|Home | About | Journals | Submit | Contact Us | Français|
Studies indicate that peroxisome proliferator-activated receptor-β/δ (PPARb/δ) can either attenuate or potentiate colon cancer. One hypothesis suggests that PPARβ/δ is upregulated by the adenomatous polyposis coli (APC)/β-CATENIN pathway and a related hypothesis suggests that PPARβ/δ is downregulated by nonsteroidal anti-inflammatory drugs (NSAIDs). The present study examined these possibilities using in vivo and in vitro models. While APC/β-CATENIN-dependent expression of CYCLIN D1 was observed in vivo and in vitro, expression of PPARβ/δ was not different in colon or intestinal polyps from wild-type or Apcmin heterozygous mice or in human colon cancer cell lines with mutations in APC and/or β-CATENIN. No difference in the level of PPARβ/δ was found in colon from wild-type or Apcmin heterozygous mice following treatment with NO-donating aspirin (NO-ASA). NSAIDs inhibited cell growth in RKO (wild-type APC) and DLD1 (mutant APC) human colon cancer cell lines but expression of PPARβ/δ was not downregulated in these cell lines in response to a broad concentration range of celecoxib, indomethacin, NS-398, or nimesulide. However, indomethacin caused an increase in PPARβ/δ mRNA and protein that was accompanied with increased expression of a known PPARβ/δ target gene. Interestingly, expression of PPARα was also increased in the human colon cancer cell lines by several NSAIDs at the highest concentration examined. Results from these studies provide additional evidence indicating that PPARβ/δ is not upregulated by the APC/β-CATENIN pathway. Further, these studies suggest that increased PPARβ/δ and/or PPARα by NSAIDs in human colon cancer cell lines could contribute to the mechanisms underlying the chemopreventive effects of NSAIDs.
Peroxisome proliferator-activated receptor-β/δ (PPARβ/δ; also referred to as PPARβ or PPARδ) is a ligand-activated transcription factor with critical roles in regulating lipid and glucose homeostasis . In response to ligand activation, PPARβ/δ modulates cellular function by directly modifying target gene expression. Since fatty acids likely act as endogenous ligands for PPARs, including PPARβ/δ, it is thought that PPARβ/δ acts as a cellular sensor. While the specific endogenous ligand(s) for PPARβ/δ remain to be identified, important physiological roles for PPARβ/δ have been established. For example, ligand activation of PPARβ/δ can decrease serum glucose concentration in diabetic mice , increase fatty acid oxidation in skeletal muscle , prevent high fat diet-induced obesity [3,4], increase serum high-density lipoprotein cholesterol , increase running endurance synergistically with exercise by increasing oxidative myofibers , prevent liver toxicity and fatty liver [7,8], and induce terminal differentiation in epithelial and neuronal cells [reviewed in 1]. For most of these effects, the essential requirement for PPARβ/δ has been established since these effects are not found in ligand-treated Pparβ/δ-null mice. In addition to direct regulation of target genes, PPARβ/δ also has potent anti-inflammatory activities that are likely mediated by PPARβ/δ interfering with other transcription factors [e.g., NFκB, AP1, STAT3; reviewed in 9]. Thus, it is not surprising that there is considerable interest in targeting PPARβ/δ for the prevention and treatment of diseases including obesity, dyslipidemias, diabetes, and cancer.
The role of PPARβ/δ in cancer remains uncertain as some studies indicate that PPARβ/δ promotes tumorigenesis while others suggest that PPARβ/δ attenuates tumorigenesis [reviewed in 1,10,11]. It was originally hypothesized that PPARβ/δ was upregulated by the adenomatous polyposis coli (APC)/β-CATENIN/transcription factor 4 (TCF4) pathway during colon cancer . This was based in part on the observed decrease in PPARβ/δ expression in a human colon cancer cell line when APC expression was increased, coupled with the reported increase in PPARβ/δ expression in a small cohort of human colon tumors as compared to normal tissue . This work led to the hypothesis that a mechanism by which nonsteroidal antiinflammatory drugs (NSAIDs) inhibit tumor growth is by suppressing cyclooxygenase-2 (COX2)-mediated formation of PPARβ/δ ligands, thereby limiting ligand activation of PPARβ/δ that putatively increased cell proliferation by upregulation of yet-to-be identified target genes . However, while some reports support the initial finding that PPARβ/δ is upregulated by the APC/β-CATENIN/TCF4 pathway, there are other studies that are not in concert with this view [reviewed in 1]. For example, disruption of APC in mouse colon does not result in an increase in PPARβ/δ expression  and increased expression of PPARβ/δ is not consistently found in human colon tumors [reviewed in 1]. A related hypothesis was also developed suggesting that NSAIDs inhibit growth of cancer cells by reducing expression of PPARβ/δ, however, not all reports support this hypothesis [reviewed in 1]. Thus, there is a pressing need to determine whether PPARβ/δ is regulated by an APC-dependent pathway and whether NSAIDs modulate PPARβ/δ expression. The present study examined the hypothesis that PPARβ/δ is regulated by the APC/β-CATENIN/TCF4 pathway and/or NSAIDs using quantitative approaches with both in vivo and in vitro models.
