Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2007 Feb 23.
Published in final edited form as: Oncogene. 2006 Feb 23;25(8):1225–1241. doi: 10.1038/sj.onc.1209160

Oxidative metabolism of linoleic acid modulates PPAR-beta/delta suppression of PPAR-gamma activity

Xiangsheng Zuo 1, Yuanqing Wu 1, Jeffrey S Morris 2, Julie B Stimmel 3, Lisa M Leesnitzer 3, Susan M Fischer 4, Scott M Lippman 1, Imad Shureiqi 1,5,*
PMCID: PMC1432095  NIHMSID: NIHMS5014  PMID: 16288226

Abstract

Peroxisome proliferator–activated receptors (PPARs) are transcription factors that strongly influence molecular events in normal and cancer cells. PPAR-beta/delta overexpression suppresses the activity of PPAR-gamma and -alpha. This interaction has been questioned, however, by studies with synthetic ligands of PPARs in PPAR-beta/delta–null cells, and it is not known whether an interaction between PPAR-beta/delta and -gamma exists, especially in relation to the signaling by natural PPAR ligands. Oxidative metabolites of linoleic and arachidonic acids are natural ligands of PPARs. 13-S-hydroxyoctadecadienoic acid (13-S-HODE), the main product of 15-lipoxygenase-1 (15-LOX-1) metabolism of linoleic acid, downregulates PPAR-beta/delta. We tested (a) whether PPAR-beta/delta expression modulates PPAR-gamma activity in experimental models of the loss and gain of PPAR-beta/delta function in colon cancer cells and (b) whether 15-LOX-1 formation of 13-S-HODE influences the interaction between PPAR-beta/delta and PPAR-gamma. We found that (a) 15-LOX-1 formation of 13-S-HODE promoted PPAR-gamma activity, (b) PPAR-beta/delta expression suppressed PPAR-gamma activity in models of both loss and gain of PPAR-beta/delta function, (c) 15-LOX-1 activated PPAR-gamma by downregulating PPAR-beta/delta , and (d) 15-LOX-1 expression induced apoptosis in colon cancer cells via modulating PPAR-beta/delta suppression of PPAR-gamma. These findings elucidate a novel mechanism of the signaling by natural ligands of PPARs, which involves modulating the interaction between PPAR-beta/delta and PPAR-gamma.

Introduction

Peroxisome proliferator-activated receptors (PPARs) are transcriptional factors that have considerable influence on important molecular events in normal and cancer cells (Kliewer et al., 1999; Michalik et al., 2004). The possibility of an interaction between various PPARs was raised by study results showing that PPAR-beta/delta (PPAR-b/d) overexpression inhibits the activities of PPAR-gamma (PPAR-g) and those of PPAR-alpha (PPAR-a) in nontransformed normal monkey kidney CV-1 cells and mouse NIH 3T3 fibroblasts (Shi et al., 2002). Other studies, however, have challenged this concept of PPAR-b/d influence on the activation of PPAR-g or PPAR-a because of findings that PPAR-g or PPAR-a activation by synthetic ligands was not significantly affected in PPAR-b/d –null cells (Matsusue et al., 2004; Peters et al., 2003). While these conflicting findings might be secondary to differences between the experimental models used (e.g., overexpression vs. knockout models), another important explanation to explore is that findings with the use of synthetic ligands might not represent what occurs biologically with the natural ligands of PPARs that are formed intracellularly. Oxidative metabolites of arachidonic and linoleic acids are well-known natural ligands of PPARs (Chawla et al., 2001) that can be used to evaluate whether such interaction between PPARs occurs intracellularly as a part of biologic signaling pathways.

To examine these important questions—whether PPAR-b/d interacts with other PPARs and whether this interaction between PPAR-b/d and other PPARs is relevant to the signaling of oxidative metabolites of polyunsaturated fatty acids—we used experimental models of both PPAR-b/d gain and loss of function in colon cancer cells that express both PPAR-g and PPAR-b/d (Gupta et al., 2000; He et al., 1999; Sarraf et al., 1998). These complementary experimental models allowed us to examine whether PPAR-b/d interaction with PPAR-g is related to experimental model differences that might have accounted for the conflicting results observed previously by others.

For a natural intracellular PPAR-b/d ligand, we used 13-S-hydroxyoctadecadienoic acid (13-S-HODE), the primary product of 15-lipoxygenase-1 (15-LOX-1) metabolism of linoleic acid, which binds to and modulates the activity of PPAR-b/d (Shureiqi et al., 2003). 13-S-HODE also activates PPAR-g when it is exogenously added in micromolar concentrations (10–30 μM) to the culture medium of cells, including colon cancer cells (Bull et al., 2003; Nagy et al., 1998; Nixon et al., 2003). Thus, the question arises whether 13-S-HODE binding to PPAR-b/d might influence 13-S-HODE activation of PPAR-g. According to the current literature, however, 15-LOX-1 endogenous formation of 13-S-HODE activates PPAR-g in mouse macrophages and fibroblasts (Huang et al., 1999) but not in human colon cancer cells (Nixon et al., 2003). Therefore, to address this conflict in the previously reported data, we initially examined whether the endogenous formation of 13-S-HODE by 15-LOX-1 can activate PPAR-g in colon cancer cells. Our findings in this study demonstrated that PPAR-b/d suppresses PPAR-g activity in both PPAR-b/d gain and loss of function models and that the oxidative metabolites of polyunsaturated fatty acids, as in the case of 13-S-HODE, modulate this interaction to activate PPAR-g.

Results

15-LOX-1 formation of endogenous 13-S-HODE increases PPAR-g activity in colon cancer cells

The adenovirus vector that carries human 15-LOX-1 cDNA (Ad-15-LOX-1) efficiently expressed 15-LOX-1 in colon cancer cells. Ad-15-LOX-1 transfection induced expression of 15-LOX-1 in HCT-116 and LoVo colon cancer cells, whereas 15-LOX-1 was not expressed in control cells transfected with the same adenoviral vector that carried the LacZ gene (Ad-LacZ) instead of 15-LOX-1 (Figures 1A and B). Expression of 15-LOX-1 increased 13-S-HODE levels in colon cancer cells (Figures 1C and D); adding linoleic acid to the culture medium enhanced 13-S-HODE production in the 15-LOX-1–transfected cells, whereas adding caffeic acid (a 15-LOX-1 inhibitor) in a concentration that specifically inhibits 15-LOX-1 enzymatic activity (Shureiqi et al., 2000) reduced their 13-S-HODE production (Figures 1C and D). 13-S-HODE formation in LoVo cells, which express 15-LOX-1, increased in a manner that was dependent on the concentrations of linoleic acid added to the medium (P < 0.0001). Caffeic acid significantly reduced 13-S-HODE formation in Ad-15-LOX-1–transfected LoVo cells (P < 0.0001). Similar effects (increased 13-S-HODE formation with Ad-15-LOX-1 transfection and decreased 13-S-HODE formation with caffeic acid) occurred in HCT-116 cells (P < 0.0001).

Fig. 1.

Fig. 1

Open in a new tab

Effects of 15-LOX-1 expression on PPAR-g activation in colorectal cancer cells. (A and B) Expression of 15-LOX-1 in (A) LoVo and (B) HCT-116 colon cancer cells via Ad-15-LOX-1 adenovirus vector (Ad-15-LOX-1) transfections. Both cell lines were transfected with Ad-15-LOX-1 or control LacZ adenovirus vector (Ad-LacZ) and analyzed by Western blotting for 15-LOX-1 expression. Lanes: S, standard recombinant human 15-LOX-1; Ad-15-LOX-1, cells transfected with Ad-15-LOX-1; Ad-LacZ, cells transfected with the control vector Ad-LacZ; 24h, cells harvested 24 h after transfection; 36h, cells harvested 36 h after transfection; 48h, cells harvested 48 h after transfection; and N, nontransfected cells. (C and D) Effects of 15-LOX-1 expression on 13-S-HODE production in LoVo and HCT-116 colon cancer cells. LoVo (C) and HCT-116 (D) cells were transfected with either Ad-15-LOX-1 or Ad-LacZ and cultured in medium supplemented with linoleic or caffeic acid as follows: LA0, no additional linoleic acid added; LA5, 5-μM linoleic acid added; LA10, 10-μM linoleic acid added; and CAF2.2, 2.2-μM caffeic acid added. Cells were harvested 48 h after transfection for measuring 13-S-HODE levels by enzyme immunoassay. Values shown are the means ± SDs of triplicate experiments. (E and F) Effects of 15-LOX-1 expression on PPAR-g activity. LoVo (E) and HCT-116 (F) cells were first transfected with either Ad-15-LOX-1 or Ad-LacZ; cells were transfected with the luciferase reporter system (AOX)3-TK-Luc 24 h later and with either pcDNA3.0-PPAR-g expression vector or control vector (empty pcDNA3.0). PPAR-g transcriptional activity was measured 24 h after the second transfection and normalized to β-galactosidase activity (relative to 100,000 β-galactosidase units). The culture medium was supplemented with linoleic acid as indicated. pcDNA3.0–PPAR-g vector transfection is denoted by +PPAR-g. Values shown are the means ± SDs of triplicate experiments. (G–J) Effects of 15-LOX-1 expression on the expression of PPAR-g endogenous target genes. LoVo (G and I) and HCT-116 (H and J) cells were transfected with Ad-15-LOX-1 or Ad-LacZ, treated with DMSO (Control) or with 2.2-μM caffeic acid and harvested 48 hour later. Expression of CD36 and Keratin 20 was measured using real-time PCR (see Methods). The relative expression levels were calculated as the values relative to that of the calibrator sample (Ad-LacZ–transfected cells). Values shown are the means ± SDs of triplicate experiments.

