Supplemental Material can be found at: http://jn.nutrition.org/content/suppl/2005/08/01/135.8.1865.D C1.

html

Nutrient-Gene Interactions

Transcriptome Analysis of Human Colon Caco-2 Cells Exposed to Sulforaphane13

Maria Traka,* Amy V. Gasper,* Julie A. Smith, Chris J. Hawkey, Yongping Bao,* ** and Richard F. Mithen*4
*Nutrition Division, Institute of Food Research, Colney Lane, Norwich, NR4 7UA, UK; Wolfson Digestive Diseases Centre, University Hospital, Nottingham, NG7 2UH, UK; and **School of Medicine, Health and Policy, University of East Anglia, Norwich, NR4 7TJ, UK
ABSTRACT Sulforaphane (SF), a dietary phytochemical obtained from broccoli, has been implicated in several physiological processes consistent with anticarcinogenic activity, including enhanced xenobiotic metabolism, cell cycle arrest, and apoptosis. In this study, we report changes in global gene expression in Caco-2 cells exposed to physiologically appropriate concentrations of SF, through the use of replicated Affymetrix array and RT-PCR experiments. After exposure to 50 mol/L SF, 106 genes exhibited a 2-fold increase in expression and 63 genes exhibited a 2-fold decrease in expression. There were fewer changes in gene expression at lower SF concentrations. The majority of these genes had not previously been shown to be modulated by SF, suggesting novel mechanisms of possible anticarcinogenic activity, including induction of differentiation and modulation of fatty acid metabolism. The changes in the expression of 10 of these genes, together with 4 additional genes of biological interest, were further quantied in independent studies with RT-PCR. These genes include several that have recently become associated with carcinogenesis, such as Kruppel-like factor (KLF)4, a gut-enriched transcription factor associated with induction of differentiation and reduction in cellular proliferation; DNA (cytosine-5-)methyltransferase 1, associated with methylation; and -methylacyl-CoA racemase (AMACR), a marker associated with the development of colon and prostate cancer. The expression of 5 of these genes [caudal type homeo box transcription factor 2 (CDX-2), KLF4, KLF5, cyclin-dependent kinase inhibitor 1A (p21), and AMACR] was additionally studied after in vitro exposure to SF of surgically resected healthy and cancerous colon tissue from each of 3 patients. The study suggests the complex effects that SF has on gene expression and highlights several potential mechanisms by which the consumption of broccoli may reduce the risk of carcinogenesis. J. Nutr. 135: 18651872, 2005. KEY WORDS: p21

Downloaded from jn.nutrition.org by guest on December 28, 2011

colon cancer

chemoprevention

sulforaphane

microarray

KLF4

AMACR

Epidemiologic studies have provided evidence that a diet rich in cruciferous vegetables may reduce the risk of cancer at several sites (1 6). Particular to crucifers among commonly consumed vegetables and fruits is the presence of glucosinolates, sulfur-containing glycosides. These compounds can degrade during and after consumption to isothiocyanates and indole compounds, both of which were shown to have biological activity consistent with anticarcinogenic properties. Among these is the isothiocyanate
1 Supported by the Biotechnology and Biological Sciences Research Council (CSG grant to the Institute of Food Research and responsive mode grant 42D/ 18003 awarded to R.F.M.) and the University of Nottingham (M.T., A.V.G., J.A.S., and C.J.H.). 2 Microarray data cited in this study were deposited with the public repository Array Express under Experiment Accession No: E-MEXP-170 (http://www. ebi.ac.uk/arrayexpress/). 3 Supplemental Tables 1 4 and Supplemental Figure 1 are available as Online Supporting Material with the online posting of this paper at http://www.nutrition. org. 4 To whom correspondence should be addressed. E-mail: richard.mithen@bbsrc.ac.uk.

sulforaphane [SF,5 1-isothiocyanato-4-(methylsulnyl) butane; CH3OSOO(CH2)4ONACAS], which is derived primarily from broccoli and contains the corresponding glycoside, 4-methylsulnylbutyl glucosinolate (GR, glucoraphanin). Although the majority of epidemiologic stud-

5 Abbreviations used: AKR, aldo-keto reductase family 1, member C1; AMACR, -methylacyl-CoA racemase; APC, adenomatous polyposis coli; ARE, antioxidant response element; ATF3, activating transcription factor 3; CDX-2, caudal type homeo box transcription factor 2; dChip, DNA-chip analyzer; DMSO, dimethyl sulfoxide; Dnmt1, DNA (cytosine-5-)-methyltransferase 1; FAP, familial adenomatous polyposis; FDR, false discovery rate; GADD45b, growth arrest and DNA-damage-inducible, ; GR, glucoraphanin; GSH, glutathione; HO, heme oxygenase 1; Hsp27, heat shock 27kD protein 1; IC50, 50% inhibitory concentration; ITC, isothiocyanate; JNK, c-Jun NH(2)-terminal kinase; KLF, Kruppel-like factor; MCM4, minichromosome maintenance decient 4; MRP2, multidrug resistance protein 2; MTS, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4sulfophenyl)-2H-tetrazolium; MTT, methythiazolyldiphenyl-tetrazolium bromide; NQO1, NAD(P)H dehydrogenase, quinone 1 or quinone reductase (QR); Nrf2, nuclear factor (erythroid-derived 2)-like 2; p21, cyclin-dependent kinase inhibitor 1A; p57, cyclin-dependent kinase inhibitor 1C; PM, Perfect Match; pNA, pnitroaniline; SF, sulforaphane; SMAR1, BTG3-associated nuclear protein; TR1, thioredoxin reductase 1.

0022-3166/05 $8.00 2005 American Society for Nutritional Sciences. Manuscript received 1 February 2005. Initial review completed 16 March 2005. Revision accepted 13 May 2005. 1865

1866

TRAKA ET AL.

