a flow rate 5 µl/min. The eluting peptides were ionized by electrospray ionization. The specific peptide ions were automatically selected and fragmented by a LCQ mass spectrometer (10). The mass spectrometer was set to switch between the MS mode and the MS/MS mode. The selected peptides were fragmented to generate a MS/MS spectrum, which contains the sequence information for a single peptide and was compared by the computer program Sequest (11) to predict spectra from sequence database. This leads to the identification of the peptide corresponding to the protein in the spot. Unambiguous protein identification was attained in a single analysis by the detection of multiple peptides derived from the protein. The MS/MS together with peptide mass fingerprinting (PMF) forms an orthogonal check for the identification.

Western blot analysis. Protein sample preparation and electrophoresis were carried out as described before (71) transfer onto a polyvinylidene fluoride membrane in 25 mM Tris and 190 mM glycine containing 5% methanol at 4°C for 2 h at 160 mA (71). The rabbit polyclonal antibody against rat kallikrein-binding protein (KBP), a serpin, was developed in the laboratory of Dr. Julie Chao (Medical University of South Carolina, Charleston, SC) (6). The antigen-antibody complexes were visualized using the Western Lighting as recommended by the manufacturer PerkinElmer Life Sciences (Boston, MA) and recorded with Hyperfilm (Amersham Biosciences, Piscataway, NJ).

Microscopy of RH and control livers. For histology and for immunolocalization studies, sections were prepared from the same RH and control livers used for the proteomic studies. For histology, the sections of RH and control livers were stained with MGP after fixation in a mixture of paraformaldehyde and glutaraldehyde (51). MGP staining was used instead of hematoxylin and eosin staining because nodule cells contain higher levels of cytoplasmic ribosomes compared with surrounding hepatic cells. Pryonin, a specific stain for RNA, dramatically distinguishes all nodule cells in the section, including those smaller foci of nodule cells sectioned at the periphery of the spherical nodules. Moreover, hematoxylin and eosin staining of nodules has previously been published (see Fig. 6 in Ref. 69). For immunolocalization studies, the following antibodies were localized in liver sections from RH and control rats according to previously published methods (50–53): 1) von Willebrand factor (vWF) (Santa Cruz Biotechnology, Santa Cruz, CA), a blood vessel marker (44); 2) KBP (kindly provided by Dr. Chao), a serpin; 3) mannosidase II (Babco, Berkeley, CA), a Golgi apparatus marker (52); and 4) GST-Pi (StressGen, Victoria, Canada), a human hepatoma marker (59, 70). Either aldehyde-fixed cryostat sections or nonfrozen aldehyde-fixed sections (10 µm sections) were exposed first to a diluted primary antibody (vWF, 1:50; KBP serpin, 1:100; mannosidase II, 1:200), followed by fluorescently labeled second antibody (donkey anti-rabbit or anti-mouse IgG-Cy 3, 1:200, Jackson Immunoresearch Laboratories, West Grove, PA). Immunostained sections were also stained with phalloidin-Alexa 488 to show actin filaments (Molecular Probes, Eugene, OR). Coimmunolocalization of KBP serpin and mannosidase II was also performed on RH and control liver sections by using a mixture of second antibodies (Cy-3-labeled donkey anti-rabbit and Cy-5 labeled donkey anti-mouse IgG for kallistatin and mannosidase II, respectively) (Jackson Immunoresearch Laboratories). For localization of GST-Pi, the second antibody was donkey anti-rabbit labeled with horseradish peroxidase, visualized by incubation in diaminobenzidine substrate (52). Controls were performed at the same time and consisted of sections not exposed to the primary antibodies, but with all other procedures the same. Images of fluorescent-labeled sections were examined with an inverted Nikon fluorescent microscope using a x60 oil immersion lens with numerical aperture 1.4 Planapo objective and a confocal microscope attached to a laser imaging system (Bio-Rad Laboratories, Hercules, CA) equipped with a krypton-argon laser. Black level, gain and laser intensity, Kalman averaging, excitation intensity, pinhole aperture, and Z-series analysis of sections were performed as described elsewhere (50, 51).


Figure 6
View larger version (67K):
[in this window]
[in a new window]

  		Fig. 6. Western blot of 2-DE gels with polycolonal antibody against serpin. Top: normal rat liver tissue homogenate. Charge train of modified serpin and unmodified serpin (arrow) is revealed in gel. Bottom: homogenate of stage 4 liver. None to low levels of serpin are detected in gel.

 

   RESULTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 GRANTS
 REFERENCES
 
2-DE maps for liver tissue proteome. Two-dimensional electrophoresis (2-DE) maps were constructed for the proteomes of normal and stage 4 nodule rat liver tissues, a prerequisite for subsequent comparative proteomic studies. The 2-DE maps for total extracts of sham and stage 4 nodule rat liver tissues are shown in Fig. 2, representing a comprehensive view of the major proteins differentially expressed in normal and stage 4 nodule rat liver tissue. However, broad-range IPGs do not provide the resolving power needed for separating proteomes from complex organisms like eukaryotes (76). The challenge of analyzing a complicated proteome is better addressed by the use of very narrow-range IPG gels for better resolution and efficient separation. In this research, not only broad pH 3–10 IPG strips were used, but also narrow-range pH 4–5, pH 5–6, and pH 6–9 IPG strips, since more protein sample (at least 5 times higher than broad IPG) can be loaded into narrow pH strips, which can significantly increase the identification of proteins with low copy numbers in cell culture tissue. Fig. 3 clearly shows a typical example. Carboxylesterase 2 could not be discriminated in the broad pH IPG strip (left). However, in the narrow-range IPG strip, it is easy to locate, identify, and confirm.