Next we examined whether 15-LOX-1 expression in colon cancer cells affects PPAR-g transcriptional activity, which we monitored with a PPAR-g–dependent luciferase reporter construct, (AOx)3-TK-Luc, with and without ectopic PPAR-g expression (Huang et al., 1999). 15-LOX-1 expression markedly increased PPAR-g activity, and this activation increased further with the addition of linoleic acid to the culture medium (Figures 1E and F). In LoVo cells, 15-LOX-1 transfection increased PPAR-g activity by a factor of 4 or more (P < 0.0001). Linoleic acid increased this activation in a concentration-dependent manner (P < 0.0001), whereas Ad-LacZ transfection, either alone or with linoleic acid supplementation, did not increase PPAR-g activity. Similar effects occurred in HCT-116 cells: 15-LOX-1 significantly increased PPAR-g activity, and linoleic acid further enhanced this activity (P < 0.0001). Transfection of both LoVo and HCT-116 cells with PPAR-g dominant-negative expression vector markedly inhibited PPAR-g activity relative to that in cells transfected with either control vector (empty vector) or the same vector that carried PPAR-g cDNA (P < 0.0001 for both LoVo and HCT-116 cells) (Supplementary Figures 1 and 2).

Fig. 2.

Fig. 2

Open in a new tab

Effects of PPAR-b/d expression on PPAR-g activation in colon cancer cells. (A and B) Effects of PPAR-b/d siRNA on PPAR-b/d expression in colon cancer cells, as measured by real-time PCR. LoVo (A) and HCT-116 (B) cells were harvested 48 h after transfection with a pool of four SMARTselected siRNA duplexes for PPAR-beta/delta (PPAR-b/d siRNA) or PPAR-a (PPAR-alpha siRNA) or with a nonspecific siRNA sequence (Non-specific siRNA). Samples were processed for PPAR-b/d mRNA measurements by real-time PCR. The relative expression levels were calculated as the values relative to that of the calibrator sample (nonspecific siRNA transfection). Values shown are the means ± SDs of triplicate experiments. (C) Effects of PPAR-b/d siRNA on PPAR-b/d protein expression, as measured by Western blotting. Lanes: 1, positive control (RKO rectal cancer cells transfected with pcDNA3.0–PPAR-b/d vector); 2, LoVo cells transfected with PPAR-b/d siRNA; 3, LoVo cells transfected with PPAR-alpha siRNA; 4, LoVo cells transfected with the nonspecific siRNA sequence; 5, mock-transfected LoVo cells (“mock” refers to transfection with the transfection medium alone); 6, HCT-116 cells transfected with PPAR-b/d siRNA; 7, HCT-116 cells transfected with PPAR-alpha siRNA; 8, HCT-116 cells transfected with the nonspecific siRNA sequence; and 9, mock-transfected HCT-116 cells. (D and E) Effects of PPAR-b/d on PPAR-g activity in colon cancer cells. LoVo (D) and HCT-116 (E) cells were first transfected with PPAR-b/d siRNA, nonspecific-siRNA, or PPAR-alpha siRNA and then transfected 24 h later with the PPAR-g–dependent luciferase reporter system (AOx)3-TK-Luc and with either a pcDNA3.0–PPAR-g expression vector (PPAR-gamma Vector) or with an empty pcDNA3.0 vector (Control Vector). PPAR-g activity was measured 24 h after the second transfections. Values shown are the means ± SDs of triplicate experiments. (F and G) Ectopic PPAR-b/d expression in KO1 cells (PPAR-b/d–null HCT-116 cells) and PPAR-g activity. KO1 cells were transfected with pcDNA3.0–PPAR-b/d vector (PPAR-b/d vector), and PPAR-b/d expression was measured by (F) real-time PCR and (G) Western blotting. (F) KO1 cells transfected with either pcDNA3.0 empty vector (KO1) or a PPAR-b/d vector (KO1-PPAR-b/d) were harvested and processed for measurement of PPAR-b/d mRNA by real-time PCR, as in A and B, except that the calibrator sample was the KO1 cells’ parental HCT-116 cells. (G) Lanes: 1, positive control (RKO cells transfected with a PPAR-b/d vector); 2, KO1 cells transfected with pcDNA3.0 empty (control) vector; and 3, KO1 cells transfected with a PPAR-b/d vector. (H) Effects of PPAR-b/d expression on PPAR-g activation in KO1 cells. KO1 cells were transfected with a PPAR-b/d vector, and PPAR-g activity was measured using the luciferase reporter system as described in D and E. Values shown are the means ± SDs of triplicate experiments. (I-K) PPAR-b/d effects on PPAR-g endogenous target gene expression. LoVo (I) and HCT-116 (J) cells were transfected with a PPAR-b/d siRNA or a PPAR-alpha siRNA or with a nonspecific siRNA sequence; cells were harvested after 48 h, and Keratin 20 expression was measured by real-time PCR (see Methods). The relative expression levels were calculated as the values relative to that of the calibrator sample (Non-specific siRNA). Values are means ± SDs of triplicate experiments. (K) Effects of reconstituting PPAR-b/d expression in HCT-116 PPAR-b/d knockout (KO1) cells on PPAR-g target genes. Samples from KO1 cells transfected with either pcDNA3.0–PPAR-b/d expression vector (KO1-PPAR-b/d) or the control vector (same vector without PPAR-b/d cDNA) (KO1) and from the parental KO1 (HCT-116) cells were processed for real-time PCR measurement of Keratin 20, as in I and J. The relative expression levels were calculated as the values relative to that of the calibrator sample (KO1). Values shown are the means ± SDs of triplicate experiments.

We also tested whether 15-LOX-1 enzymatic activity was involved in the increased PPAR-g activity after 15-LOX-1 transfection. We inhibited 15-LOX-1 enzymatic activity with caffeic acid, which significantly reduced PPAR-g activity in the 15-LOX-1–transfected cells (P < 0.0001; data not shown). 15-LOX-1 inhibition had no statistically significant effects on PPAR-g activation in Ad-LacZ–transfected cells (P = 0.55).

We later tested the effect of 15-LOX-1 expression on two endogenous PPAR-g target genes (CD36 and Keratin 20) to confirm the biologic relevance of our earlier findings. Among the PPARs, PPAR-g selectively induced Keratin 20 expression in colon cancer cells (Gupta et al., 2001). 15-LOX-1 increased CD36 and Keratin 20 expression in LoVo and HCT-116 cells, and inhibition of 15-LOX-1 enzymatic activity suppressed that increased expression. Ad-15-LOX-1 transfection of LoVo cells increased the expression of CD36 by a factor of 2.5 (P < 0.0001) (Figure 1G) and Keratin 20 by a factor of 3 (P < 0.0001) relative to that in the control (Ad-LacZ)-transfected cells (Figure 1I). Inhibition of 15-LOX-1 enzymatic activity blocked those increases in expression levels. Similarly, in HCT-116 cells, 15-LOX-1 increased the expression of both CD36 (P = 0.0003) and Keratin 20 (P = 0.0002), whereas inhibiting 15-LOX-1 enzymatic activity blocked these effects (Figures 1H and J).

15-LOX-1 expression had no effect on PPAR-g expression in either HCT-116 or LoVo cells (data not shown), demonstrating that 15-LOX-1 did not activate PPAR-g transcription by increasing the level of PPAR-g expression.

PPAR-b/d expression modulates PPAR-g activity in colorectal cancer cells

We used experimental models of both PPAR-b/d loss and gain of function in colon cancer cells that express both PPAR-g and PPAR-b/d to examine whether an interaction exists between PPAR-b/d and PPAR-g. Because colon cancer cells overexpress PPAR-b/d (Gupta et al., 2000; He et al., 1999), we initially evaluated this relationship between PPAR-b/d and PPAR-g using experimental models of PPAR-b/d downregulation. We used the siRNA approach to specifically downregulate PPAR-b/d in HCT-116 and LoVo cells; PPAR-b/d siRNA suppressed PPAR-b/d mRNA and protein expression in both cell lines (Figures 2A–C). PPAR-b/d siRNA downregulated PPAR-b/d mRNA expression by approximately 89% in LoVo cells (P < 0.0001) and 85% in HCT-116 cells (P < 0.0001) as measured by real-time polymerase chain reaction (PCR) assays (Figures 2A and B). Western blot analyses showed marked downregulation of PPAR-b/d by PPAR-b/d siRNA in both cell lines (Figure 2C). The resulting PPAR-b/d downregulation increased PPAR-g activity in both LoVo (Figure 2D) and HCT-116 (Figure 2E) cells and did so significantly more than mock (with medium only), nonspecific siRNA, or PPAR-a siRNA transfections did (P < 0.0001).

As a complementary approach, we examined the effects of reintroducing PPAR-b/d expression on PPAR-g in HCT-116 cells in which PPAR-b/d was knocked out (KO1 cells) (Park et al., 2001). Transfection of KO1 cells with PPAR-b/d expression vector was successful in reestablishing PPAR-b/d expression in KO1 cells, as measured by real-time PCR (P < 0.0001) (Figure 2F) and Western blot analyses (Figure 2G). The expression of PPAR-b/d in KO1 cells inhibited PPAR-g activation (P < 0.0001) (Figure 2H). In contrast, expression of PPAR-a through the same vector in KO1 cells failed to decrease PPAR-g activity (data not shown).

PPAR-b/d downregulation increased Keratin 20 expression in LoVo and HCT-116 colon cancer cells by factors of 3.3 (P < 0.0001) and 2.8 (P = 0.0034), respectively (Figures 2I and J). Furthermore, in KO1 cells, Keratin 20 expression levels were 5.3 times higher than they were in their parental HCT-116 cells (P < 0.0001). Transient transfection of KO1 cells with PPAR-b/d expression vector significantly reduced Keratin 20 expression by approximately 35% (P < 0.0001) (Figure 2K).

