• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Proteomics. Author manuscript; available in PMC Aug 20, 2009.
Published in final edited form as:
PMCID: PMC2729784
NIHMSID: NIHMS45040

Identification of biomarkers associated with the development of hepatocellular carcinoma in CuZn superoxide dismutase deficient mice

Abstract

To identify biomarkers associated with the development of hepatocellular carcinoma (HCC) in CuZn superoxide dismutase (CuZnSOD, Sod1) deficient mice, 2-DE followed by MS analysis was carried out with liver samples obtained from 18-month-old Sod1−/− and +/+ mice. The intra-cellular Ca binding protein, regucalcin (RGN), showed a divergent alteration in Sod1−/− samples. Whereas elevated RGN levels were observed in −/− samples with no obvious neoplastic changes, marked reduction in RGN was observed in −/− samples with fully developed HCC. GST mu1 (GSTM1), on the other hand, showed a significant increase only in the neoplastic regions obtained from Sod1−/− livers. No change in GSTM1 was observed in the surrounding normal tissues. Marked reduction was observed in two intracellular lipid transporters, fatty acid binding protein 1 (FABP1) and major urinary protein 11 and 8 (MUP 11&8), in Sod1−/− samples. Analysis of additional samples at 18−22 months of age showed a three-fold increase in enolase activities in Sod1−/− livers. Consistent with previous findings, carbonic anhydrase 3 (CAIII) levels were significantly reduced in Sod1−/− samples, and immunohistochemical analysis revealed that the reduction was not homogenous throughout the lobular structure in the liver.

Keywords: CuZn superoxide dismutase, GST mu1, Hepatocellular carcinoma, Regucalcin, Senescence marker protein 30

1 Introduction

Hepatocellular carcinoma (HCC) is the major type of primary liver tumor and is one of the most common cancers worldwide [1, 2]. The disease has a very poor prognosis with greater than 50% of the patients dying within 1 year and only 2−6% surviving up to 5-year after diagnosis [3]. HCC does not respond well to surgery or chemotherapy and often requires a radical approach such as liver transplantation to prolong the survival of the patients. Therefore, the best approach in managing HCC is prevention. Consequently, identification of reliable serum and tissue markers for screening and monitoring of patients at high risk for developing HCC may lead to earlier intervention and better prognosis.

Hepatitis B virus (HBV) and hepatitis C virus (HCV) infections are the two major etiologies of HCC, and HBV and HCV infections have been shown to increase ROS production in patients [4]. Several clinical studies have also showed a causal relationship between HCC and low levels of super-oxide dismutase (SOD) [58]. In our investigation of the long-term effects of CuZn superoxide dismutase (CuZn-SOD) deficiency in mutant animals, we observed the development of HCC in B6-Sod1 (gene designation for CuZn-SOD)−/− mice by 21 months of age [9]. In addition, liver-specific deletion of the transcription factor Nrf1, which mediates activation of oxidative stress response genes, leads to the development of HCC in adult mutant mice [10]. Taken together, the observations suggest that increased oxidative stress is a common factor in HCC development, and that studies on the hepatocarcinogenesis in Sod1−/− mice may provide important insights into the oxidative stress-mediated disease mechanism. As a first step to determine proteomic changes occurring in the oxidative stress-associated HCC development and to identify informative protein markers corresponding to the progression of hepatocarcinogenesis, we carried out 2-D/MS analysis on Sod1−/− and age-matched +/+ controls followed by immunochemical detection and transcriptional analysis of several proteins identified in this study.

