Genetic Signatures of Differentiation Induced by 1α,25-Dihydroxyvitamin D₃ in Human Colon Cancer Cells

INTRODUCTION

1α,25(OH)₂D₃ 3 is the most active metabolite of vitamin D₃, a scarce natural product that is synthesized in the organism mainly in the skin from 7-dehydrocholesterol by the action of UV sunlight (1 , 2) . In addition to its classical role in the regulation of calcium homeostasis and bone formation/resorption, 1α,25(OH)₂D₃ and several synthetic vitamin D derivatives, which show reduced calcemic activity, induce cell cycle arrest and differentiation or apoptosis in a variety of cancer cell lines (3, 4, 5) . Moreover, they have anti-invasion, antiangiogenesis, and antimetastatic activity in vivo (6, 7, 8) and are chemopreventive in animal models of colorectal and breast cancer (9, 10, 11) .

Several findings suggest that vitamin D improves colon cancer prevention and therapy. In vitro, 1α,25(OH)₂D₃ induces growth arrest and differentiation in colon cancer cells (3 , 12) . Epidemiological data indicate an inverse correlation between vitamin D dietary intake or sunlight exposure and human colorectal cancer, and it is accepted that high circulating levels of 1α,25(OH)₂D₃ associate with reduced risk of colon cancer (13 , 14) . Accordingly, several clinical trials are under way to assess the activity of various nonhypercalcemic vitamin D derivatives in patients with colorectal carcinoma and other neoplasms (2 , 15, 16, 17, 18) . 1α,25(OH)₂D₃ regulates gene expression by binding to specific receptors (VDRs) of the nuclear receptor superfamily, which are ligand-modulated transcription factors (Refs. 19 , 20 reviews). Upon ligand activation VDR binds specific nucleotide sequences (vitamin D response elements) in target genes to activate or repress their expression. Nongenomic actions and cross-talk between ligand-activated VDR and other transcription factors and signaling pathways have also been described previously (11 , 20) . Moreover, certain polymorphisms in the VDRgene have been associated with various neoplasms, including colon cancer (21 , 22) , and expression of VDR decreases during the late stages of colon carcinogenesis (23) additionally supporting the relation between 1α,25(OH)₂D₃ and cancer.

We have previously studied the mechanism of action of 1α,25(OH)₂D₃ and several analogues in human SW480 cells, a widely used model for colon cancer (24 , 25) . Despite mutations affecting TP53, K-RAS, and APC genes, these compounds inhibit the proliferation and promote the differentiation of a subline of SW480 cells expressing VDR (SW480-ADH) but not of another VDR-negative subline (SW480-R; Ref. 12 ). They inhibit the activation of the β-catenin signaling pathway by disrupting the TCF-4/β-catenin interaction and by decreasing the nuclear content of β-catenin through the induction of E-cadherin (12) . More comprehensive understanding of the molecular mechanism of 1α,25(OH)₂D₃ action may improve the clinical use and selection of patients to be treated with vitamin D derivatives. The present study was undertaken to evaluate the gene expression profiles associated with the protective effects of 1α,25(OH)₂D₃ on SW480-ADH cells, using oligonucleotide microarrays.

MATERIALS AND METHODS

Cell Culture and RNA Extraction.

The human colon cancer cell lines SW480-ADH, SW480-R and LS-147T were grown in DMEM supplemented with 10% FCS (12) . All cells were grown and harvested at 50–75% confluence no longer than 4–6 passages in culture. Treatment of cells with 10⁻⁷ m 1α,25(OH)₂D₃ [supplied by Dr. Lise Binderup, Leo Pharmaceuticals Products (Copenhagen, Denmark)] dissolved in isopropanol was performed in DMEM supplemented with charcoal-treated FCS to remove liposoluble hormones. Control cells were always treated with the corresponding concentration of isopropanol. Extraction of total RNA was performed using Trizol and purified using RNeasy columns (Qiagen, Valencia, CA).

Oligonucleotide Microarrays Hybridization, Scanning, and Scaling.

cDNA was synthesized from 10 μg of total RNA using a T7-promoter tagged oligodeoxythymidylic acid primer. RNA target was synthesized by in vitro transcription and labeled with biotinylated nucleotides (Enzo Biochem, Farmingdale, NY). Labeled target was assessed by hybridization to Test arrays (Affymetrix, Santa Clara, CA). Gene expression analysis was carried out using Affymetrix U95A human gene arrays with >12,665 features for individual known genes and ESTs. Two main response measures, the average difference and absolute call were extracted from each gene on every sample, as determined by default settings of Affymetrix Microarray Suite 5.0. Average difference was used as the primary measure of expression, and absolute call was retained as a secondary measure. Expression values of each array were multiplicatively scaled to give an average expression of 500 across the central 95–99% of all genes on the array.