RKO cells were cultured in Dulbecco’s modified Eagle medium (DMEM) (Invitrogen, Carlsbad, CA). DLD1 and LS174T cells were cultured in RPMI 1640 medium (Invitrogen). HCT116 and HT29 cells were cultured in McCoy’s 5A medium (Invitrogen). All cell lines were cultured with 10% fetal bovine serum at 37°C and 5% carbon dioxide. For quantitative Western blots, RKO or DLD1 cells were cultured for up to 48 h in culture medium containing one of the following NSAIDs: indomethacin (50, 200, or 600 µM), nimesulide (10, 100, or 500 µM), NS-398 (0.1, 1.0, 10.0, or 100 µM) or celecoxib (1.0, 10, or 100 µM). Nuclear extracts were prepared using procedures as previously described  and stored at −80°C until further analysis.
Colon tissue and small intestine polyps were dissected from female wild-type C57BL/6 or Apcmin heterozygous mice on a C57BL/6 background. Colon tissue from female wild-type C57BL/6 or Apcmin heterozygous mice on a C57BL/6 background treated with nitric oxide-donating aspirin (NO-ASA) were obtained from samples collected from a previously published study . Nuclear extracts were prepared using procedures as previously described  and stored at −80°C until further analysis.
For cell proliferation assays, RKO or DLD1 cells were plated on 12-well dishes at a density of ~20 000–40 000 cells per well and cultured in control medium for 24 h. After this 24 h culture period, cultured medium was changed to fresh medium containing DMSO (vehicle control), indo-methacin (50, 200, or 600 µM), nimesulide (10, 100, or 500 µM), NS-398 (1.0, 10, or 100 µM), or celecoxib (1.0, 10, or 100 µM). Cells were quantified every 24 h with a Z1 coulter particle counter® (Beckman Coulter, Fullerton, CA). Three independent samples for each treatment were used for each time point for every treatment, and each sample was counted in triplicate.
Thirty micrograms of nuclear extract per cell line sample or 50 µg of nuclear extract per tissue sample was resolved using SDS–polyacrylamide gels. The samples were transferred onto polyvinylidene fluoride membrane by electroblotting. The membranes were blocked with 5% dried milk in Tris-buffered saline/Tween-20 and incubated at 4°C overnight with primary antibodies. After incubation with biotinylated secondary antibody (Jackson ImmunoResearch Laboratories, West Grove, PA), immunoreactive proteins on the membrane were detected with 125I-labeled streptavidin using phosphorimaging analysis. Hybridization signals for the proteins of interest were normalized to the hybridization signal of the nuclear protein LAMIN. Minimum of three independent samples per treatment group were analyzed for all Western blots. For positive controls, lysate from COS1 cells transfected with expression vectors for mouse or human PPARα, PPARβ/δ, or PPARγ were used. The following antibodies were used: anti-LAMIN (SC-7293), anti-PPARβ/δ (H-74), anti-PPARγ (H-100, which detects both PPARγ1 and PPARγ2 isoforms), anti-cMYC (SC-42; Santa Cruz Biotechnologies, Santa Cruz, CA), or anti-CYCLIN D1 (Cell Signaling Technology, Danvers, MA). An anti-PPARβ/δ antibody  and an anti-PPARα antibody (Affinity Bioreagents, Golden, CO) were used for mouse samples.
