Resveratrol-Induced Apoptosis Is Mediated by Early Growth Response-1, Krüppel-Like Factor 4, and Activating Transcription Factor 3 Free

Current research suggests that various dietary phytochemicals function as chemopreventive and/or adjuvant chemotherapeutic agents, adding to the paradigm that a diet high in fruit and vegetable content confers protection against chronic disease (1). One such phytochemical is resveratrol (3,4,5′-trihydroxystilbene), a naturally occurring phytoalexin readily available in the diet and to which a plethora of health-promoting effects have been ascribed (2). Resveratrol has elicited much attention as a potential anticancer agent since the inhibitory effect of this compound on carcinogenic processes was first reported in 1997 (3). Subsequently, numerous studies have illustrated the antiproliferative effect of resveratrol on cancer cells, which is believed attributable to induction of cell-cycle arrest in the G₁/S or G₂/M phase and induction of apoptosis and related proteins (4, 5). More importantly, treatment with resveratrol inhibited tumorigenesis in vivo (6, 7). However, the underlying mechanism(s) involved in the antitumorigenic/carcinogenic activities of resveratrol remain poorly defined due to its capacity to modulate a multitude of signaling pathways.

Activating transcription factor 3 (ATF3), a member of the ATF/CREB family of bZIP transcription factors, is characterized as a stress-inducible or adaptive response gene (8). Much controversy exists as to the physiological role of ATF3 in tumorigenesis and is demonstrated to be a positive or negative modulator of tumor progression. Recently, a dichotomous role was reported for ATF3 in cancer development; the authors concluded that its role as a tumor suppressor or oncogene is largely dependent on cellular context and extent of malignancy (9). Yet, several lines of evidence suggest that ATF3 may function as a tumor suppressor gene in colorectal carcinogenesis. First, ATF3 expression is markedly reduced in cancer tissues, including colon, when compared to normal adjacent tissue (10, 11). Second, ATF3 is reported to mediate or enhance induction of apoptosis by compounds demonstrated to have antitumor properties (12–16). Finally, ATF3 overexpression elicits a number of cellular responses, including induction of cell-cycle arrest and inhibition of proliferation (17), induction of apoptosis in vitro and in vivo (12, 18-20), inhibition of invasion (21-23), and retardation of tumor formation in vivo (20, 22). Thus, we believe that ATF3 may play an antitumorigenic role in colorectal tumorigenesis.

In the present study, we examined the effect of resveratrol treatment on gene modulation in HCT-116 human colorectal cancer cells. We identified ATF3 as the most highly induced gene after treatment and sought to investigate the transcriptional mechanism and biological consequence of ATF3 expression in response to resveratrol. Here, we report that early growth response-1 (Egr-1) and Krüppel-like factor 4 (KLF4) mediate ATF3 transactivation by resveratrol. We show Egr-1 and KLF4 interaction in the presence of resveratrol, which may facilitate ATF3 transcriptional regulation by this compound. Furthermore, we demonstrate that induction of apoptosis by resveratrol is mediated, at least in part, by ATF3.

All human cancer cell lines were purchased from American Type Culture Collection unless otherwise stated; authentication occurred via short tandem repeat profiling, monitor of cell morphology, and karyotyping. SqCC/Y1 head and neck squamous cell carcinoma cell line was generously provided by Dr. Dong M. Shin (Emory University) and characterized previously (24). Cell lines were maintained according to established protocol. Resveratrol was purchased from Alexis Biochemicals, 3,3′-diindoylmethane (DIM) from Sigma Aldrich, and cycloheximide from Fisher Scientific. All chemicals were dissolved in dimethylsulfoxide (DMSO). ATF3, Egr-1, KLF4, and actin antibodies were purchased from Santa Cruz Biotechnology.