2 Materials and methods

2.1 Sample preparation

All mice were housed in microisolators in the animal facility atthe VA Palo Alto Health Care System. The mice were fed Harlan Teklad Global Rodent Diet 2019 (19% protein, 9% fat) from weaning till 2 months of age, and then switched permanently to a low fat diet (5% fat, Teklad Global Rodent Diet 2018). The mice were maintained on a 12 h light/dark cycle, at a constant temperature between 20 and 22°C, and given food and water ad libitum. Mice were checked daily by animal care personnel. All animal handling procedures were carried out according to the protocol approved by the Subcommittee on Animal Studies at the VA Palo Alto Health Care System. Liver samples were obtained from 3 to 6 and 18 to 22-month-old Sod1tm2Cje +/+ and −/− mice on B6 background as previously described [9]. Liver samples from three Sod1+/+ and four Sod1−/− male mice at 18 months of age were used for 2-D/MS analysis. Three of the four Sod1−/− liver samples used for 2-D/MS analysis had fully developed HCC at the time of tissue collection; therefore, tumor tissues (Sod1−/− tumor) and the surrounding normal tissues (Sod1−/− non-tumor) from these three mice were analyzed separately. All +/+ control samples had normal histo-pathology. All of the samples used for 2-D/MS analysis were included in the 18−22-month-old group in the subsequent confirmation analyses by Western blotting, enzyme assay, and real time RT-PCR. Whenever possible, the same set of samples was used for all of the confirmational assays to track changes between histopathology status, protein levels, and message levels. Livers were homogenized in three volumes (w/v = 1:3) of lysis buffer containing 50 mM Tris-HCl pH 8.0, 0.6 mM MnCl2, 2 mM citric acid, and the complete protease inhibitor cocktail (Roche, Indianapolis, IN, USA) and freeze–thawed three times between liquid nitrogen and room temperature water to disrupt the mitochondrial membrane. Tissue lysates were centrifuged at 10,000g at 4°C for 10 min and the supernatants were removed for the determination of protein concentration (BCA protein assay, Pierce, Rockford, IL) and aliquoted for storage at −80°C.

2.2 2-DE

2-DE was performed according to the procedure of Lanne et al. [11]. Briefly, proteins (200 µg) were precipitated from tissue lysates by 100% ice-cold acetone and dissolved in 150 µL of IEF buffer (8 M urea, 20 mM DTT, 2.0% w/v Chaps, 0.2% Biolytes, 2 M thiourea, and bromophenol blue). Protein concentration was determined again by Bradford method before loading. For the first-dimension electrophoresis, 150 µL (at 1 µg/µL) of sample solution was applied to an 11 cm Ready-Strip IPG strip, pH 3−10 (BioRad, Hercules, CA). Each sample was also separated with narrow range IPG strips (pH 4−7 and 7−10) for better sensitivity and resolution. The strips were soaked in the sample solution for 1 h to allow uptake of the proteins, then actively rehydrated in the Protean IEF cell (BioRad) for 12 h at 50 V. IEF was performed at the following voltage for 1 h each: 100, 200, 500, 1000; then accelerated to 8000 V in 30 min using a linear gradient and continued at 8000 V for 10 h. All the processes above were carried out at 22°C. The focused IEF strips were stored at −80°C until second-dimension electrophoresis was performed.

For second-dimension electrophoresis, thawed IPG strips were equilibrated for 15 min in 50 mM Tris-HCl pH 8.8, containing 6 M urea, 1% w/v SDS, 30% v/v glycerol, and 65 mM DTT, and re-equilibrated for 15 min in the same buffer containing 260 mM iodacetamide in place of DTT. Precast Criterion Tris-HCl gels (4−20% linear gradient, BioRad) were used to perform second-dimension electrophoresis. Protein standards were run along side the sample at 200 V for 60 min. After electrophoresis, the gels were washed in distilled water, stained in CBB for 1 h (Gelcode blue, Pierce), and destained overnight.

2.3 2-D gel analysis

The 2-D gel images were captured with a CCD camera (1280 × 960 pixels) using identical aperture setting and exposure time, and the spot intensities were analyzed using the PD Quest 2-D analysis software (version 8.0.1, BioRad). The intensity of each protein spot was assigned a value in PPM after normalization to total intensity of spots in the gel to eliminate differences in protein loading. The Sod1+/+, Sod1−/− tumor, and Sod1−/− non-tumor (the surrounding normal tissues) were analyzed as three different groups. Automatic spot matching function was used to match all the spots across all samples first; visual inspection was then carried out to make sure that spots were matched correctly. The intensity of certain unmatched spots, which may be present in one group of samples, but not in the others, was obtained manually. The intensity value of missing spots was left blank, and the statistical analysis for those spots was not carried out. Instead, the spots were noted as absence in the dataset. Significant changes in spot intensity, as a function of protein expression, were calculated using Student's t-test between different groups.

2.4 MS analysis

Protein spots were cut out with a clean large-orifice pipette tip. To be certain of the protein identities, the same spots were usually picked from multiple gels from both +/+ and −/− samples whenever possible. Based on the molecular weight, pI, and the complete absence of the protein spot in all of the Sod1−/− gels (Fig. 1), the putative CuZnSOD spot was picked from all three Sod1+/+ pH 3−10 gels and from one Sod1+/+ pH 4−7 gel. Based on 2-D/Western blot results, the spot corresponding to actin was picked from one Sod1+/+ and one Sod1−/− gel. Gel spots were treated with 50% ACN in 10 mM ammonium bicarbonate (ABC) to remove the stain from the proteins, and with 100% ACN to dry the gel pieces. The dried gel pieces were rehydrated with 10% ACN in 10 mM ABC containing 0.1 mg trypsin per sample, incubated overnight at 37°C, and the resulting peptides were recovered with 60% ACN with 0.1% trifluoroacetate.