ies cannot distinguish between consumption of different cruciferous vegetables, some studies have associated a reduction in the risk of cancer specically with consumption of broccoli (3,6). After tissue disruption, glucosinolates are hydrolyzed by the plant enzyme myrosinase to yield an unstable aglycone that rearranges to either the isothiocyanate or the nitrile derivative. Cooking crucifers may denature myrosinase, resulting in ingestion of intact glucosinolates, but isothiocyanate (ITC) metabolites are still recovered in the urine (7,8), probably due to microbial activity in the large bowel. Thus, depending upon how broccoli is processed, SF may be absorbed through the stomach, small intestine, or colon. After passive diffusion into epithelial cells, SF is rapidly conjugated with glutathione (GSH), and actively transported into the blood stream. The SF-GSH conjugate is further metabolized via the mercapturic acid pathway, and SF-N-acetylcysteine conjugates are excreted in the urine. The precise nature and concentration of the conjugates that circulate in the plasma have not been fully resolved. However, the concentration of circulating SF metabolites is likely to be on the order of a few mol/L, and excretion is complete after 24 h (7,9). Epithelial cells of the gastrointestinal tract will experience both topological exposure of SF of relatively high concentrations, followed by systemic exposure to lower concentrations. Initial interest in SF was due to its potent ability to induce phase II enzymes, such as quinone reductase (QR) and UDPglucuronosyl transferase, via antioxidant response element (ARE)-mediated transcription. To what extent the induction of detoxication enzyme in vivo accounts for the anticarcinogenic activity of SF or broccoli is uncertain. Several recent studies demonstrated that SF can induce cell cycle arrest in a variety of cell types, including prostate (10 12), lymphocyte (10,13), colon (14), and mammary (15). Our own studies with Caco-2 cells conrmed these results; exposure of Caco-2 cells to 50 mol/L SF resulted in 52% of cells in G2/M phase, compared with 15% G2/M cells in an untreated control (unpublished data). In a similar manner, several studies also reported the ability of SF to induce apoptosis in a range of cell lines (11,13,14,16,17). The mechanisms underlying these physiological processes have not been fully elucidated, although the induction of the c-Jun NH(2)-terminal kinase (JNK) signal transduction pathway and activation of caspases were implicated in apoptosis (11,18,19). Although there have been several studies that reported changes in the expression of a single or a small number of genes after exposure of colon cell cultures to SF (14,16,20), we attempted to gain an overview of changes in gene expression in Caco-2 cells via the use of Affymetrix human oligonucleotide arrays, and to investigate whether changes in the expression of a small number of genes of particular interest in colon cell cultures were also observed in surgically resected colon tissue. MATERIALS AND METHODS
Cell culture and chemicals. The human Caucasian colon adenocarcinoma cell line Caco-2 was obtained from the European Collection of Animal Cell Cultures. The cells were cultured in Eagles MEM supplemented with 1% nonessential amino acids, 2 mmol/L glutamine, 1 U/L penicillin, 1 g/L streptomycin (Invitrogen), and 10% fetal calf serum, and maintained in 5% CO2 at 37C. For cell viability and ELISA assays, cells were seeded in 96-well plates. For enzyme assays and real-time RT-PCR, cells were seeded in 6-well plates. For mRNA to be hybridized onto microarrays, cells were seeded at a density of 2 104 cells/cm2 in T75 culture asks. Cells were allowed to reach 80% conuence, 5 d in culture, at the time

of SF supplementation. SF, purchased from ICN Biomedicals, was dissolved in dimethyl sulfoxide (DMSO) so that the nal concentration of DMSO (v:v) in the medium was 0.17%. Control cells received equal amounts of vehicle DMSO. Cell passages 40 51 were used. All other chemicals were purchased from Sigma-Aldrich. MTT cell viability assay. The assay is based on the conversion of the yellow methythiazolyldiphenyl-tetrazolium bromide (MTT) to purple formazan crystals by metabolically active cells and provides a quantitative determination of viable cells (21). Cells were treated with concentrations of SF ranging from 10 to 150 mol/L in 6 replicate wells for 24 h. Formazan crystals were solubilized in DMSO and absorbance of each sample was measured at 550 nm against a 720-nm reference using a microplate reader. The experiment was repeated twice and the 50% inhibitory concentration (IC50) was calculated with the CalcuSyn software. Quantitation of apoptosis by CaspACE and cell death ELISAPLUS assay. Cells were exposed to vehicle DMSO and 1, 10, 25, and 50 mol/L SF in triplicate wells for 24 h; 20 mol/L of the caspase inhibitor Z-VAD-FMK was added to one of the wells at the same time. To conrm the assay, cells were treated for 4 and 72 h with 5 mol/L of camptothecin, which induces apoptosis in Caco-2 cells under these conditions. Caspase-3 activity was assessed using the colorimetric CaspACE Assay System (Promega), which uses a substrate labeled with the chromophore p-nitroaniline (pNA) to measure protease activity. Cleavage of the substrate by caspase-3 produced yellow pNA, which was measured spectrophotometrically at 410 nm. Apoptosis was also measured by the Cell Death Detection ELISAPLUS assay (Roche Diagnostics), which uses 1-step sandwich immunoassay to detect and quantify histone-complexed DNA fragments released from cells during apoptosis. Caco-2 cells were treated with 170 mol/L SF for 24 h in triplicate. As a positive inducer, the cells were treated with 5 mol/L camptothecin for 72 h. Absorbance was measured at 410 nm using 450 nm as a reference wavelength. RNA extraction and array hybridization. In a single experiment, cells were treated with vehicle DMSO and 1, 5, 25, and 50 mol/L SF for 24 h, and total RNA from 4 biological replicates of each treatment was isolated using the QIAGEN RNeasy Mini Kit. The quality of the resulting RNA was determined using the Agilent 2100 Bioanalyzer (Agilent Technologies). RNA samples of each of the replicates were analyzed using Affymetrix Human U133A chips (Affymetrix). Double-stranded cDNA synthesis and generation of biotin-labeled cRNA were performed according to the manufacturers protocol (Affymetrix). The nal cRNA was checked for quality before fragmentation and hybridization onto a total of 20 arrays. Fluorescence intensity for each chip was captured with an Affymetrix HP GeneArray laser confocal scanner (Agilent). Affymetrix Microarray Suite version 5.0 (MAS 5.0, Affymetrix) was used to quantitate each U133A chip (22). The 20 CEL les generated containing the summary intensities for each probe were loaded into the DNA-Chip Analyzer software (dChip), version 1.3 (23), for normalization, generation of expression values, and statistical analysis. For normalization, dChip uses the Invariant Set Normalization method, which chooses a subset of Perfect Match (PM) probes with small withinsubset rank difference in the two arrays (baseline and target array), to serve as the basis for tting a normalization curve. This curve is then used to generate new normalized values for every probe on the chip. After normalization, probe expression levels can be calculated using either the PM-only model or the PM-Mismatch difference model. Previously, comparison between the 2 models showed that applying the PM-only model consistently produced less variable results (24); therefore, it was chosen for the present study. Hierarchical clustering was performed on the genes that were found to be differentially expressed after treatment with 50 mol/L SF compared with control as well as clustering of the 20 samples. Real-time RT-PCR. We selected 10 genes that had shown a 2-fold change in expression for further study by RT-PCR in a separate experiment. Two of these [theoredoxin reducatase (TR1) and NAD(P)H quinone oxidoreductase (NQO1)] had previously been shown to be affected by SF, whereas the others had been associated with cell cycle suppression and apoptosis [growth arrest and DNA-damage-inducible, (Gadd45b); Kruppel-like factor (KLF)5; KLF4; minichromosome maintenance decient 4 (MCM4);