For deletion analysis of the ATF3 promoter, pATF3 −1850/+34 (12) was serially deleted using the Erase-a-Base System (Promega) according to manufacturer's protocol. The pATF3 −514 del Egr-1 reporter construct was previously described (25). Putative binding sites of KLF4 within the −514 bp region of the promoter were deleted using Stratagene's QuikChange II Site Directed Mutagenesis Kit with the following primers: del KLF4-A (F, 5′-CCCCCTCTCTTTCGGCCCCGCCTTGGCCCC-3′ and R, 5′-CGGGGCCGAAAGAGAGGGGGCACTGGTGATG-3′) and del KLF4-B (F, 5′-GGCCCCTCCTCCTTCCTCCGCTCCGTTCGG-3′ and R, 5′-CGGAGGAAGG- AGGAGGGGCCAAGGCGGGGC-3′). pATF3 −514/+34 del Egr-1 KLF4 reporter construct was generated using pATF3 −514 del Egr-1 and KLF4-A deletion primers as described. Deletions were confirmed by DNA sequencing. ATF3 (pCG-ATF3) expression vector was kindly provided by Dr. T. Hai (Ohio State University). KLF4 (pcDNA3.1/His/V5/KLF4) expression vector was generated using the primers F, 5′-CGAATTCTATGGCTGTCAGCGACGCG-3′ and R, 5′-CCCAAGCTTTTAAAAATG- CCTCTTCATGTGTAAGGC-3′. Egr-1 (pcDNA3.1/NEO/Egr-1), Sp1 (pCMV-Sp1), and p53 (pcDNA3.1/myc/His/p53) expression vectors were previously described (26–28).

Total RNA was isolated from HCT-116 cells treated with DMSO or resveratrol (50 μmol/L) using 5 Prime Perfect RNA Cell/Tissue Kit and 1 μg of RNA was reverse transcribed using Verso cDNA Synthesis Kit (Thermo Fisher Scientific). PCR was performed as described (25) using ATF3 (F, 5′-GTTTGAGGATTTTGCTAACCTGAC-3′ and R, 5′-AGCTGCAATCTTATTTCTTTCTCGT-3′) and GADPH (F, 5′-TCAACGGATTTGGTCGTATT-3′ and R, 5′-CTGTGGTCATGAGTCCTTCC-3′) human gene specific primers.

Cells were grown to 60% to 80% confluence in 60-mm plates, serum starved overnight, and treated with resveratrol as indicated in serum-free media. Protein lysates were isolated in RIPA buffer containing 1 mmol/L of PMSF, 1 μg/mL aprotinin, 1 μg/mL leupeptin, 0.1 nmol/L of Na₃VO₄, and 25 mmol/L of NaF. Total protein was subjected to Western blot analysis as previously described (25).

HCT-116 cells were grown to 60% to 80% confluence in 60-mm plates and pretreated with cycloheximide (10 mg/mL) for 30 minutes in serum-free media. After pretreatment, DMSO or resveratrol (50 μmol/L) was added directly to the media and incubated for 24 hours. Total RNA was isolated and reverse transcribed. Real-time RT-PCR was performed according to Fast SYBR Green Master Mix protocol (Applied Biosystems) and analyzed using MyIQ Single Color Real-time PCR Detection System (Bio-Rad Laboratories).

Transient transfections were performed using LipofectAMINE (Invitrogen) according to manufacturer's instructions. HCT-116 cells were seeded in 12-well plates at a concentration of 2.0 × 10⁵ cells per well. The next day, plasmid mixtures containing ATF3 promoter (0.5 μg) and pRL-null vector (0.05 μg) were cotransfected for 5 hours in serum-free media. For cotransfection experiments, 0.25 μg of the ATF3 promoter and 0.25 μg of expression vector were cotransfected with 0.05 μg pRL-null vector. For competition assays, plasmid mixtures were prepared as indicated in Fig. 5A. After transfection, cells were treated with DMSO or resveratrol (50 μmol/L) in serum-free media for 24 hours. Cells were harvested in 1× passive lysis buffer, and luciferase activity was measured and normalized to pRL-null luciferase activity using DualGlo Luciferase Assay Kit (Promega).

Egr-1 sense and antisense oligonucleotides were previously described (25). ATF3 and KLF4 siRNA were purchased from Santa Cruz Biotechnology; control siRNA was purchased from Ambion. HCT-116 cells were transfected with either 100 nmol/L (ATF3 and KLF4) or 200 nmol/L (Egr-1) of each construct using TransIT-TKO transfection reagent (Mirus). After transfection for 24 hours, cells were serum starved overnight and treated as indicated. Total protein was subjected to Western blot analysis as described.