Figure 2
View larger version (89K):
[in this window]
[in a new window]

  		Fig. 2. Representative 2-DE maps of the control (left) tissue and preneoplastic nodules (right) of rat liver are demonstrated. For each gel, 130 µg of total proteins were subjected to 2-DE separation, and gels (12%T) were stained and exposed to silver stain as described in MATERIALS AND METHODS. 2D gel reproducible error is <20%. Labeling nos. 01, 05, 09, 11, and 18 correspond to nos. in Table 1.

 

Figure 3
View larger version (69K):
[in this window]
[in a new window]

  		Fig. 3. Representative example of zooming-in gel in improving protein separation resolution. A broad range of immobilized pH gradients (left) is contrast to a very narrow range (right) in which separation resolution is significantly enhanced.

 
Protein expression differences identified from 2D gel spots. PMF is a highly sensitive method for the identification of proteins from acryl amide gels (29). Table 1 lists proteins identified using PMF and LC-MS/MS. All listed proteins meet the current stringent criteria for positive identifications (45, 81). The proteins in the table are identified as upregulated or downregulated from sham-operated control to nodule-enriched neoplastic liver. Under the conditions used in these experiments, three identified protein spots present in normal liver were undetectable in nodule-enriched liver. Conversely, eight identified protein spots were detected exclusively in 2D-PAGE of nodule-enriched liver. This was somewhat surprising in that the tissue used in these preliminary studies contain mixed populations of cells, e.g., hepatocytes, bile ducts, endothelial cells, and neoplastic nodular cells, among others. It is possible that the amounts of some of these proteins were below the limits of detection in the gels in which they were absent, or that their mobilities had been altered by a posttranslational processing event. However, in the case of vimentin, its expression solely in preneoplastic nodule-enriched liver, and not in normal liver, is indeed consistent with other studies of liver tumors vs. normal liver tissue (31, 59, 70).


View this table:
[in this window]
[in a new window]

  		Table 1. Comparison of proteins identified with altered expression in normal and stage 4 nodule rat liver tissue

 
Figure 4 shows the details of analysis of protein differences by 2D-PAGE and mass spectrometry. In this example, a number of lightly stained spots are detected in experimental tissue, but not controls. Two protein spots are greatly increased in nodule-enriched liver compared with control liver (Fig. 4A). These were identified by PMF to be cytokeratin 8, known to be altered in cancers (75), and actin-related protein 3 a protein name (Table 1). The MS/MS spectrum and SEQUEST analysis of one of the cytokeratin 8 peptides are also shown in Fig. 4B, exemplifying the analytical procedures used in these experiments. Other difference spots were also noted in this figure but were either mixed populations of polypeptides, below the level of detection, or otherwise outside the criteria for unambiguous identification.


Figure 4
View larger version (37K):
[in this window]
[in a new window]

  		Fig. 4. Analysis of 2D-PAGE silver-stained gel spots. A: example of comparison of a region from 1 control (1) and 1 experimental gel (2–4). The pI range is 5.6 on the left margin and 5.7 on the right margin. The relative molecular weight range is from 44,000 Da (bottom edge) to 48,000 Da. Panels 2–4 are the same gel. 3 shows in dark black the automated identification of spots present in the experimental, but undetected in the control. 4 shows in white the identification of spots that are present in the experimental with spot volumes ≥200% of the same spot in the control. The spot encircled with a dotted line was identified as actin-related protein 3, and the spot encircled with the dashed line was identified as cytokeratin 8. Volumes of the differentially expressed protein spots were compared after normalization by a method that calculates the total spot volume of spots present in all gels. The standard deviation of the mean for spot intensity was less than ±0.20. B: tandem mass spectrum of 1 peptide from the gel shown in A. The protein is identified unequivocally as cytokeratin 8, from the SEQUEST search results shown in the inset.

 
Differences in expression of a serine protease inhibitor (serpin). A region of 2D-PAGE (pH 3–10) shows what appears to be extensive modification of possibly a single polypeptide in control untreated sample (Fig. 5A, inset) that is not detected in experimental sample. Although each of the spots in the parallel tracks encircled was only lightly silver stained, in-gel digestion followed by PMF and bioinformatics searches demonstrated that each spot in this series contained rat KBP, also known as kallistatin or serpin. By using the Profound algorithm, the probability of 1 and the Z score of 2.31 place this identification in the 99th percentile in an estimated random match population (81). The heavily stained spot marked to the lower right of the parallel "charge train" of spots was identified as protein disulfide isomerase, not a potentially unmodified serpin. Expression differences were not observed for the protein disulfide isomerase polypeptide (data not shown).


Figure 5
View larger version (32K):
[in this window]
[in a new window]

  		Fig. 5. Identification of a serine protease inhibitor. A: matrix-assisted laser desorption ionization time of flight mass spectrum of trypsin-digested peptides in a serpin-containing spot. Top: 2D-PAGE of difference spots. All spots in the parallel tracks shown in the inset in the gel image at left (control) contain kallistatin. Right gel is from an experimental sample. The isoelectric points of these protein spots range from 4.2 to 5.5 (left to right), and the relative molecular weights from 50,000 to 68,000 Da (bottom edge to top edge). B: tandem mass spectrum of a tryptic peptide from the kallistatin-containing spots. The sequence is highlighted in the text below the spectrum.

 
It is common for some 2D gel spots to consist of more than one protein. The interpretation of PMF of mixtures containing peptides of several proteins is challenging and often an unequivocal result cannot be achieved. Therefore, LCQ MS/MS was utilized to confirm the identification (26, 79). Figure 5B shows an MS/MS spectrum from a peptide derived from the same spot as in MALDI time of flight (MALDI-TOF). The information contained in the spectrum was used to identify the peptide sequence as NLHVSQVVHK from rat serpin, consistent with the PMF results. Other peptide spectra from the same spot were matched to the same protein, providing unambiguous identification by LCQ analysis as well.