Data Analysis.

For U95A oligonucleotide arrays, scanned image files were visually inspected for artifacts and analyzed using Microarray Suite 5.0 (Affymetrix). Differential expression was evaluated using several measures. Final ranking to obtain genes uniform and strongly differentially expressed was determined as follows. The expression dataset was first filtered to include only those probe sets detecting genes with mean expression values that differed by at least 3.5-fold [corresponding to the increase in E-cadherin RNA levels after 4 h 1α,25(OH)₂D₃ exposure] between each pair of samples under comparison. Probes were then ranked based on the relative magnitude of the difference (t test) between the means of each comparison set. The relationship between cell lines was analyzed by hierarchical clustering using XCluster and Tree View software (26) taking only genes and ESTs displaying present call according to MAS 5.0. A nonparametric bootstrap technique was used to estimate the robustness of the clusters obtained (27) .

Northern Blot Analysis.

Northern hybridization was performed using poly(A)⁺ RNA from control and 1α,25(OH)₂D₃-treated colon cancer cell lines used in the analysis and probes generated of the genes of interest. Poly (A)⁺ RNA was purified as reported elsewhere (28) . Northern blots were performed on nylon membranes (Nytran; Schleicher and Schuell, Keene, NH) following standard protocols (29) . All probes were labeled by the random priming method. Hybridizations were carried out overnight at 65°C in 7% SDS, 500 mm sodium phosphate buffer (pH 7.2) and 1 mm EDTA, as described by Church and Gilbert (30) . Filters were washed twice for 30 min each in 1% SDS and 40 mm sodium phosphate buffer (pH 7.2) at 65°C. The sizes of respective mRNAs were calculated using RNA ladder markers (Invitrogen, Carlsbad, CA). Membranes were exposed to Hyperfilm MP films (Amersham Pharmacia Biotech, Piscataway, NJ). Complete human cDNA probes were used for protease M and NES-1 [Georgia Sotiropoulou, University of Patras, (Patras, Greece], G₀S2 [Scott Heximer, Washington University School of Medicine (St. Louis, MO)], Keratin 13 [Jośe Luis Jorcano, Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (Madrid, Spain)], FREAC-1/FoxF1 [Javier Rey, Centro de Investigaciones Biológicas (Madrid, Spain)], MRG-1 [Toshi Shioda, Massachusetts General Hospital (Charlestown, MA)], and GAPDH; a fragment (nucleotides 2209–2649) of human cDNA for E-cadherin; full-length mouse cDNAs for c-JUN, JUNB, and JUND [Rodrigo Bravo, Pharmacia (Milan, Italy)]; human clones IMAGE 2286742, 34370, and 898281 for ZNF44/KOX7, plectin, and filamin A, respectively [Orlando Domínguez, Centro Nacional de Investigaciones Oncológicas Madrid, Spain)]. Quantifications were carried out using a La Cie scanner connected to a Power Macintosh G4 computer and Adobe PhotoShop 4.0 and NIH Image programs.

Western Blot Analysis.

Whole-cell extracts were prepared by washing cell monolayers twice in PBS, and the cells were lysed by incubation in radioimmunoprecipitation assay buffer as previously described (12) , followed by centrifugation at 13,000 rpm for 10 min at 4^°C and analysis in 10% SDS-PAGE gels. Immunoblotting of cell lysates was performed by protein transfer to Immobilon P membranes (Millipore Corp., Bedford, MA) and incubation with mouse monoclonal anticytokeratin 13 antibody (ab1384, 1:50; abcam Ltd., Cambridge, United Kingdom) and goat antifilamin antiserum (F2762, 1:1000; Sigma, St. Louis, MO). Blots were developed using the enhanced chemiluminescence detection system (Amersham Pharmacia Biotech).

Histochemical Studies.