Mechanisms of PPAR-b/d modulation of PPAR-g activity

Next we investigated the mechanisms underlying the effect of PPAR-b/d expression on suppressing PPAR-g activity. PPAR-b/d downregulation, induced by siRNA transfection in HCT-116 and LoVo cells, failed to increase PPAR-g mRNA (Figures 3A and B) or protein expression (Figure 3C), thus indicating that PPAR-b/d exerts its suppressive effect on PPAR-g via mechanisms other than modulation of PPAR-g expression.

Fig. 3.

Fig. 3

Open in a new tab

PPAR-b/d effects on PPAR-g expression, ligand and DNA binding. (A and B) Effects of PAPR-d downregulation on PPAR-g RNA expression. LoVo (A) and HCT-116 (B) cells were transfected with a pool of four SMARTselected siRNA duplexes for PPAR-b/d (PPAR-b/d siRNA), a PPAR-alpha (PPAR-alpha siRNA) or with a nonspecific siRNA sequence (Non-specific siRNA) and PPAR-g RNA expression was measured by real-time PCR 48 hours later. The relative expression levels were calculated as the values relative to that of the calibrator sample (Non-specific siRNA). Values shown are the means ± SDs of triplicate experiments. (C) Effects of PPAR-b/d downregulation on PPAR-g protein expression. PPAR-b/d expression was downregulated using PPAR-b/d siRNA transfection, and then PPAR-g protein expression was measured by Western blotting. Lanes: 1, positive control (HCT-116 cells transfected with pcDNA3.0–PPAR-g); 2, HCT-116 cells transfected with PPAR-b/d siRNA; 3, HCT-116 cells transfected with PPAR-alpha siRNA; 4, HCT-116 cells transfected with the nonspecific siRNA sequence; 5, mock-transfected HCT-116 cells; 6, LoVo cells transfected with PPAR-b/d siRNA; 7, LoVo cells transfected with PPAR-alpha siRNA; 8, LoVo cells transfected with the nonspecific siRNA sequence; and 9, mock-transfected LoVo cells. (D) Effects of PPAR-b/d expression on PPAR-g activation by specific PPAR-g ligand , troglitazone. KO1 cells (PPAR-b/d–null HCT-116 cells) were transfected with either pcDNA3.0–PPAR-b/d expression vector (PPAR-b/d vector) or the empty pcDNA3.0 vector (control vector), then 24 h later with the PPAR-g–dependent luciferase reporter system (AOx)3-TK-Luc and with a pcDNA3.0–PPAR-g expression vector. Cells were treated with troglitazone (0.5 μM) 6 hours after the second transfection. PPAR-g activity was measured 24 h after the second transfections. Values represent the increase in the relative promoter activity level by troglitazone compared to control vehicle (DMSO) treated cells. Values shown are the means ± SDs of triplicate experiments. LoVo (E) and HCT-116 (F) cells were transfected with a PPAR-b/d siRNA or with a non-specific siRNA; treated with troglitazone (TZD) (1 μM) or 13-S-HODE (13.5 μM) 24 hours later and PPAR-g activity was measured at 48 h as described for figure D, values represent the increase in relative promoter activity between the TZD or 13-S-HODE and control vehicle (DMSO) treated cells (G) and (H) LoVo and HCT-116 were transfected with PPAR-b/d siRNA or non-specific siRNA and treated with troglitazone (TZD), 13-S-HODE, or linoleic acid (13.5 μM) similar to what described for figure E and F; cells were harvested after 48 h, and Keratin 20 expression was measured by real-time PCR. The relative expression levels were calculated as the values relative to that of the calibrator sample (control vehicle (DMSO) treated cells). Values are means ± SDs of triplicate experiments. (I) K01 and HCT-116 cells were transfected with 15-LOX-1 adenoviral vector (Ad-15-LOX-1 vector) then 24 h later with the luciferase reporter system (AOx)3-TK-Luc and with a pcDNA3.0–PPAR-g expression vector. PPAR-g activity was measured at 48 hours from the time of the first tasnfection. Vaules are means ± SDs of triplicate experiments. (J) Scintillation proximity assay of 13-S-HODE binding to PPAR-g. (K) Effects of PPAR-b/d on PPAR-g’s binding to a DNA recognition sequence. Gel EMSAs were performed using the PPAR-g DNA recognition sequence ACO labeled with 32P and incubated with in vitro–translated RXR-a, PPAR-b/d, and PPAR-g as indicated. “Probe only” is the negative control. Fixed amounts of RXR-a (0.5 μl of 58.3-μg/μl lysate) and PPAR-g (g; 0.5 μl of 56.4-μg/μl lysate) were used. PPAR-b/d (b/d) was added in increasing amounts (1.5, 3.0, 6.0, and 9.0 μl of 53.9-μg/μl lysate), as indicated by the black wedges. PPAR-g monoclonal antibody was added for supershift assays. DNA binding was analyzed by gel electrophoretic mobility shift and supershift assays. (L-M) PPAR-b/d effects on PPAR-g binding to a target promoter region in cells. (L) HCT-116 cells were transfected with PPAR-b/d siRNA and KO1 cells with a PPAR-b/d expression vector respectively. Cells were formaldehyde cross-linked 48 h after transfection and subjected to a ChIP assay using a PPAR-g antibody. A 230-bp fragment of the human LPL promoter was amplified by PCR as described in Methods. Lanes: 1, HCT-116 cells; 2, KO1 cells (PPAR-b/d knockout HCT-116 cells) transfected with control vector; 3, KO1 cells transfected with a PPAR-b/d expression vector; 4, HCT-116 cells transfected with the nonspecific siRNA sequence; 5, HCT-116 cells transfected with PPAR-alpha siRNA; and 6, HCT-116 cells transfected with PPAR-b/d siRNA. “PPAR-g antibody” indicates immunoprecipitations with PPAR-g antibody, and “Input” indicates total DNA before immunoprecipitation. (M) K01 cells were transfected with either PPAR-b/d vector or control vector, treated with either 13-S-HODE (13.5 μM), linoleic acid (13.5 μM) or control vehicle (DMSO), and then subjected to ChIP assay using PPAR-b/d antibody, as in Figure 3L.

We examined whether the interaction between PPAR-b/d and PPAR-g might be related to competition for binding to a common ligand since both PPARs are ligand-activated transcription factors and PPAR-b/d might inhibit PPAR-g by competing for ligands it has in common with PPAR-g. We first used a specific PPAR-g ligand, troglitazone, to examine whether the interaction between PPAR-b/d and -g would be altered in the setting of a PPAR-g–specific binding ligand. Restoration of PPAR-b/d expression by ectopic expression of PPAR-b/d in KO1 cells inhibited PPAR-g activation by troglitazone (P = 0.0017) (Figure 3D). In HCT-116 and LoVo cells, which express PPAR-b/d, downregulation of PPAR-b/d with siRNA markedly increased PPAR-g activation by troglitazone (P <0.0001) (Figures 3E and F).

We further evaluated this question in the setting of a natural PPAR ligand, 13-S-HODE, which reportedly binds to both PPAR-g and PPAR-b/d. PPAR-b/d downregulation increased PPAR-g activation with troglitazone but not with 13-S-HODE (Figures 3E and F). Troglitazone and 13-S-HODE significantly increased expression of Keratin 20, the PPAR-g–specific target gene, in LoVo and HCT-116 cells; however, PPAR-b/d downregulation significantly increased Keratin 20 expression with troglitazone but significantly decreased it in 13-S-HODE–treated cells (P = 0.0001) (Figures 3G and H). Expression of 15-LOX-1 in KO1 and HCT-116 cells increased PPAR-g activity to a higher level in HCT-116 than in KO1 cells (P = 0.0018) (Figure 3I). These data questioned whether 13-S-HODE acts as a binding ligand for PPAR-g. Scintillation proximity assays of the direct binding of 13-S-HODE to PPAR-g showed that 13-S-HODE has no binding affinity for PPAR-g (Figure 3J). These findings did not support the notion that the interaction between PPAR-b/d and PPAR-g is secondary to competition for ligand binding.

We then examined whether PPAR-b/d inhibits PPAR-g DNA binding as a transcriptional factor. PPARs bind DNA to activate transcription by forming heterodimers with retinoid X receptor alpha (RXR-a) (Gearing et al., 1993; Kliewer et al., 1992). Thus, we assessed PPAR-b/d ability to inhibit binding of the PPAR-g/RXR-a heterodimer to the PPAR-g DNA recognition sequence in the acyl-coA oxidase gene promoter (ACO) (He et al., 1999) in electrophoretic mobility-shift (EMSA) and supershift assays. The PPAR response element ACO binds efficiently to PPAR-a and PPAR-g but not to PPAR-b/d (He et al., 1999). Although PPAR-b/d binding to ACO at various concentrations (4–24 μg/μl of EMSA incubation solution) was weaker than that of PPAR-g at a much lower concentration (1.4 μg/μl of EMSA incubation solution), PPAR-b/d reduced the PPAR-g/RXR-a heterodimer’s binding to the ACO sequence at high concentrations in EMSA and supershift assays (Figure 3K), a condition that would simulate PPAR-b/d overexpression in cancer cells (He et al., 1999). We also quantitatively examined the effects of PPAR-b/d on PPAR-g binding to the peroxisome proliferator response element (PPRE) (Renard et al., 2001). PPAR-b/d alone changed the optical density reading minimally, confirming the assay’s specificity in measuring PPAR-g binding to PPRE (Supplementary Figure 3). When PPAR-b/d was added with PPAR-g, PPAR-b/d significantly decreased PPAR-g binding to the PPRE (P < 0.0001) (Supplementary Figure 3).

