Proteomics 2001, 1, 1249–1263
1249
Review Teck Keong Seow1 Rosa C. M. Y. Liang1 Chon Kar Leow2 Maxey C. M. Chung1, 3 1
Bioprocessing Technology Centre 2 Department of Surgery 3 Department of Biochemistry National University of Singapore, Republic of Singapore
Hepatocellular carcinoma: From bedside to proteomics Hepatocellular carcinoma (HCC or hepatoma) is the most common primary cancer of the liver. It is responsible for approximately one million deaths each year, mainly in underdeveloped and developing countries. The aetiological factors identified in the development of HCC included persistent infection by hepatitis B and hepatitis C viruses, and exposure to aflatoxins. Although immunization can protect individuals from being infected by the hepatitis B virus, the early detection of HCC in those who have been infected by the virus remains a challenge. Thus most HCCs present late and are not suitable for curative treatment. Hence there is a tremendous interest and urgency to identify novel HCC diagnostic marker(s) for early detection, and tumour specific disease associated proteins as potential therapeutic targets in the treatment of HCC. Screening for these HCC proteins has been facilitated by proteomics, a key technology in the global analysis of protein expression and understanding gene function. Present and earlier proteome analyses of HCC have used predominantly experimental in vitro systems. The protein expression profiles of several hepatoma cell lines such as HepG2, Huh7, SK-Hep1, and Hep3B have been compared with normal liver, and nontransformed cell lines (Chang and WRL-68), while a comprehensive proteome analysis to create a protein database was carried out for the cell line HCC-M. In the future, proteome analyses utilizing tumour tissues, which reflect the pathological state of HCC more closely, will be undertaken. This work will complement the gene expression studies of HCC which are already underway. Efforts have also been directed at the proteome analysis of hepatic stellate cells, as these cells play an important role in liver fibrosis. Since liver fibrosis is reversible but not cirrhosis, it is of considerable importance to identify therapeutic targets that can slow its progression. Keywords: Hepatocellular carcinoma / Two-dimensional gel electrophoresis / Review PRO 0135
Contents 1 2 2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . Hepatocellular carcinoma: clinical aspects . Occurrence and distribution . . . . . . . . . . . . . Aetiological factors . . . . . . . . . . . . . . . . . . . . Hepatitis B virus . . . . . . . . . . . . . . . . . . . . . . Hepatitis C virus . . . . . . . . . . . . . . . . . . . . . . Aflatoxin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cirrhosis . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1250 1250 1250 1250 1250 1251 1251 1251
Correspondence: Associate Professor Maxey C. M. Chung, Department of Biochemistry, Faculty of Medicine, National University of Singapore, Singapore 119260, Republic of Singapore E-mail:
[email protected] Fax: +65-779-1453. Abbreviations: AFP, a-fetoprotein; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; LCM, laser capture microdissection; PMF, peptide mass fingerprinting
WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001
2.2.5 2.3 2.4 3 3.1 3.2 3.2.1 3.2.1.1
Others . . . . . . . . . . . . . . . . . . . . . . . . . . . Screening and diagnosis . . . . . . . . . . . . Treatment and prevention . . . . . . . . . . . . Hepatocellular carcinoma: proteomics . Early work . . . . . . . . . . . . . . . . . . . . . . . . Proteome analysis of HCC . . . . . . . . . . . Hepatoma cell lines . . . . . . . . . . . . . . . . HepG2, Huh-7, FOCUS, SK-Hep1, Chang and WRL-68 . . . . . . . . . . . . . . . . 3.2.1.2 BEL-7404 and L-02 . . . . . . . . . . . . . . . . . 3.2.1.3 HCC-M . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1.3.1 MALDI-TOF MS . . . . . . . . . . . . . . . . . . . 3.2.1.3.2 Nanoelectrospray ionisation MS/MS . . . 3.2.1.3.3 In silico assembly of novel protein, HCC-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1.4 cDNA microarray analysis of hepatoma cell lines . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Human liver tumour tissues . . . . . . . . . . 3.2.3 Gene expression studies of HCC . . . . . .
1251 1251 1252 1253 1253 1253 1253 1253 1255 1256 1256 1256 1257 1258 1258 1259
1615-9853/01/1009–1249 $17.50+.50/0
1250
T. K. Seow et al.
3.2.4 Hepatic stellate cells . . . . . . . . . . . . . . . . . . . 3.3 Protein database of hepatocellular carcinoma . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Concluding remarks . . . . . . . . . . . . . . . . . . . 5 References . . . . . . . . . . . . . . . . . . . . . . . . . .
Proteomics 2001, 1, 1249–1263 1260 1261 1262 1262
1 Introduction Proteomics refers to the study of the proteome, which is the total protein complement of a genome. It is presently considered to be the key technology in the global analysis of protein expression and in the understanding of gene function in the postgenomic era. Unlike the genome, which is essentially static, the proteome is a dynamic entity which changes with the physiological state of the cell. While this complicates the analysis of the proteome, it offers a powerful tool in the study of differential protein expression and post-translational modifications due to alterations in physiological conditions, such as development, apoptosis, drug treatment, and diseases [1]. As such, the application of proteomics has been the most established in the clinical and biomedical fields. For example, in the study of human diseases, disease specific/associated proteins can be identified by comparing the protein profiles of normal versus diseased tissues or biological fluids. Since these proteins are potential diagnostic tools or leads for the development of drugs, proteomics has great potential in the drug discovery process [2, 3]. The disease that has been studied most widely by proteomics is cancer. Being a complex and multifactorial disease it is anticipated that several candidate proteins may be involved in tumorigenesis, and screening for these proteins has been facilitated by the proteomics approach [4]. This complements well with the earlier work on the identification of cancer-related genes by using expression genetics [5]. Cancer proteomics has recently been reviewed by Alaiya et al. [6]. The cancers covered in the review included bladder, kidney, breast, lung, ovarian and prostate, as well as leukaemia [6]. In this paper, we review recent progress made in the study of hepatocellular carcinoma (HCC), one of the commonest cancers worldwide, and a leading cause of death in Africa and Asia [7].
2 Hepatocellular carcinoma: clinical aspects 2.1 Epidemiology Hepatocellular carcinoma (HCC or hepatoma) is the most common primary cancer of the liver. It is responsible for approximately one million deaths each year [8]. HCC has
been a malignancy of the underdeveloped and developing countries but its incidence in the developed countries is on the rise. During the past three decades, the incidence of HCC has increased substantially in Japan [9] and increased slightly in the United Kingdom and France [10, 11]. Recently, El-Serag and Mason [12] reported that the incidence of HCC in the United States has increased from 1.4 per 100 000 population during the period 1976– 1980 to 2.4 per 100 000 population for the period 1991– 1995. Singapore, classified as an intermediate incidence country, has an HCC incidence of 18.9 per 100 000 population per year [13]. Depending on geographical location, HCC is four to eight times more common in males than in females. Moreover, its occurrence increases progressively with age. However, this too varies between high and low incidence countries. In high incidence countries (such as China, Mozambique) the mean age at diagnosis is in the third decade but in low incidence countries (such as United States, United Kingdom) it occurs two to three decades later.