Protein patterns of expression of identified targets were assessed for filamin, c-Jun, and β-catenin, using fixed cells. Cells were grown and treated on slides. For filamin and β-catenin, cells were rinsed four times in PBS, fixed in cold methanol for 30 s at −20°C, and rinsed in PBS. The nonspecific sites were blocked by incubation with PBS containing 1% BSA for 1 h at room temperature. Cells were incubated with mouse monoclonal antifilamin (RDI-CBL229, 1:50; Research Diagnostics, Inc., Flanders, NJ) or mouse monoclonal anti-β-catenin antibody (C19220, 1:100; Transduction Laboratories, Lexington, KY) diluted in PBS containing 1% BSA overnight at 4°C. After four washes in PBS, cells were incubated with secondary antibodies for 45 min at room temperature. For c-Jun, cells were washed twice in PBS, fixed with 3.7% formaldehyde (freshly prepared from paraformaldehyde) in PBS for 10 min at room temperature, and subsequently permeabilized with 0.5% Triton X-100 in PBS for 20 min at room temperature. Before immunostaining, fixed culture cell samples were sequentially incubated with 0.1 m glycine in PBS for 30 min, 1% BSA in PBS for 15 min, and 0.01% Tween 20 in PBS for 5 min. For immunolabeling, cells were rinsed in PBS containing 0.05% Tween 20 (PBS-Tw), incubated for 2 h at room temperature with rabbit polyclonal anti-c-Jun antibody (H-79, sc1694; Santa Cruz Biotechnology, Santa Cruz, CA; 1:200 diluted in PBS), washed in PBS-Tw, and incubated for 45 min with the secondary antibody. Cells were then washed and mounted in Vectashield (Vector Laboratories, Peterborough, United Kingdom) and sealed with nail polish. Confocal microscopy was performed with a Bio-Rad MRC-1024 laser scanning microscope, equipped with a Zeiss Axiovert 100 invert microscope (Carl Zeiss, Oberkochen, Germany).

RESULTS AND DISCUSSION

Experimental Design.

To evaluate the gene expression profiles associated with 1α,25(OH)₂D₃ treatment in human colon cancer cells we chose SW480-ADH cells expressing endogenous VDR (12) . SW480-ADH cells grow independently and responded to 1α,25(OH)₂D₃ exposure with an inhibition in their proliferation rate and a phenotypic change to a more adherent, differentiated epithelial phenotype (Fig. 1A) ⇓ . 1α,25(OH)₂D₃ also induced E-cadherin expression and the export of β-catenin from the nucleus to the plasma membrane (Fig. 1) ⇓ . Short exposure (4 h) to 1α,25(OH)₂D₃ induced substitution of the β-catenin-TCF-4 dimers by β-catenin-VDR dimers, leading to partial inhibition of the expression of β-catenin target genes (12) . E-Cadherin protein was not induced after 4 h exposure and cells remained elongated and nonadherent. Once E-cadherin protein accumulates, after 16 h of treatment, β-catenin redistributes progressively from the nucleus to the plasma membrane (Fig. 1A) ⇓ . β-Catenin target genes are then further inhibited and cells acquire a differentiated epithelial phenotype (Ref. 12 ; Fig. 1A ⇓ ).

We used oligonucleotide arrays to analyze gene expression profiles induced by 1α,25(OH)₂D₃ before and after the phenotypic change. Total RNA was obtained from control SW480-ADH cells or from cells treated with 1α,25(OH)₂D₃ for 4 or 48 h. These time points allowed us to identify early response genes putatively regulated transcriptionally (4 h) and those responding with slow kinetics that be indirectly and/or posttranscriptionally regulated (48 h). Additionally, in view of the kinetics of E-cadherin protein induction, the time points chosen might also discriminate between 1α,25(OH)₂D₃ target genes that may be E-cadherin-dependent (regulated only after 48 h of exposure) or -independent (regulated already after 4 h). Furthermore, we treated cells with the transcription inhibitor actinomycin D alone or in combination with 1α,25(OH)₂D₃ for 4 h to identify early genes transcriptionally regulated by the hormone.

We also studied the gene expression profiles induced by 1α,25(OH)₂D₃ in LS-174T colon cancer cells. This cell line contains functional VDR as shown by Western blotting and the activation of transfected vitamin D response elements after 1α,25(OH)₂D₃ treatment (12) . However, LS-174T cells do not express detectable E-cadherin protein in either basal or treated conditions (12) , and they do not differentiate in response to 1α,25(OH)₂D₃ (Fig. 1B) ⇓ . The expression profiles of this cell line served as a control not only to evaluate the correlation between gene expression and phenotype changes caused by 1α,25(OH)₂D₃ but also to estimate the contribution of E-cadherin to the gene expression profile induced by 1α,25(OH)₂D₃ in SW480-ADH cells. Control cells grown in parallel and collected at the same time points were included for both SW480-ADH and LS-174T experiments.