Downloaded from jn.nutrition.org by guest on December 28, 2011

GENES ALTERED BY SULFORAPHANE IN CACO-2 CELLS

1867

cyclin-dependent kinase inhibitor 1C (p57); activating transcription factor 3 (ATF3)], fatty acid metabolism associated with carcinogenesis [ -methylacyl-CoA racemase (AMACR)], or transcriptional activation of xenobiotic metabolizing genes [nuclear factor (erythroidderived 2)-like 2 (Nrf2)]. In addition, we quantied the expression of 4 other genes, cyclin-dependent kinase inhibitor 1A (p21), caudal type homeo box transcription factor 2 (CDX-2), multidrug resistance protein 2 (MRP2), and DNA (cytosine-5-)-methyltransferase 1 (Dnmt1). These had not met the 2-fold cutoff criterion in the array studies, but had previously been shown to be induced by SF (p21 and MRP2), were associated with KLF4 (CDX-2), or were of particular interest (Dnmt1). Cells were treated with vehicle DMSO and 1, 5, 25, and 50 mol/L SF for 24 h; total RNA from 3 biological replicates of each treatment was isolated using the QIAGEN RNeasy Mini Kit. Target mRNA was quantied using an ABI PRISM 7700 Sequence Detection System (Applied Biosystems). Primers and probes (Supplemental Table 1) were designed using ABI PRISM Primer Express software v.1.5 (Applied Biosystems). Primers were purchased from Sigma-Genosys and probes from Applied Biosystems. The probes were labeled with a 5 reporter dye (FAM, 6-carboxyuorescein) and a 3 quencher dye (TAMRA, 6-carboxytetramethylrhodamine). RTPCR reactions were carried out in a microamp optical 96-well plate in a total volume of 25 L/well, consisting of TaqMan 1-step RT-PCR master mix Reagent kit (Applied Biosystems), 20 ng total RNA, 0.25 U/ L MultiscribeTM, and concentrations of primers and probes ranging from 100 to 500 nmol/L. Reverse transcription was performed for 30 min at 48C, AmpliTaq gold activation for 10 min at 95C, followed by 40 PCR cycles of denaturation at 95C for 15 s and annealing/extension at 60C for 1 min. Reactions were carried out in triplicate and data were analyzed by the ABI PRISM 7700 Sequence Detection System software using a standard curve to quantify the mRNA amount. Standard curves were produced for each set of primers and probes using 5, 10, 20, 40, and 80 ng total RNA/reaction. Results were normalized on the basis of 18S rRNA quantication. Ex vivo studies with resected colon tissue. Human colonic tissue was obtained from surgical resections from 3 individuals. Ethical permission was granted by the Queens Medical Centre, Nottingham University Hospital NHS Trust, and written, informed consent received from the patients. Tissue was deemed to be normal if it

appeared healthy and was 5 cm from the tumor site. Tissue was washed in 1 X HBSS, then scraped and rinsed to remove mucus and blood. The healthy mucosa was stripped from the lamina propria, cut into pieces (2 mm3) and placed into 2-mL prewarmed RPMI1640 medium. Viability of tissue after a 2-h 37C exposure to 0, 25, and 50 mol/L SF was assessed with the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega), which measures the reduction of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2(4-sulfophenyl)-2H-tetrazolium (MTS) by metabolically active cells into formazan product, measured at 492 nm in a microplate reader. For studies on gene expression, healthy and cancerous tissues were incubated at 37C for 2 h in 25 mol/L SF, after which the culture medium was carefully aspirated and replaced with 1 mL RNA later (Ambion). RNA was extracted using the QIAGEN RNeasy Mini Kit and real-time RT-PCR was performed as described above. Statistical analysis. Data were expressed as means SEM or as means of the fold change SEM. For microarray analysis with dChip, 3 criteria were applied to detect differentially expressed genes: 1) a cutoff of 2.0-fold change, 2) absolute difference between the 2 group means 100, and 3) P 0.05 for Welch modied 2-sample t test, adjusted to compensate for multiple testing using false discovery rate (FDR). This analysis was subsequently repeated in R (25). Five genes present in the list generated by dChip were not conrmed by R and are marked accordingly in Supplemental Table 2. Genes showing signicant differences in expression were classied into different functional categories, based on Gene Ontology with modications (26,27). For subsequent analyses of selected genes by RT-PCR, data were analyzed by Students t tests using the MINITAB statistical software adjusted to compensate for multiple testing using FDR. Differences were considered signicant at P 0.05. Additionally, the expressions of p21, MRP2, CDX2, and Dnmt1 from microarray data in Table 1 were analyzed in the same manner as the RT-PCR data because the fold change of these 4 genes was 2.

Downloaded from jn.nutrition.org by guest on December 28, 2011

RESULTS SF is cytotoxic to Caco-2 cells, but does not induce apoptosis at 24 h. SF reduced viability of Caco-2 cells in a

TABLE 1
Comparison of changes in gene expression detected by Affymetrix array analysis and real-time RT-PCR after treatment with 50 mol/L SF1
Fold change Gene name Thioredoxin reductase 1 Growth arrest and DNA-damage-inducible, NAD(P)H dehydrogenase, quinone 1 Kruppel-like factor 5 (intestinal) Kruppel-like factor 4 (gut) Nuclear factor (erythroid-derived 2)-like 2 Activating transcription factor 3 Cyclin-dependent kinase inhibitor 1A (p21, Cip1) ATP-binding cassette, subfamily C (CFTR/MRP), member 2 Caudal type homeo box transcription factor 2 DNA (cytosine-5-)-methyltransferase 1 Minichromosome maintenance decient 4 (Saccharomyces cerevisiae) Cyclin-dependent kinase inhibitor 1C (p57, Kip2) -Methylacyl-CoA racemase
1

Accession no. NM_003330 NM_015675 AI039874 AF132818 BF514079 NM_006164 NM_001674 NM_000389 NM_000392 U51096 NM_001379 AI859865 NM_000076 AA888589

Affymetrix2 3.26 2.68 2.38 2.2 2.16 2.12 2.1 1.45 1.23 1.26 1.8 2.05 2.39 2.43 0.33 0.16 0.11 0.12 0.07 0.06 0.08 0.074 0.054 0.024 0.084 0.05 0.11 0.02

RT-PCR3,4 8.75 120.38 2.49 1.85 5.07 1.21 15.49 3.68 3.84 5.88 1.8 2.5 1.54 2.12 0.38 3.94 0.18 0.15* 0.43 0.05 0.23 0.29 0.06 0.27 0.18 0.11 0.18 0.07

TR1 Gadd45b NQO1 KLF5 KLF4 Nrf2 ATF3 p21 MRP2 CDX-2 Dnmt1 MCM4 p57 AMACR

Values are means SEM, n 4 (Affymetrix) or n 3 (RT-PCR), estimated in independent experiments; P 0.05 unless noted. * Not signicant, 0.056. Positive values represent upregulation; negative values represent downregulation. Values are mean gene expression ratios calculated using the PM-only model on normalized arrays in dChip between Caco-2 cells treated with SF and cells treated with equal amounts of vehicle DMSO. 3 Values are mean gene expression ratios calculated by real-time RT-PCR between Caco-2 cells treated with SF and cells treated with equal amounts of vehicle DMSO. 4 Data analyzed by Students t tests.
2

1868

TRAKA ET AL.

