Curcumin: a novel nutritionally derived ligand of the vitamin D receptor with implications for colon cancer chemoprevention

Abstract

The nuclear vitamin D receptor (VDR) mediates the actions of 1,25-dihydroxyvitamin D(3) (1,25D) to regulate gene transcription. Recently, the secondary bile acid, lithocholate (LCA), was recognized as a novel VDR ligand. Using reporter gene and mammalian two-hybrid systems, immunoblotting, competitive ligand displacement and quantitative real-time PCR, we identified curcumin (CM), a turmeric-derived bioactive polyphenol, as a likely additional novel ligand for VDR. CM (10(-5) M) activated transcription of a luciferase plasmid containing the distal vitamin D responsive element (VDRE) from the human CYP3A4 gene at levels comparable to 1,25D (10(-8) M) in transfected human colon cancer cells (Caco-2). While CM also activated transcription via a retinoid X receptor (RXR) responsive element, activation of the glucocorticoid receptor (GR) by CM was negligible. Competition binding assays with radiolabeled 1,25D confirmed that CM binds directly to VDR. In mammalian two-hybrid assays employing transfected Caco-2 cells, CM (10(-5) M) increased the ability of VDR to recruit its heterodimeric partner, RXR, and steroid receptor coactivator-1 (SRC-1). Real-time PCR studies revealed that CM-bound VDR can activate VDR target genes CYP3A4, CYP24, p21 and TRPV6 in Caco-2 cells. Numerous studies have shown chemoprotection by CM against intestinal cancers via a variety of mechanisms. Small intestine and colon are important VDR-expressing tissues where 1,25D has known anticancer properties that may, in part, be elicited by activation of CYP-mediated xenobiotic detoxification and/or up-regulation of the tumor suppressor p21. Our results suggest the novel hypothesis that nutritionally-derived CM facilitates chemoprevention via direct binding to, and activation of, VDR.

Fig. 1

Curcumin activates transcription of an XDR3 reporter construct in a concentration dependent fashion. Human colon cancer cells (Caco-2) were transfected with a firefly luciferase plasmid containing two copies of the distal vitamin D responsive element, XDR3, from the human cytochrome P450 (CYP) 3A4 gene. The cells were treated for 30 hours in complete media with 10⁻⁸ M 1,25D (black bar; D), 10⁻⁴ M lithocholic acid (grey bar; LCA), and two concentrations of curcumin (grey bar; CM). Ethanol was the vehicle for D and LCA, while curcumin was reconstituted in DMSO. The data were normalized for transfection efficiency as described in Methods and Materials and expressed as percent of 1,25D-stimulated transcription. The mean fold effect of each ligand was calculated with respect to the appropriate vehicle and is shown above each bar; error bars represent standard deviation, and asterisks indicate statistically significant differences (p < 0.05) from the vehicle control. The data represent the mean of three independent experiments.

Fig. 2

Evaluation of curcumin in the mammalian two hybrid system. (A) Curcumin induces heterodimerization of VDR and RXR. Human colon cancer cells, Caco-2, were transfected with expression vectors encoding RXRα bait (BD) and VDR prey (AD) fusion constructs. A firefly luciferase reporter vector (pFR-luc) and Renilla luciferase control plasmid were also introduced into the cells to measure the amount of ligand-stimulated VDR-RXR interaction and to measure transfection efficiency, respectively. The cells were treated for 30 hours in complete media with 10⁻⁸ M 1,25D (black bar; D), 10⁻⁴ M lithocholic acid (grey bar; LCA) and two concentrations of curcumin (grey bar; CM). Negative controls included the use of BD-empty and AD-empty expression vectors which resulted in low background levels of luciferase activity (not shown). (B) Curcumin induces heterodimerization of VDR and SRC-1. Caco-2 cells were transfected with expression vectors encoding SRC-1 bait (BD) and VDR prey (AD) fusion constructs. The cells were treated for 30 hours as in A. The data were normalized for transfection efficiency and expressed as a percent of 1,25D-stimulated interaction between VDR and RXR (A) or VDR and SRC-1 (B), with error bars indicating standard deviation. Differences statistically significant from control are indicated by asterisks as described in the legend to Fig. 1.

