|Home | About | Journals | Submit | Contact Us | Français|
The activation of the peroxisome proliferator‐activated receptor γ (PPARγ) that forms heterodimers with retinoid X receptors (RXRs) elicits an antineoplastic effect on colorectal cancer. It was previously reported that the accumulation of the non‐functional phosphorylated form of RXRα (p‐RXRα) interfered with its signalling and promoted the growth of hepatoma cells. In this study the effects of p‐RXRα on the ability of RXRα and PPARγ ligands to inhibit growth in colon cancer cells was examined.
The effects of the combination of the PPARγ ligand ciglitazone and the RXRα lignad 9‐cis‐retinoic acid (RA) on inhibition of cell growth in Caco2 human colon cancer cells which express high levels of p‐RXRα protein were examined
The RXRα protein was phospholylated and also accumulated in human colon cancer tissue samples as well as human colon cancer cell lines. When the phosphorylation of RXRα was inhibited by the MEK inhibitor PD98059 or by transfection with a point‐mutated RXRα, which mimicked the unphosphorylated form, the combination of 9‐cisRA and ciglitazone synergistically inhibited the cell growth and induced apoptosis. The combined treatment with these agents also caused a decrease in the expression levels of both cyclo‐oxygenase‐2 (COX‐2) and c‐Jun proteins and mRNAs. Reporter assays indicated that this combination induced the transcriptional activity of the peroxisome proliferator‐responsive element promoter and also inhibited that of the AP‐1 promoter.
A malfunction of RXRα due to phosphorylation is associated with colorectal cancer. Therefore, the inhibition of phosphorylation of RXRα and the activation of the RXR–PPARγ heterodimer by their respective ligands may be useful in the chemoprevention and/or treatment of colorectal cancer.
Peroxisome proliferator‐activated receptor γ (PPARγ) is a member of the nuclear hormone receptor superfamily that has been shown to play an important role in the regulation of lipid homeostasis, energy metabolism, inflammatory responses and the induction of apoptosis.1,2 Recent studies have indicated that the aberrant expression of PPARγ has been observed in several types of human malignancies, including colorectal cancer.3,4,5 Moreover, both in vivo and in vitro studies have demonstrated that PPARγ agonists can inhibit cell growth, cause apoptosis and thus exert antitumour effects in human colon cancer.5,6,7 These reports suggest that the activation of PPARγ by its ligand may be a potentially useful strategy for the chemoprevention and/or treatment of colorectal cancer.
PPARs are ligand‐activated transcription factors. As a result, after ligand binding, PPARγ can regulate gene expression by binding to the peroxisome proliferator‐responsive element (PPRE) in target genes as a heterodimer with retinoid X receptors (RXRs).8,9 RXR consists of three subtypes, α, β and γ.10 Among them, we previously found a malfunction of RXRα due to post‐translational modification by phosphorylation to be associated with carcinogenesis of the liver.11,12 The phosphorylated form of RXRα (p‐RXRα) lost its transactivation activity via the RXR‐responsive element and interfered with the function of remaining normal RXRα in a dominant‐negative manner, thereby promoting the growth of hepatoma cells.11 Exogenous 9‐cis‐retinoic acid (RA), a natural ligand for RXRs, induced the degradation of RXRα while also restoring the function of this receptor in human hepatoma cells.11 These findings may explain how the retinoid can inhibit the growth of hepatoma cells in both clinical13,14 and in vitro studies,15,16 thereby leading us to hypothesise that the phosphorylation of the RXRα protein may thus play a role in other types of human malignancies, including colorectal cancer.
In view of these observations, there has been considerable interest in utilising the combination of ligands for PPAR and RXR for the prevention and treatment of various types of cancer.17,18,19,20 In the present study, we examined in detail the effects of such combined treatment on the inhibition of cell growth in Caco2 human colon cancer cells which express high levels of phosphorylated RXRα (p‐RXRα) protein. We also transfected Caco2 cells with the mutant RXRα which expressed the unphosphorylated form of RXRα, and we examined the effects of the combination of these ligands on the cyclo‐oxygenase‐2 (COX‐2) expression in these cells, since COX‐2 expression has been shown to play a crucial role in the development of human colorectal cancer.21,22
9‐cisRA and PD98059 (a MEK inhibitor) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Ciglitazone (cig) was from Calbiochem (La Jolla, CA, USA). RPMI 1640 medium and fetal calf serum (FCS) were both from Invitrogen (Carlsbad, CA, USA).