dose-dependent manner. The IC50 was calculated to be 85 and 83 mol/L in 2 independent experiments (data not shown). A concentration of 50 mol/L reduced cell viability by 26% and was therefore chosen as the maximum concentration for further experiments. To determine whether SF induces apoptosis in Caco-2 cells, 2 independent methods of detection were employed on cells treated with concentrations up to 50 mol/L SF. Caspase-3 activity was not induced in cells treated with SF for 24 h compared with untreated cells. In contrast, Caco-2 cells incubated with camptothecin exhibited apoptosis as early as 4 h after treatment, which was abolished by the addition of the Z-VAD-FMK caspase inhibitor at the time of the treatment (data not shown). Similarly, induction of apoptosis was not detected by immunoassay after SF treatment (data not shown). Transcriptome analysis of gene expression after exposure to SF. Affymetrix U133A oligonucleotide arrays, which contain 22,283 probes corresponding to the best-annotated human genes, were used to detect changes in gene expression in response to treatment with SF and identify potential targets that might mediate the SF effects. When exposed to 50 mol/L SF, 106 genes exhibited a 2-fold increase in expression, whereas 63 genes exhibited a 2-fold decrease in expression (P 0.05; Supplemental Table 2). The majority of these genes had not been shown previously to be modulated by SF, and many represent potentially important targets that may help to explain the anticarcinogenic activity of SF. When exposed to 25 and 5 mol/L, the numbers of genes for which there was a 2-fold change in expression was 14 and 4, respectively (Supplemental Tables 3 and 4). The expression of these genes was also changed at 50 mol/L. No genes exhibited a 2-fold change in expression when cells were exposed to 1 mol/L. Analysis of the subset of the 169 genes that showed a 2-fold up or down change by 50 mol/L SF by hierarchical clustering revealed that all of the samples belonging to the control and 1 and 5 mol/L treatment groups appeared to cluster together with a similar expression pattern (Fig. 1), conrming that these concentrations of SF have relatively little effect on gene expression patterns as detected by Affymetrix arrays within the criteria used in this study. In contrast, there were 2 distinct clusters of samples, i.e., those treated with 25 and 50 mol/L SF, respectively. The extent of changes in the expression of the majority of genes within the 25 mol/L cluster was intermediate between those in the 50 mol/L cluster and the [control low concentration] cluster, indicating relatively little variation among the 4 replicates, and consistent and reproducible dose-dependent changes in expression on exposure to SF with the majority of the 169 genes.
FIGURE 1 Hierarchical clustering dendrogram of Caco-2 cells treated with SF (0 50 mol/L) with 169 genes that differed signicantly between control treatment (vehicle DMSO) and 50 mol/L SF treatment (P 0.05). Horizontal rows represent each individual array, specied by the concentration of SF (0, 1, 5, 25, or 50) followed by replicate number (1 4). Unsupervised hierarchical clustering of cell samples and genes was performed using the dChip software. Red color represents expression level above mean expression of a gene across all samples; black represents mean expression, and green represents expression lower than the mean.

Changes in the expression of 10 of these genes plus 4 additional genes (p21, CDX-2, MRP2, and Dnmt1) were analyzed by RT-PCR in independent experiments (Table 1). In each case, RT-PCR conrmed either a signicant increase or decrease in expression (although the precise magnitude of this change may be different), with 2 exceptions. The rst exception was the transcription factor Nrf2, for which the induction observed with the use of arrays was not conrmed by RT-PCR and supports previous ndings. The second exception was the intestinal KLF5, for which expression was indicated to be upregulated when analyzed by arrays, but reduced when analyzed by RT-PCR. Subsequent studies in independent experiments conrmed the RT-PCR result. In addition, because some studies had reported the induction of p21 by SF, the expression of p21 was investigated in independent experiments by RT-PCR; exposure to 50 mol/L SF resulted in a 3.5-fold increase in expression, as opposed to the 1.45-fold detected by arrays (Table 1). The pattern of expression identied by microarrays across all of the samples for the genes that were also veried by RT-PCR is shown in Supplemental Figure 1. Similar hierarchical sample clusters were formed as before, indicating dose-dependent expression patterns. SF-mediated changes in gene expression in surgically resected tissue. The MTS tissue viability assay showed that there was a 57.9 1.5% reduction in viability in tissue exposed to 50 mol/L SF and a 29.1 1.5% reduction when tissue was exposed to 25 mol/L SF for 2 h, compared with control samples (data not shown). This latter reduction in viability was similar to the 26% reduction in viability that was observed in the MTT assay when Caco-2 cells were exposed to 50 mol/L SF. Thus, in subsequent experiments, tissue was exposed to 25 mol/L SF. In addition, tissue was exposed to SF for only 2 h, which is more physiologically appropriate than the 24 h for which cells were exposed to SF. In each of these 3 individuals, the level of CDX-2, KLF4, p21, and KLF5 expression in untreated tissue was lower in cancerous tissue than healthy tissue (Fig. 2b d). SF had no or little effect on gene expression in healthy tissue from each of the 3 patients. Cancerous tissue from 1 of the 3 patients, patient 1, responded in a manner similar to Caco-2 cells, with an induction in both KLF4 and p21 after exposure to SF, but with little change in expression of KLF5 (Fig. 2a and b). Expression of p21 was restored to the level observed in healthy tissue. There was also an increase in the expression of p21 in cancerous tissue from patient 2, but relatively little change in expression of CDX-2, KLF4, and KLF5 (Fig. 2c). There was little change in gene expression in the tissues of patient 3 (Fig. 2d). Although expression of AMACR was reduced in Caco-2 cells by 2fold, there was little change when cancerous tissue was exposed to SF in all 3 patients (Fig. 2).

Downloaded from jn.nutrition.org by guest on December 28, 2011

GENES ALTERED BY SULFORAPHANE IN CACO-2 CELLS

1869

FIGURE 2 Expression levels of CDX-2, KLF4, p21, KLF5, and AMACR, quantied by real-time RT-PCR and normalized to 18s, after treatment of Caco-2 cells with 50 mol/L SF for 24 h (a), and after exposure of human colon healthy and cancerous resected tissue from 3 individuals to 25 mol/L SF for 2 h (b d). Bars for (a) indicate means SEM, n 3. *P 0.04, **P 0.01. Bars for (b)(d) indicate a single value of relative quantity for the genes measured. Downloaded from jn.nutrition.org by guest on December 28, 2011

DISCUSSION Exposure of cells of the gastrointestinal tract to SF results both from topological exposure after consumption of broccoli and subsequent systemic exposure via the blood stream. Broccoli typically contains 1 mol GR/g fresh weight. Thus, if one assumes that myrosinase remains active after mild cooking and 100% of GR is converted to SF, a standard 100-g portion would deliver 100 mol of SF to the GI tract. The concentration that tissue would be exposed to would depend on subsequent dilution, but local exposure is likely to exceed 50 mol/L. Alternatively, if a longer period of cooking has denatured myrosinase, intact GR is likely to pass through to the lower tract, where it is converted to SF due to microbial action. Although this may result in lower total conversion of GR into SF, local concentrations may exceed 100 mol/L. Systemic exposure postabsorption is likely to be considerably lower because SF will be rapidly diluted in the bloodstream, and is further complicated by the formation of GSH-SF conjugates and derivatives that both alter biological activity and are likely to prevent entry of SF into cells. Thus, the range of concentrations of SF that cells were exposed to was chosen to represent the probable concentrations that cells may be exposed to from either systemic exposure (1 and 5 mol/L) or topological exposure (25 and 50 mol/L). These levels of exposure are similar to those reported in several recent papers on a variety of cell lines (11,14,16,28). The majority of changes in gene expression were observed only at the higher SF concentrations. These would be found within the GI tract via topological exposure, and also within the bladder where SF and its conjugates accumulate before excretion. After exposure to SF, the IC50 was 85 and 83 mol/L in 2 independent experiments. These values were higher than reported previously for Caco-2 cells (55 mol/L) (16). However, in that earlier study, Caco-2 cells were seeded and left to adhere overnight before treatment, increasing the SF molecules available per cell and thus making them more sensitive