Western blot analysis. To confirm that the serpin polypeptide was indeed diminished in nodule-enriched rat liver and to determine whether all molecular forms, both unmodified and posttranslationally modified, were involved, Western blot analysis of the tissue was carried out (Fig. 6). Approximately two-thirds of serpin is present in the apparently posttranslationally modified form in control untreated rat tissue. The strongly immunoreactive spot in the center of the panel (Fig. 6, top, arrow) appears to be unmodified serpin. This polypeptide was not identified in the silver-stained 2D-PAGE experiments, most likely because the pattern in this region of the electrophoretic separations was very complex, even in narrow-range isoelectric point (pI) gels. All forms of serpin appear to be greatly diminished in nodule-enriched liver (bottom). Future studies of the sites and types of modifications will be carried out using immunoaffinity purified polypeptides.

Immunolocalization of serpin, mannosidase II and vWF in control and noduled livers. In control liver (Fig. 7A, and inset), colocalization of serpin and mannosidase II revealed the following: 1) serpin staining in sinusoidal endothelial cells; 2) serpin staining in hepatocytes in a structure at the trans aspect of the Golgi apparatus and in the endoplasmic reticulum; and 3) mannosidase II in the Golgi apparatus. Serpin and mannosidase II did not colocalize in the same compartments of the Golgi apparatus. Serpin is found in a distinct Golgi-associated compartment situated at the trans aspect of the Golgi apparatus, and this structure does not have the Golgi marker, mannosidase II (58, 59, 70, 72).


Figure 7
View larger version (163K):
[in this window]
[in a new window]

  		Fig. 7. Liver sections from normal liver (A, inset, and C) and noduled liver (stage 4 nodules). A and inset: coimmunolocalization of serpin and mannosidase II. In normal liver, strong serpin staining (red color) is evident in sinusoidal endothelial cells. Serpin is also found in trans Golgi-associated structure (red color; better seen in inset, arrows) of hepatocytes located at the trans aspect of the Golgi apparatus (yellow-green color). The hepatocyte Golgi apparatus, which is positive for mannosidase II (yellow-green color), contains no serpin. Nuclei (blue color) are evident. B and inset: coimmunolocalization of serpin and mannosidase II. In stage 4 noduled liver, diffuse staining of serpin is evident in the cytoplasm of the nodules cells (N) (light red color). No serpin is evident in the trans Golgi-associated structure or Golgi apparatus of either nodular cells (arrows) or surrounding hepatocytes. Mannosidase II is present in the Golgi apparatus of nodule cells and hepatocytes (yellow-green color). Differences in Golgi morphology are evident between nodule cells and hepatocytes in stage 4 nodule liver and between surrounding hepatocytes in nodule liver and hepatocytes from normal liver. Nuclei (blue color) are evident. Note dividing cells in nodule. C: coimmunolocalization of von Willebrand factor (vWF) and F-actin in normal liver. vWF (red color) is localized only in the portal vein (pv) and a small branch of the vein (arrow). F-actin (green) is localized around the periphery of hepatocytes, pv, and portal artery (pa). D: coimmunolocalization of vWF and F-actin. In stage 4 noduled liver, vWF is localized in an extensive vessel network (red color, arrows) that has developed in the parenchyma with branches of vessels into the nodule. F-actin (green color) is distributed at periphery of nodule cells, many of which are organized into acini, and at periphery of hepatocytes, which are arranged as linear cords.

 
In noduled liver (Fig. 7B, and inset), colocalization of serpin and mannosidase II revealed the following: 1) serpin staining in cytoplasm of nodular cells diffusely distributed; 2) no serpin staining in either sinusoidal endothelial cells (although a few cells express low levels) or in endothelial cells within in the nodules; 3) no serpin in Golgi apparatus, in trans Golgi compartment, in ER of nodular cells and surrounding hepatocytes; and 4) mannosidase II in Golgi apparatus of nodular cells and surrounding hepatocytes. The Golgi apparatus of the nodule cells and surrounding hepatocytes are similar to that of normal hepatocytes in that they express mannosidase II activity; however, differences in Golgi morphology are evident (Fig. 7). Immunolocalization of vWF, a marker for capillaries and blood vessels, was also carried out to determine whether neoangiogenesis occurs at an early stage of liver cancer in which an angiogenesis inhibitor (i.e., serpin) is reduced or nonfunctional. Figure 7, C and D, shows vWF localization of in normal and nodule-enriched rat liver, respectively. In normal liver, vWF staining is confined to the portal vein and no branches of the vein extend into the liver parenchyma. However, in noduled liver, a prominent vWF-positive blood vessel network has formed. vWF-positive branches are evident emanating from the portal vein. The branches are evident at the periphery of the nodules as well as branches of the network extending into the nodule (Fig. 7D).


   DISCUSSION
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
GRANTS
 REFERENCES
 
Identification of a protein in a proteomics experiment is only the first step in understanding the possible significance of the discovery (1, 18). After literature searches, cross-validation of the protein's presence is helpful, as is determination of cellular localization and investigation of biological relevance. This process is very time consuming, and extensive studies must often be carried out on each individual protein identified in the experiments. Two identified protein spots functioned as internal controls, confirming observations made previously by histochemical methods that show them to be greatly increased in preneoplastic nodule-enriched liver compared with control liver. Cytokeratin 8, known to be altered in cancers (75), is upregulated in preneoplastic nodule-enriched liver (Table 1). GST-Pi polypeptide is also present in preneoplastic nodule-enriched liver. This protein is also expressed in human hepatomas and frequently used as a marker for diagnosis of HCC (59, 70). High levels of this GST isoform are known to be localized in preneoplastic nodules (see Fig. 1C) as well as in hepatomas, with low levels in proliferating bile ductules in contrast to normal hepatocytes (49, 51, 53).