2.2 Aetiological factors 2.2.1 Hepatitis B virus Persistent viral infection is probably the most important cause of HCC worldwide. Two viruses cause almost all these tumours: hepatitis B virus (HBV) and hepatitis C virus (HCV). In the high and intermediate HCC incidence countries, the prevalence of chronic carriers of HBV is correspondingly high. The estimated number of chronic carriers world-wide is 400 million. The highest rates are in Asia and sub-Saharan Africa. The high HBV infection rates of 10–25% are due to vertical transmission from mothers to infants or horizontal spread within families to children under the age of 10 y [14]. The risk of HCC in a chronic HBV carrier is increased by a factor of 100 as compared to a noninfected individual [15] and the estimated lifetime risk of HCC in males infected at birth is 50% [14]. HBV infection results in chronic liver injury. This includes inflammation, liver regeneration, liver fibrosis and cirrhosis which predisposes to HCC. Since not all HBV-related HCC occurs on a background of liver cirrhosis, it implies that HBV, a DNA virus, may have some intrinsic hepatocarcinogenic properties. For example, following its integration into the host genome, the hepatitis B virus X protein (HBx protein) could influence the development of liver cancer by promoting the survival and growth of transformed hepatocytes. Through transactivation of cellular genes and induction of signalling pathways, HBx may lead to
Proteomics 2001, 1, 1249–1263
Proteomics of hepatocellular carcinoma
1251
changes in cell cycle progression and/or regulation as shown by the ability of HBx in inducing cell cycle progression in Chang liver cells [16]. A second mechanism in which HBx can contribute to hepatocarcinogenesis is through its ability to inhibit DNA repair. This is mediated through inhibition of global nucleotide excision repair [17] and direct interaction with DNA repair proteins/protein complexes: p53 [18] and transcription factor II H (TFII H) [19].
tation studies, 3–6% of explanted livers contained small unsuspected HCCs [26]. The chronic inflammation and cell damage in a cirrhotic liver may have provided the proliferative stimuli to promote the carcinogenic process in pre-initiated hepatocytes through growth hormones (e.g. hepatocyte growth factor, transforming growth factor a, transforming growth factor b), cytokines and growthrelated proto-oncogenes (e.g., c-fos, c-jun, c-myc).
2.2.2 Hepatitis C virus
2.2.5 Other factors
HCV infection is the cause of the rising incidence of HCC in developed countries. While parenteral transmission is the major route of transmission of the virus, up to 50% of chronic HCV carriers report no prior exposure to any known risk factors. Once acutely infected, approximately 80–85% of these individuals will become chronically infected. In at least 20% of those with chronic HCV infection, liver cirrhosis will develop during the first 10 years. On a background of cirrhosis, HCC will develop at a rate of 1–4% per year. After 20 years, 1.9–6.7% of all chronic HCV carriers will develop HCC [12]. It is estimated that 3% of the world population are HCV carriers [20]. Whether HCV is directly carcinogenic remains uncertain. HCV is an RNA virus and the integration of its nucleic acid sequences into the host genome seems less likely. Several in vitro studies by Ray et al. [21] have demonstrated that the viral core protein promotes cell growth through repression of p53 transcription, inhibits apoptotic cell death [22] and transforms rat embryo fibroblasts [23].
The association of HCC with inherited metabolic diseases probably reflects the end point of the liver disease associated with the metabolic abnormality. Those conditions where HCC occurs against a background of cirrhosis include haemochromatosis, a-1-antitrypsin deficiency, porphyria cutanea tarda, Wilson’s disease and tyrosinaemia. Glycogen storage diseases, hereditary forms of porphyria, hypercitrullinaemia, and hereditary fructose intolerance are associated with HCC in a liver without cirrhosis.
2.2.3 Aflatoxins The hepatocarcinogen aflatoxins are mycotoxins produced by Aspergillus flavus and Aspergillus parasiticus. Aflatoxins are an important cause of HCC. Improper storage of crops such as corn, peanut and rice leads to the growth of these moulds and contamination of the crops with aflatoxins. In regions where aflatoxin exposure is high, such as Qidong Province of China and sub-Saharan Africa, the incidence of HCC is correspondingly high. The liver metabolises the aflatoxins, and aflatoxin B1 induces a specific guanine-to-thymidine point mutation in codon 249 of the p53 tumour suppressor gene in hepatocytes [24]. For individuals who are concurrently exposed to aflatoxins and are chronic HBV carriers, the probability of developing HCC is increased three-fold [25].
2.2.4 Liver cirrhosis Liver cirrhosis is linked to the development of HCC. At post mortem, about 10% of patients dying with cirrhosis are found to have unsuspected HCC. In liver transplan-
Finally, other aetiological factors associated with the development of HCC include alcohol, smoking, thorium dioxide (Torotrast), oestrogens and androgens.
2.3 Screening and diagnosis As HCC can reach an advanced stage before it presents clinically, regular screening at six monthly intervals is recommended for those at risk individuals. This entails the performance of a transabdominal ultrasound scan of the liver to detect for the presence of tumour nodule(s) and the measurement of the serum tumour marker a-fetoprotein (AFP). An ideal tumour marker should be specific for that particular type of cancer, only produced by it and not by any nonmalignant conditions. It should be absent or present in indistinct amounts in normal individuals and become elevated even in the presence of a very small tumour. The biological half-life should be short and the quantity of the marker should correlate with the tumour load so that it can be used in staging the disease and to monitor response to treatment. It should also be easily and rapidly measurable by available laboratory methods. The tumour markers that have been used in patients with HCC include a-fetoprotein (AFP), serum ferritin, g-glutamyltranspeptidase isoenzyme, alkaline phosphatase, des-g-carboxy prothrombin, a-1antitrypsin, aldolase A, 5’-nucleotide phosphodiesterase, tissue polypeptide antigen and a-1-fucosidase. While AFP can only fulfil part of the requirements of an ideal tumour marker, it is still the best available marker presently and remains the most widely used test to aid in the diagnosis of and monitoring for HCC.
1252
T. K. Seow et al.
A serum AFP concentration of .500 ng/mL in a high risk individual is usually diagnostic of HCC. However, about 70% of HCC patients have levels above the reference range of .10 ng/mL but below the diagnostic level of .500 ng/mL. Furthermore, individuals without HCC but with benign chronic liver disease can have elevated AFP in this ‘grey zone’ of elevated AFP. Based on the variable affinity of AFP variants from benign liver diseases and HCC for lectins, Du et al. [27] reported the use of a lentil lectin column to differentiate HCC from benign liver diseases. However these tests are tedious and the results are sometimes not reproducible. As an alternative to testing lectin affinities, an IEF technique was applied to separate and identify HCC-specific AFP isoforms. Three AFP isoform bands were identified. Band + I was seen in patients with benign liver disease such as cirrhosis, Band + II in patients with HCC and Band + III in patients with nonseminomatous germ cell tumours [28]. The specificity of these AFP isoforms was prospectively tested on patients with chronic liver disease and patients with AFP levels .50 ng/mL. Seventeen of the twenty-six patients were found to have HCC on radiological examination at a median of 3.6 m (range 1–18 m) after the HCC-specific isoform of AFP was first detected [29]. More recently the same group of researchers have deduced the structures of the disease specific AFP glycoforms [30]. Once an underlying HCC is suspected because of a suspicious ultrasound scan, elevated AFP or both, the next radiological investigation is a triphasic computed tomography (CT) of the liver. If there is doubt after the CT, then a lipiodol CT is performed. Lipiodol CT is preferred in the East and it is the most sensitive preoperative imaging technique for detecting HCC [31]. Magnetic resonance imaging (MRI) of the liver for HCC is an alternative. While MRI is better than CT in detecting HCC, whether it is superior to lipiodol CT remains to be proven.