Figure 1
Comparison of 2-D gel patterns between Sod1+/+ and −/− liver samples. Representative gels (pH 3−10) generated from 18-month-old +/+ (top panel) and −/− (bottom panel) mice are shown. Protein spots corresponding ...

The tryptic peptides were resolved by RP chromatography followed by MS/MS analysis as described in ref. [12]. The MS data were compared to simulated proteolysis and CID of the proteins in the NR-NCBI database using the Pep-Miner software (IBM, Haifa Research Lab) [12]. All MS analyses and protein digestions were carried out at the Smoler Proteomics Center, Technion, Haifa, Israel.

2.5 Immunodetections

1-D Western blot analyses of regucalcin (RGN), fatty acid binding protein 1 (FABP1), and actin were performed as previously described [9]. The primary antibodies against RGN (Clone Regucalcin M, Cell Sciences, Canton, MA), FABP1 (AF 1565, R&D Systems, Minneapolis, MN), and actin (A 5441, Sigma, St. Louis, MO) were used at 1:250, 1:1000, and 1:500 dilution, respectively. HRP-conjugated secondary antibodies were used at 1:4000 dilution. Actin was used as a loading control and all signal intensities were normalized to that of actin. To positively identify the protein spot for actin in 2-D gels, one pH 3−7 gel each from a +/+ and a −/− sample was transferred to NC filter and subjected to Western blot analysis after second-dimension electrophoresis. The Western blot analysis was carried out as previously described [9].

Immunohistochemical (IHC) analyses of carbonic anhydrase III (CAIII) were carried out on paraffin embedded liver sections from Sod1+/+ and −/− mice as described previously [9]. Primary antibody (PA-4010, Spectral Diagnostics, Toronto, Canada) and secondary antibody (SC2768, Santa Cruz Biotechnology) were used at 1:200 and 1:1000, respectively.

2.6 Enolase assay

Enolase activities were determined by monitoring the formation of 2-phosphoglyceric acid (2-PGA) from phosphoenol pyruvate (PEP) at 240 nm according to Pal-Bhowmick et al. [13]. The enzyme activity was calculated using ε240 nm = 1400/M/cm.

2.7 Quantification of mRNA by real time PCR

Total RNA from approximately 20 mg liver tissue was extracted and purified using TRIzol® reagent (Invitrogen, Carlsbad, CA) and RNeasy column (Qiagen, Valencia, CA). RNAs were quantified with a spectrophotometer (Beckman Coulter, Fullerton, CA). After DNase I digestion (amplification grade, 18068−0158, Invitrogen), 2 µg of total RNA was reverse-transcribed by Superscript II reverse transcriptase (Invitrogen) according to the manufacturer's protocol using random oligo-d(T) as primers. Real time PCR analysis was carried out on an ABI Prism 7900HT system (Applied Bio-systems, Foster City, CA). PCR conditions were 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Each reaction contained cDNA from 10 ng of RNA, 800 nM of each primer, and SYBR® Green QPCR Master Mix (Stratagene, La Jolla, CA) in a total volume of 6 µL. Primers were designed according to the guidelines in Primer Express (Applied Biosystems) software and are listed in Supporting Information Table 1. All samples were analyzed in triplicates and were quantified using a standard curve. Standards were generated by a serial dilution of cDNAs pooled from Sod1+/+ and −/− samples. The level of each mRNA was normalized to that of ubiquitin and expressed as fold change from that of age-matched controls.

2.8 Statistical analysis

Normalized signal intensities from Western blots were analyzed by Student's t-test using Prism 4 (GraphPad, San Diego, CA) and expressed as percentage of age-matched controls. Enolase activities and quantitative RT-PCR data were also analyzed by Student's t-test. Since a subset of Sod1−/− mice in the 18−22 months group have already developed HCC at the time of tissue collection, all data analyses for this group were carried out with tumor and non-tumor samples as two separate groups first. If there was a distinctive difference between these two, they were compared separately to +/+ controls (see notations in table and figure legends). On the other hand, if there was no discernable difference between the tumor and non-tumor group, the data were combined and compared to that of +/+ controls. Likewise, within the tumor-bearing group, results from tumor tissues and the surrounding normal tissues were analyzed separately to know if there was a difference between them.