Histidine‐tagged wild‐type RXRα was constructed using pRShRXRα (kindly provided by Dr RM Evans, The Salk Institute, La Jolla, CA, USA), by PCR. The point‐mutated RXRα T82A/S260A, in which both Thr82 and Ser260 are mutated to alanine, which are putative phosphorylation sites for mitogen‐activated protein kinases (MAPKs), and thus mimicking the unphosphorylated form of RXRα protein, has been previously described.11
The Caco2, HT29, colo320, colo201, DLD‐1 and SW837 human colorectal cancer cell lines were obtained from the Japanese Cancer Research Resources Bank (Tokyo, Japan). The HCT‐116 human colorectal cancer cell line and the FHC normal human fetal colon cell line were obtained from ATCC (Manassas, VA, USA). All of the cell lines were maintained in RF10 medium containing RPMI 1640 medium supplemented with 10% FCS. The cells were cultured in an incubator with humidified air at 37°C with 5% CO2.
Four thousand Caco2 cells were seeded into multiple 35 mm diameter dishes in RF10 medium, and 24 h later they were treated with 5 μM 9‐cisRA alone, 10 μM cig alone or a combination of these agents in the presence or absence of 50 μM PD98059 for 24 h. The doses of these reagents were selected according to our previous11 study and Yang et al.20 The numbers of viable cells in replica plates were then counted using the trypan blue dye exclusion method, as previously described.23
Caco2 cells were treated with 5 μM 9‐cisRA alone, 10 μM cig alone or a combination of these agents in the presence or absence of 50 μM PD98059 for 24 h. To quantify the induction of apoptosis, a DNA fragmentation assay was performed using the Cell Death Detection ELISAPLUS kit, (Roche Diagnostics, Indianapolis, IN, USA), according to the manufacturer's instructions, as described previously.16 The TUNEL (terminal deoxynucleotidyltransferase‐mediated dUTP nick‐end labelling) method was also performed using the In situ Cell Death Detection Kit (Roche Diagnostics), which visualises apoptotic cell death by detecting DNA strand breaks in the individual cells, as described previously.23
Total cellular protein was extracted and equivalent amounts of protein were examined by a western blot analysis, as previously described.11 Polyclonal anti‐RXRα (ΔN197) and anti‐COX2 (C‐20) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Polyclonal anti‐c‐Jun (clone 3) antibody was from BD Biosciences (Franklin Lakes, NJ, USA). Monoclonal antibody against glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH) was from Chemicon International (Temecula, CA, USA). An antibody against GAPDH was used as a loading control. The intensities of the blots were quantified using the NIH Image software version 1.61.
Both RNA extraction and a reverse transcription‐PCR (RT‐PCR) analysis were performed as described previously.24 Total RNA was isolated from Caco2 cells using ISOGEN (WAKO Pure Chemical, Osaka, Japan). cDNA was amplified from 1 μg of total RNA using a High Fidelity RNA PCR Kit (TAKARA Biomedicals, Tokyo, Japan), according to the manufacturer's instructions. c‐Jun‐, COX‐2‐ and GAPDH‐specific primer sets and the amplification cycle conditions were as previously described.24,25,26 Using a thermal controller (Programmable Thermal Controller; MJ Research Inc., Watertown, MA, USA), 25‐ and 30‐cycle rounds of PCR were chosen for an optimum analysis of the expression of c‐Jun and COX‐2 mRNAs, respectively. The intensities of the mRNA bands were then quantified using the NIH Image software package version 1.61.
