Abstract

Despite some controversy, selenomethionine (SeMet)-mediated protection against colorectal cancer (CRC) might be a very promising non-cytotoxic option. However, responsive molecular targets and underlying mechanisms of SeMet-mediated chemoprevention are still unclear. Our aim was to discover new targets of SeMet-mediated chemoprevention in CRC using proteomics analysis. We found dietary SeMet supplementation before carcinoma initiation effectively suppressed polyp incidence and dysplastic lesions without any adverse effects. To determine chemopreventive targets of SeMet, we employed two-dimensional gel electrophoresis–based proteomics analysis in CRC mouse model. Pretreatment with SeMet apparently modulated the expression of 30 proteins with functions in major processes like chronic inflammation, oxidative stress and apoptosis as discovered through pathway analysis with Pathway Studio software. We validated four proteins selected from pathway analysis including prohibitin, purine nucleoside phosphorylase, annexin 2 and c-reactive protein by immunohistochemistry. 8-Hydroxy-2ʹ-deoxyguanosine (8-OHdG), a known oxidative stress marker, was decreased by SeMet treatment in CRC mice as seen by immunohistochemistry. Further network analysis was done among these new four validated proteins, 8-OHdG and colorectal cancer. These four proteins found by proteomics analysis might be considered as potential chemopreventive biomarkers of SeMet against colon cancer and can help develop and improve approaches in preventive, therapeutic and prognostic aspects.

Introduction

Being the third most commonly diagnosed cancer in males and the second in females, colorectal cancer (CRC) is among the most common worldwide cancer with over 1.2 million new cancer cases and 608 700 deaths in 2008. It accounts for 8% of all cancer deaths, making it the fourth most common cause of death from cancer. Australia, New Zealand, Europe and North America show the highest incidence rates; on the other hand, the lowest rates are found in Africa and South-Central Asia ( 1 ). Spain and a number of countries within Eastern Asia and Eastern Europe, which were historically noted to have low incidence of this cancer, are now showing a sharp rise ( 2 ). CRC requires a multistep progression and develops as a consequence of pathological transformation of normal colonic epithelium to an adenomatous polyp and finally to an invasive cancer. Genetics and environment are the two major contributing factors to the occurrence of this particular cancer. As observed for incidence, mortality rates are lower in women than in men, except in the Caribbean ( 3 ). Major risk factors including smoking, physical inactivity, obesity, red and processed meat consumption and excessive alcohol consumption are considered modifiable ( 4 ).

Apart from minimizing the risk factors, chemical substances are also used to reduce the initiation of carcinogenic processes or to reverse these processes. Regular intake of selenium as a dietary supplementation inhibits tumorigenesis and reduces cancer risk ( 5 ). Se-methylselenocysteine, methaneselenenic or methaneseleninic acid are the methylated forms of selenium able to show protective effects against tumor progression ( 6 ). Selenomethionine (SeMet) and selenocysteine are representatives of this group, which are obtained from methionine and cysteine, respectively. Inorganic selenium shows toxic effects into cells, whereas organic selenium, SeMet, shows non-toxic effects ( 7 ). It is known that selenium and selenium-containing compounds act like an antioxidant for exhibiting chemopreventive effects.

Contemporary researches on preclinical, epidemiological studies and clinical trials have proven selenium as a potential candidate for colon cancer chemoprevention ( 8 ). Methylselenol and related metabolites may play an important role in chemoprevention as their targets are both endothelial and colon cancer cells ( 9 ). Intervention studies have also revealed a nearly 50% decrease in CRC risk among patients receiving selenium supplementation ( 10 ). Using selenium in chemoprevention is advantageous due to its added anticancer effect that might be regarded as useful in combination with other anticancer agents for colon cancer treatment ( 11 ). It has been reported that using selenium as additives in food may be able to lessen colon carcinogenesis and lower the adenomatous polyps occurrence in colon ( 10 ). Taylor’s et al. (2009) previously revealed the beneficial chemopreventive effects of selenium supplementation on gastrointestinal cancer among people with nutritional deficiency in Linxian, China ( 12 ). In contrast, Scott et al . (2009) showed that selenium had no chemopreventive effects on prostate cancer in clinical trials conducted in the USA, Canada and Puerto Rico ( 13 ). This discrepancy might be explainable if we consider dietary selenium could have compensatory effects in malnourished poor populations in Linxian, China compared with its rich counterparts in USA, Canada and Puerto Rico. To overcome this controversy, we need to do more detailed and carefully designed basic, clinical and epidemiological investigations ( 14 , 15) . A previous report showed that SeMet had reduced the azoxymethane (AOM)-induced premalignant lesions through a polyamine-independent mechanism in AOM–dextran sodium sulfate (DSS) mouse model ( 16 ). Therefore, it would be meaningful to identify molecular targets responsible for SeMet-mediated inhibition of CRC carcinogenesis in a therapeutic and prognostic point of view.

In this study, we aimed to verify the chemopreventive effects of SeMet against CRC development in terms of adenomatous polyp occurrence in AOM–DSS mice, a standard animal system for investigation of colon carcinogenesis ( 17 , 18) . Consequently, we tried to identify new molecular targets, possibly acting as biomarkers involved in SeMet-mediated inhibition of colorectal carcinogenesis by proteomics analysis using two-dimensional electrophoresis (2-DE). We used Pathway Studio 8 software analysis using the 2-DE-derived proteins to find out their interrelationship in different biological processes. Based on the pathway analysis, we did validation of four selected proteins using immunohistochemistry. We also performed immunohistochemistry for a known oxidative stress biomarker 8-hydroxy-2ʹ-deoxyguanosine (8-OHdG) and prepared a network analysis among previously validated four proteins, 8-OHdG and colon cancer. To our knowledge, this is the first study demonstrating proteomics analysis to identify SeMet-mediated protein candidates in chemoprevention of colon cancer.

Materials and methods

Animal experiments

Male Crj: CD-1 (ICR) mice (Lab Animal, Korea) aged 5 weeks were used in this study. Grinded feed (Lab Animal, Korea) was applied as the basal diet throughout this study. A colonic carcinogen AOM was purchased from Sigma–Aldrich Co. (St Louis, MO). DSS with a molecular weight of 36 000–50 000 (cat. no.160110) was obtained from MP Biomedicals, LLC (Aurora, OH). DSS was prepared by dissolving in distilled water at a concentration of 1.5% (w/v) and subsequently utilized as an inflammatory agent for induction of colitis. SeMet was purchased from Pharma Se. SeMet was properly mixed with the grinded feed to apparent homogeneity. All animals were maintained according to the ethical standards of the institute.

Experimental procedure

After arrival, mice were acclimated for 7 days with free access to drinking water and basal diet. The AOM–DSS dosing of the mice was planned based on a previous study ( 19 ). Although a previous study ( 20 ) described 10 p.p.m. as the highest tolerable dose of SeMet, we tested with 15 p.p.m. initially and found no adverse signs of toxicity (body weight loss, diarrhea and lethality). The kinetics of drug-induced toxicities (body weight loss, diarrhea and lethality) were determined daily till the initiation of AOM–DSS treatment (data not shown). In this study, there were 48 mice divided into four groups (12 mice/group, as illustrated in Figure 1A ). Group 1: mice provided with only normal grinded food with absence of SeMet or AOM–DSS treatment. Group 2: mice were treated with only 15 p.p.m. of SeMet throughout the experimental period. Group 3: mice received AOM–DSS treatment only. Group 4: mice were pretreated with SeMet (15 p.p.m.) prior to AOM–DSS administration. For the AOM–DSS model, mice were injected intraperitoneally with 10mg AOM/kg body wt. One week after AOM injection, the animals were exposed to 1.5% (w/v) DSS in drinking water for an additional 1 week. The experiments were continued for 15 weeks until the mice reached 22 weeks of age. They were then killed by CO 2 euthanasia for further analysis.