2.4 Treatment and prevention Without treatment, the median survival from the time of diagnosis to death is approximately 4 m. The treatment options available to patients with HCC are surgery, systemic chemotherapy, loco-regional treatment and symptomatic relief. Of these, surgery, be it surgical resection or liver transplantation, is the only treatment which has the potential to cure. At presentation, liver resection is only feasible for 10– 15% of patients. The reasons for this low resectability rate include extensive local disease, presence of extrahepatic disease and poor functional liver reserve precluding any form of liver resection. The five year disease free survival following curative liver resection ranges from 29–58% [32]. In a selected population, treating HCC patients with liver transplantation can achieve a 75% five year disease free
Proteomics 2001, 1, 1249–1263 survival [33]. This is possible because liver transplantation replaces the diseased liver with a normal one, prevents the development of a metachronous hepatoma within the remnant cirrhotic liver after resection, relieves portal hypertension and its related complications and produces less operative mortality and morbidity than cirrhotic liver resection. However, shortage of suitable organ donors limits its application. Systemic chemotherapy, percutaneous ethanol injection into the tumour, cryotherapy or radio frequency ablation of the lesion, selective internal irradiation with yittrium-90 microspheres and transarterial chemoembolization can only palliate but not cure the disease. Apart from developing better methods to detect HCC at an earlier stage to allow the performance of curative surgery, can we prevent the development of HCC? Primary prevention with vaccination has been shown to have a dramatic effect in Taiwan. The introduction of compulsory vaccination of children against HBV has reduced the rate of chronic infection from 10% to ,1% and this has been associated with a significant reduction in the incidence of HCC in children [34]. At present no vaccine is available for HCV. In developed countries, screening all blood donors for HBV and HCV infections has been successful in the prevention of transfusion-associated HBV and HCV infection. Reducing or removing environmental carcinogens from the food chain could reduce the incidence of HCC. By ensuring agricultural products such as corn, peanuts, sorghum and rice are stored properly to avoid contamination with aflatoxin and replacing pond-ditch water contaminated with microcystin with deep well water as drinking water has reduced the risk of developing HCC. In HBV or HCV chronic infection, secondary preventive measures such as the use of interferon (IFN) alone or in combination with antiviral agents (such as ribavirin for HCV) are attempts to inhibit viral replication to decrease the activity of liver disease with the hope of reducing the risk of developing cirrhosis and HCC. For HCV, sustained virological and histological response to a-IFN treatment has been demonstrated but survival benefit has not been clearly demonstrated [20]. Once cirrhosis is established, any preventive measures (tertiary prevention) are to reduce the risk of developing HCC. Some successes in reducing the incidence of HCC in HCV cirrhotic patients with a-IFN [35] and the Chinese herbal medicine ‘Sosaiko-to’ [36] have been reported. With the development of better screening methods such as specific tumour markers, more effective antiviral therapies and implementation of preventive measures, one hopes that it may be possible to make a significant dent in the incidence of one of the most common and dreaded malignancies in the world in the near future.
Proteomics 2001, 1, 1249–1263 Primary preventive measures, such as immunization, can stop individuals from being infected with the hepatitis virus. However we are still faced with the problem of how to detect an HCC early in those who have been infected with the virus or are at risk of developing HCC. The use of serum tumour marker, such as the AFP isoforms Band + II, may pave the way for clinicians to ‘prediagnose’ HCC before it is radiologically visible. With this information of ‘microscopic’ disease, clinicians in the future can possibly treat and ‘cure’ the HCC before it is manifested clinically. By analyzing the proteome of HCC, it is hoped that one may identify: (1) novel diagnostic marker(s) which might offer better specificity and sensitivity than AFP thus giving clinicians a ‘perfect’ tumour marker to screen for and monitor the progress of HCC; and (2) specific disease associated proteins that are potential therapeutic targets in the treatment of HCC.
3 Hepatocellular carcinoma: proteomics We have concentrated our survey on approaches that utilised 2-DE as the main method of fractionation of liver proteins, since proteome analysis has had its origin from the large scale separation of proteins by 2-DE [37]. Since its first publication there was a tremendous upsurge of interest and the realization of the potential of the applications of 2-DE in clinical laboratories. This was summarized by two special issues of Clinical Chemistry [38, 39] and a monograph [40]. Since then, numerous advances have been made in 2-DE technologies, including the introduction of IPG in the first dimension [41] which was later to become one of the main technology platforms in the current field of proteomics [42–44].
3.1 Early work Review of the existing publications showed that most of the earlier work on the separation and identification of liver proteins used transformed cell lines [45, 46] and liver tissues (normal and regenerating rat livers, rat hepatomas) as starting materials [47, 48]. Some of the above work had led to the detection of several transformation and proliferation associated protein variants, but it was not possible to identify and hence assign functions to these proteins [46]. This deficiency could be overcome by obtaining partial primary sequences of these proteins, but the inability to isolate sufficient proteins for microsequencing by 2-DE had severely limited this identification process. However, subsequent introduction of IPG-based IEF [41], and increasing the protein loading to 1 mg for preparative separations without apparent loss of resolution was a major consideration in the use of
Proteomics of hepatocellular carcinoma
1253
2-DE to construct 2-DE maps of different tissues [49]. Thus, after an electroblotting step, major protein spots in a 2-DE map that were visible after Coomassie blue staining were successfully identified by Edman degradation, if they did not contain blocked N-termini [49]. Other methods that have been used for identification of proteins included internal peptide sequencing, immunoblotting, and comigration of known proteins [50]. Finally, combining silver staining [51] and MS based peptide mass fingerprint (PMF) analyses of 2-DE separated proteins resulted in the emergence of proteomics in its present form [52, 53]. Protein identification methods in proteomics have been reviewed recently by Gevaert and Vandekerckhove [54].
3.2 Proteome analysis of liver proteins Recently, several hepatoma cell lines have been used for proteome analyses with the view to understand better the underlying process of hepatocarcinogenesis. Cell lines were chosen since they were more homogeneous in nature when compared to liver tumour tissues. In addition, cell lines derived from human tumours have been used extensively as in vitro models of various diseases.