3 Results

3.1 2D/MS comparison

Sod1−/− mice on B6 background develop HCC by 21 months of age [9], and a high percentage of mice already have fully developed HCC by 18 months. Therefore, we chose 18 months as the age for our proteomic analysis. A total of 50 spots were picked from 2-D gels for MS analysis, and a total of 70 proteins were identified by a minimum of 3 peptides each from the 50 gel spots. MS analysis confirmed the identity of CuZnSOD and actin in the corresponding gel spots (data not shown). Proteins that showed a consistent difference between +/+ and −/− samples are listed in Table 1 and the normalized signal intensities presented in Fig. 2. The peptide sequences identified by MS analysis are also listed in Supporting Information Table 2. The magnitude of difference between +/+ and −/− was the largest in major urinary protein 11 and 8 (MUP 11&8), followed by CAIII, RGN, and FABP1 (Figs. 2a and b). There was no significant difference between tumor and the surrounding normal tissue in −/− samples within this group of proteins. On the other hand, a significant difference in GST mu1 (GSTM1) levels was observed between the tumor and the surrounding normal tissues in Sod1−/− livers that developed HCC. Thus, a 50% increase in GSTM1 was only observed in the neoplastic regions, and no significant difference in GSTM1 was detected between +/+ controls and −/− samples with normal histopathology (Fig. 2c).

Figure 2
Abundance of proteins that showed a consistent difference between Sod1+/+ and −/− samples. (a) MUP11&8, CAIII, and FABP1. Significant reduction of these three proteins was observed in all Sod1−/− samples regardless ...
Table 1
Changes in 2-D gel profiles in Sod1−/− livers

3.2 Western blotting and immunohistochemical confirmation of 2-D gel results

To confirm the 2-D gel results and to know age-related changes in a subset of proteins, Western blot analysis was carried out with additional liver samples collected from 3 to 6 and 18−22-month-old mice. The two ages represent Sod1−/− mice that are relatively healthy with no overt abnormalities (3−6 months) and Sod1−/− mice that are near end-stage (18−22 months) with roughly 30% of them with well-established HCC [9]. Consistent with the 2-D gel results, FABP1 showed a 70 and 55% reduction in the −/− livers at 3− 6 months and 18−22 months, respectively (Fig. 3a), regardless of the histopathology status. The data suggest that reduction in FABP1 may be directly related to CuZnSOD deficiency, and not as a consequence of hepatocarcinogenesis. RGN also showed a 56% reduction in the −/− livers at 3−6 months, but the difference was not statistically significant (Fig. 3b). In 2-D gel analysis, we observed a large deviation of RGN level between the Sod1−/− sample collected from a non-tumor-bearing mouse (normal level of RGN) and that in the Sod1−/− samples collected from tumor-bearing mice (low levels of RGN) in the Sod1−/− non-tumor group (Fig. 2b). The difference may be due to the histopathological state of the mice from which the samples were collected. To resolve this discrepancy, we collected more samples from non-tumor-bearing Sod1−/− mice and carried out Western blot to see if the relatively normal level of RGN observed in the non-tumor-bearing mouse can be reproduced in a larger set of samples. Consistent with the 2-D gel results, Sod1−/− samples with normal histopathology (18−22N) showed an average of 70% increase in RGN levels (Fig. 3b). On the other hand, the 18−22-month-old samples with obvious HCC development (18−22T) showed a 50% reduction by Western blot analysis. The divergence in RGN levels between tumor-bearing and non-tumor-bearing Sod1−/− samples is noteworthy and the difference is likely related to the development of hepatocarcinogenesis. Among the −/− liver samples with reduced RGN levels, there was no discernable difference in RGN between the tumor and the surrounding normal tissue (data not shown).

Figure 3
Quantitative Western blot analysis of FABP1 and RGN. (a) Significant reduction in FABP1 levels was observed in the liver of 3−6 and 18−22-month-old Sod1−/− mice regardless of histo-pathology status. (b) Significant increase ...

We showed the reduction of CAIII in Sod1−/− livers previously by Western blot analysis [9], and the results were confirmed by 2-D/MS analysis in this study (Fig. 2a). To further identify the reduction of CAIII at the cellular level, IHC analysis was carried out. The results showed that reduction of CAIII was not homogenous across all regions (Fig. 4). Although the majority of −/− cells had very low level, a small number of cells actually had higher than normal levels of CAIII. Nuclear localization of CAIII was also observed in a subset of cells in Sod1−/− sections.