HCT-116 cells were grown to 80% confluence. After overnight serum starvation, cells were treated with resveratrol (50 μmol/L) for 24 hours. Cells were then washed with PBS and nuclear extracts were prepared using Nuclear Extract Kit (Active Motif) according to protocol. Double-stranded oligonucleotides corresponding to the Egr-1 (F, 5′-GTGAGCGAGGGCGGGG-3′) and KLF4 (F, 5′-TCCACCCCTTCCACCCCT-3′) binding sites were synthesized and end-labeled with biotin (Operon). Mutant oligonucleotides of Egr-1 (F, 5′-GTGAGCGAAAACGGGG-3′) and KLF4 (F, 5′-TCCAAAACTTCCAAAACT-3′) were also synthesized. To ensure specific binding of Egr-1 and KLF4, recombinant proteins were generated using TNT Quick Coupled Transcription/Translation System (Promega). EMSA was performed as previously described (29).

HCT-116 cells were grown to 80% confluence and treated with DMSO or resveratrol (50 μmol/L). After 24 hours, cells were fixed with 1% formaldehyde at 37°C for 10 minutes and chromatin immunoprecipitation (ChIP) was performed as previously described (29). The region between −298 and −114 bp of the human ATF3 promoter was amplified by real-time and RT-PCR using F, 5′-CGGCTCCGGTCCTGATATGG-3′ and R, 5′-AGAACCGGCCGAACGGAGCG-3′. The 184-bp product was resolved on 2% agarose gel and visualized under UV light.

The pM/Egr-1 and pVP16/KLF4 vectors for mammalian 2-hybrid were generated. PCR fragments were amplified from pcDNA3.1/NEO/Egr-1 and pcDNA3.1/His/V5/KLF4 using Egr-1 (F, 5′-TTGAATTCGCCGCGGCCAAGGCCGAG-3′ and R, 5′-TTAAGCTTGCAAATTTCAATTGTCC-3′) and KLF4 (F, 5′-CGAATTCTATGGC- TGTCAGCGACGCG-3′ and R, 5′-CCAAAGCTTTTAAAAATGCCTCTTCATGTGTAAGGC-3′)-specific primers, respectively. After PCR, fragments were cloned into pCR2.1/TOPO vector (Invitrogen), digested with EcoRI and HindIII restriction enzymes, and cloned into pM or pVP16 vectors. Deletion clones of pM/Egr-1 were generated as described using the primers: del AD (F, 5′-CATCACCTATGGCCTAGTGAGCATGACCAACCCAC-3′and R, 5′-TCACTAGGCCATAGGTGATGGGGGGCAGTCGAGTG-3′) and del ZD2 (F, 5′-GCCCTCCCAG-TTCGCCTGCGACATCTGTGGAAGAA-3′ and R, 5′-CG CAGGCGAACTGGAAGGGCTTCTGGCCTGTGTGG-3′). Mammalian 2-hybrid assay was performed according to Matchmaker Mammalian Assay Kit 2 protocol (Clontech). Transfection occurred as described using plasmid mixtures containing 0.2 μg of pM/Egr-1 or deletion construct, 0.2 μg of pVP16/KLF4, 0.2 μg of pG5Luc, and 0.06 μg of pRL-null vector.

For immunoprecipitation of Egr-1 (WT and deletions, generated as described previously) and KLF4, each expression vector was transfected into HCT-116 cells as indicated in Fig. 5D and grown to confluence. Cells were then serum starved overnight and treated with resveratrol (50 μmol/L) for 2 hours. After washing with ice-cold PBS, cells were harvested in RIPA buffer containing inhibitors and mixed at 4°C for 15 minutes. Cell suspensions were centrifuged and protein concentrations were measured. Immunoprecipitation was performed as previously described (12).

Enzyme activity of caspase 3/7 was analyzed by Apo-ONE Homogenous Caspase-Glo 3/7 Assay (Promega) according to manufacturer's protocol. The cells were harvested in RIPA buffer containing protease and phophatase inhibitors. The same volume of caspase-Glo 3/7 reagent was added to cell lysates (30 μg protein) in 96-well plates and incubated at room temperature in the dark for 1 hour. Luminescence was measured using FLX800 microplate reader (BioTek).

SAS for Windows (v9.2; SAS Institute, Inc.) statistical analysis software was used. For multiple group comparisons, analysis of variance with Tukey's multiple comparison test was used to compare mean values. The Student's t test was used to analyze differences between samples. Results were considered statistically significant at *P < 0.05, **P < 0.01, and ***P < 0.001.