Many polypeptides related to the organization of the cytoskeleton and cellular architecture, metabolic state, and regulation were identified in this study. Changes in expression were observed in a large number of cytoskeleton-associated proteins and other polypeptides associated with cellular architecture. Of these, cytokeratins 8 and 19, for example, have been noted in other studies with other model systems (37, 53, 75). Shape changes known to accompany tumorigenesis reflect rearrangement in the cytoskeleton and its associated proteins (8). In preneoplastic nodules of the RH rat model, we have previously found a major change in the cell architecture and polarity of the nodular cells from the background hepatocytes reflected by the in situ localization of actin and by alterations in cell shape and arrangement (49, 51, 53). This change in architecture had been observed in earlier histology studies (8). The nodular cells are grouped together to form acini or glandular-like structures. This acinus arrangement is in contrast to the background hepatocytes, which are organized into linear cords of contiguous cells with one aspect (basal) of the hepatocytes facing the sinusoids and the other (apical) facing the bile canaliculus. In normal hepatocytes, actin is present underneath the entire plasma membrane with higher concentrations in the bile canalicular domain. The apical-basal polarity and the hepatic cord arrangement characteristic of normal hepatocytes are not seen with nodular cells. This observation strongly supports the notion that components of the cytoskeleton have been altered after the carcinogen treatment. Beyond changes in cell shape and tissue morphology, rearrangements in the actin cytoskeleton have been associated with tumor cell migration (64).

Others of the proteins identified have complex functions. One of the 14-3-3 family proteins was increased in expression in nodule-enriched liver. These proteins can be involved with signaling, apoptosis, or interaction with integrins or transcription factors, and have been suggested to be a part of the cellular scaffold as well (57, 73, 74). Annexins constitute a family of proteins that exhibit Ca2+-dependent binding to phospholipids and are implicated in multiple cellular interactions, such as membrane trafficking, transmembrane channel activity, inhibition of phospholipase A2, and mediation of cell-matrix interactions. The biological function of annexin V (lipocortin V) is not known, although its properties include inhibition of phospholipase A2, vascular anticoagulation, and binding of glycosaminoglycans (22). A proteomic study in normal liver did not find annexin V in parenchymal hepatocytes but did find the protein in an unidentified population of small hepatic cells (60, 61). The finding of annexin V in the liver with early nodules suggests that annexin V may be associated with populations of hepatic cells not usually present in normal liver and may indicate the beginning of the early pathology of liver cancer. Annexin V has been shown to play a role in early apoptosis. Both 14-3-3 and several annexins have been identified in proteomic studies of a human hepatocellular carcinoma cell line (43, 60, 61). Because of the complexity of their functions, orthogonal biological experiments need to be carried out to understand what precise role these and other proteins may play in development of cancer (1, 33).

Of the proteins that are present in normal liver, but are not detectable or are at very low levels of expression in nodule-enriched liver, the most striking is serpin, also known as kallistatin (6, 77). Serpin has been implicated as an inhibitor of angiogenesis and tumor growth. Angiogenesis plays an important role in tumor growth, invasion, and metastasis (2, 20, 25, 28). This process is not only a prerequisite for tumor growth but is also a major influence on the metastatic spread of malignant cells (3). The polypeptides responsible for the parallel "charge-train" of spots in the 2D gels were identified both by the peptide mass mapping and MS/MS as serpins. This pattern had been identified previously in normal rat liver (21). However, this pattern was not detectable in multiple samples analyzed from liver enriched in nodules. This observation is consistent with proteomics studies of other model systems in which maspin, another serpin, was found to be present at lower levels in human hepatocellular carcinomas with respect to human liver cell line and in hepatoma cells treated with antisense epidermal growth factor receptor sequence (78, 79). Maspin was not identified in the present experiments. Immunoblot analysis of both sample sets confirmed that both apparently modified and apparently unmodified serpin polypeptides are affected. Reduction of serpin expression is implicated in the onset of angiogenesis (4–6, 23, 30, 46, 54, 82). Immunolocalization studies of serpin paralleled the proteomic findings (i.e., presence of serpin in normal liver and low levels in nodule-enriched liver).

The observations of serpin expression differences by Western blotting and immunohistochemistry implies its important role in neoangiogenesis. Furthermore, serpin localization in the normal liver in contrast to the nodule-enriched liver, was present in a structure at the trans aspect of the Golgi apparatus and not in the mannosidase II-positive Golgi apparatus of normal hepatocytes or nodule cells. One function of the Golgi apparatus is the sequential addition of specific carbohydrate moieties (e.g., galactose, mannose, sialic acid) to amino acids of the proteins. Specific moieties are added in specific compartments of the Golgi apparatus (e.g., either in cis, medial, or trans). The nature of the serpin-positive structure needs to be more precisely defined to determine whether this structure is a part of the Golgi apparatus and therefore likely to function in glycoprotein processing or is a separate structure related spatially to the Golgi apparatus with another function (e.g., secretion). The absence of serpin in Golgi apparatus, trans Golgi compartment, and ER of nodule cells may indicate that major alterations in synthesis and processing of specific proteins and/or carbohydrates occur early in carcinogenesis. Nodule cells showed diffuse localization of serpin in the cytoplasm indicating that it may be a soluble protein. However, no serpin was evident in cytoplasmic membrane-bounded organelles (e.g., Golgi apparatus, trans Golgi structure, ER) that function in protein synthesis, glycosylation, and secretory processing. Whereas in hepatocytes of the normal liver, serpin localization is consistent with normal processing, the serpin localization in nodule cells and in hepatocytes from nodule-enriched liver indicates abnormal processing of the protein. Low levels of serpin may be present in the nodule-enriched liver but the protein may not be functioning because serpins are likely polymerized during inhibiting process, resulting in serpin aggregation (23). Our studies have detected differences in serpin levels and in posttranslational modifications between normal liver and nodule-enriched liver. The immunolocalization data is consistent with posttranslational modifications revealed by 2D-PAGE and supports the reduction of serpin, an angiogenesis inhibitor, in the noduled liver.