Figure 4
Immunohistochemical analysis of CAIII. Representative liver sections from 3- and 18-month-old Sod1+/+ and −/− mice are shown. A mosaic staining pattern is commonly observed in Sod1−/− samples, suggesting a heterogeneous ...

3.3 Enolase activities are increased in 18−22-month-old Sod1−/− livers

Increased dependence on glycolysis is common in tumor tissues [14]. To determine if enzymes in the glycolysis pathway were upregulated in Sod1−/−, liver samples from 3−6 and 18−22-month-old mice were used for a direct measurement of enolase activities. No difference in enolase activities was observed between Sod1+/+ and −/− at 3−6 months of age. However, by the time the mice reached 18−22-month-old, there was a three-fold difference between +/+ and −/− samples (Fig. 5) regardless of liver pathology. Among the +/+ samples, there was a 50% reduction from 3 to 6 months to 18−22 months. In contrast, a two-fold increase from 3−6 months to 18−22 months was observed in −/− samples. The data suggest a possible shift to glycolysis as Sod1−/− mice age.

Figure 5
Enolase activities. A three-fold increase in enolase activities was observed in 18−22-month-old Sod1−/− samples. Sample size: 3−6 months, n = 4 each; 18−22 months, n = 6 for +/+ and n = 9 for −/− ...

3.4 qRT-PCR analysis suggests transcriptional and post-transcriptional regulation of protein levels

To determine if the observed changes in 2-D analysis occurred at the transcriptional level, quantitative RT-PCR analyses were carried out on the proteins listed in Table 1. Consistent with our previous observation, message levels for CAIII (Car3) were already significantly reduced at 3−6 months (data not shown), and the reduction became more pronounced at 18−22 months in Sod1−/− samples (Fig. 6). All other proteins did not show a difference at the transcriptional level at 3−6 months of age (data not shown). However, by 18−22 months of age, Sod1−/− mice were showing a consistent decrease in the message levels in fatty acid binding protein 1 (Fabp1), regucalcin (Rgn), and major urinary protein 11&8 (Mup11&8) in the liver. On the other hand, no change in enolase (Eno1) and glutathione S-transferase mu1 (Gstm1) message levels was detected. There was no clear distinction in the message level between tumor-bearing and normal Sod1−/− samples. The concordant change between protein and message in CAIII, FABP1, and MUP 11&8 suggests a regulation at the transcriptional level. On the other hand, the lack of concordant change between protein and message in RGN and enolase suggests a posttranscriptional regulation of these two proteins in CuZnSOD deficient environment.

Figure 6
Quantitative RT-PCR analysis. Message levels in 18−22-month-old Sod1+/+ and −/− liver samples were quantified and normalized to that of ubiquitin in each sample. Sample size: n =6 for +/+ and n = 8 for −/− samples. ...

4 Discussion

To determine the effects of CuZnSOD deficiency on the proteomic changes in Sod1−/− livers and to identify informative protein markers that correlate with hepatocarcinogenesis, we carried out 2-D/MS analysis on Sod1 mutant mice at 18 months of age, a stage when a subpopulation of Sod1−/− mice already have well-established hepatocarcinogenesis. Three of the four Sod1−/− mice used for 2-D/MS analysis have developed HCC at the time of tissue collection. This enabled us to compare between tumor and its surrounding normal tissue from the same mutant mouse.