RKO and DLD1 cells were cultured as described above in medium containing 0.1% DMSO (control), GW0742 (0.1, 1.0, or 10 µM) and/or NSAIDs. Total mRNA was isolated using TRIZOL and following manufacturer’s recommended protocol (Invitrogen). The mRNA-encoding PPARβ/δ, adipose differentiation-related protein (ADRP), and angiopoietin-like 4 (ANGPTL4) were quantified using quantitative real-time polymerase chain reaction (qPCR). The cDNA was generated using 2.5 µg total RNA with MultiScribe Reverse Transcriptase kit (Applied Biosystems, Foster City, CA). Primers were designed for real-time PCR using the Primer Express software (Applied Biosystems). The sequence and GenBank accession number for the forward and reverse primers used to quantify mRNAs were: PPARα (L02932.1) forward, 5′-TGGACGAGTCTCCCAGTGG-3′ and reverse, 5′-CCCCGCAGATTCTACATTCG-3′, PPARβ/δ (AY919140) forward, 5′-GACAGTGACCTGGCCCTATTCA-3′ and reverse, 5′-AGGATGGTGTCCTGGATAGCCT-3′, ADRP (NM_000122) forward, 5′-CTGCTCTTCGCCTTTCGCT-3′, and reverse, 5′-ACCACCCGAGTCACCACACT-3′, and ANGPTL4 (NM_020581) forward, 5′-TTCTCGCCTACCAGAGAAGTTGGG-3′ and reverse, 5′-CATCCACAGCACCTACAACAGCAC-3′. The mRNA was normalized to the gene encoding GAPDH (BC083149) using the following primers: forward, 5′-GGTGGAGCCAAAAGGGTCAT-3′ and reverse, 5′-GGTTCACACCCATCACAAACAT-3′. Real-time PCR reactions were carried out using SYBR green PCR master mix (Finnzymes, Espoo, Finland) in the iCycler and detected using the MyiQ Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA). The following conditions were used for PCR: 95°C for 15 s, 94°C for 10 s, 60°C for 30 s, and 72°C for 30 s and repeated for 45 cycles. The PCR included a no template control reaction to control for contamination and/or genomic amplification. All reactions had >85% efficiency. Relative expression levels of mRNA were normalized to GAPDH and analyzed for statistical significance using one-way analysis of variance (Prism 5.0).
To critically examine the role of APC in modulating PPARβ/δ, expression of PPARβ/δ was quantitatively examined in mouse colon from wild-type and Apcmin heterozygous mice, and intestinal polyps from the latter genotype. No difference in the expression of PPARβ/δ was found in colon between either genotype, and no change in expression of PPARβ/δ was noted in the polyps as compared to normal colon tissue (Figure 1A). Expression of PPARγ1 and PPARγ2 was very low as compared to PPARβ/δ, and no difference in PPARγ1 or PPARγ2 levels was found between either genotype or in polyps (Figure 1A). PPARα was not detected in any of the samples (Figure 1A). Interestingly, while expression of cMYC was not different between genotypes or between colon and polyp samples, increased expression of CYCLIN D1 was observed in the polyp samples as compared to colon from both wild-type and Apcmin heterozygous mice (Figure 1A). These data are consistent with past reports showing no difference in PPARβ/δ expression between wild-type and Apcmin heterozygous mice as determined by Western blotting [16–18].
To examine the role of APC in modulating PPARβ/δ expression in human colon cancer models, human colon cancer cell lines with different mutations in APC and β-CATENIN were utilized (Table 1). RKO human colon cancer cells were used as a control as they have wild-type APC and β-CATENIN alleles, while the other cell lines (DLD1, HT29, LS174T, HCT116) exhibit enhanced β-CATENIN/TCF4 signaling due to either mutant APC or gain-of-function mutations in β-CATENIN. Indeed, expression of CYCLIN D1 was significantly greater in DLD1, HT29, LS174T, and HCT116 cells as compared to RKO controls (Figure 1B), consistent with enhanced β-CATENIN/TCF4 signaling since CYCLIN D1 is a known target of this pathway. Expression of cMYC was markedly higher in LS174T cells as compared with all other cells lines (Figure 1B). Despite evidence of enhanced APC/β-CATENIN/TCF4 signaling in DLD1, HT29, LS174T, and HCT116 cells, expression of PPARβ/δ was not different between any of the five humancolon cancer cell lines (Figure 1B). Expression of PPARα was not detected in any of the cell lines, and expression of PPARγ1 was similar between all of the cell lines (Figure 1B). Expression of PPARγ2 was higher in HT29 and LS174T cells as compared to the other cells lines (Figure 1B). Collectively, results from this analysis are consistent with results from the mouse models showing no change in PPARβ/δ expression due to APC-dependent upregulation.