3.2.1 Hepatoma cell lines 3.2.1.1 HepG2, Huh-7, FOCUS, SK-Hep1, Chang and WRL-68 In an earlier study, Wirth et al. [55] analysed and compared the 2-DE maps of hepatic proteins between normal adult liver and two nontransformed, Chang and WRL-68, and four human hepatoma derived cell lines, HepG2, Huh-7, FOCUS and SK-Hep1. Using 60 commonly expressed human liver proteins identified from these samples as a basis of comparison semiquantitatively, it was found that the proteins in Chang and WRL-68 cells showed the greatest resemblance to those found in normal human liver. On the other hand, marked differences in protein expression were observed between normal liver and each of the transformed cell lines, HepG2, FOCUS, Huh-7 and SK-Hep1 [55] (Table 1). In a separate study, Sanchez et al. [56] also presented 2-DE maps of human liver, HepG2 cells and its secreted proteins (HepG2sp) in the SWISS-2DPAGE database. Proteins from these micropreparative 2-DE gels were mainly identified by N-terminal sequencing [55] and by a combination of immunoblotting, microsequencing, amino acid composition, pI and Mr [56]. These methods of protein analysis limited the number of proteins that could be identified as they were not sensitive enough to analyse proteins that were visible by silver staining. Nevertheless,
1254
T. K. Seow et al.
Proteomics 2001, 1, 1249–1263
Table 1. Differentially expressed proteins/genes among normal adult liver, nontransferable human derived, and human patina derived cell lines Proteins
Normal liver
15-Hydroxyprostaglandin ↑ dehydrogenase 3-Hydroxyisobutyrate ↑ dehydrogenase a2-Macroglobina) – AFP – – Aggrecan 1a) – Angiotensin receptor 1a) Calmodulin – ↑ Calreticulin precursor (Calreticulum) (Calregulin) (CRP55) (ERP60) (HACBP) Carbamoyl phosphate synthase ↑ – Caveolin 1a) Cytochrome c oxidase ↓ polypeptide VA – Cytochrome p450 reductasea) Elongation factor 1B (EF-1B) ↑ Endoplasmin ↓ Ephrin-A1 – ↑ ER-60 (58 kDa microsomal protein) (Phosphatidylinositol 4,5-bisphosphate phosphodiesterase) F1 ATPase b subunit ↑ A-FABP (adipocyte lipid-binding – protein) E-FABP (psoriasis-associated – FABP homolog) Liver fatty acid binding protein ↑ Glutamate dehydogenase ↑ – Glutamate receptor, ionotropic, N-methyl D-aspartate 2Ca) Glutathione peroxidase ↑ Glutathione S-transferse P ↓ (GST P) (class PI) GSTP1-P1 Glyceraldehyde 3-phosphate ↓ dehydrogenase Growth arrest-specific 1a) – Haptoglobin A chain precursor ↑ ↓ Heat shock protein 60 (Mitochondrial protein 2) (Ribosomal protein A) HSP27 (Stress responsive – protein 27) (SRP27) Heat shock protein 70 ↑ Heat shock protein 90 ↑ Homo sapiens full length – insert cDNA clone, YW24E06a) Human common acute – lymphoblastic leukemia antigen (CALLA)a)
CHANGd) THLE5be)
L-02
BEL7404
HepG2f) Huh7f)
SKHep1e)
WRL68d)
FOCUS
Hep3Bf)
n. d
–
–
–
n. d
n. d
n. d
n. d
n. d
–
↓
–
–
–
↓
↓
↓
↓
↓
–
– – – – – ↑
↓ ↓ ↑ ↓ – –
– – – – ↓ ↓
– – – – ↑ ↑
↑ ↑ ↓ ↑ – ↑
↑ ↑ ↓ ↑ – ↑
– – – – – ↑
– – – – – ↑
– – – – – ↑
↑ ↑ ↓ ↑ – –
↑ – ↓
– ↑ –
– – –
– – –
n. d ↓ ↑
n. d ↓ ↓
n. d – ↓
↓ – ↓
n. d – ↓
– ↓ –
– ↓ ↑ – ↓
↓ – – ↓ –
– – – – –
– – – – –
↑ ↓ ↑ ↑ ↓
↑ ↓ ↑ ↑ ↓
– ↓ ↑ – ↓
– ↓ ↑ – ↓
– ↓ n. d – ↓
↑ – – ↑ –
↓ –
– –
– ↑
– n. d
↑ –
↑ –
↑ –
↑ –
↑ –
– –
–
–
↑
n. d
–
–
–
–
–
–
n. d ↓ –
– – ↓
– – –
– – –
n. d. ↓ ↑
n. d ↓ ↑
n. d ↓ –
n. d ↓ –
n. d ↓ –
– – ↑
↑ ↓
– –
– ↓
– ↑
n. d n. d
↑ ↑
↑ n. d
↑ ↑
↑ n. d
– –
↓
–
–
–
↑
↓
↑
↓
↑
–
– ↓ ↓
↑ – –
– – –
– – –
↓ ↓ ↑
↓ ↓ ↓
– ↓ ↓
– ↓ ↓
– ↓ ↓
↓ – –
–
–
↓
↑
–
–
–
–
–
–
↓ n. d –
– – ↓
– – –
– – –
n. d ↑ ↑
n. d n. d ↑
↓ ↑ –
↑ n. d –
↓ n. d –
– – ↑
–
↑
–
–
↓
↓
–
–
–
↓
Proteomics 2001, 1, 1249–1263
Proteomics of hepatocellular carcinoma
1255
Table 1. Continued Proteins
Normal liver
CHANGd) THLE5be)
Human MOP1a) Inosine-5’-monophosphate dehydrogenase 2 (IMPDH-II) Maspin precursor Meltrin gamma, A disintergrin and metalloproteinase domain 9 (ADAM-9)a) Mitochondrial enoyl CoA hydratase Monoamine oxidase Ba) NADH dehydrogenase (ubiquinone) NF-IL6a) Peptidyl-prolyl cistrans isomerase Ab) Peroxiredoxin 2 (Natural killer cell enhancing factor B) Peroxiredoxin 3 (MER5 homologous protein) Phospatidylethanol-aminebinding protein Pyruvate carboxylase Reticulocalbin homolog-human Serotransferrin Short chain acyl-CoA dehydrogenase Spectrin b-chain Superoxide dismutase (Mn) TGF-b2a) Thioredoxin Transcriptional regulator ISGF3-g subunita) Translationally controlled tumor proteinc) Transmembrane 4 superfamily member 2a) Triosephosphate isomerase Tubulin beta-1 chain Ubiquitin
– –
– –
– –
↑ = upregulation ↓ = downregulation n. d = not detected
L-02
BEL7404
HepG2f) Huh7f)
SKHep1e)
WRL68d)
FOCUS
Hep3Bf)
↑ –
– ↓
– ↑
↓ –
↓ –
– –
– –
– –
↓ –
– –
– ↑
↑ –
↓ –
– ↓
– ↓
– –
– –
– –
– ↓
↑
↓
–
–
–
n. d
↓
↓
↓
↓
–
– ↑
– n. d
↓ –
– –
– –
↑ n. d
↑ n. d
– n. d
– n. d
– n. d
↑ –
– ↓
– ↑
↑ –
– –
– –
↓ ↓
↓ ↓
– ↑
– ↑
– ↓
↓ –
–
–
–
↑
↓
–
–
–
–
–
–
↑
↓
–
–
–
n. d
↓
↓
↓
↓
–
↑
↓
–
–
–
n.