Gene Expression Analyses of SW480-ADH Cells Treated with 1α,25(OH)₂D₃.

We compared the expression profiles of SW480-ADH cells treated with 1α,25(OH)₂D₃ for 4 or 48 h versus control cells. As in previous studies (12) , the response to 1α,25(OH)₂D₃ was verified by studying E-cadherin gene expression by Northern blot analysis and β-catenin localization assessed by immunofluorescence (data not shown).

Tables 1 ⇓ (n = 2) and 2 ⇓ (n = 1) summarize the expression changes at 4 and 48 h of 1α,25(OH)₂D₃ exposure taking a 3.5-fold difference cutoff. These preliminary data show that most target genes were up-regulated, and only a few were down-regulated. Little overlap was found between the lists of genes regulated at 4 and 48 h of 1α,25(OH)₂D₃ treatment. Those whose expression changed at both time points appear in italics. Remarkably, some genes that were up-regulated at 4 h became down-regulated at 48 h such as insulin-like growth factor binding protein 2, Ha-RAS, calgizzarin, and ubiquitin-activating enzyme E1. These findings indicate that the wide regulatory effects of 1α,25(OH)₂D₃ are exerted through the control of a primary set of genes, which in turn modulate others in a cascade.

Several genes regulated by 1α,25(OH)₂D₃ or analogues in other cell types have been identified (indicated with asterisks in Tables 1 ⇓ and 2 ⇓ ), such as 1α,25(OH)₂D₃ 24-hydroxylase, E-cadherin, insulin-like growth factor-binding protein-3, TGF type I receptor, GADD45, and protease M (30, 31, 32, 33) .

Many genes involved in transcription were found overexpressed after 1,25(OH)₂D₃ exposure. HIRA and ZNF44/KOX7 were the only such genes up-regulated at both time points, whereas the TIF1β corepressor gene was up-regulated at 4 h but down-regulated at 48 h. Three members of the AP-1 family (c-JUN, JUNB, and JUND) and several zinc-finger and homeobox proteins (ASCL2, FREAC-1/FoxF1, LD5–1, and ZNF44/KOX7) were up-regulated. Some target genes encode proteins involved in chromatin remodeling such as the metastasis-associated gene (MTA-1) and HIRA (34 , 35) or transcription activator or repressors (TIF1β, ATF3, C/EBPγ, GATA-2, TGIF, and GCF-2). These results indicate a major regulatory role of 1α,25(OH)₂D₃ on gene transcription.

A large subset of regulated genes may be involved in the epithelial differentiation induced by 1α,25(OH)₂D₃ treatment. These include cytoskeletal and small GTP proteins and related, several cell-cell or cell-matrix adhesion proteins (plectin, zyxin, filamin, keratins 13 and 15, laminin B3, and β-myosin) and E-cadherin. 1α,25(OH)₂D₃ also regulated a group of genes that participate in signaling pathways such as receptor and nonreceptor kinases, phosphatases, and small GTP proteins and their regulators. Several genes with antiproliferative or proapoptotic effect were induced upon treatment, observations that support the antitumoral role of 1α,25(OH)₂D₃. Among them, the TP53 tumor suppressor, myc-associated zinc-finger protein, protease M, SEL1, several insulin-like growth factor binding proteins, and the proapoptotic BAX, DAP-1, and PAR-4 genes. Certain genes thought to play a role in tumorigenesis such as l-MYC oncogene were repressed by 1α,25(OH)₂D₃, as were several putative tumor suppressor genes: SYK, known to be down-regulated in breast cancer (36) ; NF-2 and NM23H1; and also TOB-1, a negative cell cycle regulator. The expression of several proteases thought to play a role in tumorigenesis was induced by 1α,25(OH)₂D₃. Protease M and NES-1 (normal epithelial cell-specific 1 gene or kallikrein 10) are down-regulated in tumor cells and so thought to contribute to the maintenance of the cell phenotype (37, 38, 39) .