Recent studies using immunohistochemical analysis suggest that PPARβ/δ expression was increased 10-fold in Apcmin heterozygous mouse colon as compared to wild type . Further, it was also suggested that the chemopreventive effects of the NSAID, NO-ASA, was due in part to reduced expression of PPARβ/δ caused by NO-ASA . Since it was recently shown that the background immunoreactivity of a highly specific anti-PPARβ/δ antibody is very high , samples of mouse colon from wild-type and Apcmin heterozygous mice treated with and without NO-ASA were obtained from a previously published study , and used to examine PPARβ/δ expression using quantitative Western blotting. In contrast to the previous report , no difference in the level of PPARβ/δ in colon samples was observed between wild-type and Apcmin heterozygous mice, and NO-ASA had no effect on PPARβ/δ expression (Figure 1C). No differences in the expression of PPARα or PPARγ1 were detected between genotypes or treatment (Figure 1C).
Some studies suggest that the cell growth inhibitory activity of NSAIDs may be due in part to inhibition of PPARβ/δ expression while others do not [reviewed in 1]. Thus, expression of PPARβ/δ protein was quantified in RKO and DLD1 human colon cancer cell lines treated with NSAIDs. A significant decrease in cell proliferation was observed in both RKO and DLD1 cells after 48 h of culture with 500 µM nimesulide and by 72 h this effect was greater in DLD1 cells (Figure 2A). A significant decrease in cell proliferation was observed in both RKO and DLD1 cells after 48 h of culture by 600 µM indomethacin (Figure 2B). Inhibition of cell growth was also observed after 72 h following culture in 200 µM indomethacin in both cell lines (Figure 2B). In response to celecoxib, inhibition of cell proliferation was observed in both RKO and DLD1 after 48 h by a concentration of 100 µM and after 72 h by concentrations greater than 1.0 µM (Figure 2C). Significant inhibition of cell proliferation was found in RKO cells after 72 h of culture in either 10.0 or 100 µMNS-398, and this effect was not observed in similarly treated DLD1 cells (Figure 2D). The decreases in cell proliferation by NSAIDs were likely due to increased cell death as determined by trypan blue exclusion (data not shown) and gross cell morphology (Supplemental Figure 1). These results show that RKO and DLD1 cells exhibit significant decreases in cell proliferation between 48 and 72 h following treatment with the indicated concentrations of NSAIDs and provide the rationale for the concentrations of NSAIDs that were used to evaluate the effect of NSAIDs on PPARβ/δ expression and function.
If NSAIDs-dependent inhibition of human colon cancer cell growth is mediated by downregulation of PPARβ/δ, then decreased expression of this receptor should precede the observed decreases in cell proliferation (Figure 2). Previous work by others also showed that decreased expression of PPARβ/δ by NSAIDs occurs between 24 and 48 h in human colon cancer cell lines [20,21]. Thus, expression of PPARβ/δ was examined in RKO and DLD1 cells treated with NSAIDs after 24 h. Expression of PPARβ/δ was either unchanged or increased by NSAIDs in both RKO and DLD1 cells (Figure 3). For example, PPARβ/δ was increased in RKO cells ~2-fold to 3-fold with 50–600 µM indomethacin or 1.4-fold with 500 µM nimesulide. In DLD1 cells, PPARβ/δ was increased ~2-fold treated with 500 µM nimesulide or ~2-fold with 100 µM celecoxib (Figure 3). The changes in PPARβ/δ protein were also reflected by similar changes in mRNA expression after 24 h of treatment (Figure 5A). Importantly, no decrease in PPARβ/δ protein or mRNA was observed in response to any of the NSAIDs after 24 h of treatment (Figure 3, Figure 5A). Interestingly, expression of PPARα was increased by 2-fold to 6-fold in both RKO and DLD1 cells following exposure to 200 or 600 µM indomethacin or 100 µM celecoxib (Figure 3). Increased expression of PPARα was also noted in DLD1 cells treated with 500 µM nimesulide (Figure 3). The changes in PPARα protein were also reflected by similar changes in mRNA expression after 24 h of treatment (Figure 5A). No changes in the level of PPARγ were observed in either RKO or DLD1 cells treated with any of the NSAIDs (Figure 3). Expression of PPARs was also examined after 48 h (Figure 4) using concentrations of NSAIDs shown to cause inhibition of cell proliferation (Figure 3). Expression of PPARβ/δ was increased by 200 µM indomethacin in RKO cells but not in DLD1 cells (Figure 4). No other changes in the expression of PPARβ/δ were noted in response to 500 µM nimesulide, 10 µM celecoxib, 100 µM NS-398, or 150 µM sulindac (Figure 4). While no change in PPARγ1 expression was observed in either cell type by NSAIDs, expression of PPARα was significantly higher in DLD1 cells following treatment with 500 µM nimesulide, 100 µM NS-3998, or 150 µM sulindac in RKO cells following treatment with 100 µM NS-398 (Figure 4).