Fig. 1.

( A ) Experimental design for analyzing the chemopreventive effect of SeMet against AOM/DSS-induced colon carcinogenesis. ( B ) Group 1: mice given normal food and no AOM–DSS or SeMet. Group 2: mice were treated with 15 p.p.m. of SeMet throughout the experimental period without AOM–DSS. Group 3: mice received only a single intraperitoneal injection of 10mg/kg body wt of AOM followed by a 1 week exposure to 1.5% (w/v) DSS in drinking water in the absence of SeMet supplementation. Group 4: mice were initially provided SeMet (15 p.p.m.) containing feed prior to AOM–DSS (same dose as Group 3). All mice were between 5 weeks of age at the beginning of the experiment. ( C ) Representative appearance of colon tumors in mice of different groups (control, SeMet, AOM–DSS, SeMet–AOM–DSS treated) at the time of killing (22 weeks). ( D ) Polyp analysis was performed by scoring based on diameter of each polyp (as explained in Materials and methods) showing dramatically lower size of polyps in SeMet–AOM–DSS-treated mice than that in AOM–DSS-treated mice. There was no apparent polyps present in the SeMet-treated mice. * P < 0.05 and ** P < 0.01.

Fig. 1.

( A ) Experimental design for analyzing the chemopreventive effect of SeMet against AOM/DSS-induced colon carcinogenesis. ( B ) Group 1: mice given normal food and no AOM–DSS or SeMet. Group 2: mice were treated with 15 p.p.m. of SeMet throughout the experimental period without AOM–DSS. Group 3: mice received only a single intraperitoneal injection of 10mg/kg body wt of AOM followed by a 1 week exposure to 1.5% (w/v) DSS in drinking water in the absence of SeMet supplementation. Group 4: mice were initially provided SeMet (15 p.p.m.) containing feed prior to AOM–DSS (same dose as Group 3). All mice were between 5 weeks of age at the beginning of the experiment. ( C ) Representative appearance of colon tumors in mice of different groups (control, SeMet, AOM–DSS, SeMet–AOM–DSS treated) at the time of killing (22 weeks). ( D ) Polyp analysis was performed by scoring based on diameter of each polyp (as explained in Materials and methods) showing dramatically lower size of polyps in SeMet–AOM–DSS-treated mice than that in AOM–DSS-treated mice. There was no apparent polyps present in the SeMet-treated mice. * P < 0.05 and ** P < 0.01.

Scoring of polyp analysis

After killing, mice colons were collected and polyps inside the colons were evaluated by scoring ( Figure 1D ). Based on polyp size by measuring its diameter, five points were given to polyps about 5cm in size, three points were scored to polyps about 3cm in size, two points to the polyps about 1cm in size and one point to polyps of about 0.5cm.

Histological analysis

The mouse colon and liver specimens were collected and fixed in 10% neutral-buffered formalin and then routinely paraffin embedded. The paraffin sections (10 μm thickness) were affixed to micro-slides (MUTO-GLASS, Japan) and air-dried overnight at 37°C. Then, the sections were dewaxed in xylene and rehydrated through a descending-graded alcohol series. Finally, the sections were subjected to hematoxylin and eosin (H&E) staining.

Protein extraction

Colon tissue (non-polyp) was collected from all groups, immediately washed using cold homogenization buffer A (50mM Tris–HCl [pH 7.5], 2mM ethylenediaminetetraacetic acid, 150mM NaCl and 0.5mM dithiothreitol) and cut into small pieces. The same colon samples were divided for two-dimensional polyacrylamide gel electrophoresis analysis and immunohistochemsitry.

Two-dimensional polyacrylamide gel electrophoresis

Homogenization of tissue samples was done in buffer (50mM Tris–HCl [pH 7.5], 0.25M sucrose, 5mM magnesium acetate, 0.2mM ethylenediaminetetraacetic acid and 0.5mM dithiothreitol) with Halt™ protease inhibitor cocktail (Thermo Fisher Scientific, Rockford, IL) on ice using a sample grinding kit (GE Healthcare Life Science, Uppsala, Sweden).Then, we centrifuged at 13 000 r.p.m. for 30min (4°C) and 10% trichloroacetic acid was added to the supernatant to precipitate the proteins. The precipitates were redissolved in rehydration buffer (8M urea, 2% [(3-cholamidopropyl)-dimethyl-ammonio]1-propanesulfonate, 50mM dithiothreitol and 0.2% IPG buffer) for two-dimensional polyacrylamide gel electrophoresis. Protein concentrations were determined using a BCA protein assay kit (Thermo Fisher Scientific). Then, we separated 200 µg of protein with an Immobiline Dry Strip (pH 4–7, 18 cm; GE Healthcare Life Science). The second dimension of the separation was carried out on a 12% acrylamide gel for 7 h in an Ettan Dalt II system (10 mA/gel; 1 h, 40 mA/gel; >6 h) (GE Healthcare Life Science). Gels were stained using silver staining technique and image analysis and spot detection of these gels were performed using Progenesis SameSpots software (version. 4.1; Nonlinear Dynamics, Newcastle, UK). The gel images were aligned by automated calculation of alignment vectors. The automatic analysis was performed on all the aligned images using the analysis wizard. The aligned images were grouped to reflect the biological grouping and were used to generate a reference (master) image using the Progenesis SameSpots software. The reference image was used to normalize and quantify the spot volumes.

Nano-high-performance liquid chromatography–electrospray ionization–quadrupole ion trap-mass spectrometer and protein identification

For mass spectrometer (MS) analysis, briefly we cut spots of interest from the 2D gels containing samples from mice of all groups and digested with trypsin. Protein identification was performed using a nano liquid chromatography/MS system consisting of a surveyor high-performance liquid chromatography system (Thermo Scientific, Waltham, MA) and electrospray ionization–quadrupole ion trapMS (LCQ Deca XP-Plus; Thermo Finnigan, San Jose, CA) equipped with a nano-ESI source. About 10 μl of tryptic peptides were loaded by the auto sampler onto a C18 trap column (intradermally; 300 μm, length 5mm, particle size 5 μm; LC Packings, Amsterdam, Netherlands) for desalting and concentration at a flow rate of 20 μl/min. The trapped peptides were then back-flushed and separated on a home-made C18 reversed-phase capillary column (75 μm silica tube, length 150mm, particle size 5 μm). The pump flow rate was split 1:100 for a column flow rate of 150 μl/min. Mobile phase A was 0.5% acetic acid and 0.02% formic acid in water, and B was 0.5% acetic acid and 0.02% formic acid in 80% acetonitrile. Samples were introduced into the column and eluted with a gradient of 5–5–20–50–60–80–100% of mobile phase B for 0–15–18–50–55–60–62min, respectively. MS and MS/MS spectra were obtained using a heated capillary temperature of 220°C, an ESI voltage of 2.5 kV and a collision energy setting of 35%. Data-dependent peak selection of the three most abundant ions in the mass spectra was used. Dynamic exclusion was enabled with a maximum repeat count of two, a repeat duration of 0.5 min and a 3min exclusion duration. MS/MS mass peak lists were analyzed for b and y ions using SEQUEST (version 3.3.1; Thermo Finnigan) software. SEQUEST was used for the identification of proteins using the IPI database. The SEQUEST results were filtered using the following parameters: a mass tolerance of 2.0Da for the precursor ion and 1.0Da for the fragment ions, one missed cleavage per peptide was allowed and modifications of proteins were not taken into account. The validity of peptide/spectrum matches was assessed using the SEQUEST-defined parameters, the cross-correlation score (Xcor) and the normalized difference in Xcors. Matched peptide sequences were required to pass the following filters for identification: (i) the uniqueness scores of the matches’ normalized difference in Xcors were at least 0.1 and (ii) minimum Xcor values ≥1.90, ≥2.20 and ≥3.75 for singly, doubly and triply charged ions, respectively.