We also examined whether PPAR-b/d affects PPAR-g binding to target gene promoters in vivo. This analysis involved chromatin immunoprecipitation (ChIP) assays in experimental models of PPAR-b/d loss of function (KO1 cells and HCT-116 cells transfected with PPAR-b/d siRNA) and PPAR-b/d gain of function (PPAR-b/d ectopic expression in KO1 cells). No amplification product of a PPAR-g target region in the human lipoprotein lipase (LPL) promoter was detected with immunoprecipitated chromatin from parental HCT-116 cells or KO1 cells in which PPAR-b/d was ectopically expressed (Figure 3L, lanes 1 and 3). In contrast, the same fragment of the LPL promoter was amplified from immunoprecipitated chromatin of KO1 cells and their parental HCT-116 cells in which PPAR-b/d was downregulated by PPAR-b/d siRNA (Figure 3L, lanes 2 and 6). Immunoprecipitated chromatin from HCT-116 cells transfected with either nonspecific siRNA or PPAR-a siRNA showed no amplification of the same LPL promoter region (Figure 3L, lanes 4 and 5). Troglitazone promoted PPAR-g binding to LPL (Supplementary Figure 4). Furthermore, in a ChIP assay using immunoprecipitation with a PPAR-b/d antibody, PPAR-b/d binding to LPL was detected when PPAR-b/d expression was restored in KO1 cells (Figure 3M). 13-S-HODE, but not its parent compound (linoleic acid), inhibited PPAR-b/d binding to LPL (Figure 3M).

Fig. 4.

Fig. 4

Open in a new tab

Effects of PPAR-b/d on 15-LOX-1–induced PPAR-g activation in colorectal cancer cells. (A) Relationship between 15-LOX-1 and PPAR-b/d expression. LoVo and HCT-116 cells were transfected with either Ad-15-LOX-1 or Ad-LacZ and with either pcDNA3.0–PPAR-b/d (PPAR-b/d Vector) or pcDNA3.0 empty vector (Control Vector), and PPAR-b/d expression was measured by Western blotting. Lanes: S, standard PPAR-b/d positive control (RKO rectal cancer cells transfected with pcDNA3.0–PPAR-b/d vector); the other lanes are as labeled. (B and C) Effects of PPAR-b/d and 15-LOX-1 expression on PPAR-g binding to the PPRE in LoVo and HCT-116 cells. LoVo (B) and HCT-116 (C) cells were first transfected with Ad-15-LOX-1 or Ad-LacZ and then 6 h later, with PPAR-b/d or control empty vectors. Cells were harvested 36 h after the second transfection, and the nuclear extracts were prepared and assayed (6 μg nuclear protein/well) for PPAR-g binding to the PPRE using specific PPAR-g antibody in an enzyme-linked immunosorbent–based assay. Values shown are the means ± SDs. LoVo (D) and (E) HCT-116 were transfected with either Ad-lacZ or Ad-15-LOX-1. Ad-lacZ transfected cells were treated with either DMSO (control) or 13-S-HODE (13.5 μM). PPAR-g binding to PPRE was measured as in Figures B and C. (F) LoVo cells and (G) HCT-116 cells were transfected with either Ad-15-LOX-1 or Ad-LacZ and then with PPAR-b/d, empty vector (Control Vector), PPAR-g dominant negative vector, or PPAR-gamma expression vector as described in Supplementary Data Figure 1 and 2. PPAR-g activity was measured using the PPAR-g–dependent luciferase reporter system (AOx)3-TK-Luc. Values shown are the means ± SDs of triplicate experiments. (H and I) Effects of PPAR-b/d expression on 15-LOX-1–induced expression of PPAR-g endogenous target genes. LoVo (H) and HCT-116 (I) cells were transfected with Ad-15-LOX-1 and then with control vector, PPAR-b/d vector, PPAR-g dominant negative vector, or PPAR-gamma expression vector (PPAR-g vector). Keratin 20 expression was measured using real-time PCR (see Methods), and its relative expression levels were calculated as the values relative to that of the calibrator sample (Ad-15-LOX-1 + control vector). Values shown are the means ± SDs of triplicate experiments.

15-LOX-1 modulates PPAR-g activity via downregulating PPAR-b/d

15-LOX-1 downregulates PPAR-b/d expression and activity in colon cancer cells (Shureiqi et al., 2003) but increases PPAR-g activation, as we observed earlier in this study. To examine whether these two effects of 15-LOX-1 are mechanistically linked, we ectopically overexpressed PPAR-b/d in HCT-116 and LoVo cells to overcome its downregulation by 15-LOX-1 and found that the expression of PPAR-b/d compensated for that downregulation (Figure 4A). 15-LOX-1 expression significantly increased (P = 0.0034 for LoVo; P = 0.01 for HCT-116), whereas PPAR-b/d blocked PPAR-g binding to the PPRE in both cell lines (P = 0.0028 for LoVo; P = 0.0013 for HCT-116) (Figures 4B and C).

We also examined whether the effects of 15-LOX-1 on PPAR-g activation were mediated through its main product, 13-S-HODE. We supplemented Ad-LacZ (control)–transfected colon cancer cells with either 13-S-HODE or linoleic acid and compared PPAR-g binding to the PPRE in these cells with that in cells expressing 15-LOX-1 via Ad-15-LOX-1. We found that 13-S-HODE, but not linoleic acid (the parent compound before 15-LOX-1 metabolism), significantly increased PPAR-g binding to the PPRE in both HCT-116 and LoVo cells (P = 0.007 and P = 0.004, respectively). 13-S-HODE significantly increased PPAR-g binding to the PPRE in Ad-LacZ–transfected (compared with control-treated) HCT-116 and LoVo cells (P = 0.0006 and P = 0.04, respectively) (Figures 4D and E). The level of increase in PPAR-g binding to PPRE was similar to levels in Ad-15-LOX-1–transfected cells. These findings indicated that 15-LOX-1 effects on PPAR-g are mediated through 13-S-HODE formation.

PPAR-b/d also inhibited the 15-LOX-1–mediated increase of PPAR-g activity, as measured using the (AOx)3-TK-Luc reporter system (Figures 4F and G). PPAR-b/d expression reduced 15-LOX-1–associated PPAR-g activation in LoVo cells by 65%, and PPAR-g dominant-negative expression reduced it by 70% (Figure 4F) (P < 0.0001 for both comparisons). Similar results occurred in HCT-116 cells: PPAR-b/d expression reduced 15-LOX-1–associated activation of PPAR-g by 52%, and PPAR-g dominant-negative expression reduced it by 56% (Figure 4G) (P < 0.0001 for both comparisons). PPAR-b/d and PPAR-g dominant-negative expression reduced 15-LOX-1 activation of PPAR-g to similar levels in both cell lines (Figures 4F and G).

15-LOX-1 increased the expression of Keratin 20 and CD36 in LoVo and HCT-116 cells, whereas PPAR-b/d ectopic expression blocked those effects of 15-LOX-1 (Figures 4H and I for Keratin 20; data not shown for CD36). Similarly, ectopic PPAR-g expression increased Keratin 20 expression in both cell lines, whereas PPAR-b/dominant negative suppressed 15-LOX-1’s ability to increase that expression.

15-LOX-1 induces apoptosis via downregulation of PPAR-b/d and activation of PPAR-g

The expression of 15-LOX-1 inhibited the growth of HCT-116 and LoVo colorectal cancer cells by inducing apoptosis, as confirmed by caspase-3 activity and DNA fragmentation assays (Figures 5A–D). 15-LOX-1 expression increased casapase-3 activity by a factor of approximately 2 relative to that in LacZ-transfected cells (Figures 5C and D). Casapase-3 activity was significantly higher in Ad-15-LOX-1–transfected cells than it was in Ad-LacZ–transfected cells (P < 0.002 for LoVo; P = 0.0265 for HCT-116). Furthermore, ectopic expression of PPAR-b/d or PPAR-g dominant negative significantly inhibited apoptosis induction in HCT-116 and LoVo cells (Figures 5E and F). 15-LOX-1 expression significantly increased caspase-3 activity, whereas PPAR-b/d overexpression or PPAR-g dominant-negative expression significantly reduced caspase-3 activity in 15-LOX-1–transfected cells (P = 0.0072 for LoVo; P = 0.0215 for HCT-116). Caspase-3 activity was similar in cells co-transfected with Ad-15-LOX-1 and either PPAR-b/d or PPAR-g dominant-negative vectors (P = 0.4166 for LoVo; P = 0.5512 for HCT-116).

Fig. 5.

Fig. 5

Open in a new tab

15-LOX-1 induction of apoptosis and the interaction between PPAR-b/d and PPAR-g in colon cancer cells. (A) Light microscopic images of LoVo cells (left) and HCT-116 cells (right) after transfection with either control adenoviral vector (Ad-LacZ) or 15-LOX-1 expression adenoviral vector (Ad-15-LOX-1); “Mock” refers to nontransfected cells. Photomicrographs were taken 50 h after transfection (original magnification ×200). (B) Effects of 15-LOX-1 expression on apoptosis induction in LoVo and HCT-116 cells, as measured by DNA laddering. Cells were harvested 50 h after transfection. Lanes: DNA Ladder, standard DNA ladder; Ad-15-LOX-1, cells transfected with Ad-15-LOX-1; Ad-LacZ, cells transfected with Ad-LacZ; Mock, nontransfected cells. (C and D) Effects of 15-LOX-1 expression on apoptosis induction in LoVo and HCT-116 cells, as measured by the caspase-3 enzymatic activity assay. LoVo (C) and HCT-116 (D) cells were transfected with either Ad-LacZ or Ad-15-LOX-1. Cells were harvested 48 h later and processed for measuring caspase-3 activity levels. Values shown are the mean ± SDs of triplicate experiments. (E and F) Effects of PPAR-b/d and PPAR-g on 15-LOX-1–induced apoptosis. LoVo (E) and HCT-116 (F) cells were transfected with either Ad-15-LOX-1 or Ad-LacZ, and 12 h later, they were transfected with PPAR-b/d expression vector (PPAR-b/d Vector), PPAR-g dominant negative expression vector, or control vector. Cells were harvested 36 h after the second transfection and processed for the caspase-3 enzymatic activity assay as in C and D. (G and H) Conceptual models of the 15-LOX-1 signaling pathway through PPAR-b/d and PPAR-g in colon cancer cells illustrate (G) the effects of PPAR-b/d expression on PPAR-g activity in cancer cells in which PPAR-b/d is overexpressed and 15-LOX-1 expression is lost and (H) the effects of restoring 15-LOX-1 expression on the interaction between PPAR-b/d and PPAR-g.