↓
↓
↓
↓
–
↑ – ↑ ↑
n. d – n. d ↓
– – – –
– ↓ – –
– ↑ – –
n. d – n. d ↓
n. d – n. d ↓
n. d – n. d ↓
n. d – n. d ↓
n. d – n. d ↓
– – – –
↑ ↑ – ↑ –
↓ n. d – ↑ –
– – ↓ – ↓
– – – – –
– – – – –
n. d n. d ↑ ↑ ↑
n. d n. d ↑ ↑ ↑
↓ ↓ – ↓ –
↓ ↓ – ↓ –
n. d ↓ – ↑ –
– – ↑ – ↑
↓
↑
–
–
–
↑
↑
↑
↑
↑
–
–
–
↓
–
–
↑
↑
–
–
–
↑
↑ – ↓
↑ – ↑
– – –
– ↑ –
– ↓ –
↑ – ↑
↑ – ↑
↑ – ↑
↑ – ↑
↓ – ↑
– – –
a) genes differentially experessed in cDNA microarray analysis [72] b) isoform at 7.0/16 (Mr/pI) c) isoform at 6.9/26 (Mr/pI)
the results derived from this study were very useful as it was an initial attempt to create a 2-DE database for liver proteins, and which also allowed comparison to be made with the corresponding results from the SWISS-2D PAGE which is accessible through the WWW network on the ExPASy molecular biology server [55, 56].
d) Non transformed cell line: Chang, WRL-68 e) NonAFP producing hepatoma: THLE-5b, SK-Hep1 f) AFP-producing hepatomas: HepG2, Huh7, Hep3B
3.2.1.2 BEL-7404 and L-02 A recent communication by Yu et al. [57] reported the identification of differentially expressed proteins between a human hepatoma (BEL-7404) and normal (L-02) liver cell line by 2-DE and liquid chromatography-ion trap-MS
1256
T. K. Seow et al.
(LC-IT-MS). Ninety-nine proteins spots were found to show significant differences in expression between the two cell lines, but only 12 of which were identified from a Coomassie blue stained preparative gel. Out of these, seven protein spots had higher expression levels in the hepatoma cell line BEL-7404, while the remaining five were up-regulated in the normal cell line L-02 (Table 1). On the basis of their identities, the functional implication of these differentially expressed proteins in hepatocarcinogenesis was discussed. Some of these proteins, such as reticulocalbin, are demonstrated to be present in hepatoma cells for the first time, while adipocyte-type fatty acid-binding protein (A-FABP) had earlier been found to be associated with the progression of human bladder transitional cell carcinomas [58].
3.2.1.3 HCC-M The most comprehensive proteomic analysis of a hepatoma cell line, HCC-M, was carried out by Seow et al. [59]. HCC-M, was a HBV surface antigen-positive derived cell line [60]. An integrated approach consisting of 2-DE, MALDI-TOF and MS/MS, bioinformatics, and molecular biology was used to characterize the expressed proteins of this cell line (Fig. 1).
3.2.1.3.1 MALDI-TOF MS In this study, about 2000 spots were visualized by silver staining [51], and 408 spots were directly excised from the 2-DE gel for MALDI-TOF MS analysis. Out of these,
Proteomics 2001, 1, 1249–1263 301 spots gave good spectra, but only 272 spots were positively identified from the protein database search. However, the 272 spots were the products of 192 genes as a single gene often gave rise to multiple protein spots on a 2-DE gel. These spots represent protein variants arising from alternative splicing, post-translational modification, protein degradation, or isoforms encoded by copies of a duplicated gene. These results have been tabulated and mapped onto the 2-DE map of HCC-M [59] (Fig. 2).
3.2.1.3.2 Nanoelectrospray ionisation MS/MS Among the unidentified spots, 29 yielded good MALDITOF spectra but had no confident protein database search results. These spots were subsequently analysed by nanoelecrospray ionisation (nESI) MS/MS, which is the method of choice for high sensitivity MS/MS analyses [61]. Most of the spots were subjected to ‘peptide tag sequencing’ [62] to identify the proteins [63]. For spots that still did not return matches from the databases, de novo sequencing [64] was performed followed by subsequent searches in the protein and nucleotide databases with the NCBI Basic Local Alignment Search Tool (BLAST). One of the possible reasons why these proteins were not identified was probably due to the absence of the protein entry in the database at the time of the search. For example, two of the proteins identified, human N-acetylneuraminic acid phosphate synthase [65] and human inorganic
Figure 1. Flowchart of the integrated approach utilised in the proteomics analysis of the HCC-M cell line. More than 270 of the protein spots were identified using MALDI-MS analysis while 29 of the spots were identified by MS/MS analysis. The novel HCC-1 protein was characterised using bioinformatics and functional genomics tools.
Proteomics 2001, 1, 1249–1263
Proteomics of hepatocellular carcinoma
1257
Figure 2. Analytical 2-DE map of the HCC-M cell line from Seow et al. [59]. Approximately 120 mg of protein was focused on a pH 3–10 nonlinear IPG strip before being separated on a 10% polyacrylamide gel. Identified proteins are indicated according to their respective SWISS-PROT or NCBI accession numbers. Asterisked labels indicate comigrating proteins.
pyrophosphate, were only entered into the NCBI protein database in February and June 2000 respectively [63]. Initially, the former protein was matched to an unnamed protein from a teratocarcinoma cell line, and was predicted to contain sialic acid and antifreeze domains by bioinformatics tools. Interestingly, in the recent publication of the human genome in Nature [66], a human sialic acid synthase was reported to contain a domain homologous to polar fish antifreeze protein III. It was also apparent that most of the proteins in this category were located in the lower molecular mass region of the 2-DE gel (Fig. 3). It has been suggested that these lower molecular mass proteins might not have generated sufficient number of tryptic peptides for unambiguous identification by MALDI-TOF analysis [67]. Finally, these proteins might be novel or might possess post-translational modifications which made it difficult to identify based on PMF data alone.
3.2.1.3.3 In silico assembly of novel protein, HCC-1 The role of bioinformatics in a novel protein discovery from the HCC-M cell line was illustrated by the in silico assembly of HCC-1. HCC-1 is a protein spot with a pI of 6.2 and mass of 35 kDa (Fig. 4), and did not have
any match with proteins from the protein database. Three tryptic peptides were selected for de novo sequencing, and used to search the EST database [68]. The Cap3 assembly (TigemNET) software was used to assemble to a putative protein of 210 amino acids (Fig. 5). The assembled putative protein sequence was validated by the masses of three additional peptides which were present in the original MALDI-TOF spectrum (Fig. 5). Subsequently, using two DNA primers designed from two peptide fragments, rapid amplification of cDNA ends (RACE) was carried out to confirm the assembled sequence of the putative protein as well [69]. Bioinformatics tools were then used to predict the structure and function of the putative protein. It was suggested that the first 42 amino acid residue peptide fragment of HCC-1 is a SAP (a putative bihelical nucleic acid binding motif) domain with identity matches to heterogeneous nuclear ribonucleoproteins (hnRNP) from various vertebrate species including human. A nuclear localization signal was also predicted at residue position 197–201 of the protein. Finally, various molecular biology techniques were used to characterize the novel gene/protein with the view to establish its potential role in hepatocarcinogenesis [69]. It is apparent from the above that an inte-
1258
T. K. Seow et al.
Proteomics 2001, 1, 1249–1263
Figure 3. Preparative 2-DE map of the HCC-M cell line from Ou et al. [63]. Approximately 300 mg of protein was focused on a pH 3–10 nonlinear IPG strip before being separated on a 10% polyacrylamide gel. Identified proteins are indicated according to their respective SWISS-PROT or NCBI accession numbers. Asterisked labels indicate comigrating proteins. Protein spots labelled in red were identified using MS/MS analysis while those labelled in black were identified by MALDI-MS [59].
grated proteomics approach is a very powerful means of discovering new proteins and predicting their structures and functions (Fig. 1).