Protein pathway analysis

We used Pathway Studio 8 software (Ariadne Genomics, Rockville, MD) to study functional interactions and possible pathways among proteins in our data. This software helps to interpret biological significance from protein expression data, build and analyze pathways, and find protein interaction maps and their relation to cell processes ( 21 ). Pathway Studio uses an inbuilt RESNET database for molecular interactions based on usual language processing of scientific abstracts in PubMed. RESNET helps one to make a new pathway of interest for a number of proteins or genes depending on the information present till now. It also utilizes all existing pathways database and makes a connection among all of the proteins in a new specific pathway.

Immunohistochemistry

Paraffin sections were deparaffinized and rehydrated, and endogenous peroxidase was quenched with 0.3% H 2 O 2 in methanol for 20 min. Sections were incubated in primary antibodies overnight in 4ºC with occasional special pretreatment as mentioned in Supplementary Table , available at Carcinogenesis Online. Sections were incubated for 30 min with biotin-conjugated secondary antibodies, washed in phosphate-buffered saline and incubated with streptavidin-biotinylated horseradish peroxidase (Vector labs) for 30 min. After washing in phosphate-buffered saline, visualization of immune complexes was performed by a 10 min incubation with a 3,3ʹ-diaminobenzidine substrate solution containing 1.8 × 10 3 % (v/v) H 2 O 2 . After the visualization with 3,3ʹ-diaminobenzidine substrate, the sections were washed twice in phosphate-buffered saline and counterstained with Gill’s hematoxylin. Antiprohibitin (PHB; H-80) (sc-28259), antipurine nucleoside phophorylase (PNP) (sc-135163), anti-c-reactive protein (CRP; H-90) (sc-30047), antiannexin II (H-50) (sc-9061) antibodies were purchased from Santa Cruz Biotechnology and anti 8-OHdG (MOG-100P) was purchased from JaICA. Antivascular endothelial growth factor antibody was purchased from Abcam (ab46154) and antiepidermal growth factor receptor antibody was purchased from Cell Signaling (D38B1). The immunoreactivity was scored using a method described previously ( 22 ). Sections were scored by examination of staining intensity and by estimating the percentage (P) of tumor cells showing characteristic staining in nine randomly selected fields at ×200 magnifications and expressed as an average of scores ranging between 1 and 300 in at least three samples from each group. This score was later transformed to a percent score with 300 being 100% and graphical data were constructed.

Western blot

Colon tissue protein was extracted using a PRO-PREP TM protein extraction kit (cat. no. 17081) and quantified by the BCA method. Western blot gels were prepared according to protein size, and 50–70 µg protein was loaded onto the gel. We used running buffer (10× Tris/glycine/sodium dodecyl sulfate) from Bio-Rad (cat. no. 161–0732; Hercules, CA) and manually prepared the transfer buffer (25mM Tris, 192mM glycine, 10% methanol). Similar antibodies for four target proteins were used for immunohistochemistry. For loading control, we used horseradish peroxidase-conjugated anti-β-actin (Sigma; cat. no. A3854).

Results

Significant reduction of AOM–DSS-induced polyps upon SeMet dietary administration

Colonic polyposis can be considered as one of the plausible predictors of CRC progression and outcome based on evidence showing modest correlation between polyp size and mortality among CRC patients ( 23 ). To investigate the chemopreventive activity of SeMet on CRC development, we examined whether or not SeMet supplementation can reduce colonic polyp incidence and size in a mouse model of inflammation-related colon carcinogenesis. In this model, colonic carcinogen AOM in combination with a colitis-inducing agent, DSS, was applied to promote tumor development within a short-term period ( 24 ). After we treated four groups of mice, as depicted in Figure 1A , colonic polyps were observed and scored according to their size ( Figure 1C and D ). As expected, polyps with the highest score were present in the AOM–DSS-treated mice. There was no apparent polyp formation in the SeMet-treated mice similar to control. The incidence of polyp formation in the SeMet–AOM–DSS-treated mice was significantly lower than that in the AOM–DSS-treated mice, indicating that dietary consumption of SeMet had an inhibitory effect on colon cancer progression. Dose-dependent experiment using SeMet was also done and found that higher SeMet treatment (15 p.p.m.) gives better result than lower SeMet treatment (5 p.p.m.) with less occurrence of polyps (data not shown).

Histopathological observation of AOM–DSS-induced polyps and normal mucosa in mice treated with SeMet and immunohistochemistry for known oxidative marker

Examination with H&E staining confirmed histopathological alterations of the colonic mucosa, including dysplasia, neoplastic lesions and crypt loss, of AOM–DSS-treated mice ( Figure 2A.III ). Dietary supplementation of SeMet before AOM–DSS introduction notably decreased these lesions ( Figure 2A.IV ). Expectedly, control mice and SeMet-treated mice showed normal colonic mucosa without any significant inflammation ( Figure 2A.I and II ). Taken together, the results obviously show that SeMet pretreatment conferred the ability to suppress neoplasm and inflammation in colonic tissue. Although the mice were killed only after 15 weeks of AOM–DSS injection giving insufficient time for full-blown liver metastasis to occur but clusters of glandula-like cells with higher vascularity than surrounding tissue ( 25 , 26) were seen specially near capillaries of AOM–DSS-treated mice ( Figure 2A.VII ) liver sections with H&E staining. Although SeMet–AOM–DSS mice ( Figure 2A.VIII ) showed only some hyperplasia and the other two groups—control ( Figure 2A.V ) and SeMet only ( Figure 2A.VI ) mice—showed normal appearance of liver tissue. Immunohistochemistry was carried out on the colon tissue for probable SeMet targets like the oxidative stress biomarker 8-OHdG and the results were further scored as mentioned in Materials and methods( Figure 2B and C ). It was found that 8-OHdG was decreased in SeMet-treated AOM–DSS mice than colon tissue of AOM–DSS mice. We also performed immunohistochemistry for other biomarkers such as vascular endothelial growth factor and epidermal growth factor receptor ( Supplementary Figure 2 , available at Carcinogenesis Online). They showed similar pattern of expression as 8-OHdG.

Fig. 2.

( A ) Cross sections of colon tissue from mice (representative images) showing a significantly less advanced stage of carcinogenesis in mouse treated with SeMet before AOM–DSS treatment ( IV ) than in the AOM–DSS only group ( III ), which showed partial destruction of epithelial architecture, submucosal edema and predominant lymphocytic infiltration with H&E staining. No destruction was observed in control and SeMet-treated groups ( I and II , respectively). Histology of mouse liver tissue from animals treated with AOM–DSS only ( VII ) demonstrated diffuse microscopic glandula-like arrangement with increased vascularity indicating micro metastasis, in contrast, livers of SeMet-treated animals ( VI ) and control animals ( V ) showing almost normal appearance of the hepatic sinusoids except for some hyperplasia of hepatocytes in SeMet–AOM–DSS-treated mice ( VIII ). Images of colon tissue were taken at ×200 magnification and scale bar = 100 µm. Liver section pictures were taken at ×400 and scale bar = 50 µm. ( B ) Immunohistochemical staining for the 8-OHdG in colonic tissue showing positive staining in majority of cells in a homogeneous pattern from AOM–DSS group ( BIII ) but SeMet–AOM–DSS tissue ( BIV ) shows a mixture of positively and negatively stained cells. Control group ( BI ) and SeMet group ( BII ) do not show any significant change in 8-OHdG staining. ( C ) Graphical presentations were done for expression of 8-OHdG, where y -axis represents immunoreactivity score in percentage (%) as described in Materials and methods. * P < 0.05.