To determine if any functional changes in PPARβ/δ activity are observed by NSAIDs, analysis of known PPARβ/δ target gene expression was performed. ADRP and ANGPTL4 are well-characterized PPARβ/δ target genes [22–25]. Interestingly, there was a significant difference in the response to the PPARβ/δ ligand GW0742 between RKO and DLD1 cells. In RKO cells, no change in the expression of ANGPTL4 mRNA was observed following treatment with GW0742 between 0.1 and 10 µM while an increase in ADRP mRNA was observed in these cells with these concentrations of GW0742 (Figure 5B). In contrast, in DLD1 cells, no change in the expression of ADRP mRNA was observed following treatment with GW0742 between 0.1 and 10 µM while an increase in ANGPTL4 mRNA was observed in these cells with these concentrations of GW0742 (Figure 5B). Thus, ADRP mRNA was used as a marker for PPARβ/δ activity in RKO cells while ANGPTL4 mRNA was used as a marker for PPARβ/δ activity in DLD1 cells. As compared to control, a significant increase in the level of mRNA encoding ADRP was found in RKO cells treated for 24 h with 600 µM indomethacin, 500 µM nimesulide, and 100 µM celecoxib, but no decrease in expression of this PPARβ/δ target gene was observed with the other treatments (Figure 5B). A significant increase in the level of mRNA encoding ANGPTL4 was found in DLD1 cells treated for 24 h with 600 µM indomethacin and 500 µM nimesulide as compared to control, but no decrease in expression of this PPAR target gene was observed with the other treatments as compared to control (Figure 5B). The increase in expression of the PPARβ/δ target genes ADRP and ANGPTL4 correlate well with the observed increases in expression of PPARβ/δ observed in response to NSAIDs at both the mRNA (Figure 5A) and protein level (Figure 3 and Figure 4).
Results from these studies provide evidence indicating that PPARβ/δ is not upregulated by the APC/β-CATENIN/TCF4 pathway. No difference in the level of PPARβ/δ protein was found between wild-type and Apcmin heterozygous mouse colon, or between colon and intestinal polyps in Apcmin heterozygous mice. Importantly, this is in contrast to the increase in CYCLIN D1 expression noted in intestinal polyps as compared to colon in both wild-type and Apcmin heterozygous mice. The lack of change in PPARβ/δ protein between wild-type and Apcmin heterozygous mouse colon and/or intestinal tumors is consistent with three previous reports [16–18]. Results from the present studies also show that the expression of PPARβ/δ is not different between five different human colon cancer cell lines with different mutations in APC and/or β-CATENIN (Table 1). Notably, as compared to RKO cells that have wild-type APC and β-CATENIN alleles, expression of PPARβ/δ was not different in DLD1, HT29, LS174T, or HCT116 cells. In contrast, expression of CYCLIN D1 was higher in DLD1, HT29, LS174T, or HCT116 cells as compared to RKO cells consistent with the known mutational spectrum in APC and β-CATENIN in these human colon cancer cell lines (Table 1). Thus, despite strong evidence that APC/β-CATENIN/TCF4 signaling is enhanced in DLD1, HT29, LS174T, or HCT116 cells as compared to RKO cells, PPARβ/δ expression was not different. Collectively, data from an in vivo model of dysregulated APC (Apcmin heterozygous mice) and in vitro models of dysregulated APC indicate that PPARβ/δ is not upregulated by the APC/β-CATENIN/TCF4 pathway. This is consistent with a large number of other studies including the lack of change in PPARβ/δ expression when APC is disrupted in the colon or when other APC-related signaling molecules are repressed [reviewed in 1]. The present studies significantly extend the current understanding of potential roles of PPARβ/δ in colon cancer because to date, no comprehensive and quantitative analysis of PPARβ/δ protein expression in related models have been reported. Essentially all reports suggesting that PPARβ/δ is upregulated by the APC/β-CATENIN/TCF4 pathway are based largely on mRNA measurement data [reviewed in 1]. Given the results from the present studies, it is clear that there remains a need for careful quantification of PPARβ/δ protein and function (e.g., PPARβ/δ target gene expression) from human samples comparing normal epithelium with tumors from patients with mutations in APC and/or β-CATENIN signaling genes since examination of mRNA may not reflect protein expression.