Proteomic analysis has identified one protein that is an inhibitor of angiogenesis, namely KBP serpin, that is detected in low levels in nodule tissue compared with normal liver by proteomics. Moreover, immunocytochemistry has validated the proteomic data and showed localization to specific hepatic cells and subcellular organelles. In addition, these studies provide possible mechanisms for the differential expression of serpin in normal liver and early stage of liver carcinogenesis. At an early stage of liver cancer, a decrease in angiogenesis inhibitors may shift the normal physiological balance of angiogenesis factors resulting in the proliferation of new blood vessels. Our studies provide direct evidence that neoangiogenesis occurs at an early stage of liver carcinogenesis well before hepatoma development (49). In an earlier study, measurement of blood flow in rat liver with early-stage nodules was investigated and found to be reduced compared with the surrounding liver as measured by infusion of microspheres; however, the study did not demonstrate neoangiogenesis in relation to either nodules or surrounding liver (68). Whether expression of other anti-angiogenesis proteins are downregulated in all cell types or only in specific cells will be studied in future experiments. The modifications that result in the multiplicity of protein spots will also be analyzed with immunoaffinity purified protein.

A major hindrance to understanding the process of carcinogenesis is the paucity of specific markers for new populations that appear in the initiation and promotion stages of the disease, and that may differ from those at advanced stages. The RH rat provides a model in which the progression to tumors passes through synchronous stages, each well characterized in morphology and immunohistochemistry. Early-stage neoplastic lesions as manifested by preneoplastic hyperplastic nodules were selected for these initial studies because nodules can be distinguished without staining or other treatment that would interfere with protein analysis. Furthermore, early dysplastic nodules in human HCC and preneoplastic nodules in rat liver that represent early-stage neoplasm have similar hyperplastic morphology (36). These features enable future detailed studies of the differences in protein expression using laser capture microdissection before proteomic analysis for nodules cells as well as other hepatic cells type (e.g., stromal cells, bile ductule/oval cells). Since changes in stromal cells as well as tumor cells are thought to be important, the ability to dissect and enhance these differences is essential. This is further underscored by some of the findings in the present study. High levels of GST-Pi are localized exclusively in preneoplastic nodule cells, and low levels of GST-Pi are found in proliferating bile ductules (51, 53). Cytokeratin 19 has been demonstrated by immunofluorescence microscopy to be localized only in bile ductule cells (53).

Protein expression patterns are not necessarily equivalent to mRNA expression patterns (7, 27). Although the present studies are a prelude to more intensive proteomics analysis, some comparison can be made to results from published microarray studies of hepatocellular carcinoma and carcinogen-induced hepatoma. The levels of vimentin mRNA were found to be diminished in two such studies (32, 42), corresponding to the reduction in vimentin polypeptide found in our experiments. Jeong and colleagues (32) also observed a downregulation in levels of 14-3-3 mRNA. However, this group noted a downregulation of expression of cytokeratin 19 mRNA. This is in contrast to what is observed in our present study and in other histological and proteomic studies. Addition of other proteomic technologies to the methods used in the present work will include laser capture microdissection of individual cell types, multidimensional chromatographic separations to improve the resolution and dynamic range of detection, and quantitative mass spectrometry procedures. The studies will also be expanded to stages of the RH model that appear before and after the nodules, so that a larger set protein expression patterns will be generated to discriminate important functional changes in the development of liver cancer. In summary, our studies have shown the feasibility of investigating a liver carcinogenesis model with properties similar to human liver cancer progression by proteomics technology. The goal is to discover alterations in protein composition and levels and posttranslational modifications between normal and liver cancer stages with the possibility of finding liver cancer markers.


   GRANTS
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
REFERENCES
 
This study was supported by grants from the National Institutes of Health: CA-055011 (R. H. Angeletti), CA-06576 (P. M. Novikoff), and P30 KD-41296 (Marion Bessin Liver Center of the Albert Einstein College of Medicine).


   ACKNOWLEDGMENTS
 
The authors thank Dr. Julie Chao of the Medical University of South Carolina for the gift of serpin antiserum and Michael Cammer, Analytical Imaging Center, for expert assistance in confocal microscopy.


   FOOTNOTES
 

Address for reprint requests and other correspondence: R. H. Angeletti, Laboratory for Macromolecular Analysis and Proteomics, 1300 Morris Park Ave., Bronx, NY 10461 (e-mail: angelett{at}aecom.yu.edu)

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.