Reporter assays were performed as described previously.11 The PPRE‐luciferase reporter plasmid PPRE3‐TK‐LUC was kindly provided by the late Dr K Umesono (Kyoto University, Kyoto, Japan).27 The AP‐1 promoter luciferase reporter plasmid pAP‐1‐Luc was purchased from Stratagene (La Jolla, CA, USA). Caco2 cells were co‐transfected with a combination of wild‐type or a mutant hRXRα‐expressing plasmid (300 ng/35 mm dish) with PPRE3‐TK‐LUC or pAP‐1‐Luc reporter (750 ng/35 mm dish), along with pRL‐CMV (Renilla luciferase, 100 ng/35 mm dish; Promega, Madison, WI, USA) as an internal standard to normalise the transfection efficiency. Transfections were performed using Lipofectamine Plus reagent (Invitrogen) according to the manufacturer's protocol. After exposure of the cells for 24 h to the transfection mixture, the cells were treated with vehicle, 5 μM 9‐cisRA, 10 μM cig or the combination of these agents for 24 h. Thereafter, the cell lysates were prepared and the luciferase activity of each cell lysate was determined using a dual‐luciferase reporter assay system (Promega). Changes in the firefly luciferase activity were calculated and plotted after normalisation with changes in Renilla luciferase activity in the same sample.
Colon cancer and its surrounding non‐cancerous colon tissues were obtained by surgical resections from eight patients. This study was approved by the Ethics Committee of the Gifu University School of Medicine, and all of the patients gave their written informed consent.
The data are expressed as the mean (SD). Statistical significance of the difference in the mean values was assessed with one‐way analysis of variance (ANOVA), followed by Sheffe's t test.
In our initial study we examined whether or not the RXRα protein is constitutively phosphorylated in a series of human colorectal cancer cell lines and the FHC normal human fetal colon cell line (fig 11).). The anti‐RXRα (ΔN197) antibody, which can be regarded as the specific antibody for the phosphorylated form of RXRα protein, was utilized for the western blot analysis.11,28 We found that the level of the p‐RXRα protein was constitutively increased in all seven colorectal cancer cell lines (Caco2, HT29, Colo201, Colo320, DLD‐1 HCT‐116 and SW837) that we examined in this study. On the other hand, the p‐RXRα protein was not detected in the FHC cell line (fig 11,, lanes 1). We also found a marked decrease in the level of this protein when the cells were treated with the MEK inhibitor PD98059 alone (fig 11,, lanes 3) or the combination of 9‐cisRA plus PD98059 (fig 11,, lanes 4) in all of these colon cancer cells.
We next examined whether 9‐cisRA and the PPARγ agonist cig inhibit the phosphorylation of RXRα in Caco2 cells, which express both the RXRα and PPARγ mRNAs.29 There was no remarkable change in the expression level of p‐RXRα when the cells were treated with 9‐cisRA alone, cig alone or the combination of these ligands in the absence of PD98059 (fig 22,, lanes 1–4). However, when the cells were treated with these ligands in the presence of PD98059, there was a decrease in the expression level of p‐RXRα proteins (fig 22,, lanes 6 and 7). This decrease was more apparent when the cells were treated with a combination of all three agents (fig 22,, lane 8).
COX‐2 may thus play a critical role in the development of colorectal cancer.21 We therefore next examined whether the combined treatment with 9‐cisRA plus cig inhibits the expression of COX‐2 and c‐Jun, a component of AP‐1 which plays a role in the expression of COX‐2 at the level of transcription,30,31 in Caco2 colon cancer cells (fig 33).). In the presence of PD98059, 9‐cisRA or cig caused a small decrease in the levels of COX‐2 (fig 3A3A,, lanes 6 or 7, lower panel) and c‐Jun (fig 3B3B,, lanes 6 or 7, lower panel) proteins. The combination of these three agents further inhibited the expression of these proteins (fig 3A and BB,, lanes 8, lower panel). The decreases in COX‐2 and c‐Jun proteins were paralleled by similar decreases in their mRNAs (fig 3A and BB,, upper panels). Moreover, as shown in fig 3C and DD,, similar effects for the expression levels of both COX‐2 and c‐Jun proteins and mRNAs were also observed in mutant RARα T82A/S260A‐transfected Caco2 cells which do not express the p‐RXRα protein (data not shown). We previously reported that MAPKs are involved in RXRα phosphorylation in human HCC cells.11,15 Therefore, these findings suggest that the phophorylation of the RXRα protein by MAPKs may thus interfere with the function of the ligands for RXRα and PPARγ and that the induction of the unphosphorylated (ie, functional) form of RXRα intensifies the responses to ligands of RXRα and PPARγ, thereby affecting the expression of target molecules, including COX‐2 and c‐Jun.