to xenobiotics, whereas in this study, Caco-2 cells were seeded and left to grow until they became 80% conuent, 5 d later, before treatment with SF. To investigate changes in gene expression in Caco-2 cells exposed to SF, we undertook a study of global gene expression with the use of Affymetrix arrays, and then conrmed the changes in expression of a subset of genes in independent experiments with RT-PCR. In general, the magnitude of the change in expression of this subset of genes was less when quantied with arrays than with RT-PCR (Table 1), consistent with previous studies (29). As expected, there was upregulation of genes involved in xenobiotic metabolism and changes in the expression of genes involved with cell cycle control and DNA synthesis, discussed in detail below. Of particular interest were changes in the expression of KLF4, which is a transcription factor regulating cell cycle and differentiation, and p21 (Table 1, Fig. 2a). There were also changes in the expression of several other genes that may be of potential importance in chemopreventative mechanisms. These include induction of tissue inhibitor of metalloproteinase 3 (2.5-fold), which codes for a protein that inhibits matrix metalloproteinases and is a potent inhibitor of angiogenesis (30), nonsteroidal anti-inammatory drugactivated gene (2.7-fold), which codes for a proapoptotic and antitumorigenic protein and is also induced by other dietary phytochemicals (31,32), and downregulation of formyltetrahydrofolate synthase ( 2.1-fold) and Dnmt1 ( 1.8-fold), both of which may have consequences for maintenance of methylation status. These genes, and several others in Supplemental Table 2, warrant further studies. SF was reported to induce apoptosis after 24 h in a variety of mammalian cells at concentrations similar to those used in this study (10,11,13,14,16,28,33,34). A previous report of SFinduced apoptosis in Caco-2 cells was based on visual inspection of nuclei after DAPI (4 ,6-diamidino-2-phenylindole) staining, but neither the extent of apoptosis nor the concen-

1870

TRAKA ET AL.

tration of SF required to induce apoptosis was reported (16). In contrast, we have no evidence of induction of apoptosis at 24 h, based on both caspase activity and the Cell Death Detection ELISAPLUS assay. Similarly, although previous studies showed that SF can induce caspase-3 activity (15,28,34), we have no evidence of any such activity. Several studies provided evidence for the activation of the JNK signal transduction pathway by SF and other ITCs (18,19,3537) to be the mechanism of caspase-3 induction and subsequent apoptosis. From our gene array analyses, we identied 2 potential mechanisms that may interfere with JNK/caspase-3 induced apoptosis. First, the expression of GADD45b was shown to lead to downregulation of JNK signaling induced by tumor necrosis factor (38). Second, a more likely potential mechanism of suppression of apoptosis post-JNK activation is through induction of ATF3, which induces the antiapoptotic factor heat shock 27kD protein 1 (Hsp27) (39). Both ATF3 and Hsp27 were signicantly upregulated in our study (Table 1 and Supplemental Table 2). Hsp27 was shown to inhibit apoptosis and caspase activation through negative regulation of the mitochondrial proteins cytochrome c and Smac, and other means (40,41). This mechanism would be consistent with known mechanisms of apoptosis induced by SF and other ITCs (14,19). In addition to these 2 potential antiapoptotic mechanisms, induction of heme oxygenase and p21, as observed in our current study, was also associated with inhibition of apoptosis (42). As expected, there was upregulation of several genes associated with ARE-mediated transcription, notably NQO1, TR1, aldo-ketoreductase (AKR), and heme oxygenase-1 (HO). NQO1, TR1, and AKR were reported to be induced by SF, in both cell cultures and mammalian tissue (16,43 46). These genes and HO are likely to be upregulated via the activation of the transcription factor Nrf2 following its release from the Kelch-like ECH-associated protein 1 (Keap1) after exposure to SF has altered the cell redox potential (47 49). Sulforaphane upregulated the expression of several genes of known importance in suppression of the cell cycle. In particular, there was signicant upregulation of the cyclin-dependent kinase inhibitor p21, previously reported to have been upregulated in SF-treated prostate cell lines and HT29 colon cells (12,20). Second, there was upregulation of GADD45b, which was shown to interact with p21 in suppressing cell proliferation (50). SF also enhanced the expression of BTG family, member 2 (PC3TIS21/BTG2), BTG3-associated nuclear protein (SMAR1), and CDC28 protein kinase 2 (CKSHS2), all of which have been associated with regulation of the cell cycle (5153). SMAR1 has also been associated with delay in tumor growth (51). Consistent with inhibition of cell cycle, there was a decrease in expression of members of the minichromosome maintenance family (MCM4 and MCM7), all of which are associated with DNA synthesis (54). To understand how these genes themselves are regulated, it was of interest to observe the induction of the KLF4 transcription factor, in contrast to the reduction in expression of KLF5. KLF4 is expressed in the postmitotic cells lining the villi in the small intestine and in the middle-to-upper gland region in the colon (55,56); it is associated with the region of cell differentiation, whereas KLF5 is expressed in the proliferative cell compartment of the adult gut epithelium (57). Consistent with these respective activities, in Caco-2 cells, we observed an induction of KLF4 with SF, but a reduction in KLF5 (Table 1, Fig. 2a). KLF4 has recently received considerable attention due to its probable role in regulating several processes associated with cell cycle regulation and differentiation (58). Transcriptome proling of a cell culture line with inducible KLF4

conrmed that expression of KLF4 regulates several genes concerned with inhibition of cell cycle progression and DNA synthesis, including p21 and MCM4 (59). Notably, the inhibition of cell proliferation of cancer cell lines due to expression of KLF4 is not associated with apoptosis (60). KLF4 was reported to be itself regulated via expression of CDX-2 and APC (61,62), as illustrated in Figure 3. Consistent with its potential bioactivity, there is reduced expression of KLF4 in the intestinal tissue of adenomatous polyposis coli (APC)min mice (63), and also within intestinal adenomas of patients with familial adenomatous polyposis (FAP) in whom there is a mutation in APC. Although SF enhanced KLF4 expression, it reduced expression of CDX-2 by 5.9-fold (Table 1, Fig. 2a), suggesting that SF is mediating the expression of KLF4 independently of CDX-2 and APC. Thus, SF may be able to restore KLF4 expression (and associated downstream genes such as p21) within cells even when the APC-CDX-2 transcriptional pathway is impaired. Because both alleles of APC are mutated in 85% of familial and sporadic colorectal tumors (64), the ability of SF to at least partially restore KLF4 activity may be important. Furthermore, the association between SF and KLF4 and p21 induction links, for the rst time, the activity of this dietary anticarcinogen with one of the major pathways of intestinal carcinogenesis, that of APC downregulation or mutation (Fig. 3). The lower levels of p21 and KLF4 expression in cancerous compared with adjacent healthy surgically resected colon tissue is consistent with an increase in cellular proliferation and with previous studies that reported a reduction in KLF4 expression within intestinal adenomas of FAP patients (65). Although some caution is required in interpretation of these results because gene expression in morphologically normal tissue from cancer patients may have perturbations in gene expression compared with normal subjects (66), the extent of these perturbations is still likely to be less than within proliferating tissue itself. Exposure of tissue to SF supported the results from cell culture, i.e., there was an induction of KLF4 and/or p21, without an induction of CDX-2, in at least some of the individuals. However, these data suggested that p21 could be induced independently of KLF4. Although these

Downloaded from jn.nutrition.org by guest on December 28, 2011

FIGURE 3 Hypothetical schematic representation of major chemoprotective pathways probably regulated by SF. Figures and arrows indicate changes in gene expression (bold by RT-PCR, normal by array analyses) after SF exposure.