In concordance with its growth inhibitory action, 1α,25(OH)₂D₃ down-regulated several genes involved in DNA replication (thymidine kinase, thymidylate synthase) and cell cycle (cyclin F, CDC21) and some histone genes. Supporting the proposed cross-talk between 1α,25(OH)₂D₃ and the TGF-β (40) , the expression of at least three genes of this signaling pathway (TGF-β type I receptor, SARA, SMAD6) increased, although another (TGIF corepressor gene) decreased upon 1α,25(OH)₂D₃ addition. Moreover, the level of expression of MDMX gene encoding a RING finger ubiquitin ligase that inhibits small mothers against decapentaplegic homologue-(SMAD)-induced transactivation (41) and controls the ubiquitination of p53 protein was repressed. Given the critical roles played by the TGF-β in controlling cell proliferation, matrix production and interaction and the acquisition of the transformed phenotype, the regulation of this pathway might contribute to the beneficial effect of 1α,25(OH)₂D₃. Besides MDMX, the genes coding the ubiquitin-conjugating enzyme UbcH2 and the ubiquitin ligase Cbl-b were also induced by 1α,25(OH)₂D₃, whereas those for ubiquitin carrier protein E2-EPF and ubiquitin-activating enzyme E1 were up-regulated at 4 h but down-regulated at 48 h. Cbl-b mediates down-regulation of receptor tyrosine kinases and, therefore, could be involved in the control of the proliferative response to growth factors (42) . These results suggest a broad effect of 1α,25(OH)₂D₃ on proteasome-mediated protein turnover.

Tables 1 ⇓ and 2 ⇓ do not include TCF-4-β-catenin target genes, although 1α,25(OH)₂D₃ has been shown to regulate them in an opposite way than β-catenin (12) . The explanation is that the regulation by 1α,25(OH)₂D₃ of c-MYC, CD44, vascular endothelial growth factor, engrailed-2, and Brachyury, which appeared in the microarray analysis, did not reach the 3.5-fold cutoff used in the present study. This result agrees with the regulation found previously (12) .

Besides its effects on genes involved in processes that are related to cell transformation, survival, or oncogenesis, 1α,25(OH)₂D₃ regulates genes involved in basic cellular functions such as the control of the cellular redox status and intermediary metabolism, RNA splicing and translation, or protein turnover and folding. Some of these genes might potentially be attributed to posttranscriptional regulatory actions of 1α,25(OH)₂D₃.

Effect of Actinomycin D on 1α,25(OH)₂D₃ Treatment.

We compared the expression profiles of SW480 cells treated with 1α,25(OH)₂D₃ alone or in the presence of actinomycin D for 4 h. A few genes showed substantial level of expression under actinomycin D exposure. Certain genes that were up-regulated by 1α,25(OH)₂D₃ alone (Table 1) ⇓ , including 1α,25(OH)₂D₃ 24-hydroxylase, E-cadherin, cystatin D, protease M, JUNB, and GATA-2, were inhibited under combined treatment with 1α,25(OH)₂D₃ and actinomycin D. These genes are candidates to be regulated by 1α,25(OH)₂D₃ at the transcription level. In contrast, those genes unaffected by actinomycin D are probably regulated posttranscriptionally or indirectly.

Expression Profiles Associated with 1α,25(OH)₂D₃ in LS-174T Cells.

We also evaluated the expression profiles of LS-174T cells that express comparable VDR levels to SW480-ADH cells but do not undergo differentiation after 48 h of exposure to 1α,25(OH)₂D₃. The number of expression changes regulated by 1α,25(OH)₂D₃ in LS-174T cells was lower than in SW480-ADH cells (Table 3 ⇓ ; n = 1). The comparison between the two cell lines revealed that only four of the genes regulated at least 3.5-fold by 1α,25(OH)₂D₃ in SW480-ADH cells listed in Table 2 ⇓ were also regulated in LS-174T cells: 1α,25(OH)₂D₃ 24-hydroxylase, protease M, Bilirubin UDP-glycorosyltransferase isoenzyme 2 and ZFAB. This finding shows the correlation between genotype and phenotype and the differences between the two cell lines and suggests that putative cell-specific genetic and/or epigenetic alterations define the response to 1α,25(OH)₂D₃ of colon carcinoma cells.

Hierarchical Clustering.

The application of a bootstrapping technique provided robustness to the clusters identified based on time exposure to 1α,25(OH)₂D₃ (Fig. 2A) ⇓ . Two main clusters were observed. The first was associated with 4 h of 1α,25(OH)₂D₃ exposure and the second with longer exposure (48 h). This finding supports the differential phenotype associated with time of exposure of SW480-ADH cells to 1α,25(OH)₂D₃ (12) , which prompted us to perform this study. It was also possible to segregate the effect of actinomycin D from cells exposed only to 1α,25(OH)₂D₃. Interestingly, the expression profiles of each control cell type at 4 or 48 h was more similar to that of the paired 1α,25(OH)₂D₃-treated cells than those belonging to the other control cell type. The gene expression patterns of cells after 4 h of exposure were significantly different from those obtained after 48 h of treatment. It was possible to segregate the two different cell types under study within the 48 h exposure cluster, and the differences between treated and control cells were smaller in LS-174T cells. Overall, the hierarchical clustering confirmed the hypothesis that the phenotypic differences between SW480-ADH and LS-174T cells upon 1α,25(OH)₂D₃ exposure are genetically regulated. Growth did not affect cell phenotype: the increase in cell confluency during the 48-h incubation period did not induce differentiation of control cells, as compared with 1α,25(OH)₂D₃-treated ones (Fig. 2B) ⇓ .