Fig. 2.

( A ) Cross sections of colon tissue from mice (representative images) showing a significantly less advanced stage of carcinogenesis in mouse treated with SeMet before AOM–DSS treatment ( IV ) than in the AOM–DSS only group ( III ), which showed partial destruction of epithelial architecture, submucosal edema and predominant lymphocytic infiltration with H&E staining. No destruction was observed in control and SeMet-treated groups ( I and II , respectively). Histology of mouse liver tissue from animals treated with AOM–DSS only ( VII ) demonstrated diffuse microscopic glandula-like arrangement with increased vascularity indicating micro metastasis, in contrast, livers of SeMet-treated animals ( VI ) and control animals ( V ) showing almost normal appearance of the hepatic sinusoids except for some hyperplasia of hepatocytes in SeMet–AOM–DSS-treated mice ( VIII ). Images of colon tissue were taken at ×200 magnification and scale bar = 100 µm. Liver section pictures were taken at ×400 and scale bar = 50 µm. ( B ) Immunohistochemical staining for the 8-OHdG in colonic tissue showing positive staining in majority of cells in a homogeneous pattern from AOM–DSS group ( BIII ) but SeMet–AOM–DSS tissue ( BIV ) shows a mixture of positively and negatively stained cells. Control group ( BI ) and SeMet group ( BII ) do not show any significant change in 8-OHdG staining. ( C ) Graphical presentations were done for expression of 8-OHdG, where y -axis represents immunoreactivity score in percentage (%) as described in Materials and methods. * P < 0.05.

Identification of molecular targets for SeMet-mediated CRC prevention by analysis of proteomics

Comparative proteomics analysis was carried out in the colonic tissue specimens using 2-DE and MS of the four groups. Altered proteins between control and AOM–DSS mice were as expected (data not shown). Altered expressions of 76 protein spots were analyzed between AOM–DSS- and SeMet-treated AOM–DSS mice. Thirty proteins were identified based on the 2-DE analysis. Some of the protein isomers also gave simultaneous expression into different spot positions ( Supplementary Figure 1 and Table 1 , available at Carcinogenesis Online). Among 30 proteins, nine proteins were upregulated and 21 proteins were downregulated in SeMet-treated AOM–DSS mice.

Analysis of differentially expressed proteins by Pathway Studio 8 software

First, the software was used to analyze the network of the differentially expressed proteins with an intention to find regulatory pathways active in SeMet-mediated prevention in CRC. The direct interaction pathway analysis using the Pathway Studio 8 software based on an integrated bioinformatics approach displayed specific proteins that are common binding partners, targets and regulators for these proteins. Additionally, it showed the cellular processes involvement and their localization ( Figure 3 ). This analysis provided an approach to compare different kinds of interactions, including protein–protein interactions and physiological interaction in SeMet-pretreated CRC. In Pathway Studio analysis, we used all 30 proteins but 27 proteins were found to be interconnected in a pathway we made including all 9 upregulated proteins and 18 downregulated proteins in SeMet-treated AOM–DSS mice. Our observation shows these proteins changed their expression level due to the effects of SeMet and AOM–DSS on cell processes including cell proliferation, apoptosis, cell survival, cell growth, cell death, reactive oxygen species generation, oxidative stress, inflammation, immune response and their cellular localizations. We have found that SeMet modulated upregulated proteins, which include PHB, PNP, isocitratrate dehydrogenase 3α, eukaryotic translation initiation 5A, proteasome activator complex subunit 1, inorganic pyrophosphatase 1, β-actin, annexin 7 (ANXA7) and ANXA3, and downregulated proteins, which include ANXA1, ANXA2, cofilin 1 (CFL1), CFL2, CRP1, destrin (DSTN), glutathione transferase ω1, hypoxanthine–guanine phosphoribosyltransferase 1, tropomyosin α-1 chain, l -lactate dehydrogenase A chain, nucleoside diphosphate kinase B, peroxiredoxin 1, peroxiredoxin 4, phosphoglycerate mutase 2, S-formylglutathione hydrolase, triosephosphate isomerase 1, transaldolase and ubiquinol cytochrome c reductase 1. These up- and downregulated proteins were actively involved throughout the cell processes involving other small molecules, transcription factors, ligands and so on ( Figure 3 and Table I ). After building the pathway, we found that among the upregulated proteins, PHB and PNP, and downregulated proteins, ANXA2 and CRP, were playing major role in modulating major events. Therefore, after validating by immunohistochemistry ( Figure 4 ), we made another pathway ( Figure 5 ) manually with the help Pathway Studio 8 software. Here, we have mechanistic approach to relate oxidative stress marker 8-OHdG and colorectal neoplasm with our previously validated major four proteins ( Figure 4 ) obtained from proteomics analysis. We found all these four proteins are directly or indirectly related with 8-OHdG and colorectal cancer ( Figure 5 ), through cell processes involving apoptosis, oxidative stress, cytokinesis and phosphatidylinositol 3-kinase and DNA topoisomerase 1.

Fig. 3.

Biological networks generated by pathway analyses in Pathway Studio software using proteins identified by 2D gel proteomics analysis show proteins (red), cell processes (yellow) and their localizations (blue circle = nucleus, orange-ellipsoid = mitochondria, orange parallel lines = cell membrane, yellow structure = endoplasmic reticulum) and their interrelationships. Proteins that were increased and decreased in SeMet–AOM–DSS compared with AOM–DSS only are indicated by purple and blue shadows, respectively.

Fig. 3.

Biological networks generated by pathway analyses in Pathway Studio software using proteins identified by 2D gel proteomics analysis show proteins (red), cell processes (yellow) and their localizations (blue circle = nucleus, orange-ellipsoid = mitochondria, orange parallel lines = cell membrane, yellow structure = endoplasmic reticulum) and their interrelationships. Proteins that were increased and decreased in SeMet–AOM–DSS compared with AOM–DSS only are indicated by purple and blue shadows, respectively.

Table I.

2D proteomics analysis of differentially expressed proteins in SeMet-treated AOM–DSS mice compared with AOM–DSS mice