GENES ALTERED BY SULFORAPHANE IN CACO-2 CELLS

1871

preliminary ex vivo studies require repeating and expanding, they point toward the potential of SF to perturb expression of KLF4 and p21 in cancerous tissue, indicating that SF may be important not only in preventing initiation of carcinogenesis, but also in reducing cellular proliferation. Additionally, SF altered expression of the AMACR, a marker upregulated in prostate carcinoma but also recently identied as being highly expressed in colon adenomas and carcinomas compared with normal colon and nonneoplastic polyps (6770). There was a signicant decrease in AMACR expression after exposure of Caco-2 cells to SF. In contrast, SF did not seem to consistently alter the expression in the individuals tested in the present study. However, upregulation in cancerous tissue compared with normal tissue was found in only 1 individual, suggesting that our sample is not representative of colon cancer with high AMACR. This article is the rst report of transcriptome analysis of a cell line exposed to SF; we identied 169 genes that are highly likely to be altered in expression after exposure to SF, and conrmed the changes in expression of a subset of these with RT-PCR. Several of these genes, such as KLF4, CDX-2, Dnmt1, and AMACR, which are involved in diverse biological processes that have been implicated in carcinogenesis, warrant further study. ACKNOWLEDGMENTS
We thank Jim Bacon for assistance with RT-PCR and Dr. Robert Foxall for assistance with statistical analyses.

LITERATURE CITED
1. Ambrosone, C. B., McCann, S. E., Freudenheim, J. L., Marshall, J. R., Zhang, Y. & Shields, P. G. (2004) Breast cancer risk in premenopausal women is inversely associated with consumption of broccoli, a source of isothiocyanates, but is not modied by GST genotype. J. Nutr. 134: 1134 1138. 2. Fowke, J. H., Chung, F. L., Jin, F., Qi, D., Cai, Q., Conaway, C., Cheng, J. R., Shu, X. O., Gao, Y. T. & Zheng, W. (2003) Urinary isothiocyanate levels, brassica, and human breast cancer. Cancer Res. 63: 3980 3986. 3. Joseph, M. A., Moysich, K. B., Freudenheim, J. L., Shields, P. G., Bowman, E. D., Zhang, Y., Marshall, J. R. & Ambrosone, C. B. (2004) Cruciferous vegetables, genetic polymorphisms in glutathione S-transferases M1 and T1, and prostate cancer risk. Nutr. Cancer 50: 206 213. 4. London, S. J., Yuan, J. M., Chung, F. L., Gao, Y. T., Coetzee, G. A., Ross, R. K. & Yu, M. C. (2000) Isothiocyanates, glutathione S-transferase M1 and T1 polymorphisms, and lung-cancer risk: a prospective study of men in Shanghai, China. Lancet 356: 724 729. 5. Seow, A., Yuan, J. M., Sun, C. L., Van den Berg, D., Lee, H. P. & Yu, M. C. (2002) Dietary isothiocyanates, glutathione S-transferase polymorphisms and colorectal cancer risk in the Singapore Chinese Health Study. Carcinogenesis 23: 20552061. 6. Lin, H. J., Probst-Hensch, N. M., Louie, A. D., Kau, I. H., Witte, J. S., Ingles, S. A., Frankl, H. D., Lee, E. R. & Haile, R. W. (1998) Glutathione transferase null genotype, broccoli, and lower prevalence of colorectal adenomas. Cancer Epidemiol. Biomark. Prev. 7: 647 652. 7. Conaway, C. C., Getahun, S. M., Liebes, L. L., Pusateri, D. J., Topham, D. K., Botero-Omary, M. & Chung, F. L. (2000) Disposition of glucosinolates and sulforaphane in humans after ingestion of steamed and fresh broccoli. Nutr. Cancer 38: 168 178. 8. Getahun, S. M. & Chung, F. L. (1999) Conversion of glucosinolates to isothiocyanates in humans after ingestion of cooked watercress. Cancer Epidemiol. Biomark. Prev. 8: 447 451. 9. Ye, L., Dinkova-Kostova, A. T., Wade, K. L., Zhang, Y., Shapiro, T. A. & Talalay, P. (2002) Quantitative determination of dithiocarbamates in human plasma, serum, erythrocytes and urine: pharmacokinetics of broccoli sprout isothiocyanates in humans. Clin. Chim. Acta 316: 4353. 10. Fimognari, C., Nusse, M., Berti, F., Iori, R., Cantelli-Forti, G. & Hrelia, P. (2002) Cyclin D3 and p53 mediate sulforaphane-induced cell cycle delay and apoptosis in non-transformed human T lymphocytes. Cell Mol. Life Sci. 59: 2004 2012. 11. Chiao, J. W., Chung, F. L., Kancherla, R., Ahmed, T., Mittelman, A. & Conaway, C. C. (2002) Sulforaphane and its metabolite mediate growth arrest and apoptosis in human prostate cancer cells. Int. J. Oncol. 20: 631 636. 12. Wang, L., Liu, D., Ahmed, T., Chung, F. L., Conaway, C. & Chiao, J. W. (2004) Targeting cell cycle machinery as a molecular mechanism of sulforaphane in prostate cancer prevention. Int. J. Oncol. 24: 187192. 13. Fimognari, C., Nusse, M., Cesari, R., Iori, R., Cantelli-Forti, G. & Hrelia, P.