Validation of Results.

We screened 17 available cDNAs corresponding to genes regulated by 1α,25(OH)₂D₃ in the microarray study. Filters containing RNA from control SW480-ADH cells or cells treated with 1α,25(OH)₂D₃ for different periods of time were probed. Thirteen of the 17 genes were validated (Fig. 3) ⇓ . There was overall concordance between data from the microarrays screening and Northern blotting. This analysis confirmed the induction at 4 h of JUNB, JUND, filamin, and plectin (Fig. 3A) ⇓ at 48 h of FREAC-1/FoxF1, c-JUN, MRG1, and keratin-13 (Fig. 3B) ⇓ and at both time points of E-cadherin, G₀S2, NES-1, protease M, and ZNF-44/KOX7 (Fig. 3C) ⇓ . In addition, the negative regulation of SPROUTY-2 was confirmed but at a different time (48 h) from that suggested by the microarray screening (4 h; data not shown). Three other genes, TP53, NMP200, and PMS2, did not show differential expression. As a negative control, we confirmed that 1α,25(OH)₂D₃ did not induce FREAC-1/FoxF1 and NES-1 expression in VDR-negative SW480-R cells or in LS-174T cells in which the microarray analysis showed no regulation (Table 3) ⇓ . In addition, protease M, regulated in LS-174T cells according to the microarray analysis (Table 3) ⇓ , was validated by Northern blotting and remained uninduced in SW480-R cells (Fig. 3D) ⇓ . In agreement with the results of the microarray analysis indicated above, the inhibition of 1α,25(OH)₂D₃ action by actinomycin D confirmed the transcriptional regulation of several genes such as JUNB, FREAC-1/FoxF1, and protease M (Fig. 3E) ⇓ . The translation inhibitor cycloheximide failed to block 1α,25(OH)₂D₃ action on these three genes, suggesting that this does not require protein synthesis de novo.

We also examined whether the induction by 1α,25(OH)₂D₃ was restricted to the RNA level or extended to differential protein expression. 1α,25(OH)₂D₃ led to increases in filamin A and keratin-13 protein levels as assessed by Western blotting (Fig. 4A ⇓ , top panel). As control, neither of these two proteins nor E-cadherin increased in LS-174T or SW480-R cells upon treatment (Fig. 4A ⇓ , bottom panel). Immunofluorescence analysis confirmed the increase in filamin A expression after 1α,25(OH)₂D₃ treatment and also a change in its subcellular distribution from a homogeneous cytosolic localization to a highly preferential presence in the periphery of the epithelioid islands formed upon 1α,25(OH)₂D₃ treatment (Fig. 4B) ⇓ . Likewise, 1α,25(OH)₂D₃ increased the nuclear content of c-Jun protein (Fig. 4B) ⇓ .

Taken together, these results confirmed those obtained in the gene array screening. The gene expression patterns regulated by 1α,25(OH)₂D₃ in colon SW480-ADH cells have similarities with those found in 1α,25(OH)₂D₃- or EB1089-treated head and neck cancer cells (32 , 33) . The differences could be attributed to the sets of genes (coregulators and others) expressed and the signaling pathways activated and alterations present in both cell types.

In conclusion, this study has revealed novel molecular targets associated with 1α,25(OH)₂D₃ exposure in human colon cancer cells. Results suggest a pleiotropic regulatory role for 1α,25(OH)₂D₃. 1α,25(OH)₂D₃ regulates a series of genetics events involved in inhibiting cell proliferation, inducing cell adhesion, and modulating apoptosis. As a whole, it induces a phenotypic change toward a normal epithelial phenotype. Although additional research is warranted toelucidate the role of 1α,25(OH)₂D₃ target genes in colon tumorigenesis and treatment, data shown here support the ongoing clinical studies using nonhypercalcemic vitamin D₃ derivatives for the prevention and treatment of colon cancer.