   REFERENCES
TOP
 ABSTRACT
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Baldwin MA. Protein identification by mass spectrometry: issues to be considered. Mol Cell Proteomics 3: 1–9, 2004.[Free Full Text]
  2. Bergers G, Benjamin LE. Tumorigenesis and the angiogenic switch. Nat Rev Cancer 3: 401–410, 2003.[CrossRef][ISI][Medline]
  3. Bissell MJ, Radisky D. Putting tumors in context. Nat Rev Cancer 1: 46–54, 2001.[CrossRef][Medline]
  4. Borgono CA, Diamandis EP. The emerging roles of human tissue kallikreins in cancer. Nat Rev Cancer 4: 876–890, 2004.[CrossRef][ISI][Medline]
  5. Borgono CA, Michael IP, Diamandis EP. Human tissue kallikreins: physiologic roles and applications in cancer. Mol Cancer Res 2: 257–280, 2004.[Abstract/Free Full Text]
  6. Chao J, Chai KX, Chen LM, Xiong W, Chao S, Woodley-Miller C, Wang LX, Lu HS, Chao L. Tissue kallikrein-binding protein is a serpin. I. Purification, characterization, and distribution in normotensive and spontaneously hypertensive rats. J Biol Chem 265: 16394–16401, 1990.[Abstract/Free Full Text]
  7. Chen G, Gharib TG, Huang CC, Taylor JM, Misek DE, Kardia SL, Giordano TJ, Iannettoni MD, Orringer MB, Hanash SM, Beer DG. Discordant protein and mRNA expression in lung adenocarcinomas. Mol Cell Proteomics 1: 304–313, 2002.[Abstract/Free Full Text]
  8. Chowdhury NR, Saber MS, Lahiri P, Mackenzie PI, Novikoff PM, Becker FR, Chowdury JR. Expression of specific UDP-glucuronosyltransferase isoforms in carcinogen-induced preneoplastic rat liver nodules. Hepatology 13: 38–46, 1991.[ISI]
  9. Cui JF, Liu YK, Li JZ, Shen HL, Song HY, Dai Z, Yu YL, Zhang Y, Sun RX, Chen J, Tang ZY, Yang PY. Identification of metastasis candidate proteins among HCC cell lines by comparative proteome and biological function analysis of S100A4 in metastasis in vitro. Proteomics 6: 5953–5961, 2006.[CrossRef][ISI][Medline]
  10. Ducret A, Van Oostveen I, Eng JK, Yates JR 3rd, Aebersold R. High throughput protein characterization by automated reverse-phase chromatography/electrospray tandem mass spectrometry. Protein Sci 7: 706–719, 1998.[Abstract]
  11. Eng J, McCormack A, Yates JR 3rd. An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J Am Sci Mass Spectrum 5: 976–989, 1994.[CrossRef]
  12. Enomoto K, Farber E. Kinetics of phenotypic maturation of remodeling of hyperplastic nodules during liver carcinogenesis. Cancer Res 42: 2330–2335, 1988.
  13. Farazi PA, DePinho RA. Hepatocellular carcinomar pathogenesis: from genes to environment. Nat Rev Cancer 6: 674–687, 2006.[CrossRef][ISI][Medline]
  14. Farber E. The multistep nature of cancer development. Cancer Res 44: 4217–4223, 1984.[Free Full Text]
  15. Farber E, Cameron RG, Laishes B, Lin JC, Medline A, Ogawa K, Solt DB. Physiological and molecular markers during carcinogenesis. In: Carcinogens: Identification and Mechanisms of Action, edited by Griffin AC and Shaw CR. New York: Raven, 1979, p. 319–335.
  16. Farber E, Rubin E. Cellular adaptation in the origin and development of cancer. Cancer Res 51: 2751–2761, 1991.[Free Full Text]
  17. Farber E, Sarma SR. Hepatocarcinogenesis: a dynamic cellular perspective. Lab Invest 56: 4–22, 1987.[ISI][Medline]
  18. Feng JT, Shang S, Beretta L. Proteomics for the early detection and treatment of hepatocellular. Oncogene 25: 3810–3817, 2006.[CrossRef][ISI][Medline]
  19. Fisher RA. Statistical Methods and Scientific Inference. New York: Harfner, 1973.
  20. Folkman J. Tumor angiogenesis: therapeutic implications. N Engl J Med 285: 1182–1186, 1971.[ISI][Medline]
  21. Fountoulakis M, Suter L. Proteomic analysis of the rat liver. J Chromatogr B Analyt Technol Biomed Life Sci 782: 197–218, 2002.[CrossRef][ISI][Medline]
  22. Gerke V, Creutz CE, Moss SE. Annexins: linking Ca2+ signaling to membrane dynamics. Nat Rev Cell Biol 6: 449–461, 2005.[CrossRef]
  23. Gettins PG. Serpin structure, mechanism, function. Chem Rev 102: 4751–4804, 2002.[CrossRef][ISI][Medline]
  24. Gharahdaghi F, Weinberg CR, Meagher DA, Imai BS, Mische SM. Mass spectrometric identification of proteins from silver-stained polyacrylamide gel: a method for the removal of silver ions to enhance sensitivity. Electrophoresis 20: 601–605, 1999.[CrossRef][ISI][Medline]
  25. Griffioen AW, Molema G. Angiogenesis: potentials for pharmacologic intervention in the treatment of cancer, cardiovascular diseases, and chronic inflammation. Pharmacol Rev 52: 237–268, 2000.[Abstract/Free Full Text]
  26. Gygi SP, Han DK, Gingras AC, Sonenberg N, Aebersold R. Protein analysis by mass spectrometry and sequence database searching: tools for cancer research in the post-genomic era. Electrophoresis 20: 310–319, 1999.[CrossRef][ISI][Medline]
  27. Gygi SP, Rochon Y, Franza BR, Aebersold R. Correlation between protein and mRNA abundance in yeast. Mol Cell Biol 19: 1720–1730, 1999.[Abstract/Free Full Text]