Protein name Gene symbol Protein ID Spot number Expression in SeMet/AOM–DSS 
Annexin 3 Anxa3 IPI00132722.8 40 Increased 
Annexin 7 Anxa7 IPI00114017.2 57 Increased 
β-actin Actb IPI00110850.1 39 Increased 
Eukaryotic translation initiation 5A Eif5a IPI00108125.4 10 Increased 
Inorganic pyrophosphatase 1 Ppa1 IPI00110684.1 38 Increased 
Isoform 1 of isocitrate dehydrogenase (nicotinamide adenine dinucleotide) subunit α Idh3a IPI00459725.2 41 Increased 
PHB Phb IPI00133440.1 34 Increased 
Proteasome activator complex subunit 1 Psme1 IPI00124223.1 32 Increased 
PNP Pnp IPI00315452.5 33 Increased 
Aldose reductase Akr1b3 IPI00223757.4 49 Decreased 
Alcohol dehydrogenase Akr1a4 IPI00466128.3 50 Decreased 
Annexin 1 Anxa1 IPI00230395.5 51 Decreased 
Annexin 2 Anxa2 IPI00468203.3 48 Decreased 
Cofilin 1 Cfl1 IPI00407543.2 Decreased 
Cofilin 2 Cfl2 IPI00266188.6 Decreased 
CRP Crp1 IPI00314936.1 14 Decreased 
Destrin Dstn IPI00127942.4 Decreased 
Glutathione transferase ω1 Gsto1 IPI00114285.1 25 Decreased 
Hypoxanthine–guanine phosphoribosyltransferase 1 Hprt1 IPI00284806.8 29 Decreased 
Isoform 1 of tropomyosin α-1 chain Tpm1 IPI00123316.1 43 Decreased 
l -Lactate dehydrogenase A chain  Ldha IPI00319994.6 47 Decreased 
Nucleoside diphosphate kinase B Nme2 IPI00127417.1 Decreased 
Peroxiredoxin 1 Prdx1 IPI00121788.1 16, 21 Decreased 
Peroxiredoxin 4 Prdx4 IPI00116254.1 16, 21 Decreased 
Phosphoglycerate mutase 2 Pgam2 IPI00230706.5 24 Decreased 
Proteasome subunit β type 1 precursor Psmb1 IPI00113845.1 18 Decreased 
S-formylglutathione hydrolase Esd IPI00109142.4 46 Decreased 
Triosephosphate isomerase 1 Tpi1 IPI00467833.5 19, 23 Decreased 
Transaldolase Taldo1 IPI00124692.1 52 Decreased 
Ubiquinol cytochrome c reductase 1 Uqcrfs1 IPI00133240.1 20 Decreased 
Protein name Gene symbol Protein ID Spot number Expression in SeMet/AOM–DSS 
Annexin 3 Anxa3 IPI00132722.8 40 Increased 
Annexin 7 Anxa7 IPI00114017.2 57 Increased 
β-actin Actb IPI00110850.1 39 Increased 
Eukaryotic translation initiation 5A Eif5a IPI00108125.4 10 Increased 
Inorganic pyrophosphatase 1 Ppa1 IPI00110684.1 38 Increased 
Isoform 1 of isocitrate dehydrogenase (nicotinamide adenine dinucleotide) subunit α Idh3a IPI00459725.2 41 Increased 
PHB Phb IPI00133440.1 34 Increased 
Proteasome activator complex subunit 1 Psme1 IPI00124223.1 32 Increased 
PNP Pnp IPI00315452.5 33 Increased 
Aldose reductase Akr1b3 IPI00223757.4 49 Decreased 
Alcohol dehydrogenase Akr1a4 IPI00466128.3 50 Decreased 
Annexin 1 Anxa1 IPI00230395.5 51 Decreased 
Annexin 2 Anxa2 IPI00468203.3 48 Decreased 
Cofilin 1 Cfl1 IPI00407543.2 Decreased 
Cofilin 2 Cfl2 IPI00266188.6 Decreased 
CRP Crp1 IPI00314936.1 14 Decreased 
Destrin Dstn IPI00127942.4 Decreased 
Glutathione transferase ω1 Gsto1 IPI00114285.1 25 Decreased 
Hypoxanthine–guanine phosphoribosyltransferase 1 Hprt1 IPI00284806.8 29 Decreased 
Isoform 1 of tropomyosin α-1 chain Tpm1 IPI00123316.1 43 Decreased 
l -Lactate dehydrogenase A chain  Ldha IPI00319994.6 47 Decreased 
Nucleoside diphosphate kinase B Nme2 IPI00127417.1 Decreased 
Peroxiredoxin 1 Prdx1 IPI00121788.1 16, 21 Decreased 
Peroxiredoxin 4 Prdx4 IPI00116254.1 16, 21 Decreased 
Phosphoglycerate mutase 2 Pgam2 IPI00230706.5 24 Decreased 
Proteasome subunit β type 1 precursor Psmb1 IPI00113845.1 18 Decreased 
S-formylglutathione hydrolase Esd IPI00109142.4 46 Decreased 
Triosephosphate isomerase 1 Tpi1 IPI00467833.5 19, 23 Decreased 
Transaldolase Taldo1 IPI00124692.1 52 Decreased 
Ubiquinol cytochrome c reductase 1 Uqcrfs1 IPI00133240.1 20 Decreased 
Fig. 4.

( A ) Validation of selected candidate proteins from 2D proteomics analysis using immunohistochemistry. PHB ( I–IV ) is highly increased in SeMet-treated mice colon tissue ( II ) and quite reduced in AOM–DSS tissue ( III ) as expected but some amount of positive staining can be seen in SeMet–AOM–DSS-treated tissue ( IV ). PNP also shows the same pattern of staining ( V–VIII ). ANXA2 staining ( IX–XII ) is increased in AOM–DSS tissue ( XI ) but reduced in SeMet–AOM–DSS tissue ( XII ) and no stain in control ( IX ) and SeMet ( X ) only–treated groups. CRP ( XIII–XVI ) staining pattern is similar to ANXA2. Allpictures were taken at ×400 magnification and scale bar = 50 µm. ( B ) Statistical presentations of protein expression against PHB ( I ), PNP ( II ), ANXA2 ( III ) and CRP ( IV ) antibody in colon tissue. The y -axis represents immunoreactivity score in percentage (%) as described in Materials and methods. * P < 0.05.

Fig. 4.

( A ) Validation of selected candidate proteins from 2D proteomics analysis using immunohistochemistry. PHB ( I–IV ) is highly increased in SeMet-treated mice colon tissue ( II ) and quite reduced in AOM–DSS tissue ( III ) as expected but some amount of positive staining can be seen in SeMet–AOM–DSS-treated tissue ( IV ). PNP also shows the same pattern of staining ( V–VIII ). ANXA2 staining ( IX–XII ) is increased in AOM–DSS tissue ( XI ) but reduced in SeMet–AOM–DSS tissue ( XII ) and no stain in control ( IX ) and SeMet ( X ) only–treated groups. CRP ( XIII–XVI ) staining pattern is similar to ANXA2. Allpictures were taken at ×400 magnification and scale bar = 50 µm. ( B ) Statistical presentations of protein expression against PHB ( I ), PNP ( II ), ANXA2 ( III ) and CRP ( IV ) antibody in colon tissue. The y -axis represents immunoreactivity score in percentage (%) as described in Materials and methods. * P < 0.05.

Fig. 5.

Manually prepared illustration of biological networks found through pathway analyses with the help of Pathway Studio software using 8-OHdG protein, CRC and validated major four proteins (PHB, PNP, ANXA2 and CRP) from 2D proteomics analysis. Proteins that were increased and decreased in SeMet–AOM–DSS group were compared with AOM–DSS only and are indicated by red and green color, respectively.

Fig. 5.

Manually prepared illustration of biological networks found through pathway analyses with the help of Pathway Studio software using 8-OHdG protein, CRC and validated major four proteins (PHB, PNP, ANXA2 and CRP) from 2D proteomics analysis. Proteins that were increased and decreased in SeMet–AOM–DSS group were compared with AOM–DSS only and are indicated by red and green color, respectively.

Immunohistochemistry for verification of newly found markers

We also verified the newly found marker proteins PHB, PNP, ANXA2 and CRP with immunohistochemistry on the colon tissue of mice from respective groups and found that the results were similar to their protein expression patterns in 2-DE analysis. Upregulated proteins showed increased expression and downregulated proteins showed decreased expression in SeMet-treated AOM–DSS mice compared with AOM–DSS mice colon tissue ( Figure 4A and B ). To further verify, we also performed western blot from colon tissue lysate and immunoblotted with relevant antibodies to check the expression level of the proteins of interest. It also showed similar pattern as immunohistochemistry data ( Supplementary Figure 3 , available at Carcinogenesis Online).