The mechanism(s) underlying the antitumorigenic properties of resveratrol in colorectal cancer remain mostly unclear. To investigate how resveratrol alters gene expression in colorectal cancer cells, microarray analysis was performed using HCT-116 cells treated with resveratrol (50 and 100 μmol/L). Among the genes upregulated by the treatment, ATF3 was identified as most highly induced (Table S1 and S2). We confirmed resveratrol-induced ATF3 transcript using RT-PCR (Fig. 1A). We next determined whether resveratrol increases ATF3 expression at the protein level. As shown in Fig. 1B and 1C, resveratrol increased ATF3 in a concentration- and time-dependent manner in HCT-116 cells. Furthermore, increased ATF3 expression was observed in other colorectal (Fig. 1D) and non-colorectal (Fig. 1E) cancer cells after treatment; however, we did not observe ATF3 induction in LoVo and SqCC/Y1 cells. Absence of ATF3 induction by resveratrol in these cells is most likely due to high endogenous expression of ATF3 (LoVo) or the lack of proper signaling pathway(s) needed for ATF3 increase by resveratrol (SqCC/Y1). Because resveratrol increased ATF3 expression at the mRNA and protein level, we sought to characterize ATF3 as a molecular target of this compound at the transcription level.

To identify the transcriptional binding site responsible for resveratrol-mediated ATF3 expression, HCT-116 cells were transfected with different promoter constructs spanning the −1850 to +34 bp promoter region (Fig. 2A). Resveratrol treatment resulted in a significant increase in ATF3 promoter activity for all reporter constructs; however, the greatest induction was observed within the −514 and +34 bp region, suggesting that a major response element(s) may be present. This region was further analyzed using 3 independent transcription factor search engines (Genomatix-MatInspector, TFSearch, and Transcription Element Search System), and 5 possible factors were commonly identified: p53, Sp1, Egr-1, KLF4, and ATF/CREB (Fig. 2B).

To investigate the role of identified cis-acting elements in resveratrol-mediated transcriptional regulation of ATF3, cotransfection experiments were performed utilizing expression vectors of the aforementioned factors and ATF3 promoter constructs as indicated (Fig. 2C and D). Expression of each vector, with the exception of ATF3, resulted in increased pATF3 −514/+34 luciferase activity after resveratrol treatment compared to pcDNA3.1 empty vector control (Fig. 2C). Interestingly, cells overexpressing KLF4 and Egr-1 showed the greatest increase in resveratrol-induced ATF3 promoter activity, resulting in a 4.1- and 4.3-fold induction, respectively. A similar pattern was observed for the −132 to +34 bp region (Supplementary Fig. S1); however, cotransfection of these expression vectors with pATF3 −84/+34 reporter construct had no effect on promoter activity after treatment (Fig. 2D). These results identify putative resveratrol response-elements located within the −514 to −84 bp region of the ATF3 promoter and suggest that both Egr-1 and KLF4 may play a role in ATF3 regulation by resveratrol.

To clarify the importance of Egr-1 and KLF4 in resveratrol-mediated activation of ATF3, internal deletion clones lacking the Egr-1 and KLF4 binding sites were generated and transfected into HCT-116 cells. Deletion of Egr-1 or KLF4-A binding site markedly reduced ATF3 promoter activity in response to resveratrol treatment when compared with wild type (Fig. 2E). Deletion of the KLF4-B binding site did not change resveratrol-induced transactivation. Because deletion of binding sites corresponding to either Egr-1 or KLF4-A resulted in significant decrease in resveratrol-induced promoter activity, a deletion clone lacking both these sites was generated. Resveratrol-induced activity was dramatically reduced in cells transfected with the double deletion promoter construct compared to wild type (Fig. 2E). Moreover, suppression of either endogenous Egr-1 or KLF4 decreased ATF3 expression in the presence of resveratrol (Fig. 2F), confirming that both Egr-1 and KLF4 contribute to resveratrol-induced ATF3 expression.