Both RGN and GSTM1 showed important correlations with hepatocarcinogenesis. RGN levels were markedly reduced in −/− liver samples with well-established HCC, but elevated in −/− liver samples with no overt histopatho-logical abnormalities (Table 1, Figs. 2b and and3b).3b). The reduction in RGN was not limited to tumors; normal adjacent tissues also showed a similar level of reduction. Notably, reduction of RGN in clinical HCC samples has been demonstrated in two separate proteomic studies [15, 16], and substantial reduction in RGN expression has been shown in GST plascental form (GST-P) positive preneoplastic foci in a rat model of diethyl nitrosamine (DEN)/2-acetylaminofluorene (2-AAF)/partial hepatectomy (PH)-induced HCC [17]. Therefore, RGN level may be a useful biomarker for following the development of HCC in human population predisposed to the development of HCC and in mouse models of hepatocarcinogenesis. How does a change in RGN level relate to the pathogenesis of HCC? RGN is a Ca binding protein and is important for intracellular Ca homeostasis and in signaling pathways controlling glucose utilization and lipid production [1820]. RGN is sensitive to oxidative modification and a more than two-fold increase in the carbonyl content has been reported in old mice [21]. RGN is also known as senescence marker protein 30 (SMP30) [22]. Over-expression of RGN in the rat hepatoma cell line H4-II-E was shown to suppress the expression of oncogenes c-myc, Haras, and c-src, and increase the expression of tumor suppressor genes p53 and Rb [23, 24]. Recent studies in hepatoma cell lines also showed growth suppression effects by RGN [25, 26]. Therefore, upregulation of RGN may provide the benefit of suppressing hepatocarcinogenesis in Sod1−/− mice, whereas reduced expression of RGN may have the opposite effect. A recent report showed that SMP30 functions as a gluconolactonase, which is involved in the de novo synthesis of ascorbic acid [27, 28]. Consequently, SMP30 knockout mice are severely vitamin C deficient, containing only 6−8% of the tissue and plasma vitamin C levels found in wild-type controls even though vitamin C is supplemented (55 mg/kg) in the mouse chow [27]. Furthermore, when SMP30 knockout mice are fed a vitamin C-deficient diet, they develop symptoms of scurvy [27]. It is possible that upregulation of RGN in Sod1−/− livers helps to increase the tissue level of vitamin C and thus, helps to alleviate the oxidative stress imposed by CuZnSOD deficiency. Upregulation of RGN in Sod1−/− livers may also lead to a better-regulated intracellular calcium homeostasis, reduced cell death, and suppression of oncogene expressions. These modulations then help to prevent the development of hepatocarcinogenesis. On the other hand, reduced RGN in Sod1−/− livers would likely lead to vitamin C deficiency and further exacerbate the extent of oxidative stress. Reduced levels of RGN may also sensitize hepatocytes to apoptosis, enhance tumor-related gene expression, and ultimately, facilitate the development of hepatocarcinogenesis.

GSTM1 protein levels were only increased in the tumors of tumor-bearing Sod1−/− livers (Table 1, Figs. 1 and and2c),2c), and no change in GSTM1 was detected in the adjacent normal tissues. This is reminiscent of the expression pattern of another GST isoform, GST-P, in preneoplastic foci in rat chemical hepatocarcinogenesis [29]. GSTM1 belongs to a large group of detoxifying proteins that inactivate reactive intermediates by conjugating glutathione (GSH) to the reactive intermediates [30]. Both constitutive and inducible expression of GSTM1 has been shown to be controlled by the transcription factor Nrf2 [31]. Interestingly, GSTM1 was also shown to be transcriptionally upregulated by the Myb onco-protein [32]. In addition, GSTM1 has been shown to inhibit MEK kinase 1 (MEKK1) activity and suppress MEKK1-medaited apoptosis [33]. Therefore, increased GSTM1 levels in tumor tissues may be a consequence of oncogenic induction, and increased GSTM1 may play a role in tumor survival.

Other proteins identified in this study did not show an association with hepatocarcinogenesis in Sod1−/−, and the changes were most likely a consequence of CuZnSOD deficiency. However, alteration in these proteins may eventually contribute to different aspects of hepatocarcinogenesis. These proteins include enolase, FABP1, MUP 11&8, and CAIII. Alterations of these proteins tend to occur earlier and in larger magnitude. Enolase activities were increased by three-fold in 18−22-month-old Sod1−/− mice (Fig. 5). Similarly, a recent proteomic study with HCV-associated HCC samples showed a positive correlation between a-enolase levels and the severity of the tumors [34]. Upregulation of enzymes in the glycolysis pathway is common in various types of tumors [14], and may reflect the metabolic shift in tumor cells in response to the hypoxic growth environment [35]. In addition to being part of the glycolysis pathway, enolase, and a number of other glycolytic enzymes have been localized in the nucleus [36] and shown to play a role in transcriptional regulation. Several studies have identified the Myc promoter-binding protein-1 (MBP-1), a transcriptional suppresser of Myc, as an alternative translation product of Eno1 [36]. Since Eno1 is upregulated by Myc overexpression [37], enolase-mediated transcriptional suppression of Myc likely serves as a negative feedback on Myc-induced glycolysis.

Consistent with our previous finding [9], total protein levels of CAIII was reduced in Sod1−/− livers due to transcriptional downregulation of Car3 (Table 1, Figs. 2a and and6).6). However, IHC analysis revealed a mosaic expression pattern, with some cells expressing higher than control level and most cells totally devoid of CAIII in Sod1−/− lives (Fig. 4). The significance of CAIII reduction on hepatocarcinogenesis in Sod1−/− mice is not clear. CAIII null mutant mice have been generated, and no obvious phenotype has been attributed to CAIII deficiency [38]. However, expression of Car3 in cells lacking the protein protects the cells from hydrogen peroxide-induced apoptosis [39]. Therefore, CAIII deficiency in Sod1−/− mice may further reduce cellular resistance to oxidative stress and amplify the deleterious effects of CuZn-SOD deficiency.