PPARγ regulates gene expression by binding to the PPRE element in target genes as a heterodimer with RXRs and thus controlling growth of cells.8 We then examined the effects of the combination of 9‐cisRA plus cig on the transcriptional activity of the PPRE and AP‐1 promoters, using transient transfection luciferase reporter assays. As shown in fig 4A4A,, left panel, we found that in mutant RARα T82A/S260A‐transfected Caco2 cells, the treatment with this combination caused a synergistic increase in the transcriptional activity of the PPRE luciferase reporter activity. Treatment with either 9‐cisRA alone or cig alone also caused significant inhibition of the AP‐1 luciferase reporter activity, and this inhibition was thus enhanced when the cells were treated with a combination of these agents (fig 4A4A,, right panel). Moreover, in the presence of PD98059, treatment with 9‐cisRA or cig significantly inhibited the growth of Caco2 cells, and the combined treatment with all three agents caused even stronger growth inhibition (fig. 4B4B,, lanes 6–8, left panel). Similar growth inhibitory effects by treatment with these ligands were also observed in mutant RARα T82A/S260A‐transfected Caco2 cells (fig 4B4B,, lanes 6–8, right panel). In addition, as shown in fig 4C4C,, the DNA fragmentation assays (left panel) and TUNEL (terminal deoxynucleotidyltransferase‐mediated dUTP nick‐end labelling) methods (right panel) indicated that, in the presence of PD98059, the combination of 9‐cisRA plus cig caused a significant induction of apoptosis in Caco2 cells (lane 8 of left panel and photograph 8 of right panel). These findings indicate that under conditions in which the phosphorylation of RXRα is reduced, the suppression of cell growth by combined treatment with 9‐cisRA plus cig is associated with apoptosis.
We next examined whether the p‐RXRα protein is present in human colorectal cancer tissues. Figure 55 shows the results of a western blot analysis of the nuclear extracts prepared from eight pairs of human colorectal cancer tissue and adjacent normal colon epithelial tissue. We thus found that the RXRα protein was highly phosphorylated in six of the eight samples of colorectal cancer tissues when compared with their corresponding normal colon epithelial tissue (fig 55),), although there were no significant differences in the expression levels of the unphosphorylated form of RXRα proteins between cancer tissues and adjacent normal colon epithelial tissues (data not shown).
The present studies provide us with the first evidence that the RXRα protein is phospholylated in both a series of human colon cancer cell lines and primary human colon cancer tissue samples, although the levels of expression of p‐RXRα did not increase in FHC normal colonic epithelial cells ((figsfigs 1 and 55).). These findings are consistent with our previous findings in human hepatoma cell lines and in hepatoma tissue samples,11 and suggest that, as in hepatoma, a malfunction of RXRα due to phosphorylation might be associated with the development of colorectal cancer. In addition, we also found that the combination of 9‐cisRA and cig caused a synergistic inhibition in the growth of human colon cancer Caco2 cells (fig 4B4B)) and this was associated with the induction of apoptosis (fig 4C4C)) and inhibition of the expression of both COX‐2 and c‐Jun proteins and mRNAs (fig 33).). The combination of these agents had a synergistic effect in increasing the PPRE activity and decreasing the AP‐1 activity (fig 4A4A).). These preferable effects were observed when the phosphorylation of RXRα protein was either inhibited by MEK inhibitor or transfected with point‐mutated RXRα, which mimicked the unphosphorylated form. Therefore, the inhibition of the phosphorylation of this protein appears to play a critical role in inducing the growth inhibitory effect presented in this study.