Discussion

We found that (a) 15-LOX-1 expression increased PPAR-g activity in colon cancer cells, (b) PPAR-b/d expression decreased PPAR-g activation (likely through modulation of PPAR-g DNA binding,) without altering PPAR-g expression or ligand binding in colon cancer cells, (c) 15-LOX-1 downregulation of PPAR-b/d expression contributed significantly to 15-LOX-1 activation of PPAR-g in colon cancer cells, and (d) 15-LOX-1 induced apoptosis in colorectal cancer cells by modulating the suppressive effects of PPAR-b/d on PPAR-g.

Our results demonstrated for the first time that endogenous formation of 13-S-HODE by 15-LOX-1 promotes PPAR-g activation in human colon cancer cells. Our following findings confirm that 15-LOX-1 activation of PPAR-g occurred via 13-S-HODE production: (a) expressed 15-LOX-1 significantly increased 13-S-HODE formation in colon cancer cells, (b) adding linoleic acid (the substrate of 13-S-HODE) significantly increased PPAR-g activation in colon cancer cells that express, but not in those that do not express, 15-LOX-1, and (c) adding caffeic acid at concentrations known to specifically inhibit 15-LOX-1 enzymatic activity (Shureiqi et al., 2000) reduced both 13-S-HODE formation and PPAR-g activation. These results indicate that 15-LOX-1 expression induced PPAR-g activation via its enzymatic activity (i.e., production of 13-S-HODE).

These findings concur with prior results obtained from activating PPAR-g via exogenous 13-S-HODE in various cell lines, including colon cancer cell lines (Bull et al., 2003; Nagy et al., 1998), and obtained from expressing 15-LOX-1 in mouse cell lines (Huang et al., 1999). However, our results that endogenous formation of 13-S-HODE in colon cancer cells activates PPAR-g counters previously reported findings in colon cancer cell lines (Nixon et al., 2003). This discrepancy may result from the use of different assays. The previous study (Nixon et al., 2003) measured 15-LOX-1 effects on activation of the PPAR-g ligand-binding domain, whereas we directly measured PPAR-g transcriptional activity. We have confirmed that our measured 15-LOX-1 effects on PPAR-g activity via the luciferase-reporter system (AOx)3-TK-Luc were specific for PPAR-g activity by co-expressing 15-LOX-1 with PPAR-g, PPAR-g dominant negative, or both. Ectopic PPAR-g expression increased the measured luciferase activity through the (AOx)3-TK-Luc system; this increase in luciferase activity was thus attributed to PPAR-g. The increased luciferase activity resulting from ectopic PPAR-g expression was further amplified (3 to 5 fold) by the expression of 15-LOX-1, indicating that 15-LOX-1 increased PPAR-g activity. The measured increase in PPAR-g activity by 15-LOX-1 expression through the (AOx)3-TK-Luc system was strongly suppressed by the ectopic expression of PPAR-g dominant negative; this finding further confirmed that 15-LOX-1 expression specifically increased PPAR-g activity.

Furthermore, we found that 15-LOX-1 expression markedly increased the endogenous expression of the PPAR-g target genes CD36 and Keratin 20 in colon cancer cells. PPAR-g is the only member of the PPAR family that promotes Keratin 20 expression in colon cancer cells (Gupta et al., 2001). These findings also confirmed that 15-LOX-1 expression in colon cancer cells increases PPAR-g activity. We previously found that ectopic 15-LOX-1 expression in colorectal cancer cells inhibited PPAR-b/d activity (Shureiqi et al., 2003). Therefore, our findings indicate that 15-LOX-1 expression has a differential effect on PPARs, promoting PPAR-g activity while inhibiting PPAR-b/d.

CD36 can induce apoptosis in non-endothelial cells (Febbraio et al., 2001; Jimenez et al., 2000; Rusinol et al., 2000), and Keratin 20 expression is associated with intestinal epithelial cell differentiation (Calnek & Quaroni, 1993; Moll et al., 1993). Therefore, our findings that 15-LOX-1 promoted the expression of PPAR-g target genes such as CD36 and Keratin 20, which contribute to cell maturation and apoptosis, indicate possible downstream molecular events that are modulated by 15-LOX-1 signaling via PPAR-g to inhibit tumorigenesis (Nixon et al., 2004; Shureiqi et al., 2003; Shureiqi & Lippman, 2001; Shureiqi et al., 1999; Girnun et al., 2002; Osawa et al., 2003; Sarraf et al., 1998).

PPAR-b/d inhibited the activities of PPAR-g and PPAR-a in experimental models of PPAR-b/d overexpression in nontransformed normal monkey kidney CV-1 cells and mouse NIH 3T3 fibroblasts (Shi et al., 2002). In contrast, exogenous synthetic ligands of the PPARs produced no significant alterations in PPAR-g activity in experimental models of PPAR-b/d or PPAR-a knockout (Matsusue et al., 2004; Peters et al., 2003). These contradictory results may reflect differences between the experimental models used, but the exact causes remained to be defined. Therefore, we examined whether PPAR-b/d expression can influence PPAR-g activity in models of PPAR-b/d loss and gain of function. We found that PPAR-b/d downregulation with the siRNA approach increased PPAR-g activation in colon cancer cells. The effects of PPAR-b/d expression on PPAR-g activity were specific to PPAR-b/d because we found that downregulation of PPAR-a expression failed to influence PPAR-g activity.

In complementary experiments, we examined the effects of reconstituting PPAR-b/d expression on PPAR-g activity in PPAR-b/d–null HCT-116 cells (derived from the parental HCT-116 cells used in our PPAR-b/d downregulation experiments). We found that PPAR-b/d expression in PPAR-b/d–null HCT-116 cells inhibited PPAR-g activity, supporting our current findings that downregulation of PPAR-b/d increased PPAR-g activity in the parental HCT-116 and LoVo cells. Our results concur with those of Shi et al. (Shi et al., 2002), who showed that overexpression of PPAR-b/d inhibits the activity of PPAR-g. Furthermore, our findings showed for the first time that downregulation of PPAR-b/d increases PPAR-g activity. Therefore, in models of both PPAR-b/d loss and gain of function, PPAR-b/d expression suppressed PPAR-g activity. The biologic significance of the interaction between PPAR-b/d and PPAR-g was demonstrated by the ability of this interaction to influence the expression of a specific endogenous PPAR-g target gene, Keratin 20. Downregulation of PPAR-b/d expression in LoVo and HCT-116 cells that express PPAR-b/d significantly increased Keratin 20 expression. In concurrence with these findings, Keratin 20 expression levels in KO1 cells were approximately 5.3 times higher than in their parental HCT-116 cells, and reconstitution of PPAR-b/d expression in KO1 cells significantly reduced the expression of Keratin 20.

We examined the following possible mechanisms for the interaction between PPAR-b/d and PPAR-g: a) PPAR-b/d modulates PPAR-g expression; b) PPAR-b/d competes for ligand binding with PPAR-g; and c) PPAR-b/d interferes with PPAR-g DNA binding. Downregulation of PPAR-b/d expression failed to increase PPAR-g expression, indicating that PPAR-b/d inhibition of PPAR-g activation does not occur through modulation of PPAR-g expression. Regarding the possibility that the interaction might be mediated through PPAR-b/d competing with PPAR-g for ligand binding, we found that PPAR-g activation by troglitazone, a specific PPAR-g ligand (Gupta et al., 2001), was inhibited in association with PPAR-b/d expression and increased in association with PPAR-b/d downregulation. These finding indicates that PPAR-b/d inhibition of PPAR-g activity is unlikely to be mediated via competition for ligand binding because PPAR-g activation by a specific PPAR-g ligand was influenced by the expression of PPAR-b/d, which does not bind to troglitazone.

We further examined the possibility of competition between PPAR-b/d and -g for ligand binding with regard to 13-S-HODE, a natural ligand that has been reported to bind to both PPAR-g (Nagy et al., 1998) and PPAR-b/d (Shureiqi et al., 2003). Unlike the case with troglitazone, downregulation of PPAR-b/d failed to increase PPAR-g activation by 13-S-HODE. This finding is directly opposed to what would be expected if the PPAR-b/d and PPAR-g interaction were mediated via binding to common ligands, in which case decreased PPAR-b/d expression should increase PPAR-g binding and activation by 13-S-HODE. This finding also raised the possibility that 13-S-HODE does not activate PPAR-g through direct binding but does so indirectly through PPAR-b/d downregulation. Indeed, KO1 cells that lack PPAR-b/d expression had significantly lower PPAR-g activation with 15-LOX-1 expression than did their parental colon cancer cell line, which expresses PPAR-b/d. This further supported the concept that 13-S-HODE, the main product of 15-LOX-1, depends on PPAR-b/d downregulation, at least in part, to activate PPAR-g indirectly. We previously reported that the Kd for 13-S-HODE binding to PPAR-b/d is 10.8 μM (Shureiqi et al, PNAS, 2003). The 13-S-HODE binding constant to PPAR-g has not been formally determined but has been estimated to be in the 10–20-μM range (Nagy et al., 1998). We measured the 13-S-HODE binding constant to PPAR-g in a scintillation proximity assay and found that 13-S-HODE has a negligible affinity to PPAR-g even in concentrations reaching 100 μM. These findings further support the concept that 13-S-HODE does not activate PPAR-g through direct binding but more likely through an indirect mechanism such as downregulating PPAR-b/d.