  28. Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86: 353–364, 1996.[CrossRef][ISI][Medline]
  29. Handley J. Software for MS protein identification. Anal Chem 74: 159A–162A, 2002.[Medline]
  30. Hotary KB, Allen ED, Brooks PC, Datta NS, Long MW, Weiss SJ. Membrane type I matrix metalloproteinase usurps tumor growth control imposed by the three-dimensional extracellular matrix. Cell 114: 33–45, 2003.[CrossRef][ISI][Medline]
  31. Hu L, Lau SH, Tzang CH, Wen JM, Wang W, Xie D, Huang MH, Wang Wu MC Y, Huang JF, Zeng WF, Sham JST, Yang M, Guan XY. Association of vimentin overexpression and hepatocellular carcinoma metastasis. Oncogene 23: 298–302.
  32. Jeong JS, Lee SH, Jung KJ, Choi YC, Park WY, Kim IH, Kim SS. Hepatotoxin N-nitrosomorpholine-induced carcinogenesis in rat liver: ex vivo exploration of preneoplastic and neoplastic hepatocytes. Exp Mol Pathol 74: 74–83, 2003.[CrossRef][ISI][Medline]
  33. Jiang D, Ying W, Lu Y, Wan J, Zhai Y, Liu W, Zhu Y, Qiu Z, Qian X, He F. Identification of metastasis-associated proteins by proteomic analysis and functional exploration of interleukin-18 in metastasis. Proteomics 3: 724–737, 2003.[CrossRef][ISI][Medline]
  34. Kammula US. Predicting outcomes after percutaneous ethanol injection for small hepatocellular cancer. Cancer J 12: 175–177, 2006.[ISI][Medline]
  35. Kim J, Kim SH, Lee SU, Ha GH, Kang DG, Ha NY, Ahn JS, Cho HY, Kang SJ, Lee YJ, Hong SC, Ha WS, Bae JM, Lee CW, Kim JW. Proteome analysis of human liver tumor tissue by two-dimensional gel electrophoresis and matrix assisted laser desorption/ionization-mass spectrometry for identification of disease-related proteins. Electrophoresis 23: 4142–4156, 2002.[CrossRef][ISI][Medline]
  36. Kojiro M, Roskams T. Early hepatocarcinoma and dysplastic nodules. Semin Liver Dis 25: 133–142, 2005.[CrossRef][ISI][Medline]
  37. Le Naour F, Brichory F, Misek DE, Brechot C, Hanash SM, Beretta L. A distinct repertoire of autoantibodies in hepatocellular carcinoma identified by proteomic analysis. Mol Cell Proteomics 1: 197–203, 2002.[Abstract/Free Full Text]
  38. Lee CL, Hsiao HH, Lin CW, Wu SP, Huang SY, Wu CY, Wang AHJ, KH. Strategic shotgun proteomics approach for efficient construction of an expression map of targeted protein families in hepatoma cell lines. Proteomics 3: 2472–2486, 2003.[CrossRef][ISI][Medline]
  39. Liang CRMY, Leow CK, Neo JCH, Tan GS, Lo SL, Lim JWE, Seow TK, Lai PBS, Chung MCM. Proteome analysis of human hepatocellular carcinoma tissues by two-dimensional difference gel electrophoresis and mass spectrometry. Proteomics 5: 2258–2271.
  40. Liang CRMY, Neo JCH, Lo SL, Tan GS, Seow TK, Chung MCM. Proteome database of hepatocellular carcinoma. J Chromatogr B Analyt Technol Biomed Life Sci 771: 303–328, 2002.[CrossRef][ISI][Medline]
  41. Li C, Hong Y, Tang YX, Zhou Hu, Ai JH, Li SJ, Zhang L, Xia QZ, Wu JR, Wang HY, Zeng R. Accurate qualitative and quantitative proteomics analysis of clinical hepatocellular carcinoma using laser capture microdissection coupled with isotope-coded affinity tag and two-dimensional liquid chromatography mass spectrometry. Mol Cell Proteomics 3: 399–409, 2004.[Abstract/Free Full Text]
  42. Li Y, Li Y, Tang R, Xu H, Qiu M, Chen Q, Chen J, Fu Z, Ying K, Xie Y, Mao Y. Discovery and analysis of hepatocellular carcinoma genes using cDNA microarrays. J Cancer Res Clin Oncol 238: 369–379, 2002.
  43. Lim SO, Park SJ, Kim W, Park SG, Kim HJ, Kim YI, Sohn TS, Noh JH, Jung G. Proteome analysis of hepatocellular carcinoma. Biochem Biophys Res Commun 291: 1031–1037, 2002.[CrossRef][ISI][Medline]
  44. Little D, Said JW, Siegel RJ, Fealy M, Fishbein MC. Endothelial cell markers in vascular neoplasms: an immunohistochemical study comparing factor VIII-related antigen, blood group specific antigens, 6 keto-PGF1 alpha and Ulex europaeus 1 lectin. J Pathol 149: 89–95, 1986.[CrossRef][ISI][Medline]
  45. Luo Q, Nieves E, Kzhyshkowska J, Angeletti RH. Endogenous transforming growth factor-beta receptor-mediated Smad signaling complexes analyzed by mass spectrometry. Mol Cell Proteomics 7: 1245–1260, 2006.
  46. Miao RQ, Agata J, Chao L, Chao J. Kallistatin is a new inhibitor of angiogenesis and tumor growth. Blood 100: 3245–3252, 2002.[Abstract/Free Full Text]
  47. Murray CJL, Lopez AD. Mortality by cause for eight regions of the world: Global burden of disease study. Lancet 349: 1269–1276, 1997.[CrossRef][ISI][Medline]
  48. Novikoff P, Ikeda T, Hixson DC, Yam A. Characterizations of and interactions between bile ductule cells and hepatocytes in early stages of rat hepatocarcinogenesis induced by ethionin. Am J Path 139: 1351–1368, 1991.[Abstract]
  49. Novikoff PM, Annamaneni P, Bielawski M. Angiogenesis in early stages of hepatoma development from carcinogen induced rat hepatocarcinogenesis model (Abstract). Proc Am Assoc Cancer Res 45: 830A, 2004.