Discussion

The results from this study clearly confirmed that SeMet, as a dietary supplement, effectively suppressed AOM–DSS-induced colon carcinogenesis in mice without any significant deleterious effect. As reported, selenium exhibits beneficial effects against certain cancer although it might not be effective in all types of cancer giving rise to the recent conflict of opinions ( 13 , 27–30 ). SeMet (a selenium analog of the amino acid methionine, Met) is characterized by the absence of cytotoxicity or genotoxicity as it can be readily oxidized and reduced ( 31 , 32) . Epidemiologic observations and laboratory research have suggested that dietary selenium reduces the risk of colon cancer ( 33 , 34) . We found that SeMet treatment has reduced polyp occurrence and dysplasia ( Figures 1C , D and 2A ) in AOM–DSS mouse. However, our understanding of its effects on molecular targets still needs to be clarified. As seen in Figure 1 , prolonged treatment with SeMet throughout the 15 week experiment did not cause any adverse effects consistent with previous research, where SeMet has been reported to have antioxidant efficacy with no oxidative DNA damage ( 31 ). We studied already known oxidative stress marker related to colon carcinogenesis, which has been identified as probable targets of SeMet chemoprotective activity 8-OHdG ( 35 , 36) and found that they were decreased with SeMet treatment as expected ( Figure 2 ).

Using 2-DE-based proteomics analysis, we investigated protein profile changes relating to the chemopreventive efficacy of SeMet against colon carcinogenesis induced in an AOM–DSS mouse model following experimental example of Yeo et al. ( 37 ). We found 9 upregulated and 21 downregulated proteins in SeMet-treated AOM–DSS mice ( Table I ). Upregulated proteins include ANXA3, ANXA7, β-actin, EIF5A, inorganic pyrophosphatase 1, isocitratrate dehydrogenase 3α, PHB, proteasome activator complex subunit 1, PNP and so on and downregulated proteins include ANXA1, ANXA2, CFL1, CFL2, CRP, DSTN, glutathione transferase ω1and so on ( Table I ). Then, we performed pathway analysis to find the interrelationship of these proteins and their relation to colon cancer.

Pathway analysis serves as the important step toward an improved understanding of the relationship of studied proteins in SeMet-pretreated colorectal cancer. Protein network analysis using Pathway Studio 8 software revealed that SeMet exerted its chemopreventive activity in the initial stages by affecting molecular events mainly involved in cell proliferation, cell differentiation, oxidative stress, DNA repair and apoptosis ( Figure 3 ). Similarly, a previous report suggested that the protective effects of selenium compounds against various cancers during the postinitiation phase of carcinogenesis are mediated via modulation of several processes, including immunological response, apoptotic cell death, DNA repair, oxidative stress response, carcinogen metabolism, tumorigenesis and angiogenesis ( 38 ). As presented in Figure 3 , in AOM–DSS mouse pretreated with SeMet, we demonstrated that the upregulated proteins which cooperated, directly and/or indirectly, with the downregulated oncogenic proteins in contributing SeMet-related chemoprotection against tumor formation. Our results revealed that major pathways might be modulated through dietary uptake of SeMet in AOM–DSS mouse. From this analysis, we have been able to find that SeMet treatment might repress inflammation, metastasis and oncogenic proteins, and simultaneously increased tumor suppressor proteins in AOM–DSS mouse.

This is the first evidence of newly identified candidate proteins, plausible to be considered as potential biomarkers, in response to SeMet, which confer protection against CRC development and progression. In SeMet-pretreated AOM–DSS mice, there was an increase ( Table I , Figure 4 ) in the expression of PHB, a highly conserved protein across species, which has been shown to modulate cell proliferation, apoptosis, transcription, mitochondrial protein folding and works as a cell-surface receptor ( 39 ). This result was also validated by immunohistochemistry, and SeMet treatments showed an increased expression of this protein in AOM–DSS mouse model ( Figure 4 ). Although little is known regarding the regulation and role of PHB during intestinal inflammation or carcinogenesis, it has already been suggested as a candidate for tissue-based detection of gastric cancer ( 40 ). In addition, we also found an upregulated level of PNP, which catalyzes the reversible conversion of a purine riboside to the corresponding base. Inosine is converted to hypoxanthine and guanosine to guanine. These latter two reactions occur in human metabolism during the so-called ‘purine salvage pathway’, which converts purines and their ribo- and deoxyribonucleosides to the corresponding mononucleotides. PNP has been recognized to be increased in some cases of colon carcinogenesis ( 41 ) but the correlation with extent of CRC is still not resolved ( 42 ). Studies in serum level of colon cancer patients receiving chemotherapy show that PNP increased after receiving medication, which might indicate its role in the repair process ( 43 ). In this study, mice treated with SeMet only showed a high increase of PNP even without the presence of cancer. In human serum of cancer patients, PNP was increased several folds than normal people ( 43 ) contradicting our finding, which shows a decreased PNP in colon tissue from AOM–DSS mice than control mice and a further increase in SeMet–AOM–DSS-treated mice emphasizing the fact that the mechanism of PNP activity in colon cancer needs critical attention and further clarification as it might prove to be an important tissue biomarker. In our validation study by immunohistochemistry, we found similar trend of expression of this protein where SeMet has increased PNP in AOM–DSS colon carcinogenesis ( Table I , Figure 4 ). Similarly, SeMet supplementation in AOM–DSS mice enhanced the expression level of eukaryotic translation initiation 5A. Saini et al. ( 44 ) used molecular genetics and biochemical studies to show that EIF5A promotes translation elongation. This protein acts as a positive regulator of p53 and p53-dependent apoptosis ( 45 ). Indeed, our previous work has suggested that SeMet promotes p53-mediated cellular response in an in vitro cell system ( 46 ). We also found that the expression of ANXA3 and ANXA7 was upregulated upon SeMet pretreatment. Previous reports have demonstrated the tumor suppressor functions of ANXA3 and ANXA7 in human cancers ( 47 ).

In contrast, we have also detected several downregulated proteins in SeMet-pretreated mice ( Table I ) and those were actively involved in pathway analysis ( Figure 3 ). The levels of ANXA1 and ANXA2 were reduced in SeMet-pretreated AOM–DSS mice. Annexins are calcium- and phospholipid-binding proteins with important roles in signal transduction, cellular differentiation, proliferation and thus tumorigenesis ( 47 ). In agreement with a previous comparative proteomics analysis by Duncan et al. ( 48 ), ANXA1 and ANXA2 are overexpressed during colorectal cancer. Emoto et al. ( 49 ) showed that ANXA2 is overexpressed in advanced carcinoma and it might be linked to the development and metastatic spread of CRC. In our validation study by immunohistochemistry, we also found AOM–DSS-induced colon cancer increased ANXA2 level, and SeMet treatment reduced ANXA2 level in AOM–DSS cancer mouse ( Table I , Figure 4 ). Further study is required to determine the underlying mechanisms and/or responsive pathways of SeMet in reduction of ANXA proteins. Additionally, we also identified another downregulated protein called CRP. Recently, Johnson et al. ( 50 ) reported CRP-mediated tumor metastasis behavior in kidney cancer patients. Interestingly, CRP also is strongly related to chronic inflammation, which is a major factor of colon carcinogenesis ( 51 ). Thus, CRP might be a potential target for the modulation of colon carcinogenesis. Our validation by immunohistochemistry also showed increased expression of CRP protein in colon carcinogenesis mouse model, but decreased in SeMet-treated AOM–DSS mouse model ( Table I , Figure 4 ). Our findings showed the reduction of the protein complex ADF/cofilin family, including CFL1, CFL2 and DSTN, in SeMet-treated AOM–DSS mouse model. These proteins play distinct roles in governing cell migration and invasion in colon cancer based on their differential abilities ( 52 ). The results also intriguingly showed the downregulation of the peroxiredoxin protein family, including PRDX1 and PRDX4. PRDX proteins play dual roles as ‘gatekeepers’ toward oxidative damage in normal tissue or as ‘promoters’ for survival in proliferating malignant cells. Overexpression of PRDX1 has been detected in various human carcinomas of colorectal cells ( 53 ). On the other hand, PRDX1 is required to suppress chronic inflammation ( 54 ). Indeed, SeMet has been suggested to play a role as an antioxidant to inflammatory diseases ( 55 ) and the serum level of selenium has been identified to have strong correlation with inflammation and mortality by an undergoing clinical trial for its effectiveness in inflammation in the elderly ( 56 ).