We next examined the possibility that the interaction between PPAR-b/d and PPAR-g is mediated by PPAR-b/d modulating PPAR-g DNA binding. For this analysis, we used in vitro gel shift and super shift assays and in vivo ChIP assays in colon cancer cells with and without PPAR-b/d expression. For the in vitro EMSA and supershift assays, we used the PPAR response element ACO, which preferentially binds more to PPAR-a and PPAR-g than to PPAR-b/d (He et al., 1999). PPAR-b/d at various concentrations (4–24 μg/μl of EMSA incubation solution) bound weakly to ACO, whereas PPAR-g bound strongly to ACO at a much lower concentration (1.4 μg/μl of EMSA incubation solution), thus confirming PPAR-b/d low affinity for binding to ACO. When the same respective concentrations of PPAR-b/d and PPAR-g were combined, PPAR-b/d inhibited PPAR-g DNA binding at high concentrations in both gel-shift and supershift assays, simulating what occurs in cancer cells that overexpress PPAR-b/d. We also examined the effects of PPAR-b/d expression on PPAR-g binding to a target promoter region in vivo. This analysis involved a previously validated ChIP assay to measure PPAR-g binding to the LPL promoter (Fajas et al., 2002b). PPAR-b/d expression inhibited the binding of PPAR-g to its target gene promoter region in HCT-116 cells that naturally express PPAR-b/d and in PPAR-b/d knockout HCT-116 (i.e., KO1) cells in which PPAR-b/d expression was reconstituted. Furthermore, PPAR-g binding to the same promoter region was restored when PPAR-b/d expression was transiently downregulated by PPAR-b/d siRNA in HCT-116 or when PPAR-b/d was permanently deleted in KO1 cells. In later studies, using ChIP assays, we found that PPAR-b/d can bind to the PPAR-g target DNA sequence in the LPL promoter and that this binding can be inhibited by 13-S-HODE but not by its parent compound, linoleic acid. These findings support the conclusion that PPAR-b/d inhibits PPAR-g binding to its DNA recognition sites and that this interaction can be modulated through 13-S-HODE, which downregulates PPAR-b/d expression.

Our findings that PPAR-b/d suppressed PPAR-g activity conflict with those of another group, who studied the effects of synthetic PPAR ligands in nontransformed PPAR-b/d–null cells (Matsusue et al., 2004; Peters et al., 2003). These differences might be attributed to variation in the experimental models used, including the basal levels of PPAR-b/d expression in different cells (e.g., in transformed vs. nontransformed cells). Cancer cells, in contrast with normal cells, commonly overexpress PPAR-b/d (Gupta et al., 2000; He et al., 1999; Stephen et al., 2004), and the interaction between PPAR-b/d and PPAR-g may depend on the degree of PPAR-b/d expression: high levels of PPAR-b/d may be required to competitively interfere with PPAR-g activity such as DNA binding. Therefore, knockout models of PPAR-b/d in normal cells may not be optimal for studying the interaction between PPAR-b/d and PPAR-g, which more likely occurs when PPAR-b/d is overexpressed. The previous studies used only a knockout model of PPAR-b/d in nontransformed cells treated only with synthetic ligands; whereas we used several complementary experimental systems of PPAR-b/d loss and gain of function in colon cancer cells and tested synthetic and natural ligands to study the effects of modulating PPAR-b/d expression on the interaction between PPAR-b/d and PPAR-g. These systems included the genetic knockout of PPAR-b/d, e.g., in KO1 cells (versus their parental HCT-116 cells); the downregulation of PPAR-b/d by the siRNA approach in colon cancer cells; and the reconstitution of PPAR-b/d expression in KO1 cells. Our results in these complementary experimental systems confirmed that PPAR-b/d expression suppressed PPAR-g activity.

We have also observed differences in the mechanisms of PPAR-g activation between a synthetic ligand, troglitazone, and a natural PPAR ligand such as 13-S-HODE. The ability of 13-S-HODE to activate PPAR-g indirectly via PPAR-b/d downregulation demonstrates a novel mechanism for natural ligand activation of PPAR-g indirectly through modulation of PPAR-b/d expression.

Taken together, our results indicate that PPAR-b/d can negatively affect PPAR-g activity and that this inhibitory effect might be mediated through PPAR-b/d inhibition of PPAR-g binding to its DNA recognition sequences. This ability of PPAR-b/d expression to influence PPAR-g activity should be considered when interpreting results of studies of the biologic effects of PPAR-b/d overexpression or downregulation in experimental models.

Adding 13-S-HODE to cell culture medium activated PPAR-g through PPAR-b/d downregulation. We also examined whether restoring 15-LOX-1 expression to colon cancer cells, which leads to 13-S-HODE formation intracellularly, activates PPAR-g through PPAR-b/d downregulation. The ectopic overexpression of PPAR-b/d blocked 15-LOX-1 from downregulating PPAR-b/d expression and inhibited 15-LOX-1 from activating PPAR-g. The ability of PPAR-b/d overexpression to suppress PPAR-g activation by 15-LOX-1 was similar to that of the expression of a PPAR-g dominant-negative sequence. This similarity supports the concept that 15-LOX-1 increases PPAR-g activity by modulating the suppressive effects of PPAR-b/d on PPAR-g. Furthermore, the ability of 15-LOX-1 to promote the induction of PPAR-g endogenous target genes such as CD36 and Keratin 20 was inhibited by expression of PPAR-b/d. PPAR-g dominant-negative expression in this experimental system produced similar effects to those of PPAR-b/d overexpression, further supporting the conclusion that 15-LOX-1 signaling through PPAR-g to influence molecular/gene transcription events in cells is influenced by 15-LOX-1 effects on PPAR-b/d.

The biologic significance of 15-LOX-1 effects on the interaction between PPAR-b/d and PPAR-g in colon cancer cells was further supported by our findings that 15-LOX-1 expression induced apoptosis and that these effects depended on modulation of the interaction between PPAR-b/d and PPAR-g. We previously found that colorectal cancer cells lack 15-LOX-1 expression and that exogenous replacement of the 15-LOX-1 product 13-S-HODE in colorectal cancer cells induces apoptosis (Shureiqi et al., 1999). The findings from our new studies demonstrate for the first time that genetically reconstituting enzymatically active 15-LOX-1 by ectopic expression in colorectal cancer cells induces apoptosis. Furthermore, the ability of 15-LOX-1 to induce apoptosis in these cells was inhibited by the ectopic expression of either PPAR-g dominant negative or PPAR-b/d. These findings demonstrate, also for the first time, that 15-LOX-1 induction of apoptosis can be mediated by modulating PPAR-b/d suppressive effects on PPAR-g.

In conclusion, our present and previous findings support a new model of the 15-LOX-1 signaling pathway in colon cancer cells (Figures 5G and H), which is an example of the ability of oxidative metabolism of polyunsaturated fatty acids to modulate the interaction between PPAR-b/d and PPAR-g. In colon cancer cells, 15-LOX-1 expression is lost and PPAR-b/d is overexpressed, resulting in the suppression of PPAR-g transcriptional activity (Figure 5G). However, restoring 15-LOX-1 expression produces 13-S-HODE, which downregulates PPAR-b/d expression and thus promotes PPAR-g activity (Figure 5H). As with any conceptual model, our proposed model may somewhat oversimplify highly complex interactive biologic events. Future studies will be needed to further elucidate how the signaling of polyunsaturated fatty acid oxidative metabolic pathways modulates the interaction between PPARs to influence important biologic events such as apoptosis in cells.

Materials and Methods

Materials

We obtained the HCT-116 parental PPAR-b/d–wild-type (PPAR-b/d +/+) colon cancer cell line, the HCT-116 PPAR-b/d–null cell line (PPAR-b/d −/−; HCT-116 cells in which PPAR-b/d was knocked out [KO1 cells]) (Park et al., 2001), rabbit anti-human PPAR-b/d antibody, rabbit polyclonal antiserum to recombinant human 15-LOX-1, and 13-S-HODE and linoleic acid (formulated in dimethylsulfoxide [DMSO]) as described previously (Shureiqi et al., 2003). The LoVo colon cancer cell line was obtained from the American Type Culture Collection (Manassas, VA). Mouse anti-human PPAR-g monoclonal antibody (E-8) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Other reagents, molecular-grade solvents, and chemicals were obtained from commercial manufacturers or as specified.

Cell cultures

HCT-116 and KO1 cells were grown in McCoy modified medium, and LoVo cells were grown in RPMI 1640 medium at 37 °C in 5% CO2. Both media contained 10% FBS and were supplemented with 1% penicillin–streptomycin as described previously (Shureiqi et al., 2003). Cells were treated with 13.5-μM 13-S-HODE or linoleic acid as described previously (Shureiqi et al., 2003).

Construction of 15-LOX-1 adenoviral vector and transfection of colorectal cancer cells

15-LOX-1 cDNA was subcloned into a pAd/TERT-GV 16-plasmid shuttle vector (Shureiqi et al., 2003). The pAd/TERT-GV 16-plasmid shuttle vector with the 15-LOX-1 cDNA was used to construct the adenovirus vector that carries human 15-LOX-1 cDNA in an AdEasy vector (Qbiogene, Inc., Carlsbad, CA), Ad-15-LOX-1. Ad-15-LOX-1 was amplified in human 293 cells, purified, and titrated. HCT-116 and LoVo cells were transfected with Ad-15-LOX-1 or a control vector, Ad-LacZ (the same adenovirus vector but with LacZ cDNA inserted in place of 15-LOX-1), at concentrations of 500 viral particles/cell (HCT-116 cells) and 1,000 viral particles/cell (LoVo cells).

Plasmids

Full-length PPAR-g-1 cDNA (Invitrogen Life Technologies, Carlsbad, CA), PPAR-a cDNA (Dr. Steven D. Clarke, Louisiana State University, Baton Rouge, LA), and PPAR-b/d cDNA (OriGene Technologies, Inc., Rockville, MD) were subcloned into pcDNA3.0 vectors to construct the PPAR-g, PPAR-a, and PPAR-b/d vectors, respectively. PPAR-g-1 dominant-negative pcDNA3.0 vector (PPAR-g dominant-negative vector) was provided by Professor V.K.K. Chatterjee (University of Cambridge, Cambridge, UK) (Gurnell et al., 2000). Dr. Christopher K. Glass (University of California, San Diego, San Diego, CA) provided a PPAR-g–dependent reporter construct, (AOx)3-TK-Luc (Huang et al., 1999). pSV-β-galactosidase was purchased from Promega Biosciences, Inc. (San Luis Obispo, CA).