Reduced FABP1 levels were observed at as early as 3 months of age in Sod1−/− liver samples (Table 1, Figs. 2a and and3a).3a). FABP1, also called L-FABP for its dominance in the liver, is a cytosolic protein and acts as intracellular transporter of fatty acids between membranes [40]. Deletion of FABP1 in mutant mice causes problem in branched-chain fatty acid metabolism [41]. A recent study showed the antioxidant function of FABP1, and increased expression of FABP1 in a liver cell line reduced cellular damage from hypoxia/reoxygenation [42]. Therefore, similar to CAIII deficiency, reduced FABP1 may contribute to a further increase in oxidative damage in Sod1−/− mice. Marked reduction of MUP 11&8 was consistently observed in Sod1−/− samples (Table 1, Figs. 1 and and2a),2a), and the reduction was controlled at the transcriptional level (Fig. 6). MUP 11&8 belong to the lipocalin family and, similar to FABP1, are responsible for transporting small hydrophobic molecules such as lipids and volatile pheromones [43]. The concomitant reduction of FABP1 and MUP 11&8 suggests a dramatic change in intra-cellular lipid transport in Sod1−/− mice.

In summary, alterations in the levels or activities of several proteins were identified in the liver of Sod1−/− mice by 2-D/MS analysis. RGN and GSTM1 showed a distinctive correlation with hepatocarcinogenesis. Reduction of RGN was associated with the development of HCC, whereas increased RGN levels were associated with normal histo-pathology in Sod1−/− livers. On the other hand, increased levels of GSTM1 were only observed in the neoplastic region of tumor-bearing Sod1−/− livers. The difference in RGN levels is particularly important as it showed a clear divergence between tumor-bearing and non-tumor-bearing Sod1−/− samples. Further studies will be required to determine the mechanism controlling RGN levels in the CuZn-SOD deficient environment and how changes in RGN levels affect the development of HCC.

Supplementary Material

1

Supplementary Table 1 Primer sequences for quantitative RT-PCR

Supplementary Table 2 Peptide sequences identified in the MS analysis

Acknowledgments

This work was supported by funding from the National Institutes of Health (AG24400), the Stanford Cancer Council, and the Digestive Disease Center at Stanford University. Mass spectrometry analysis was carried out at the Technion Smoler Proteomics Center, Israel.

Abbreviations

CAIII
carbonic anhydrase 3
CuZnSOD
CuZn superoxide dismutase
FABP1
fatty acid binding protein 1
GSTM1
GST mu1
HCC
hepatocellular carcinoma
IHC
immunohistochemistry
MUP 11&8
major urinary protein 11 and 8
RGN
regucalcin
SMP30
senescence marker protein 30