To determine whether Egr-1 and KLF4 bind to the ATF3 promoter, EMSA was performed using biotin-labeled oligonucleotides containing 1 to 2 copies of the corresponding promoter binding site. First, we examined the effect of resveratrol on promoter binding by these factors using nuclear extracts prepared as described in Materials and Methods. Preincubation of resveratrol-treated nuclear extracts with Egr-1 and KLF4 oligonucleotides resulted in DNA/protein complex formation that was competed out with addition of 10× and 100× molar excess of unlabeled wild-type oligonucleotides (Fig. 3A and B, left, lanes 3–4), whereas preincubation with unlabeled mutant oligonucleotides had no affect on DNA/protein complex formation (Fig. 3A and B, left, lanes 5–6). This suggests that both Egr-1 and KLF4 are able to bind to the ATF3 promoter. Secondly, we verified the specificity of the binding sites for Egr-1 and KLF4 using in vitro translated proteins mixed with their respective oligonucleotides. As shown in Fig. 3A and B, both Egr-1 and KLF4 bind to their specific binding sites as evidenced by the formation of shift bands that were competed out by 10× and 100× molar excesses of unlabeled oligonucleotides (Fig. 3A and B, right, lanes 10–11). Addition of mutant oligonucleotides did not affect binding of Egr-1 or KLF4 to the ATF3 promoter (Fig. 3A and B, right, lanes 12–13). Finally, a ChIP assay was performed to further confirm that Egr-1 and KLF4 bind to the ATF3 promoter after resveratrol treatment. As shown in Fig. 3C, immunoprecipitation of the chromatin/protein complex with Egr-1 and KLF4 resulted in the enrichment of these proteins at the ATF3 promoter (left) and the visualization of a 184-bp band of the amplified promoter (right). In conjunction with previous results shown in Fig. 2, these data suggest that both Egr-1 and KLF4 bind to the ATF3 promoter and activate transcription in the presence of resveratrol. However, Egr-1 may contribute more to resveratrol-induced ATF3 expression, since resveratrol enhances Egr-1 binding capacity to the ATF3 promoter.

We then examined requirement of de novo protein synthesis for resveratrol-mediated ATF3 expression because Egr-1 and KLF4 appear to be required for induction of ATF3. HCT-116 cells were treated with cycloheximide and resveratrol (50 μmol/L). In the presence of cycloheximide, resveratrol was unable to increase ATF3 expression, suggesting that ATF3 increase by resveratrol is dependent on de novo protein synthesis (Fig. 4A). We next determined whether resveratrol could alter expression of Egr-1 and KLF4. Resveratrol treatment increased Egr-1 and KLF4 mRNA transcript (Supplementary Fig. S2) and protein expression (Fig. 4B). It should be noted that induction of both Egr-1 and KLF4, which occurred at approximately 1 hour of treatment and continued until 3 hours, preceded that of ATF3, which began after 6 hours. Together, these results are compatible with the notion that observed increase in ATF3 expression by resveratrol requires synthesis of Egr-1 and KLF4.

Because both Egr-1 and KLF4 are involved in resveratrol-mediated activity, we examined whether these factors coordinate or compete in ATF3 expression. HCT-116 cells were cotransfected with pATF3 −514/+34 and expression vectors and treated as described. As observed previously, both Egr-1 and KLF4 increased ATF3 promoter activity in the presence of resveratrol (Fig. 5A). Moreover, coexpression of these factors resulted in enhanced transactivation of ATF3 at both the basal level and after resveratrol treatment.

Previous studies have shown interaction between 2 Zn finger transcription factors (30) and that this interaction facilitates transcriptional activation (31). To investigate the potential for Egr-1and KLF4 interaction, mammalian 2-hybrid assay was performed using pM/Egr-1 and pVP16/KLF4 vectors. Egr-1 and KLF4 bind to each other at basal level and this interaction is enhanced after resveratrol treatment (7.2-fold vs. 52-fold increase; Fig. 5B). Increased Egr-1 and KLF4 interaction was also observed after DIM treatment. DIM increases ATF3 expression (13). To exclude the potential for direct binding of pVP16/KLF4 to the DNA binding domain of the pM empty vector, the vectors were cotransfected and mammalian 2-hybrid assay was performed; no interaction was detected, allowing one to infer the interaction of Egr-1 and KLF4 is genuine (Supplementary Fig. S3).

We next sought to tentatively identify the region that may facilitate this interaction and generated pM/Egr-1 deletion constructs lacking (1) 133 to −159 amino acids (aa) of the activation domain and (2) 370 to 395 aa of the Zn finger domain (Fig. 5C, top). Both the partial deletion of the activation domain and deletion of the second Zn finger domain of Egr-1 resulted in decreased interaction with KLF4 (Fig. 5C, bottom). We confirmed mammalian 2-hybrid assay by Egr-1 and KLF4 immunoprecipitation experiments (Fig. 5D). Egr-1 and KLF4 are coimmunoprecipitated with either antibody, whereas Egr-1 deletion clones had diminished binding capacity to KLF4. Together, these data suggest that Egr-1 and KLF4 physically interact, which is increased by resveratrol, and that the Egr-1 activation or Zn finger domain may mediate this interaction.