PPAR-g activity assays

PPAR-g transcriptional activity was measured by transfecting cells with the PPAR-g–dependent reporter construct (AOx)3-TK-Luc (0.4 μg/well of a 24-well plate) (Huang et al., 1999) using Lipofectamine 2000 (Invitrogen). The pSV-β-galactosidase vector (0.2 μg/well of a 24-well plate) was cotransfected for normalization to β-galactoside activity. Cells were harvested at the indicated times and lysed, and then luciferase activity was measured using a luciferase assay kit (Promega). Luciferase activity levels were normalized to β-galactosidase activity, which was measured using a commercial kit (Invitrogen).

Small interfering RNA (siRNA) transfection

Cells were cultured to 50%–70% confluence and then transfected with 100 nmol of a pooled mixture of four SMARTselected siRNA duplexes (SMARTpool; Dharmacon, Inc., Lafayette, CO) for PPAR-b/d, PPAR-a, or a nonspecific control siRNA (siGLO RISC-Free siRNA; Dharmacon) using Lipofectamine 2000.

Quantitative real-time reverse-transcriptase PCR (RT-PCR)

Total RNA was extracted from cells using TRI reagent (Molecular Research Center Inc., Cincinnati, OH). The isolated RNA was size fractionated by electrophoresis in a 1% agarose–formaldehyde gel, stained with ethidium bromide, and confirmed to be of adequate quality (clear RNA bands for 18S, 28S, and 5S; 28S:18S of 2:1). Extracted RNA was quantified using an RNA quantitation kit (RiboGreen; Molecular Probes, Inc., Eugene, OR). A 500-ng aliquot of each RNA sample was reverse transcribed in a 20-μl reaction using the iScript cDNA synthesis kit (Bio-Rad Laboratories, Hercules, CA). The real-time PCR was carried out in 25 sl of a reaction mixture containing 1 sl of cDNA (25 ng/sl), 12.5 sl of 2× TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, CA), and 10.25 sl of primer and probe mixture (Applied Biosystems). Real-time PCR assays were performed in triplicate using a 7300 real-time PCR system (Applied Biosystems) with the following program: 50°C for 2 min, 95°C for 10 min, 40 cycles at 95°C for 15 s and at 60°C for 1 min. A sequence detection program calculated a threshold cycle number (CT) at which the probe cleavage–generated fluorescence exceeded the background signal (Pfaffl, 2001).

Measurement of relative RNA expression level

We calculated the relative RNA expression level using a comparative CT method (Pfaffl, 2001). The sets of gene primer and probe for the target genes (PPAR-g, PPAR-b/d, CD36, and Keratin 20) were confirmed to have amplification efficiency equal to that of the internal reference gene (HPRT1). The relative expression level of an individual target gene was normalized to that reference gene and to a calibrator sample that was run on the same plate. The normalized relative expression level of a target gene in an individual sample was calculated using the following formula:

(Etarget)ΔCT   target   (calibrator-sample)(Ereference)ΔCT   reference   (calibrator-sample)

in which the real-time PCR efficiency of the target gene transcript is donated by Etarget and that of the reference gene transcript, by Ereference (Pfaffl, 2001). Thus, the relative RNA expression level of a gene is a unitless number relative to that of the calibrator sample (Yu et al., 2003). Relative RNA expression calculations were performed using commercial software (SDS V1.2; Applied Biosystems).

Scintillation proximity assays of 13-S-HODE binding to PPAR-g

These assays were performed as described previously (Shureiqi et al., 2003) with use of [3H]rosiglitazone as the radiolabeled ligand for PPAR-g.

In vitro transcription/translation

The full-length proteins PPAR-g, PPAR-b/d, and RXR-a were produced by in vitro transcription–coupled translation of PCR products using TNT T7 Quick-coupled transcription and translation systems (Promega) and PCR-generated DNA templates. The PCR primers used were as described previously (He et al., 1999). PPAR-g, PPAR-b/d, and pCMV6-XL4-RXR-a (OriGene) vectors containing full-length cDNA of the three genes (PPAR-g, PPAR-b/d, and RXR-a) were used as PCR templates. The translation products were verified by Western blotting.

Gel EMSA

To study the mechanisms of PPAR-b/d’s effects on PPAR-g’s transcriptional activity, the PPAR-g DNA recognition sequence from the acyl-coA oxidase gene promoter ACO was used (He et al., 1999). The double-stranded oligonucleotide of the ACO sequence was formed by annealing 5′-GCGGACCAGGACAAAGGTCACGTTC-3′ and 5′-CGAACGTGACCTTTGTCCTGGTCCG-3′. The resulting double-stranded oligonucleotide was 32P-labeled with T4 polynucleotide kinase (Promega). RXR-a (0.5 μl of 58.3-μg/μl lysate), PPAR-g (0.5 μl of 56.4-μg/μl lysate), and/or PPAR-b/d (1.5 μl, 3.0 μl, 6.0 μl, or 9.0 μl of 53.9-μg/μl lysate) derived from in vitro–translated proteins were mixed with the probe (40,000 cpm) and gel-shift binding buffer (Promega) for 25 min on ice. The DNA–protein complexes were identified as being resolved from the free probe by electrophoresis at 4°C on 5% polyacrylamide gel in 1× Tris borate–EDTA buffer (pH 8.3). For gel supershift assays, 2 μl of 10× PPAR-g monoclonal antibody (Santa Cruz) was added to each sample and incubated at room temperature for 20 min and then on ice for another 25 min.

Chromatin immunoprecipitation

To study PPAR-g binding to its target genes in vivo, we used a previously validated ChIP assay for PPAR-g binding to the human LPL promoter (Fajas et al., 2002a; Fajas et al., 2002b). The effects of PPAR-b/d loss and gain of expression on PPAR-g DNA binding in vivo were assessed by transfecting KO1 cells (PPAR-b/d knockout HCT-116 cells) with either PPAR-b/d expression vector or control vector and by downregulating PPAR-b/d expression in the parental HCT-116 cells by PPAR-b/d siRNA transfection as described above. Cells were cross linked 48 h after the designated transfections by adding formaldehyde to the culture medium to a final concentration of 1% and incubating for 10 min at 37°C. ChIP assays were performed using a commercial assay kit according to the manufacturer’s protocol (Upstate Cell Signaling Solutions, Waltham, MA). Chromatin was immunoprecipitated using a specific monoclonal anti–PPAR-g antibody (Santa Cruz). The following primers were used to amplify a 230-bp fragment of the human LPL promoter: 5′-GGGCCCCCGGGTAGAGTGG-3′ (sense) and 5′-CACGCCAAGGCTGCTTATGTGACT-3′ (antisense) using the following parameters: 94 °C for 3 min and then 94 °C for 20 s, 60 °C for 30 s, and 71.5 °C for 70 s for 25 cycles. The PCR conditions were optimized for the primer set to ensure that the yield of product was within the linear range (data not shown).

Quantitative assay for PPAR-g DNA binding

PPAR-g DNA binding was quantified using a PPAR-g transcription factor assay kit (TransAM PPAR-g kit; Active Motif Co., Carlsbad, CA) (Renard et al., 2001). A PPAR-g–specific antibody is used in this enzyme-linked immunosorbent assay to specifically measure PPAR-g binding to an oligonucleotide containing the PPRE. Nuclear extracts were prepared, and PPAR-g binding to the PPRE was assayed according to the manufacturer’s protocol.

Western blot analysis

For Western blotting, protein samples were prepared and subjected to SDS-PAGE under reducing conditions as described previously (Shureiqi et al., 2000). After transfer, blots were probed with a solution of rabbit polyclonal antibody to human 15-LOX-1 (1:2,000), PPAR-b/d (1:500), or a monoclonal antibody to human PPAR-g-1 (1:300) and then analyzed using the enhanced chemiluminescence method.

13-S-HODE measurements

Cells were harvested, and 13-S-HODE was extracted from both the culture medium and cell lysates and then measured using a commercially available enzyme immunoassay kit (Assay Designs, Inc., Ann Arbor, MI) (Shureiqi et al., 2000).

Assessments of apoptosis

Apoptosis was measured by cell imaging and by DNA fragmentation and caspase-3 activity assays. Inverse light (phase-contrast) microscopy was used to assess gross evidence of growth inhibition and apoptosis. Cell morphology images were taken under light microscopy 60 h after the cells’ transfection with Ad-15-LOX-1 or Ad-LacZ. Floating and attached cells were harvested 48 h after transfection with Ad-15-LOX-1 or Ad-LacZ, and the DNA fragmentation assay was performed as described previously (Shureiqi et al., 2003). For the caspase-3 activity assay, cells were transfected first with Ad-15-LOX-1 or Ad-LacZ and then 12 h later with pcDNA3.0–PPAR-g dominant-negative vector, pcDNA3.0–PPAR-b/d, or pcDNA3.0 empty vector (control vector). Cells were harvested 36 h after the second transfection, and caspase-3 activity was measured using a commercially available kit (BD Biosciences Clontech, Palo Alto, CA) according to the manufacturer’s protocol.