Interestingly, we found that oxidative stress marker 8-OHdG is closely related with the expression of the four major proteins, PHB, PNP, ANXA2 and CRP, found in proteomics analysis by interacting with them directly or indirectly ( Figure 5 ). These putative responsive pathways might be crucial for early stage response to SeMet-mediated chemoprevention in colon carcinogenesis mouse model.

Previously, it has been shown that SeMet combined with other antioxidants had a positive effect on survivability ( 57 ). It may be due to the effect of SeMet, which might extend the life span in our colon carcinoma with SeMet treatment. As we have tried to find out new biomarkers in SeMet-treated colon cancer mouse on initial stage of colon cancer progression, we had withdrawn SeMet before the carcinogen initiation. In this study, we did not find any mortality in treated mouse during the experiments and when we killed the mouse, cancer progression was at early stage. We considered early stage to kill the mouse because we wanted to differentiate effects of SeMet from carcinogen-treated mice. If we continue experiment until the final stage of cancer progression, we might not be able to see the impact of SeMet as the entire carcinogen-treated mice would have had full-blown carcinoma. Thereby, we would not be able to find out any biomarkers difference, which was our main focus.

In conclusion, our observations indicate an inhibitory effect of SeMet on CRC development and progression in an inflammatory colon carcinogenesis model. This is the first proteomics analysis demonstrating the candidates responsible for SeMet-mediated chemoprevention. We cannot totally rule out the fact that these differences in protein expression may be just due to the difference in developmental stages of CRC but SeMet treatment showed decrease in overall carcinoma progression. Our network analysis emphasized that potential target proteins have an important role in the pathways related to cell proliferation, differentiation, migration, oxidative stress and apoptosis. This study in a mouse system might provide insights on SeMet-mediated colon cancer prevention, although further investigation is necessary to determine the detailed mechanisms underlying SeMet’s chemopreventive activity for its utilization in a practical manner in the treatment and prevention of colon cancer.

Supplementary material

Supplementary Figures 1–3 can be found at http://carcin.oxfordjournals.org/

Funding

Korea Ministry of Environment as ‘The Ecoinnovation Project’ (412-112-011); Korea Healthcare Technology R&D Project, Ministry of Health and Welfare (A060768); Basic Science Research Program through National Research Foundation of Korea (NRF), Ministry of Education, Science and Technology, Republic of Korea (2010-0027630).

Acknowledgements

We are grateful for the kind assistance regarding experimental planning and suggestions from Dr Takuji Tanaka, (Professor and Chairman, Department of Oncologic Pathology, Kanazawa Medical University, Japan). We also thank Dr Jong Bok Seo for his kind guidance from Seoul Branch, Korea Basic Science Institute, Seoul.

Conflict of Interest Statement: No competing financial interests exist.

Abbreviations:

    Abbreviations:
  • ANXA1

    annexin 1

  • AOM

    azoxymethane

  • CFL1

    cofilin 1

  • CRC

    colorectal cancer

  • 2-DE

    two-dimensional electrophoresis

  • DSTN

    destrin

  • DSS

    dextran sodium sulfate

  • H&E

    hematoxylin and eosin

  • MS

    mass spectrometer

  • 8-OHdG

    8-hydroxy-2ʹ-deoxyguanosine

  • PHB

    prohibitin

  • PNP

    purine nucleoside phosphorylase

  • SeMet

    selenomethionine

  • Xcor

    cross- correlation score.