(2002) Growth inhibition, cell-cycle arrest and apoptosis in human T- cell leukemia by the isothiocyanate sulforaphane. Carcinogenesis 23: 581586. 14. Gamet-Payrastre, L., Li, P., Lumeau, S., Cassar, G., Dupont, M. A., Chevolleau, S., Gasc, N., Tulliez, J. & Terce, F. (2000) Sulforaphane, a naturally occurring isothiocyanate, induces cell cycle arrest and apoptosis in HT29 human colon cancer cells. Cancer Res. 60: 1426 1433. 15. Jackson, S. J. & Singletary, K. W. (2004) Sulforaphane: a naturally occurring mammary carcinoma mitotic inhibitor, which disrupts tubulin polymerization. Carcinogenesis 25: 219 227. 16. Bonnesen, C., Eggleston, I. M. & Hayes, J. D. (2001) Dietary indoles and isothiocyanates that are generated from cruciferous vegetables can both stimulate apoptosis and confer protection against DNA damage in human colon cell lines. Cancer Res. 61: 6120 6130. 17. Kim, B. R., Hu, R., Keum, Y. S., Hebbar, V., Shen, G., Nair, S. S. & Kong, A. N. (2003) Effects of glutathione on antioxidant response element-mediated gene expression and apoptosis elicited by sulforaphane. Cancer Res. 63: 7520 7525. 18. Chen, Y. R., Wang, W., Kong, A. N. & Tan, T. H. (1998) Molecular mechanisms of c-Jun N-terminal kinase-mediated apoptosis induced by anticarcinogenic isothiocyanates. J. Biol. Chem. 273: 1769 1775. 19. Hu, R., Kim, B. R., Chen, C., Hebbar, V. & Kong, A. N. (2003) The roles of JNK and apoptotic signaling pathways in PEITC-mediated responses in human HT-29 colon adenocarcinoma cells. Carcinogenesis 24: 13611367. 20. Parnaud, G., Li, P., Cassar, G., Rouimi, P., Tulliez, J., Combaret, L. & Gamet-Payrastre, L. (2004) Mechanism of sulforaphane-induced cell cycle arrest and apoptosis in human colon cancer cells. Nutr. Cancer 48: 198 206. 21. Vistica, D. T., Skehan, P., Scudiero, D., Monks, A., Pittman, A. & Boyd, M. R. (1991) Tetrazolium-based assays for cellular viability: a critical examination of selected parameters affecting formazan production. Cancer Res. 51: 25152520. 22. Affymetrix, I. O. (2004) Statistical Algorithms Description Document. http://www.affymetrix.com/support/downloads/manuals/data_analysis_fundamentals_manual.pdf [accessed Jan. 25, 2004]. 23. Li, C. & Wong, W. H. (2001) Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc. Natl. Acad. Sci. U.S.A. 98: 3136. 24. Han, E. S., Wu, Y., McCarter, R., Nelson, J. F., Richardson, A. & Hilsenbeck, S. G. (2004) Reproducibility, sources of variability, pooling, and sample size: important considerations for the design of high-density oligonucleotide array experiments. J. Gerontol. A Biol. Sci. Med. Sci. 59: 306 315. 25. R Development Core Team (2004) R: a language and environment for statistical computing. http://www.R-project.org. [accessed Dec 10, 2004]. 26. Ashburner, M., Ball, C. A., Blake, J. A., Botstein, D., Butler, H., Cherry, J. M., Davis, A. P., Dolinski, K., Dwight, S. S. et al. (2000) Gene ontology: tool for the unication of biology. The Gene Ontology Consortium. Nat. Genet. 25: 2529. 27. The Gene Ontology http://www.geneontology.org [accessed Dec 10, 2004]. 28. Gingras, D., Gendron, M., Boivin, D., Moghrabi, A., Theoret, Y. & Beliveau, R. (2004) Induction of medulloblastoma cell apoptosis by sulforaphane, a dietary anticarcinogen from Brassica vegetables. Cancer Lett. 203: 35 43. 29. Yuen, T., Wurmbach, E., Pfeffer, R. L., Ebersole, B. J. & Sealfon, S. C. (2002) Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays. Nucleic Acids Res. 30: e48. 30. Qi, J. H., Ebrahem, Q., Moore, N., Murphy, G., Claesson-Welsh, L., Bond, M., Baker, A. & Anand-Apte, B. (2003) A novel function for tissue inhibitor of metalloproteinases-3 (TIMP3): inhibition of angiogenesis by blockage of VEGF binding to VEGF receptor-2. Nat. Med. 9: 407 415. 31. Baek, S. J., Kim, J. S., Jackson, F. R., Eling, T. E., McEntee, M. F. & Lee, S. H. (2004) Epicatechin gallate-induced expression of NAG-1 is associated with growth inhibition and apoptosis in colon cancer cells. Carcinogenesis 25: 24252432. 32. Wilson, L. C., Baek, S. J., Call, A. & Eling, T. E. (2003) Nonsteroidal anti-inammatory drug-activated gene (NAG-1) is induced by genistein through the expression of p53 in colorectal cancer cells. Int. J. Cancer 105: 747753. 33. Misiewicz, I., Skupinska, K. & Kasprzycka-Guttman, T. (2003) Sulforaphane and 2-oxohexyl isothiocyanate induce cell growth arrest and apoptosis in L-1210 leukemia and ME-18 melanoma cells. Oncol. Rep. 10: 20452050. 34. Singh, A. V., Xiao, D., Lew, K. L., Dhir, R. & Singh, S. V. (2004) Sulforaphane induces caspase-mediated apoptosis in cultured PC-3 human prostate cancer cells and retards growth of PC-3 xenografts in vivo. Carcinogenesis 25: 8390. 35. Chen, Y. R., Han, J., Kori, R., Kong, A. N. & Tan, T. H. (2002) Phenylethyl isothiocyanate induces apoptotic signaling via suppressing phosphatase activity against c-Jun N-terminal kinase. J. Biol. Chem. 277: 39334 39342. 36. Xu, K. & Thornalley, P. J. (2001) Signal transduction activated by the cancer chemopreventive isothiocyanates: cleavage of BID protein, tyrosine phosphorylation and activation of JNK. Br. J. Cancer 84: 670 673. 37. Yu, R., Jiao, J. J., Duh, J. L., Tan, T. H. & Kong, A. N. (1996) Phenethyl isothiocyanate, a natural chemopreventive agent, activates c-Jun N-terminal kinase 1. Cancer Res. 56: 2954 2959. 38. De Smaele, E., Zazzeroni, F., Papa, S., Nguyen, D. U., Jin, R., Jones, J., Cong, R. & Franzoso, G. (2001) Induction of gadd45beta by NF-kappaB downregulates pro-apoptotic JNK signalling. Nature (Lond.) 414: 308 313. 39. Nakagomi, S., Suzuki, Y., Namikawa, K., Kiryu-Seo, S. & Kiyama, H. (2003) Expression of the activating transcription factor 3 prevents c-Jun N-