3.2.1.4 cDNA microarray analysis of hepatoma cell lines Recently, several reports have described the classification of human cancers based on gene expression profiling using cDNA microarrays [70, 71]. Of particular relevance to this review is the paper by Kawai et al. [72] which detailed the cDNA microarray and cluster analysis of seven hepatoma cell lines. Among the seven cell lines analysed, five AFP producing cell lines (Huh 7, Hep3B, HepG2, Huh 6, and PLC/PRF/5) have common gene expression profiles when compared to the AFP-negative cell lines (SK-Hep1 and HLE) and cancer cells of nonhepatocyte origin (KMBC and HeLa). In addition, HepG2, Huh7, and Hep3B, had higher expressions of AFP and shared common gene expression profiles even when compared with other AFP producing cells. From the expression profiles of these three AFP-positive cell lines,
18 of the genes (Table 1) showed altered expression levels (11 up-regulated and 7 down-regulated). Five of these genes, including AFP were previously shown to be involved in HCC, while six others were linked with other types of cancer [72]. This result implies that the classification of hepatoma cell lines by gene expression profiling is possible, but whether this is applicable to clinical samples of HCC is not established at present.
3.2.2 Human liver tumour tissues Almost all proteomics research on HCC has hitherto used cell lines to study protein differential expression and to create a database of liver proteins. This is understandable as cell cultures are relatively homogeneous. However, ultimately, more meaningful results will derive from the study of differential protein expression in human tumours. Unfortunately sample preparation using solid tissues presents numerous problems, especially with regards to tissue heterogeneity [6, 44]. Methods such as flow cytometry and antibody purification [73] have been used to alleviate this problem, but a more generalized and simple
Proteomics 2001, 1, 1249–1263
Proteomics of hepatocellular carcinoma
1259
Figure 4. Preparative 2-DE map of the HCC-M cell line with the novel HCC-1 protein circled. The observed and theoretical pI and Mr of the protein are indicated in the table. The theoretical values were calculated based on the amino acid sequence of the protein.
method is still lacking. Thus, some investigators have continued to use tumour tissues without prior purification of tumour cells [74, 75]. However, a recent report on the use of a laser-capture microdissection (LCM) to pick specific cell types from normal and cancerous tissues for 2-DE looked promising [76]. The only shortcoming of the method is that it required 13 h to pick sufficient cells for one 2-DE experiment. The widespread use of the technique will probably hinge on overcoming this slow process of cell picking, or that more sensitive techniques are developed specifically for proteome analysis. In our work, we have selected matched sample pairs (normal versus diseased) from the same liver for solubilization before 2-DE. Figures. 6A and 6B show the 2-DE patterns of a well and a poorly differentiated liver tumour versus tissue specimens from the normal region of the same liver, respectively. Our current efforts are directed at the analysis and identification of the differentially expressed protein spots from these liver tissue samples.
3.2.3 Gene expression studies of HCC Several recent reports on differential gene expression studies of HCC have been published. Most notable were those by Kondoh et al. [77], Shuda et al. [78] and Tanaka et al. [79]. These authors used surgically resected primary HCCs and adjacent nontumorous portions of the liver from patients, and hepatoma cell lines (Huh7, HepG2, and HLF) for their studies. Eight cDNAs encoding galectin 4 (Gal-4), UGT2B4 (UDP-glucuronosyltransferase), ribosomal phosphoprotein P0 (rpP0), dek, insulin-like growth factor binding protein (IGFBP) 1, vitronectin, retinoic acidinduced gene (RIG-E), and CYP3A4 (cytochrome P450 nifedipine oxidase) were found to be differentially expressed between HCC and the matched nontumorous liver tissues [77]. Further characterization studies suggested that high expression levels of these genes could be correlated with the malignant potential of HCC. In another study, Shuda et al. [78] examined the expression
1260
T. K. Seow et al.
Proteomics 2001, 1, 1249–1263
Figure 5. MALDI spectrum and amino acid sequence (inset) of the novel HCC-1 protein from HCC-M cell line. The three peptides indicated by blue asterisks were subjected to de novo sequencing by MS/MS. The sequences of the peptides are boxed and in blue (inset), and were utilised in the in silico assembly of the protein. The peptides indicated by green asterisks were the three additional peptides with masses that matched the underlined and green-coloured sequences (inset) of the assembled sequence.
of translation factor mRNAs in HCC, and suggested that the coordination and specific activation of the translator genes might be involved in the process of liver carcinogenesis. Finally, the recent paper by Tanaka et al. [79] described the identification of four genes (antisecretory factor-1, AF, gp96, DAD1, and CDC34) that were upregulated specifically along with HCC progression. The mRNA levels of these four genes were also high in the HCC cell lines (Huh7, HepG2 and HLF). Thus overexpression of these genes could be an important manifestation of HCC phenotypes, and should shed light on the molecular basis of hepatocellular carcinogenesis [79]. From the proteomics perspective, the above results should provide useful clues when the HCC proteome is being analysed. Identification of these genes at the protein level (provided they are above the detection limit) would confirm their enhanced expression in HCC. This is important since the level of mRNA expression does not always represent the amount of active protein in the cell [80].
3.2.4 Hepatic stellate cells Hepatic stellate cells (HSC) (previously called Ito cell, lipocyte, perisinusoidal cell or fat-storage cell), which comprise 15% of the total number of resident liver cells, are the major
producers of extracellular matrix proteins in the liver [81]. They play an important role in the development of liver fibrosis. For example, following liver injury of any aetiology (Section 2.2), HSCs undergo a response called “activation”, which is the transition of quiescent cells into proliferative, fibrogenic, and contractile myofibroblasts [81]. Since liver fibrosis is reversible but not cirrhosis (end-stage consequence of fibrosis), efforts to understand fibrosis have concentrated on identifying therapeutic targets that can slow its progression. A recent report described the proteome analysis of rat hepatic stellate cells [82]. The analysis was performed on cellular and secreted proteins of normal (quiescent), in vitro and in vivo activated HSCs. A total of 312 proteins (from the products of 156 different genes) were identified from the 2-DE gels, but only 43 of them exhibited differential expression when the cells were activated in vivo and/or in vitro. A further 27 of these proteins showed similar changes in vivo and in vitro, including calcyclin, calgizzarin, and galectin-1 (which are up-regulated) and liver carboxylase 10 and serine protease inhibitor 3 (down-regulated) [82]. Finally, 16 of them showed differences in expression levels between the in vivo and in vitro activated stellate cells. This protein data would undoubtedly serve as a useful resource in understanding the underlying basis of the activation of HSC, a key step in liver fibrogenesis.
Proteomics 2001, 1, 1249–1263
Proteomics of hepatocellular carcinoma
1261
Figure 6. 2-DE maps of matched pairs of tissues from HCC patients. (A) Tissues from a patient with well differentiated HCC. Panel A1 is the 2-DE map of the nontumorous region of the liver and Panel A2 is the 2-DE map of the tumour. (B) Tissues from a patient with poorly differentiated HCC. Panel B1 is the 2-DE map of the nontumorous region of the liver and Panel B2 is the 2-DE map of the tumour.