We have identified ATF3 as a molecular target of resveratrol; however, the biological consequence(s) of ATF3 induction by resveratrol in colorectal tumorigenesis is unknown. As described in the introduction, pharmaceuticals and phytochemicals' increase of ATF3 mediated or enhanced apoptosis by these compounds (12-16). To determine if this is true for resveratrol, we measured caspase 3/7 enzymatic activity for resveratrol-treated HCT-116 and HT29 colorectal cancer cells. As depicted in Fig. 6A, resveratrol treatment increased apoptosis in a dose-dependent manner. Furthermore, ATF3 overexpression in these cells also increased apoptosis, which depend on caspase activation (Fig. 6B). Together, these data suggest that both resveratrol and ATF3 expression induce apoptosis in our model system. Next, to investigate whether resveratrol-induced apoptosis is mediated by ATF3, ATF3 expression was knocked down by siRNA followed by resveratrol treatment for 24 hours. As depicted in Fig. 6C, knockdown of ATF3 significantly abrogated caspase 3/7 enzymatic activity by resveratrol. This suggests that ATF3 plays a role in resveratrol-induced apoptosis.

Resveratrol, a dietary phytoalexin readily available in the diet, has garnered much attention as a potential chemopreventive and/or chemotherapeutic agent. Numerous studies, utilizing in vitro and in vivo model systems, have illustrated resveratrol's capacity to inhibit the stages of carcinogenesis (initiation, promotion, and progression) and modulate a multitude of signaling pathways associated with cellular growth and division, apoptosis, angiogenesis, and metastasis (32). However, the underlying molecular mechanism(s) involved in the antitumorigenic/carcinogenic activities of resveratrol, especially in colorectal cancer, are complex and remain poorly defined. This study sought to investigate the effect of resveratrol on gene modulation in HCT-116 human colorectal cancer cells in order to identify a novel target that may mediate the anticancer activities of this compound and determine the mechanism involved in its regulation. Our data showed ATF3 was indeed upregulated at both the mRNA and protein (Fig. 1) level by the compound.

Recently, a dichotomous role in cancer development was reported for ATF3; the authors of that study demonstrated both a tumor suppressive (early-stage tumorigenesis) and oncogenic (late-stage tumorigenesis) role in breast cancer cell lines derived from similar genetic backgrounds with varying degrees of malignancy (9). Similarly, a duality of function was demonstrated for ATF3 in cancer studies of the prostate (19, 33) and of the colon (22, 34). Yet, as described in the introduction, several lines of evidence suggest ATF3 behaves as a negative regulator of tumorigenesis. ATF3 expression is induced by several compounds demonstrated to possess antitumorigenic properties (12–16). The results presented here add to the list of compounds that induce ATF3 with mechanistic data.

To elucidate the molecular mechanism by which resveratrol induced ATF3 expression, the promoter region spanning −1850 to +34 bp was assessed by luciferase assay in response to resveratrol. We found that resveratrol transactivated the ATF3 promoter and identified putative response-elements located within the −514- to −84 bp region (Fig. 2A). Cotransfection experiments suggested involvement of the transcription factors Egr-1 and KLF4 (Fig. 2C). Egr-1 and KLF4 participation in resveratrol-mediated activation of ATF3 was confirmed by deletion (Fig. 2E) and knockdown (Fig. 2F) experiments. Interestingly, simultaneous deletion of both Egr-1 and KLF4-A binding sites from the ATF3 promoter dramatically reduced transactivation by resveratrol, indicating that the combination of these 2 factors is important for ATF3 transcriptional regulation by resveratrol. Yet, we cannot exclude the involvement of other cofactors, such as p53 and Sp1 (Fig. 2C), in the compound's mediated activity. ATF3 regulation by p53 is documented in the literature (35), which is increased by resveratrol (4) and suggests a potential role for p53 in resveratrol-mediated ATF3 activation. Furthermore, Sp1 was recently demonstrated to regulate ATF3 expression in colon cancer cells (36); however, we did not observe Sp1 induction by resveratrol (data not shown). Here, we focused on Egr-1 and KLF4 because both increased resveratrol-mediated ATF3 activation and were modulated by the compound.