References

1.
Jemal
A.
et al
2011
Global cancer statistics
.
CA. Cancer J. Clin
  ,
61
,
69
90
.
2.
Center
M.M.
et al
2009
International trends in colorectal cancer incidence rates
.
Cancer Epidemiol. Biomarkers Prev.
  ,
18
,
1688
1694
.
3.
Ferlay
J.
et al
2010
Estimates of worldwide burden of cancer in 2008: GLOBOCAN 2008
.
Int. J. Cancer
  ,
127
,
2893
2917
.
4.
Triantafillidis
J.K.
et al
2009
Colorectal cancer and inflammatory bowel disease: epidemiology, risk factors, mechanisms of carcinogenesis and prevention strategies
.
Anticancer Res.
  ,
29
,
2727
2737
.
5.
Tinggi
U
.
2008
Selenium: its role as antioxidant in human health
.
Environ. Health Prev. Med.
  ,
13
,
102
108
.
6.
Brigelius-Flohé
R
.
2008
Selenium compounds and selenoproteins in cancer
.
Chem. Biodivers.
  ,
5
,
389
395
.
7.
Letavayová
L.
et al
2008
Toxicity and mutagenicity of selenium compounds in Saccharomyces cerevisiae
.
Mutat. Res.
  ,
638
,
1
10
.
8.
Nelson
M.A.
et al
2005
Studies into the anticancer effects of selenomethionine against human colon cancer
.
Ann. N. Y. Acad. Sci.
  ,
1059
,
26
32
.
9.
J.
et al
2005
Selenium and cancer chemoprevention: hypotheses integrating the actions of selenoproteins and selenium metabolites in epithelial and non-epithelial target cells
.
Antioxid. Redox Signal.
  ,
7
,
1715
1727
.
10.
Marshall
J.R
.
2008
Prevention of colorectal cancer: diet, chemoprevention, and lifestyle
.
Gastroenterol. Clin. North Am.
  ,
37
,
73
82
.
11.
Valdiglesias
V.
et al
2010
In vitro evaluation of selenium genotoxic, cytotoxic, and protective effects: a review
.
Arch. Toxicol.
  ,
84
,
337
351
.
12.
Qiao
Y.L.
et al
2009
Total and cancer mortality after supplementation with vitamins and minerals: follow-up of the Linxian General Population Nutrition Intervention Trial
.
J. Natl. Cancer Inst.
  ,
101
,
507
518
.
13.
Lippman
S.M.
et al
2009
Effect of selenium and vitamin E on risk of prostate cancer and other cancers: the Selenium and Vitamin E Cancer Prevention Trial (SELECT)
.
JAMA
  ,
301
,
39
51
.
14.
Herszényi
L.
et al
2008
Chemoprevention of colorectal cancer: feasibility in everyday practice?
Eur. J. Cancer Prev.
  ,
17
,
502
514
.
15.
Brozmanová
J.
et al
2010
Selenium: a double-edged sword for defense and offence in cancer
.
Arch. Toxicol.
  ,
84
,
919
938
.
16.
Baines
A.T.
et al
2000
The effects of dietary selenomethionine on polyamines and azoxymethane-induced aberrant crypts
.
Cancer Lett.
  ,
160
,
193
198
.
17.
Tanaka
T.
et al
2003
A novel inflammation-related mouse colon carcinogenesis model induced by azoxymethane and dextran sodium sulfate
.
Cancer Sci.
  ,
94
,
965
973
.
18.
Krehl
S.
et al
2012
Glutathione peroxidase-2 and selenium decreased inflammation and tumors in a mouse model of inflammation-associated carcinogenesis whereas sulforaphane effects differed with selenium supply
.
Carcinogenesis
  ,
33
,
620
628
.
19.
Suzuki
R.
et al
2006
Strain differences in the susceptibility to azoxymethane and dextran sodium sulfate-induced colon carcinogenesis in mice
.
Carcinogenesis
  ,
27
,
162
169
.
20.
Yan
L.
et al
1999
Dietary supplementation of selenomethionine reduces metastasis of melanoma cells in mice
.
Anticancer Res.
  ,
19
,
1337
1342
.
21.
Nikitin
A.
et al
2003
Pathway studio–the analysis and navigation of molecular networks
.
Bioinformatics
  ,
19
,
2155
2157
.
22.
Charafe-Jauffret
E.
et al
2004
Immunophenotypic analysis of inflammatory breast cancers: identification of an ‘inflammatory signature’
.
J. Pathol.
  ,
202
,
265
273
.
23.
Simons
B.D.
et al
1992
Relationship of polyps to cancer of the large intestine
.
J. Natl. Cancer Inst.
  ,
84
,
962
966
.
24.
Kim
M.
et al
2009
Zerumbone, a tropical ginger sesquiterpene, inhibits colon and lung carcinogenesis in mice
.
Int. J. Cancer
  ,
124
,
264
271
.
25.
Chambers
A.F.
et al
2002
Dissemination and growth of cancer cells in metastatic sites
.
Nat. Rev. Cancer
  ,
2
,
563
572
.
26.
Varghese
H.J.
et al
2002
Activated ras regulates the proliferation/apoptosis balance and early survival of developing micrometastases
.
Cancer Res.
  ,
62
,
887
891
.
27.
Jao
S.W.
et al
1996
Effect of selenium on 1,2-dimethylhydrazine-induced intestinal cancer in rats
.
Dis. Colon Rectum
  ,
39
,
628
631
.
28.
Suzana
S.
et al
2009
Relationship between selenium and breast cancer: a case-control study in the Klang Valley
.
Singapore Med. J.
  ,
50
,
265
269
.
29.
Chiang
E.C.
et al
2009
Defining the optimal selenium dose for prostate cancer risk reduction: insights from the u-shaped relationship between selenium status, DNA damage, and apoptosis
.
DoseResponse.
  ,
8
,
285
300
.
30.
Whanger
P.D
.
2004
Selenium and its relationship to cancer: an update
.
Br. J. Nutr.
  ,
91
,
11
28
.
31.
Stewart
M.S.
et al
1999
Selenium compounds have disparate abilities to impose oxidative stress and induce apoptosis
.
Free Radic. Biol. Med.
  ,
26
,
42
48
.
32.
Suryo Rahmanto
A.
et al
2011
Catalytic activity of selenomethionine in removing amino acid, peptide, and protein hydroperoxides
.
Free Radic. Biol. Med.
  ,
51
,
2288
2299
.
33.
Kim
J.H.
et al
2011
Effects of selenium on colon carcinogenesis induced by azoxymethane and dextran sodium sulfate in mouse model with high-iron diet
.
Lab. Anim. Res.
  ,
27
,
9
18
.
34.
Bhattacharya
A.
et al
2011
Magnetic resonance and fluorescence-protein imaging of the anti-angiogenic and anti-tumor efficacy of selenium in an orthotopic model of human colon cancer
.
Anticancer Res.
  ,
31
,
387
393
.
35.
Kondo
S.
et al
2000
Overexpression of the hOGG1 gene and high 8-hydroxy-2’-deoxyguanosine (8-OHdG) lyase activity in human colorectal carcinoma: regulation mechanism of the 8-OHdG level in DNA
.
Clin. Cancer Res.
  ,
6
,
1394
1400
.
36.
Seo
Y.R.
et al
2002
Selenomethionine induction of DNA repair response in human fibroblasts
.
Oncogene
  ,
21
,
3663
3669
.
37.
Yeo
M.
et al
2010
Loss of SM22 is a characteristic signature of colon carcinogenesis and its restoration suppresses colon tumorigenicity in vivo and in vitro
.
Cancer
  ,
116
,
2581
2589
.
38.
Nadiminty
N.
et al
2008
Mechanisms of selenium chemoprevention and therapy in prostate cancer
.
Mol. Nutr. Food Res.
  ,
52
,
1247
1260
.
39.
Theiss
A.L.
et al
2011
The role and therapeutic potential of prohibitin in disease
.
Biochim. Biophys. Acta
  ,
1813
,
1137
1143
.
40.
Kang
X.
et al
2008
Prohibitin: a potential biomarker for tissue-based detection of gastric cancer
.
J. Gastroenterol.
  ,
43
,
618
625
.
41.
Roessler
M.
et al
2005
Identification of nicotinamide N-methyltransferase as a novel serum tumor marker for colorectal cancer
.
Clin. Cancer Res.
  ,
11
,
6550
6557
.
42.
Sanfilippo
O.
et al
1994
Relationship between the levels of purine salvage pathway enzymes and clinical/biological aggressiveness of human colon carcinoma
.
Cancer Biochem. Biophys.
  ,
14
,
57
66
.
43.
Roberts
E.L.
et al
2004
Plasma purine nucleoside phosphorylase in cancer patients
.
Clin. Chim. Acta
  ,
344
,
109
114
.
44.
Saini
P.
et al
2009
Hypusine-containing protein eIF5A promotes translation elongation
.
Nature
  ,
459
,
118
121
.
45.
Li
A.L.
et al
2004
A novel eIF5A complex functions as a regulator of p53 and p53-dependent apoptosis
.
J. Biol. Chem.
  ,
279
,
49251
49258
.
46.
Seo
Y.R.
et al
2002
Selenomethionine regulation of p53 by a ref1-dependent redox mechanism
.
Proc. Natl. Acad. Sci. U.S.A.
  ,
99
,
14548
14553
.
47.
Mussunoor
S.
et al
2008
The role of annexins in tumour development and progression
.
J. Pathol.
  ,
216
,
131
140
.
48.
Duncan
R.
et al
2008
Characterisation and protein expression profiling of annexins in colorectal cancer
.
Br. J. Cancer
  ,
98
,
426
433
.
49.
Emoto
K.
et al
2001
Annexin II overexpression correlates with stromal tenascin-C overexpression: a prognostic marker in colorectal carcinoma
.
Cancer
  ,
92
,
1419
1426
.
50.
Johnson
T.V.
et al
2010
C-reactive protein as a clinically useful biomarker of metastasis of renal cell carcinoma
.
Mol. Diagn. Ther.
  ,
14
,
191
193
.
51.
Eisenhardt
S.U.
et al
2009
C-reactive protein: how conformational changes influence inflammatory properties
.
Cell Cycle
  ,
8
,
3885
3892
.
52.
Estornes
Y.
et al
2007
Differential involvement of destrin and cofilin-1 in the control of invasive properties of Isreco1 human colon cancer cells
.
Int. J. Cancer.
  ,
121
,
2162
2171
.
53.
Rho
J.H.
et al
2008
Proteomic expression analysis of surgical human colorectal cancer tissues: up-regulation of PSB7, PRDX1, and SRP9 and hypoxic adaptation in cancer
.
J. Proteome Res.
  ,
7
,
2959
2972
.
54.
Kisucka
J.
et al
2008
Peroxiredoxin1 prevents excessive endothelial activation and early atherosclerosis
.
Circ. Res.
  ,
103
,
598
605
.
55.
Volp
A.C.P.
et al
2010
Selenium antioxidant effects and its link with inflammation and metabolic syndrome
.
Rev. Nutr.
  ,
23
,
581
590
.
56.
Walston
J.
et al
2006
Serum antioxidants, inflammation, and total mortality in older women
.
Am. J. Epidemiol.
  ,
163
,
18
26
.
57.
Hertz
N.
et al
2009
Improved survival in patients with end-stage cancer treated with coenzyme Q(10) and other antioxidants: a pilot study
.
J. Int. Med. Res.
  ,
37
,
1961
1971
.