Statistical analyses

For analyses involving the simultaneous consideration of two factors, we performed two-way analyses of variance. Our analyses proceeded as follows: we first tested the interaction effect, and if it was significant, we subsequently performed specific comparisons to investigate which differences were driving this effect, using the Bonferroni adjustment to adjust for the multiple testing problem. This means that if we performed k comparisons, an individual comparison would not be considered significant unless its P value was less than 0.05/k. If the interaction effect was not significant, we tested the individual main effects. Then, if those were significant, we determined which pairwise comparisons were significant, again adjusting for multiplicities using the Bonferroni correction. For analyses involving single factors, we performed a one-way analysis of variance. If the overall analysis of variance test was significant, then we performed pairwise comparisons, adjusting for multiplicities using the Bonferroni correction. All tests were two-sided and conducted at the P ≤ 0.05 level. All quantitative analyses were done on the log-transformed data because we found that the log transformation decoupled the relationship between the mean and variance and accommodated the normal-distributional assumptions underlying the methods. Data were analyzed using SAS software (SAS Institute, Cary, NC).

Supplementary Material

Supplementary

Acknowledgments

We thank Dr. Reuben Lotan for his critical review of the manuscript and helpful comments. We also thank Karen Phillips from the Department of Scientific Publications at The University of Texas M. D. Anderson Cancer Center for editing the manuscript. In addition, we acknowledge the technical assistance of Dongning Chen. This work was supported in part by the National Cancer Institute, National Institutes of Health, Department of Health and Human Services R01 grant CA104278 (to I.S.); the American Cancer Society Scholar Award RSG-04-020-01-CNE (to I.S.); National Institute of Environmental Health Sciences, NIH Center grant ES07784; and funding from The Jerry and Maury Rubenstein Foundation.

References

  1. Bull AW, Steffensen KR, Leers J, Rafter JJ. Carcinogenesis. 2003;24:1717–22. doi: 10.1093/carcin/bgg131. [DOI] [PubMed] [Google Scholar]
  2. Calnek D, Quaroni A. Differentiation. 1993;53:95–104. doi: 10.1111/j.1432-0436.1993.tb00649.x. [DOI] [PubMed] [Google Scholar]
  3. Chawla A, Repa JJ, Evans RM, Mangelsdorf DJ. Science. 2001;294:1866–70. doi: 10.1126/science.294.5548.1866. [DOI] [PubMed] [Google Scholar]
  4. Fajas L, Egler V, Reiter R, Hansen J, Kristiansen K, Debril MB, Miard S, Auwerx J. Dev Cell. 2002a;3:903–10. doi: 10.1016/s1534-5807(02)00360-x. [DOI] [PubMed] [Google Scholar]
  5. Fajas L, Landsberg RL, Huss-Garcia Y, Sardet C, Lees JA, Auwerx J. Dev Cell. 2002b;3:39–49. doi: 10.1016/s1534-5807(02)00190-9. [DOI] [PubMed] [Google Scholar]
  6. Febbraio M, Hajjar DP, Silverstein RL. J Clin Invest. 2001;108:785–91. doi: 10.1172/JCI14006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Gearing KL, Gottlicher M, Teboul M, Widmark E, Gustafsson JA. Proc Natl Acad Sci U S A. 1993;90:1440–4. doi: 10.1073/pnas.90.4.1440. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Girnun GD, Smith WM, Drori S, Sarraf P, Mueller E, Eng C, Nambiar P, Rosenberg DW, Bronson RT, Edelmann W, Kucherlapati R, Gonzalez FJ, Spiegelman BM. Proc Natl Acad Sci U S A. 2002;99:13771–6. doi: 10.1073/pnas.162480299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Gupta RA, Brockman JA, Sarraf P, Willson TM, DuBois RN. J Biol Chem. 2001;276:29681–7. doi: 10.1074/jbc.M103779200. [DOI] [PubMed] [Google Scholar]
  10. Gupta RA, Tan J, Krause WF, Geraci MW, Willson TM, Dey SK, DuBois RN. Proc Natl Acad Sci U S A. 2000;97:13275–80. doi: 10.1073/pnas.97.24.13275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gurnell M, Wentworth JM, Agostini M, Adams M, Collingwood TN, Provenzano C, Browne PO, Rajanayagam O, Burris TP, Schwabe JW, Lazar MA, Chatterjee VK. J Biol Chem. 2000;275:5754–9. doi: 10.1074/jbc.275.8.5754. [DOI] [PubMed] [Google Scholar]
  12. He TC, Chan TA, Vogelstein B, Kinzler KW. Cell. 1999;99:335–45. doi: 10.1016/s0092-8674(00)81664-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Huang JT, Welch JS, Ricote M, Binder CJ, Willson TM, Kelly C, Witztum JL, Funk CD, Conrad D, Glass CK. Nature. 1999;400:378–82. doi: 10.1038/22572. [DOI] [PubMed] [Google Scholar]
  14. Jimenez B, Volpert OV, Crawford SE, Febbraio M, Silverstein RL, Bouck N. Nat Med. 2000;6:41–48. doi: 10.1038/71517. [DOI] [PubMed] [Google Scholar]
  15. Kliewer SA, Lehmann JM, Willson TM. Science. 1999;284:757–60. doi: 10.1126/science.284.5415.757. [DOI] [PubMed] [Google Scholar]
  16. Kliewer SA, Umesono K, Noonan DJ, Heyman RA, Evans RM. Nature. 1992;358:771–4. doi: 10.1038/358771a0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Matsusue K, Peters JM, Gonzalez FJ. Faseb J. 2004;18:1477–9. doi: 10.1096/fj.04-1944fje. [DOI] [PubMed] [Google Scholar]
  18. Michalik L, Desvergne B, Wahli W. Nat Rev Cancer. 2004;4:61–70. doi: 10.1038/nrc1254. [DOI] [PubMed] [Google Scholar]
  19. Moll R, Zimbelmann R, Goldschmidt MD, Keith M, Laufer J, Kasper M, Koch PJ, Franke WW. Differentiation. 1993;53:75–93. doi: 10.1111/j.1432-0436.1993.tb00648.x. [DOI] [PubMed] [Google Scholar]
  20. Nagy L, Tontonoz P, Alvarez JG, Chen H, Evans RM. Cell. 1998;93:229–40. doi: 10.1016/s0092-8674(00)81574-3. [DOI] [PubMed] [Google Scholar]
  21. Nixon JB, Kamitani H, Baek SJ, Eling TE. Prostaglandins Leukot Essent Fatty Acids. 2003;68:323–30. doi: 10.1016/s0952-3278(03)00023-1. [DOI] [PubMed] [Google Scholar]
  22. Nixon JB, Kim KS, Lamb PW, Bottone FG, Eling TE. Prostaglandins Leukot Essent Fatty Acids. 2004;70:7–15. doi: 10.1016/j.plefa.2003.06.001. [DOI] [PubMed] [Google Scholar]
  23. Osawa E, Nakajima A, Wada K, Ishimine S, Fujisawa N, Kawamori T, Matsuhashi N, Kadowaki T, Ochiai M, Sekihara H, Nakagama H. Gastroenterology. 2003;124:361–7. doi: 10.1053/gast.2003.50067. [DOI] [PubMed] [Google Scholar]
  24. Park BH, Vogelstein B, Kinzler KW. Proc Natl Acad Sci U S A. 2001;98:2598–603. doi: 10.1073/pnas.051630998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Peters JM, Aoyama T, Burns AM, Gonzalez FJ. Biochim Biophys Acta. 2003;1632:80–9. doi: 10.1016/s1388-1981(03)00065-9. [DOI] [PubMed] [Google Scholar]
  26. Pfaffl MW. Nucleic Acids Res. 2001;29:e45. doi: 10.1093/nar/29.9.e45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Renard P, Ernest I, Houbion A, Art M, Le Calvez H, Raes M, Remacle H. Nucleic Acids Res. 2001;29:E21. doi: 10.1093/nar/29.4.e21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Rusinol AE, Yang L, Thewke D, Panini SR, Kramer MF, Sinensky MS. J Biological Chem. 2000;275:7296–7303. doi: 10.1074/jbc.275.10.7296. [DOI] [PubMed] [Google Scholar]
  29. Sarraf P, Mueller E, Jones D, King FJ, DeAngelo DJ, Partridge JB, Holden SA, Chen LB, Singer S, Fletcher C, Spiegelman BM. Nat Med. 1998;4:1046–52. doi: 10.1038/2030. [DOI] [PubMed] [Google Scholar]
  30. Shi Y, Hon M, Evans RM. Proc Natl Acad Sci U S A. 2002;99:2613–8. doi: 10.1073/pnas.052707099. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Shureiqi I, Chen D, Lee JJ, Yang P, Newman RA, Brenner DE, Lotan R, Fischer SM, Lippman SM. J Natl Cancer Inst. 2000;92:1136–42. doi: 10.1093/jnci/92.14.1136. [DOI] [PubMed] [Google Scholar]
  32. Shureiqi I, Jiang W, Zuo X, Wu Y, Stimmel JB, Leesnitzer LM, Morris JS, Fan HZ, Fischer SM, Lippman SM. Proc Natl Acad Sci U S A. 2003;100:9968–73. doi: 10.1073/pnas.1631086100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Shureiqi I, Lippman SM. Cancer Res. 2001;61:6307–12. [PubMed] [Google Scholar]
  34. Shureiqi I, Wojno KJ, Poore JA, Reddy RG, Moussalli MJ, Spindler SA, Greenson JK, Normolle D, Hasan AA, Lawrence TS, Brenner DE. Carcinogenesis. 1999;20:1985–95. doi: 10.1093/carcin/20.10.1985. [DOI] [PubMed] [Google Scholar]
  35. Stephen RL, Gustafsson MC, Jarvis M, Tatoud R, Marshall BR, Knight D, Ehrenborg E, Harris AL, Wolf CR, Palmer CN. Cancer Res. 2004;64:3162–70. doi: 10.1158/0008-5472.can-03-2760. [DOI] [PubMed] [Google Scholar]
  36. Yu J, Marsh S, Ahluwalia R, McLeod HL. Cancer Res. 2003;63:6170–3. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary
NIHMS5014-supplement-Supplementary.pdf (232.8KB, pdf)

RESOURCES