Egr-1 and KLF4 belong to a family of immediate early response genes whose expression is transiently induced in response to various environmental stimuli (37, 38). Each encodes a transcription factor containing 3 carboxyl C2H2-type Zn finger motifs that coordinate the expression of genes associated with cell proliferation, differentiation, and apoptosis (38–40). In addition, Egr-1 and KLF4 are suggested to act as master regulatory proteins involved in cell fate decisions (41, 42). Nonetheless, as with ATF3, much controversy exists as to the role of Egr-1 and KLF4 in cancer development and their biological function appears largely context dependent. Several studies have demonstrated that Egr-1 and KLF4 expression facilitates tumor progression in vivo (43–45); however, there is ample evidence supporting a tumor-suppressive role for both transcription factors (29, 46–51).

Our data indicates that Egr-1 and KLF4 are able to bind to the ATF3 promoter (Fig. 3) and that their biosynthesis is necessary for resveratrol-mediated activity (Fig. 4). Moreover, these transcription factors cooperate in ATF3 transactivation (Fig. 5A). From this, 2 inferences can be made: (1) synergism between Egr-1 and KLF4 promote ATF3 activation, which is enhanced by resveratrol, and (2) interaction between the 2 facilitates transcriptional regulation of ATF3 by the compound. Indeed, the results ascertained by mammalian 2-hybrid assay and immunoprecipitation experiments corroborate Egr-1 and KLF4 interaction (Fig. 5B and D). Furthermore, resveratrol and DIM increased Egr-1 and KLF4 interaction. These results imply that a similar mechanism may contribute to induction of ATF3 expression by these and possibly other phytochemicals that increase ATF3 expression.

Current research suggests the importance of C2H2 type Zn finger domains in protein–protein interaction (52). Consistently, we found the Zn finger domain of Egr-1 as a possible region responsible for interaction with KLF4. Interestingly, our data also identified the activation domain of Egr-1 as responsible for mediating the interaction with KLF4 (Fig. 5C and D). To our knowledge, this is the first report identifying the involvement of the activation domain in protein–protein interaction of 2 Zn finger transcription factors. Thus, the activation domain contributes not only to the initiation of transcription but also contribute to protein–protein interactions with other transcription factors; interaction via the activation domain or the Zn finger domain leads to enhanced compound-induced transcription.

Results identify ATF3 as a molecular target of resveratrol whose regulation is mediated by Egr-1 and KLF4 interaction. However, the biological consequence(s) of ATF3 induction by the compound is unknown in colorectal tumorigenesis. To answer this question, we examined induction of apoptosis by resveratrol and ATF3 in colorectal cancer cells because, as described, ATF3 mediated or enhanced apoptosis by compounds with known antitumor activities. As depicted in Figure 6A and B, resveratrol and ATF3 overexpression increased caspase 3/7 enzymatic activity in HCT-116 and HT29 cancer cells. Furthermore, knockdown of ATF3 expression resulted in reduced apoptosis by resveratrol and suggests that ATF3 plays a role in resveratrol-induced apoptosis. However, further study is required to determine whether ATF3 contributes to resveratrol anticancer effects in vivo.

In conclusion, we found the stress-inducible and/or adaptive response gene ATF3 as most highly induced gene after resveratrol treatment in human colorectal cancer cells. Here, we report for the first time the involvement of resveratrol in transcriptional regulation of ATF3 and that this regulation is mediated by the C2H2 type Zn finger transcription factors Egr-1 and KLF4. We demonstrate that Egr-1 and KLF4 interact with each other in the presence of resveratrol, which may facilitate ATF3 transcriptional regulation by this compound. Furthermore, increased ATF3 expression by resveratrol facilitates induction of apoptosis by the compound (Fig. 6D).

No potential conflicts of interest were disclosed.

We thank Misty Bailey for her critical reading of this manuscript. We also thank Dr. Luis Miguel Lembcke for his technical assistance.

American Cancer Society grant CNE-111611, NIH grant RO1CA108975, The University of Tennessee Center of Excellence in Livestock Diseases and Human Health (S.J. Baek), and in part by NIEHS/NIH intramural research program (T.E. Eling).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.