A general question posed by the present study is the molecular mechanisms by which the combination of 9‐cisRA and cig synergistically induce these diverse cellular and biochemical effects in colon cancer cells. A hypothetical scheme that addresses this question is shown in fig 66.. In normal colonic epithelial cells, activated PPARγ forms a heterodimer with RXRα and binds to the PPRE in promoter regions of target genes.8 However, in colon cancer cells, the Ras/MAPK signalling pathway is highly activated and phosphorylates RXRα, thus impairing the functions of this receptor. The MEK inhibitor PD98059 or the point‐mutated RXRα T82A/S260A prevent the phosphorylation of RXRα, thus causing an increase in the unphosphorylated form of RXRα protein and restoring the heterodimeric activity with PPARγ in the presence of their lignads, 9‐cisRA and cig. A synergistic effect on target gene expression was originally reported with the optimal interaction of RXRα–PPARγ heterodimers.8 Tsao et al32 also reported that in HEK293 cells the transcriptional activity of PPRE was additively induced by treatment with a PPARγ activator plus 9‐cisRA, and RXRα accumulation through inhibiting its degradation by the proteasome system contributes to the enhancement of PPARγ/RXR activation. It is of interest that the PPAR and RXR ligands have been shown to recruit subsets of transcriptional coactivators to the receptor complex differentially, thus leading to an enhanced transcriptional activation and cellular effects.33 The findings of these reports support our finding that the accumulation of the unphosphorylated form (ie, functional form) of RXRα enhances the transcriptional activity of PPRE (fig 4A4A,, left panel) and it may also explain the growth inhibitory effects of the combined treatment with RXR and PPAR ligands, which are observed in various types of cancer cells.17,18,19,20
There is considerable evidence that COX‐2 may play a critical role in the development of colorectal cancer and might, therefore, be an important molecular target for colorectal cancer prevention and treatment.21,22 In this study, we found that the combined use of ligands for RXRα and PPARγ inhibited COX‐2 expression in Caco2 cells (fig 3A and CC).). This finding may be of significance because the inhibition of COX‐2 expression is one of the mechanisms by which the PPARγ pathway induces apoptosis.20 Presumably, the inhibition of the transcriptional activity of the AP‐1 promoter by the combined action of these agents (fig 4A4A,, right panel) contributes to the inhibition of COX‐2 expression because AP‐1 plays an important role in activating the expression of this molecule.30,31 However, it is likely that some other unknown mechanisms also play a role. For example, in the 2,4,6‐trinitrobenzene sulphonic acid (TNBS)‐induced rat colitis model, administration of a PPARγ agonist reduced the expression of COX‐2 and nuclear transcription factor NF‐κB p65 proteins.34 Both NF‐κB and AP‐1 play an important role in regulating COX‐2 expression via the NF‐κB‐binding site in the COX‐2 promoter.35 Moreover, TNBS‐induced colitis was significantly reduced by the administration of both PPARγ and RXR agonists, and this beneficial effect was also reflected by the reduction in the NF‐κB DNA binding activity in the colon.36 PPARγ agonist has been shown to inhibit chemically induced colitis and an early phase of colitis‐related colon carcinogenesis in the rodent model.37 It also seems to be of interest that the inhibition of the β‐catenin pathway, which promotes the development of colon cancer, by non‐steroidal anti‐inflammatory drugs, requires a high level of expression of RXRα and PPARγ.38 As a result, the activation of the RXR–PPARγ heterodimer by the co‐administration of their ligands is thus considered to protect against colonic inflammation and tumorigenesis. In conclusion, these previous reports, together with the results of our in vitro mechanistic studies on colon cancer cells as described herein, suggest that combination therapy with both RXR and PPARγ ligands might therefore be clinically useful in the prevention and/or treatment of both colon cancer and colonic inflammation, due to their synergistic effects.39
cig - ciglitazone
COX - cyclo‐oxygenase
FCS - fetal calf serum
GAPDH - glyceraldehyde‐3‐phosphate dehydrogenase
MAPK - mitogen‐activated protein kinase
PPAR - peroxisome proliferator‐activated receptor
PPRE - peroxisome proliferator‐responsive element
p‐RXRα - phosphorylated RXR, retinoid X receptor α
RA - retinoic acid
RT‐PCR - reverse transcription‐PCR
RXR - retinoid X receptor
TNBS - 2,4,6‐trinitrobenzene sulphonic acid
TUNEL - terminal deoxynucleotidyltransferase‐mediated dUTP nick‐end labelling
This work was supported in part by Grants‐in‐Aid from the Ministry of Education, Science, Sports and Culture of Japan (No. 17015016 to HM and No. 18790457 to MS).
Competing interests: None.