Downloaded from jn.nutrition.org by guest on December 28, 2011

1872

TRAKA ET AL.
56. Shields, J. M., Christy, R. J. & Yang, V. W. (1996) Identication and characterization of a gene encoding a gut-enriched Kruppel-like factor expressed during growth arrest. J. Biol. Chem. 271: 20009 20017. 57. Conkright, M. D., Wani, M. A., Anderson, K. P. & Lingrel, J. B. (1999) A gene encoding an intestinal-enriched member of the Kruppel-like factor family expressed in intestinal epithelial cells. Nucleic Acids Res. 27: 12631270. 58. Chen, X., Johns, D. C., Geiman, D. E., Marban, E., Dang, D. T., Hamlin, G., Sun, R. & Yang, V. W. (2001) Kruppel-like factor 4 (gut-enriched Kruppel-like factor) inhibits cell proliferation by blocking G1/S progression of the cell cycle. J. Biol. Chem. 276: 3042330428. 59. Chen, X., Whitney, E. M., Gao, S. Y. & Yang, V. W. (2003) Transcriptional proling of Kruppel-like factor 4 reveals a function in cell cycle regulation and epithelial differentiation. J. Mol. Biol. 326: 665 677. 60. Dang, D. T., Chen, X., Feng, J., Torbenson, M., Dang, L. H. & Yang, V. W. (2003) Overexpression of Kruppel-like factor 4 in the human colon cancer cell line RKO leads to reduced tumorigenecity. Oncogene 22: 3424 3430. 61. Dang, D. T., Mahatan, C. S., Dang, L. H., Agboola, I. A. & Yang, V. W. (2001) Expression of the gut-enriched Kruppel-like factor (Kruppel-like factor 4) gene in the human colon cancer cell line RKO is dependent on CDX2. Oncogene 20: 4884 4890. 62. Stone, C. D., Chen, Z. Y. & Tseng, C. C. (2002) Gut-enriched Kruppellike factor regulates colonic cell growth through APC/beta-catenin pathway. FEBS Lett. 530: 147152. 63. Ton-That, H., Kaestner, K. H., Shields, J. M., Mahatanankoon, C. S. & Yang, V. W. (1997) Expression of the gut-enriched Kruppel-like factor gene during development and intestinal tumorigenesis. FEBS Lett. 419: 239 243. 64. Ichii, S., Horii, A., Nakatsuru, S., Furuyama, J., Utsunomiya, J. & Nakamura, Y. (1992) Inactivation of both APC alleles in an early stage of colon adenomas in a patient with familial adenomatous polyposis (FAP). Hum. Mol. Genet. 1: 387390. 65. Dang, D. T., Bachman, K. E., Mahatan, C. S., Dang, L. H., Giardiello, F. M. & Yang, V. W. (2000) Decreased expression of the gut-enriched Kruppel-like factor gene in intestinal adenomas of multiple intestinal neoplasia mice and in colonic adenomas of familial adenomatous polyposis patients. FEBS Lett. 476: 203207. 66. Chen, L. C., Hao, C. Y., Chiu, Y. S., Wong, P., Melnick, J. S., Brotman, M., Moretto, J., Mendes, F., Smith, A. P. et al. (2004) Alteration of gene expression in normal-appearing colon mucosa of APC(min) mice and human cancer patients. Cancer Res. 64: 3694 3700. 67. Jiang, Z., Wu, C. L., Woda, B. A., Iczkowski, K. A., Chu, P. G., Tretiakova, M. S., Young, R. H., Weiss, L. M., Blute, R. D., Jr. et al. (2004) AlphaMethylacyl-CoA racemase: a multi-institutional study of a new prostate cancer marker. Histopathology 45: 218 225. 68. Wu, C. L., Yang, X. J., Tretiakova, M., Patton, K. T., Halpern, E. F., Woda, B. A., Young, R. H. & Jiang, Z. (2004) Analysis of alpha-methylacyl-CoA racemase (P504S) expression in high-grade prostatic intraepithelial neoplasia. Hum. Pathol. 35: 1008 1013. 69. Zhou, M., Chinnaiyan, A. M., Kleer, C. G., Lucas, P. C. & Rubin, M. A. -Methylacyl-CoA racemase: a novel tumor marker over-expressed in (2002) several human cancers and their precursor lesions. Am. J. Surg. Pathol. 26: 926 931. 70. Jiang, Z., Fanger, G. R., Banner, B. F., Woda, B. A., Algate, P., Dresser, K., Xu, J., Reed, S. G., Rock, K. L. & Chu, P. G. (2003) A dietary enzyme: alpha-methylacyl-CoA racemase/P504S is overexpressed in colon carcinoma. Cancer Detect. Prev. 27: 422 426.

terminal kinase-induced neuronal death by promoting heat shock protein 27 expression and Akt activation. J. Neurosci. 23: 51875196. 40. Chauhan, D., Li, G., Hideshima, T., Podar, K., Mitsiades, C., Mitsiades, N., Catley, L., Tai, Y. T., Hayashi, T. et al. (2003) Hsp27 inhibits release of mitochondrial protein Smac in multiple myeloma cells and confers dexamethasone resistance. Blood 102: 3379 3386. 41. Concannon, C. G., Gorman, A. M. & Samali, A. (2003) On the role of Hsp27 in regulating apoptosis. Apoptosis 8: 6170. 42. Liu, Z. M., Chen, G. G., Ng, E. K., Leung, W. K., Sung, J. J. & Chung, S. C. (2004) Upregulation of heme oxygenase-1 and p21 confers resistance to apoptosis in human gastric cancer cells. Oncogene 23: 503513. 43. Brooks, J. D., Paton, V. G. & Vidanes, G. (2001) Potent induction of phase 2 enzymes in human prostate cells by sulforaphane. Cancer Epidemiol. Biomark. Prev. 10: 949 954. 44. Hintze, K. J., Keck, A. S., Finley, J. W. & Jeffery, E. H. (2003) Induction of hepatic thioredoxin reductase activity by sulforaphane, both in Hepa1c1c7 cells and in male Fisher 344 rats. J. Nutr. Biochem. 14: 173179. 45. Zhang, J., Svehlikova, V., Bao, Y., Howie, A. F., Beckett, G. J. & Williamson, G. (2003) Synergy between sulforaphane and selenium in the induction of thioredoxin reductase 1 requires both transcriptional and translational modulation. Carcinogenesis 24: 497503. 46. Thimmulappa, R. K., Mai, K. H., Srisuma, S., Kensler, T. W., Yamamoto, M. & Biswal, S. (2002) Identication of Nrf2-regulated genes induced by the chemopreventive agent sulforaphane by oligonucleotide microarray. Cancer Res. 62: 5196 5203. 47. Dinkova-Kostova, A. T., Holtzclaw, W. D., Cole, R. N., Itoh, K., Wakabayashi, N., Katoh, Y., Yamamoto, M. & Talalay, P. (2002) Direct evidence that sulfhydryl groups of Keap1 are the sensors regulating induction of phase 2 enzymes that protect against carcinogens and oxidants. Proc. Natl. Acad. Sci. U.S.A. 99: 11908 11913. 48. McMahon, M., Itoh, K., Yamamoto, M. & Hayes, J. D. (2003) Keap1dependent proteasomal degradation of transcription factor Nrf2 contributes to the negative regulation of antioxidant response element-driven gene expression. J. Biol. Chem. 278: 2159221600. 49. Kwak, M. K., Egner, P. A., Dolan, P. M., Ramos-Gomez, M., Groopman, J. D., Itoh, K., Yamamoto, M. & Kensler, T. W. (2001) Role of phase 2 enzyme induction in chemoprotection by dithiolethiones. Mutat. Res. 480 481: 305315. 50. Vairapandi, M., Balliet, A. G., Fornace, A. J., Jr., Hoffman, B. & Liebermann, D. A. (1996) The differentiation primary response gene MyD118, related to GADD45, encodes for a nuclear protein which interacts with PCNA and p21WAF1/CIP1. Oncogene 12: 2579 2594. 51. Kaul, R., Mukherjee, S., Ahmed, F., Bhat, M. K., Chhipa, R., Galande, S. & Chattopadhyay, S. (2003) Direct interaction with and activation of p53 by SMAR1 retards cell-cycle progression at G2/M phase and delays tumor growth in mice. Int. J. Cancer 103: 606 615. 52. Corrente, G., Guardavaccaro, D. & Tirone, F. (2002) PC3 potentiates NGF-induced differentiation and protects neurons from apoptosis. Neuroreport 13: 417 422. 53. Egan, E. A. & Solomon, M. J. (1998) Cyclin-stimulated binding of Cks proteins to cyclin-dependent kinases. Mol. Cell Biol. 18: 3659 3667. 54. Lindner, K., Gregan, J., Montgomery, S. & Kearsey, S. E. (2002) Essential role of MCM proteins in premeiotic DNA replication. Mol. Biol. Cell 13: 435 444. 55. Shie, J. L., Chen, Z. Y., OBrien, M. J., Pestell, R. G., Lee, M. E. & Tseng, C. C. (2000) Role of gut-enriched Kruppel-like factor in colonic cell growth and differentiation. Am. J. Physiol. 279: G806 G814.

Downloaded from jn.nutrition.org by guest on December 28, 2011