3.3 Protein database of hepatocellular carcinoma One of the major goals of proteomics is to establish a protein database (proteome) for the tissue, cell or model organism under consideration [83]. Unfortunately, to date, there is as yet no complete proteome of an organism or a
cell. This is most probably a result of the limitations of the present technology in which it is not yet possible to identify all the proteins in a given cell [83]. However, a protein database even though incomplete and limited in extent, is a useful resource for researchers working in the same field. In this spirit, Sanchez et al. [56] established the SWISS-2DPAGE database for several tissues, cell lines,
1262
T. K. Seow et al.
and body fluids. Presently the most exhaustive database is that of the yeast protein database [84] and the human 2D-PAGE of the Danish Centre for Human Genome Research [85]. In the case of HCC, several groups have attempted to map and identify either the expressed proteins in liver [55, 56], cell lines [56, 59] or the differentially expressed proteins/genes of several different HCC cell lines [55, 57, 72]. The identities of the differentially expressed proteins are summarized in Table 1. Amongst these differentially expressed proteins some are known to be implicated in HCC, while some are linked to some other type of cancers [72]. It is hoped the comparison of these proteins with those of the yet to be published differentially expressed proteins in liver tumour tissues (both well and poorly differentiated types) will facilitate the discovery and identification of novel proteins that may be involved in hepatocarcinogenesis.
4 Concluding remarks Cell lines derived from human tumours have been extensively used as in vitro models of various diseases. In the study of HCC thus far, almost all proteome analyses have used hepatoma cell lines to unravel the process of hepatocarcinogenesis. The reason for this is obvious as human tumour tissues are well known to be heterogeneous. However, recent improvements in sample preparation, such as the ability to pick specific types of cells from tumours using LCM, may lead to an increased use of solid tissue samples for proteome analyses. The alternative is to use matched tissues from surgically resected primary HCCs and adjacent nontumorous portions of the liver for solubilization and analysis. With the recent introduction of several new proteomics technologies such as isotope-coded affinity tags (ICAT) [86], protein chips [87, 88], and immuno-detection amplified by T7 RNA polymerase (IDAT) which has single-cell resolution [89], the screening for novel diagnostic markers for early detection and therapeutic targets for treatment of HCC will be expected to accelerate greatly in the foreseeable future. This will complement nicely the current work of gene expression studies of HCC that are being undertaken by many laboratories. We wish to acknowledge the skillful assistance and contribution of all the staff in the Proteomics Laboratory of the Bioprocessing Technology Centre in this HCC project. Received May 25, 2001
Proteomics 2001, 1, 1249–1263
5 References [1] Pennington, S. R., Wilkins, M. R., Hochstrasser, D. F., Dunn, M. J., Trends Cell Biol. 1997, 7, 168–173. [2] Page, M. J., Amess, B., Rohlff, C., Stubberfield, C., Parekh, R., Drug Discov. Today 1999, 4, 55–62. [3] Wang, J. H., Hewick, R. M., Drug Discov. Today 1999, 4, 129–133. [4] Hochstrasser, D. F., in: Wilkins, M. R., Williams, K. L., Appel, R. D., Hochstrasser, D. F. (Eds.), Proteome Research: New Frontiers in Functional Genomics, Springer, New York 1997, pp. 187–219. [5] Sager, R., Proc. Natl. Acad. Sci. USA 1997, 94, 952–955. [6] Alaiya, A. A., Franzen, B., Auer, G., Linder, S., Electrophoresis 2000, 21, 1210–1217. [7] Schafer, D. F., Sorrell, M. F., Lancet 1999, 353, 1253–1257. [8] Di Bisceglie, A. M., Simpson, L. H., Lotze, M. T., Hoofnagle, J. H., J. Clin. Gastroenterol. 1994, 19, 222–226. [9] Okuda, K., Fujimoto, I., Hanai A., Urano, Y., Cancer Res. 1987; 47, 4967–4972. [10] Taylor-Robinson, S. D., Foster, G. R., Arora, S., Hargreaves, S., Thomas, H. C., Lancet 1997, 350, 1142–1143. [11] Deuffic, S., Poynard, T., Buffat, L., Valleron, A. J., Lancet 1998, 351, 214–215. [12] El-Serag, H. B., Mason, A. C., New Engl. J. Med. 1999, 340, 745–750. [13] Chia, K. S., Seow, A., Lee, H. P., Shanmugaratnam, K., Singapore Cancer Registry Report No. 5, National University of Singapore, Singapore 2000. [14] Moradpour, D., Wands, J. R., in: Zakim, D., Boyer, T. D. (Eds.), Hepatology: A Textbook of Liver Diseases, 3rd Ed., 2, W. B. Saunders, Philadelphia 1996, pp. 1490–1512. [15] Beasley, R. P., Hwang, L. Y., Lin, C. C., Chien, C. S., Lancet 1981, 2, 1129–1133. [16] Benn, J., Schneider, R. J., Proc. Natl. Acad. Sci. USA 1995, 92, 11215–11219. [17] Sitterlin, D., Lee, T. H., Prigent, S., Tiollais, P., et al., J. Virol. 1997, 71, 6194–6199. [18] Wang, X. W., Forrester, K., Yeh, H., Feitelson, M. A., et al., Proc. Natl. Acad. Sci. USA 1994, 91, 2230–2234. [19] Qadri, I., Conaway, J. W., Conaway, R. C., Schaack, J., Siddiqui, A., Proc. Natl. Acad. Sci. USA 1996, 93, 10578–10583. [20] Boyer, N., Marcellin, P., J. Hepatol. 2000, 32 Suppl. 1, 98– 112. [21] Ray, R. B., Steele, R., Meyer, K., Ray, R., J. Biol. Chem. 1997, 272, 10983–10986. [22] Ray, R. B., Meyer, K., Ray, R., Virology 1996, 226, 176–182. [23] Ray, R. B., Lagging, L. M., Meyer, K., Ray, R., J. Virol. 1996, 70, 4438–4443. [24] Ozturk, M., Bressac, B., Puisieux, A., Kew, M., et al., Lancet 1991, 338, 1356–1359. [25] Sun, Z., Lu, P., Gail, M. H., Pee, D., et al., Hepatology 1999, 30, 379–383. [26] Figueras, J., Jaurrieta, E., Valls, C., Benasco, C., et al., Hepatology 1997, 25, 1485– 489. [27] Du, M. Q., Hutchinson, W. L., Johnson, P. J., Williams, R., Cancer 1991, 67, 476–480. [28] Ho, S., Cheng, P., Yuen, J., Chan, A., et al., Brit. J. Cancer 1996, 73, 985–988. [29] Johnson, P. J., Leung N., Cheng, P., Welby, C., et al., Br. J. Cancer 1997, 75, 236–240. [30] Johnson, P. J., Poon, T. C. W., Hjelm, N. M., Ho, C. S., et al., Br. J. Cancer 2000, 83, 1330–1337. [31] Choi, B. I., Lee, H. J., Han, J. K., Choi, D. S., et al., Am. J. Roentgenol. 1997, 168, 219–224.