  50. Novikoff PM, Cammer M, Tao L, Oda H, Stockert RJ, Wolkoff AW, Satir P. Three dimensional organization of rat hepatocyte cytoskeleton: relation to the asialoglycoprotein endocytosis pathway. J Cell Sci 109: 21–32, 1996.[Abstract]
  51. Novikoff PM, Yam A. Stem cells and rat liver carcinogenesis: contributions of confocal and electron microscopy. J Histochem Cytochem 46: 613–626, 1998.[Abstract/Free Full Text]
  52. Novikoff PM, Tulsiani DRP, Touster O, Yam A, Novikoff AB. Immunocytochemical localization of {alpha}-D-mannosidase II in the Golgi apparatus of rat liver. Proc Natl Acad Sci USA 80: 4364–4368, 1983.[Abstract/Free Full Text]
  53. Novikoff PM, Yam A, Oikawa I. Blast-like cell compartment in carcinogen-induced proliferating bile ductules. Am J Pathol 148: 1473–1492, 1996.[Abstract]
  54. Paliouras M, Diamandis EP. The kallikrein world: an update on the human tissue kallikreins. Biol Chem 387: 643–652, 2006.[CrossRef][ISI][Medline]
  55. Parent R, Beretta L. Proteomics in the study of liver pathology. J Hepatol 43: 177–183, 2006.[CrossRef][ISI]
  56. Parkin DM, Pisani P, Ferlay J. Estimates of the worldwide incidence of 25 major cancers in 1990. Int J Cancer 80: 827–841, 1999.[CrossRef][ISI][Medline]
  57. Rosenquist M. 14-3-3 proteins in apoptosis. Braz J Med Biol Res 36: 403–408, 2003.[ISI][Medline]
  58. Sabatini D, Kreibich G, Morimato I, Adesnik M. Mechanisms for the incorporation of proteins in membranes and organelles. J Cell Biol 92: 1–22, 1982.[Free Full Text]
  59. Sato K. Glutathione transferases as markers of preneoplasia and neoplasia. Adv Cancer Res 52: 205–217, 1989.[ISI][Medline]
  60. Seow TK, Liang RCMY, Leow C, Chung MCM. Hepatocellular carcinoma: from bedside to proteomics. Proteomics 1: 1249–1263, 2001.[CrossRef][ISI][Medline]
  61. Seow TK, Ong SE, Liang RCMY, Ren EC, Chan L, Ou K, Chung MCM. Two-dimensional electrophoresis map of the human hepatocellular carcinoma cell line, HCC-M, and identification of the separated proteins by mass spectrometry. Electrophoresis 21: 1787–813, 2000.[CrossRef][ISI][Medline]
  62. Shalhoub P, Kern S, Girard S, Beretta L. Proteomic-based approach for the identification of tumor markers associated with hepatocellular carcinoma. Dis Markers 17: 217–223, 2001.[ISI][Medline]
  63. Sherman MS. Hepatocellular carcinoma: epidemiology, risk factors, and screening. Semin Liver Dis 25: 143–154, 2005.[CrossRef][ISI][Medline]
  64. Shestakova EA, Wyckoff J, Jones J, Singer RH, Condeelis J. Correlation of beta-actin messenger RNA localization with metastatic potential in rat adenocarcinoma cell lines. Cancer Res 59: 1202–1205, 1999.[Abstract/Free Full Text]
  65. Shevchenko A, Wilm M, Vorm O, Mann M. Mass spectrometric sequencing of proteins silver-stained polyacrylamide gels. Anal Chem 68: 850–858, 1996.[Medline]
  66. Smith MW, Yue ZN, Geiss GK, Sadovnikova NY, Carter VS, Boix L, Lazaro CA, Rosenberg GB, Bumgarner RE, Fausto N, Bruix J, Katze MG. Identification of novel tumor markers in hepatitis C virus-associated hepatocellular carcinoma. Cancer Res 63: 859–864, 2003.[Abstract/Free Full Text]
  67. Solt D, Farber E. New principles for the analysis of chemical carcinogenesis. Nature 263: 701–703, 1976.[CrossRef]
  68. Solt DB, Hay JB, Farber E. Comparison of the blood supply to diethylnitrosoamine-induced hyperplastic nodules and hepatomas and to the surrounding liver. Cancer Res 37: 1686–1691, 1977.[Abstract/Free Full Text]
  69. Solt DB, Medline A, Farber E. Rapid emergence of carcinogen-induced hyperplastic lesions in a new model for the sequential analysis of liver carcinogenesis. Am J Pathol 88: 595–610, 1977.[ISI][Medline]
  70. Takematsu M, Mera Y, Ito N, Satoh K, Sato K. Relative merits of immunocytochemical demonstration of placental A, B and C forms of glutathione S-transferase and histochemical demonstration of {gamma}-glutamyl transferase as markers of altered foci during liver carcinogenesis in rats. Carcinogenesis 6: 1621–1626, 1985.[Abstract/Free Full Text]
  71. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76: 4350–4354, 1979.[Abstract/Free Full Text]
  72. Tulsiani DRP, Hubbard SC, Robbins PW, Touster O. {alpha}-D-Mannosidases of rat liver Golgi membranes. J Biol Chem 257: 3660–3668, 1982.[Abstract/Free Full Text]
  73. Tzivion G, Avruch J. 14-3-3 proteins: active cofactors in cellular regulation by serine/threonine phosphorylation. J Biol Chem 277: 3061–3064, 2002.[Free Full Text]
  74. Tzivion G, Shen YH, Zhu J. 14-3-3 proteins; bringing new definitions to scaffolding. Oncogene 20: 6331–6338, 2001.[CrossRef][ISI][Medline]
  75. Van Eyken P, Sciot R, Callea F, Ramaekers F, Schaart G, Desmet VJ. A cytokeratin-immunohistochemical study of hepatoblastoma. Human Pathol 21: 302–308, 1990.[CrossRef][ISI][Medline]
  76. Westbrook JA, Yan JX, Wait R, Welson SY, Dunn MJ. Zooming-in on the proteome: very narrow-range immobilized pH gradients reveal more protein species and isoforms. Electrophoresis 22: 2865–2871, 2001.