Expression of PPARγ1 and PPARγ2 was not different between wild-type and Apcmin heterozygous mice, but expression of PPARγ2 was markedly higher in HT29 cells and LS174T cells as compared to RKO, DLD1, and HCT116 cells. The lack of change in PPARγ expression between wild-type and Apcmin heterozygous mice is in contrast to previous work by others showing increased expression of PPARγ in Apcmin heterozygous mouse colon as compared to wild type . However, the increased expression of PPARγ2 found in HT29 and LS174T cells is consistent with the findings that activation of β-CATENIN/TCF4 signaling can increase PPARγ expression in SW480 cells . The significance of these observations remains to be determined.
A related important finding from the present studies is the lack of change in PPARβ/δ protein in the colon from mice treated with NO-ASA, a relatively unique form of an NSAID. A previous report suggested that expression of PPARβ/δ was ~10-fold higher in the colonic nuclei of Apcmin heterozygous mice as compared to wild type and that treatment of NO-ASA caused downregulation of PPARβ/δ expression in the colon and polyps . This putative downregulation was thought to contribute to the mechanisms underlying the chemopreventive effects of NO-ASA. However, results from the present studies quantitatively demonstrate that expression of PPARβ/δ protein is not higher in nuclear extracts from colon of Apcmin heterozygous mice as compared to wild type and that NO-ASA did not cause downregulation of PPARβ/δ in the colon. This discrepancy is likely explained by differences in the method of PPARβ/δ detection. The publication reporting these effects based their conclusion on immunohistochemical analysis of PPARβ/δ , whereas the present analysis used quantitative Western blotting using nuclear extracts from tissues from the same experiment. It was recently shown that there is considerable background immunoreactivity of even a highly specific anti-PPARβ/δ antibody  indicating that immunohistochemical analysis of PPARβ/δ is likely not suitable for quantifying PPARβ/δ. This is significant because there are a number of reports that used immunohistochemical analysis to examine PPARβ/δ expression in colon cancer models. For example, immunohistochemical analysis of PPARβ/δ suggests that expression of PPARβ/δ is increased in some human colon cancer patients but no quantified statistical analysis of Western blots was provided . Additionally, another report suggested that expression of PPARβ/δ was higher in flat dysplastic adenomas from Apcmin heterozygous mice but Western blot analysis showed no change in the expression of PPARβ/δ in adenomas as compared to normal mucosa despite significant upregulation of CYCLIND1. Results from the present studies and the possible nonspecific immunoreactivity of anti-PPARβ/δ antibodies indicate that analysis of PPARβ/δ in colon cancer models should always include quantitative Western blotting. In addition to quantifying expression of PPARβ/δ, analysis of known PPARβ/δ target genes should also be performed in order to establish a functional role for potential differences in receptor expression since PPARβ/δ functions as a ligand-dependent transcription factor.