5 References

1. Murray CJ, Lopez AD. Lancet. 1997;349:1436–1442. [PubMed]
2. Levy L, Renard CA, Wei Y, Buendia MA. Ann. N. Y. Acad. Sci. 2002;963:21–36. [PubMed]
3. Hoofnagle JH. Gastroenterology. 2004;127:S319–S323. [PubMed]
4. Valgimigli M, Valgimigli L, Trere D, Gaiani S, et al. Free Radic. Res. 2002;36:939–948. [PubMed]
5. Casaril M, Corso F, Bassi A, Capra F, et al. Int. J. Clin. Lab. Res. 1994;24:94–97. [PubMed]
6. Huang C, Wu M. Zhonghua Yi Xue Za Zhi. 1990;70:138–139. [PubMed]
7. Inagaki T, Katoh K, Takiya S, Ikuta K, et al. Gastroenterol. Jpn. 1992;27:382–389. [PubMed]
8. Liaw KY, Lee PH, Wu FC, Tsai JS, Lin-Shiau SY. Am. J. Gastroenterol. 1997;92:2260–2263. [PubMed]
9. Elchuri S, Oberley TD, Qi W, Eisenstein RS, et al. Oncogene. 2005;24:367–380. [PubMed]
10. Xu Z, Chen L, Leung L, Yen T, et al. Proc. Natl. Acad. Sci. USA. 2005;102:4120–4125. [PMC free article] [PubMed]
11. Lanne B, Potthast F, Hoglund A, Brockenhuus von Lowenhielm H, et al. Proteomics. 2001;1:819–828. [PubMed]
12. Beer I, Barnea E, Ziv T, Admon A. Proteomics. 2004;4:950–960. [PubMed]
13. Pal-Bhowmick I, Sadagopan K, Vora HK, Sehgal A, et al. Eur. J. Biochem. 2004;271:4845–4854. [PubMed]
14. Altenberg B, Greulich KO. Genomics. 2004;84:1014–1020. [PubMed]
15. Kim W, Oe Lim S, Kim JS, Ryu YH, et al. Clin. Cancer Res. 2003;9:5493–5500. [PubMed]
16. Kim J, Kim SH, Lee SU, Ha GH, et al. Electrophoresis. 2002;23:4142–4156. [PubMed]
17. Suzuki S, Asamoto M, Tsujimura K, Shirai T. Carcinogenesis. 2004;25:439–443. [PubMed]
18. Yamaguchi M. Int. J. Mol. Med. 2005;15:371–389. [PubMed]
19. Fujita T, Shirasawa T, Inoue H, Kitamura T, Maruyama N. J. Gastroenterol. Hepatol. 1998;13:S124–S131. [PubMed]
20. Nakashima C, Yamaguchi M,J. ell. Biochem. 2006;99:1582–1592. [PubMed]
21. Chaudhuri AR, de Waal EM, Pierce A, Van Remmen H, et al. Mech. Ageing Dev. 2006;127:849–861. [PubMed]
22. Fujita T. Biochem. Biophys. Res. Commun. 1999;254:1–4. [PubMed]
23. Tsurusaki Y, Yamaguchi M. J. Cell. Biochem. 2003;90:619–626. [PubMed]
24. Tsurusaki Y, Yamaguchi M. Int. J. Mol. Med. 2004;14:277–281. [PubMed]
25. Yamaguchi M, Daimon Y. J. Cell. Biochem. 2005;95:1169–1177. [PubMed]
26. Yamaguchi M, Kobayashi M, Uchiyama S. J. Cell. Biochem. 2005;96:543–554. [PubMed]
27. Kondo Y, Inai Y, Sato Y, Handa S, et al. Proc. Natl. Acad. Sci. USA. 2006;103:5723–5728. [PMC free article] [PubMed]
28. Linster CL, Van Schaftingen E. FEBS J. 2007;274:1–22. [PubMed]
29. Sato K, Kitahara A, Satoh K, Ishikawa T, et al. Gann. 1984;75:199–202. [PubMed]
30. Hayes JD, Flanagan JU, Jowsey IR. Annu. Rev. Pharmacol. Toxicol. 2005;45:51–88. [PubMed]
31. Chanas SA, Jiang Q, McMahon M, McWalter GK, et al. Biochem. J. 2002;365:405–416. [PMC free article] [PubMed]
32. Bartley PA, Keough RA, Lutwyche JK, Gonda TJ. Oncogene. 2003;22:7570–7575. [PubMed]
33. Ryoo K, Huh SH, Lee YH, Yoon KW, et al. J. Biol. Chem. 2004;279:43589–43594. [PubMed]
34. Takashima M, Kuramitsu Y, Yokoyama Y, Iizuka N, et al. Proteomics. 2005;5:1686–1692. [PubMed]
35. Leo C, Giaccia AJ, Denko NC. Semin. Radiat. Oncol. 2004;14:207–214. [PubMed]
36. Kim JW, Dang CV. Trends Biochem. Sci. 2005;30:142–150. [PubMed]
37. Kim JW, Zeller KI, Wang Y, Jegga AG, et al. Mol. Cell. Biol. 2004;24:5923–5936. [PMC free article] [PubMed]
38. Kim G, Lee TH, Wetzel P, Geers C, et al. Mol. Cell. Biol. 2004;24:9942–9947. [PMC free article] [PubMed]
39. Raisanen SR, Lehenkari P, Tasanen M, Rahkila P, et al. FASEB J. 1999;13:513–522. [PubMed]
40. Martin GG, Danneberg H, Kumar LS, Atshaves BP, et al. J. Biol. Chem. 2003;278:21429–21438. [PubMed]
41. Atshaves BP, McIntosh AL, Payne HR, Mackie J, et al. Am. J. Physiol. Cell Physiol. 2005;288:C543–C558. [PubMed]
42. Wang G, Gong Y, Anderson J, Sun D, et al. Hepatology. 2005;42:871–879. [PubMed]
43. Beynon RJ, Hurst JL. Biochem. Soc. Trans. 2003;31:142–146. [PubMed]
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...