Proteomics 2001, 1, 1249–1263 [32] Lau, J. W. Y., Leow, C. K., in: Leong, A. S. Y., Liew C. T, Lau, J. W. Y., Johnson, P. J. (Eds.), Hepatocellular Carcinoma. Contemporary Diagnosis, Investigation and Management, Arnold, London 1999, pp. 147–172. [33] Wall, W. J., Marotta, P. J., Liver Transpl. 2000, 6 Suppl 2, S16–S22. [34] Chang, M. H., Chen, C. J., Lai, M. S., Hsu, H. M., et al., New Engl. J. Med. 1997, 336, 1855–1859. [35] Mazzella, G., Accogli, E., Sottili, S., Festi, D., et al., J. Hepatol. 1996, 24, 141–147. [36] Oka, H., Yamamoto, S., Kuroki, T., Harihara, S., et al., Cancer 1995, 76, 743–749. [37] O’Farrell, P. H., J. Biol. Chem. 1975, 250, 4007–4021. [38] Young, D. S., Anderson, N. G., Clin. Chem. 1982 28, 737– 1092. [39] Anderson, L., Anderson, N. G., Clin. Chem. 1984 30, 1897– 2108. [40] Celis, J. E., Bravo, R. (Eds.), Two Dimensional Gel Electrophoresis of Proteins, Academic Press, New York 1984. [41] Bjellqvist, B., Ek, K., Righetti, P. G., Gianazza, E., et al., J. Biochem. Biophys. Methods 1982, 6, 317–339. [42] Wilkins, M. R., Williams, K. L., Appel, R. D., Hochstrasser, D. F. (Eds.), Proteome Research: New Frontiers in Functional Genomics, Springer, New York 1997. [43] Rabilloud, T. (Ed.), Proteome Research: Two-Dimensional Gel Electrophoresis and Identification Methods, Springer, New York 1999. [44] Hanash, S. M., Electrophoresis, 2000, 21, 1202–1209. [45] Wirth, P. J., Luo, L., Fujimoto, Y., Bisgaard, H. C., Electrophoresis 1992, 13, 305–320. [46] Zeindl-Eberhart, E., Rabes, H. M., Carcinogenesis 1992, 13, 1177–1183. [47] Kudofuku, T., Sato, T., J. Chromatogr. 1985, 343, 51–58. [48] Wirth, P. J., Rao, M. S., Evarts, R. P., Cancer Res. 1987, 47, 2839–2851. [49] Hanash, S. M., Strahler, J. R., Neel, J. V., Hailet, N., et al., Proc. Natl. Acad. Sci. USA 1991, 88, 5709–5713. [50] Honore, B., Leffers, H., Madsen, P., Celis, J. E., Eur. J. Biochem. 1993, 218, 421–430. [51] Shevchenko, A., Wilm, M., Vorm, O., Mann, M., Anal. Chem. 1996, 68, 850–858. [52] Henzel, W. J., Billeci, T. M., Stults, J. T., Wong, S. C., Proc. Natl. Acad. Sci USA 1993, 90, 5011–5015. [53] Mann, M., Hojrup, P., Roepstorff, P., Biol. Mass Spectrom. 1993, 22, 338–345. [54] Gevaert, K., Vandekerckhove, J., Electrophoresis 2000, 21, 1145–1154. [55] Wirth, P. J., Hoang, T. N., Benjamin, T., Electrophoresis 1995, 16, 1946–1960. [56] Sanchez, J.-C., Appel, R. D., Golaz, O., Pasquali, C., et al., Electrophoresis 1995, 16, 1131–1151. [57] Yu, L.-R., Zeng, R., Shao, X.-X., Wang, N., et al., Electrophoresis 2000, 21, 3508–3068. [58] Celis, J. E., Ostergaard, M., Basse, B., Celis, A., et al., Cancer Res. 1996, 56, 4782–4790. [59] Seow, T. K., Ong, S.-E., Liang, R. C. M. Y., Ren, E. C., et al., Electrophoresis 2000, 21, 1787–1813. [60] Watanabe, T., Morizane, T., Tsuchimoto, K., Inagaki, Y., et al., Int. J. Cancer 1983, 32, 141–146.
Proteomics of hepatocellular carcinoma
1263
[61] Wilm, M., Shevchenko, A., Houthaeve, T., Breit, S., et al., Nature 1996, 379, 466–469. [62] Mann, M., Wilm, M., Anal. Chem. 1994, 66, 4390–4399. [63] Ou, K., Seow, T. K., Liang, R. C. M. Y., Ong, S.-E., Chung, M. C. M., Electrophoresis 2001, 22, 2804–2811. [64] Shevchenko, A., Chernushevich, I., Ens, W., Standing, K. G., et al., Rapid Commun. Mass Spectrom. 1997, 11, 1015– 1024. [65] Lawrence, S. M., Huddleston, K. A., Pitts, L. R., Nguyen, N., et al., J. Biol. Chem. 2000, 275, 17869–17877. [66] International Human Genome Sequencing Consortium, Nature 2001, 409, 860–921. [67] Andersen, J. S., Mann, M., FEBS Lett. 2000, 480, 25–31. [68] Mann, M., Pandey, A., Trends Biochem. Sci. 2001, 26, 54–61. [69] Choong, M. L., Tan, L. K., Lo, S. L., Ren, E. C., et al., FEBS Lett. 2001, 496, 109–116. [70] Golub, T. R., Slonim, D. K., Tamayo, P., Huard, C., et al., Science 1999, 286, 531–537. [71] Ross, D. T., Scherf, U., Eisen, M. B., Perou, C. M., et al., Nat. Genet. 2000, 24, 227–235. [72] Kawai, H. F., Kaneko, S., Honda, M., Shirota, Y., Kobayashi, K., Hepatology 2001, 33, 676–691. [73] Page, M. J., Amess, B., Townsend, R. R., Parekh, R., et al., Proc. Natl. Acad. Sci. USA 1999, 96, 12589–12594. [74] Sarto, C., Marocchi, A., Sanchez, J. C., Giannone, D., et al., Electrophoresis 1997, 18, 599–604. [75] Bini, L., Magi, B., Marzocchi, B., Arcuri, F., et al., Electrophoresis 1997, 18, 2832–2841. [76] Banks, R. E., Dunn, M. J., Forbes, M. A., Stanley, A., et al., Electrophoresis 1999, 20, 689–700. [77] Kondoh, N., Wakatsuki, T., Ryo, A., Hada, A., et al., Cancer Res. 1999, 59, 4990–4996. [78] Shuda, M., Kondoh, N., Tanaka, K., Ryo, A., et al., Anticancer Res. 2000, 20, 2489–2494. [79] Tanaka, K., Kondoh, N., Shuda, M., Matsubara, O., et al., Biochim. Biophys. Acta 2001, 1536, 1–12. [80] Anderson, L., Seilhammer, J., Electrophoresis 1997, 18, 533–537. [81] Friedman, S. L., J. Biol. Chem. 2000, 275, 2247–2250. [82] Kristensen, D. B., Kawada, N., Imamura, K., Miyamoto, Y., et al., Hepatology 2000, 32, 268–277. [83] Haynes, P. A., Gygi, S. P., Figeys, D., Aebersold, R., Electrophoresis 1998, 19, 1862–1871. [84] Crawford, M. E., Cusick, M. E., Garrels, J. I., Proteomics : A Trends Guide 2000, July, 17–21. [85] Celis, J. E., Ostergaard, M., Jensen, N. A., Gromova, I., et al., FEBS Lett. 1998, 430, 64–72. [86] Gygi, S. P., Rist, B., Gerber, S. A., Turecek, F., et al., Nat. Biotechnol. 1999, 17, 994–999. [87] Merchant, M., Weinberger, S. R., Electrophoresis 2000, 21, 1164–1177. [88] Nelson, R. W., Nedelkov, D., Tubbs, K. A., Electrophoresis 2000, 21, 1155–1163. [89] Zhang, H.-T., Kacharmina, J. E., Miyashiro, K., Greene, M. I., Eberwine, J., Proc. Natl. Acad. Sci. USA 2001, 98, 5497– 5502.