Another major finding from the present studies is the lack of downregulation of PPARβ/δ by NSAIDs in human colon cancer cell lines. There are conflicting reports with some suggesting that NSAIDs decrease expression of PPARβ/δ while others suggest that NSAIDs increase or have no effect on the expression of PPARβ/δ in cancer models [reviewed in 1]. The focus of the present studies was on RKO and DLD1 human colon cancer cell lines for several reasons including the fact that RKO cells express COX1/2 while DLD1 cells do not (this allows for distinguishing between COX1/2-dependent vs. COX1/2-independent effect), the fact that RKO cells have wildtype APC and β-CATENIN while DLD1 cells have a mutation in APC (this allows the ability to determine whether PPARβ/δ is upregulated by the APC/β-CATENIN/TCF4 pathway) and a previous report suggesting that a variety of NSAIDs (indomethacin, sulindac, NS-398, and celecoxib) cause reduced expression of PPARβ/δ in the same cell lines . Comprehensive and quantitative analysis of PPARβ/δ expression following treatment with NSAIDs at concentrations that effectively inhibited cell growth, and at time points preceding this inhibition, demonstrates that no decrease in expression or function of PPARβ/δ protein is observed in either RKO or DLD1 cells as compared to controls. This is indeed surprising since the same cell lines and time point (48 h) were examined using concentrations of indomethacin (200 µM vs. 150 µM), celecoxib (10 µM vs. 12.5 µM), and NS-398 (100 µMvs. 120 µM)that are comparable between the present study and former study , respectively. Others have recently suggested that NSAIDs (indomethacin and sulindac) induce apoptosis in colon cancer cell lines by PPARβ/δ-dependent suppression of 14-3-3ε . However, no decrease in PPARβ/δ protein expression was noted in DLD1 cells or HT29 cell (data not shown) by either NSAID. The reason for this and other discrepancies in the literature cannot be determined from the present studies. There are a number of problems with some of the reports suggesting NSAIDs cause a decrease in PPARβ/δ [reviewed in 1], including the lack of positive controls, no independent replicates, and the absence of quantification and statistical analysis. Examination of changes in PPARβ/δ-dependent gene expression in the present study strongly supports the Western blot analysis as no evidence of decreased PPARβ/δ activity was observed following treatment with NSAIDs. Future studies examining NSAID-induced changes in PPARβ/δ should include analysis of PPARβ/δ target gene expression to yield more convincing evidence rather than relying on receptor expression alone.
While no evidence of decreased PPARβ/δ following NSAIDs in human colon cancer cell lines was noted, increased expression of PPARβ/δ and PPARα was observed for several NSAIDs. For example, increased expression of PPARβ/δ was observed in RKO cells by indomethacin after 24 and 48 h of treatment. This is consistent with mRNA expression as well. More importantly, the increase in expression of PPARβ/δ by indomethacin is also reflected by increased expression of known PPARβ/δ target genes. This is of interest because previous studies have also noted an increase in PPARβ/δ mRNA expression by indomethacin treatment in HCT116 cells  or SW-480 cells . This suggests that increased expression of PPARβ/δ, not decreased expression might be mechanistically linked to the observed inhibition of cell proliferation induced by indomethacin. This finding is supported by another report showing enhanced cell proliferation by silencing PPARβ/δ expression in HCT116 cells . Expression of PPARα was also found to be significantly higher in RKO and DLD1 cells treated with the higher doses of NSAIDs. There is evidence that activation of PPARα can inhibit colon tumorigenesis based on in vitro and in vivo analysis . Thus, the significance of the observed increase in PPARα expression and whether the increase in PPARα function could mediate the cell growth modulatory effects of NSAIDs should also be examined based on these findings.
Collectively, results from these studies strongly suggest that the previously described mechanisms suggesting that the APC/β-CATENIN/TCF4 pathway upregulates PPARβ/δ and that NSAIDs inhibit cell growth by decreasing PPARβ/δ expression should be evaluated more rigorously. Importantly, evidence of direct modulation of PPARβ/δ-dependent gene expression coupled with quantified statistical analysis of PPARβ/δ expression should be examined for this purpose. Indeed, results from the present studies strongly suggest that increased expression of PPARβ/δ by indomethacin could contribute to indomethacin-induced inhibition of cell proliferation in human colon cancer cell lines. This idea is supported by a large body of literature demonstrating PPARβ/δ-dependent inhibition of cell proliferation [reviewed in 1].
We gratefully acknowledge Drs. Andrew Billin and Timothy Willson for providing GW0742. This work was supported in part by CA129467 (M.L.C.), CA97999, CA124533 (J.M.P.), and CA92423 (B.R.).
Additional Supporting Information may be found in the online version of this article.