Fig. 3

Specificity test of curcumin binding to the retinoid X receptor and glucocorticoid receptor. Human Caco-2 cells were transfected with appropriate firefly luciferase reporter constructs and nuclear receptor expression plasmids, as well as Renilla luciferase to normalize for transfection efficiency. The cells were treated with the indicated combinations of ligands for 24 hours and a dual luciferase assay was performed. A total of three independent experiments were carried out with each reporter construct. (A) Transcriptional activation by VDR. The cells received VDR expression vector and the luciferase reporter plasmid containing two copies of the XDR3. The results were normalized for transfection efficiency and plotted as a percent of 1,25D-stimulated transcription. (B) Transcriptional activation by RXR. The cells received an RXR expression vector and the luciferase reporter plasmid containing the RXRE from the rat cellular retinol binding protein, type II, gene. The results were normalized for transfection efficiency and plotted as a percent of rexinoid LG101305 (Rex)-stimulated transcription. (C) Transcriptional activation by GR. A GR expression vector and the luciferase reporter plasmid containing two copies of the GRE from the rat tyrosine aminotransferase gene were introduced into the cells. The results were normalized for transfection efficiency and plotted as a percent of dexamethasone (Dex)-stimulated transcription. Error bars represent standard deviation.

Fig. 4

Evaluation of VDR mRNA and protein levels in 1,25D- and curcumin-treated cells. (A) Real-time PCR. Rat osteoblast-like osteosarcoma cells (ROS 17/2.8) cells were treated with 10⁻⁷ M 1,25D, 5×10⁻⁵ M CM, or a combination of both compounds for 24 hours in order to evaluate potential regulation of VDR mRNA. Relative levels of VDR mRNA were measured using quantitative real-time PCR as described in Methods and Materials. (B) Western blots. ROS 17/2.8 were treated as in A, followed by preparation of cell lysates in a solution of 2% SDS, 0.125 M Tris-HCl, pH 6.8 and 20% glycerol. The protein content of the lysates was determined using a BCA assay, and 100 μg of total protein from each sample were run on a 5%–15% gradient SDS-polyacrylamide gel, followed by Western blotting with an antibody (9A7γ) directed against VDR (see Methods and Materials). This result is representative of three independent experiments.

Fig. 5

Ability of curcumin to compete with 1,25D for binding to VDR. Competition curves display the concentration range in which curcumin is able to compete for binding to VDR with ≈ 4.0×10⁻¹⁰ M [³H]1,25D. Dexamethasone (Dex) is a GR ligand and has no appreciable binding to VDR, thus serving as a non-competing negative control. Lithocholic acid (LCA) was included as a positive (competing) control. This plot was generated in Prism4 (GraphPad Software, Inc.) and is representative of three independent experiments.

Fig. 6

Curcumin regulates CYP24 (25-hydroxyvitamin D 24-hydroxylase) in human Caco-2 cells in a dose-dependent fashion. (A) Results of a real-time PCR experiment using human CYP24 primers and cDNA prepared from Caco-2 cells incubated for 4 h with the indicated concentration of CM or ethanol vehicle. Results are plotted as ΔRn vs. cycle number where ΔRn is the reporter dye signal normalized to the passive reference dye, and Rn is Rn from which the baseline dye signal has been subtracted. (B) Induction of CYP24 mRNA by the indicated ligands as compared to the ethanol control. Data were obtained using Ct values from real time PCR as described in A; error bars represent triplicate determinations ± SD. (C) Western blotting of CYP24 protein detected in cell lysates from Caco-2 cells incubated with the indicated ligand.

Fig 7

1,25D and curcumin modulate Caco-2 cell migration. Caco-2 cells were grown in MEM to 90% confluency. Cells were starved in serum-free media for 24 h, and a linear lesion (“scratch”) was generated by removing the monolayer with a sterile cell scraper along a line drawn on each plate. The cells were then incubated with MEM+1% FBS and either ethanol (EtOH), 10⁻⁷ M 1,25D (+1,25D), or 10⁻⁴ M curcumin (+CM) for 24 h. Cell migration was assessed under an inverted phase contrast microscope (magnification: 40X).