Prostaglandin E₂ Enhances Intestinal Adenoma Growth via Activation of the Ras-Mitogen-Activated Protein Kinase Cascade

Colorectal cancer develops according to a complex and multistep process which involves genetic alterations and progressive changes in signaling pathways that regulate intestinal epithelial cell proliferation, differentiation, and apoptosis. A large body of data supports a role of the cyclooxygenase-2 (COX-2) enzyme in colorectal cancer progression (1, 2) and indicates that inhibition of COX-2 reduces tumor growth (3). Moreover, it is well established that cyclooxygenase enzymes play a key role in intestinal adenoma formation in Apc^min mice, a model frequently employed for studying colorectal cancer (4), and in Apc^Δ716 mutant mice (5). Tumor number decreases when Apc mutant mice are treated with COX-2-selective inhibitors (5). In addition, in vitro studies confirm that expression of COX-2 is correlated with intestinal epithelial cell proliferation and invasiveness, a reduction in the apoptotic rate, and induction of proangiogenic factors such as vascular endothelial growth factor (6–9).

COX-2 is an inducible gene which is regulated by a number of factors, including serum, growth factors, proinflammatory cytokines, hormones, oncogenes, or tumor promoters (10). Various extracellular stimuli regulate COX-2 expression through a mitogen-activated protein kinase (MAPK)–dependent pathway. For example, transforming growth factor α, INFγ, and platelet-derived growth factor induce COX-2 expression via activation of the extracellular signal–regulated kinase (ERK) signal transduction pathway in normal human epidermal keratinocytes, squamous carcinoma cells, and NIH 3T3 cells, respectively (11, 12).

COX-2 is the key enzyme in the conversion of arachidonic acid to prostaglandins, such as PGE₂, PGD₂, PGF₂α, and PGI₂ (10, 13–15). PGE₂ is the most abundant prostaglandin found in human colorectal cancers, premalignant lesions, and cells derived from a number of solid malignancies (16–19). PGE₂ exerts its actions either in autocrine or in paracrine fashion via binding to G-protein coupled receptors (EP1-4), which belong to the family of rhodopsin-type receptors. Genetic studies using mice lacking the PGE₂ cell surface receptors EP1, EP2, or EP4 point to an important role for all three receptors in intestinal polyp formation (20–22). Moreover, EP1 or EP4 receptor antagonists decrease the incidence of intestinal adenomas in both the Apc^min and the carcinogen-treated mouse models (20, 21). These results provide strong evidence that PGE₂ plays a pivotal role in regulating intestinal adenoma formation. We recently reported that PGE₂ treatment increases adenoma burden in Apc^min mice (23). In addition, in vitro studies have shown that PGE₂ enhances clonogenicity and increases invasiveness of LS-174T carcinoma cells by activating the EP4-PI3k-Akt signaling cascade (24) and that PGE₂ promotes integrin αvβ3–dependent endothelial cell adhesion and spreading (25). However, all of the mechanisms responsible for the effects of PGE₂ on intestinal adenoma growth are not known.

Ras mutations are found in a wide variety of human malignancies and in about 50% of colorectal carcinomas (26). In rodents, AOM-induced colonic carcinogenesis involves activation of the K-Ras gene (27). For example, K-Ras mutations were identified in 14 of 84 AOM-induced colonic tumors. Most importantly, a subset of tumors (18 of 70) had significantly higher activation of wild-type K-Ras compared with controls, suggesting that the activation of wild-type Ras is also involved in AOM-induced colonic carcinogenesis. Moreover, forced expression of constitutively active Ras (mutant Ras) up-regulates COX-2 expression and enhances cell proliferation in a variety of cell culture models (28–31). Therefore, we examined whether PGE₂ activates endogenous wild-type Ras that is known to regulate cell proliferation.

It is well known that activation of Ras triggers the downstream signaling pathways such as the Raf/MAPK kinase (MEK)/ERKs and PI3K/Akt pathways. In the Raf/MEK/ERKs pathway, activated Ras recruits Raf to the plasma membrane, which leads to phosphorylation of MEK, a dual specificity kinase that phosphorylates the Thr-X-Tyr motif in the activation loop of ERK. Upon activation, ERK translocates to the nucleus and regulates the activity of many transcription factors including Elk-1 (32, 33). Elk-1 is a member of the ternary complex factor family of Ets domain proteins that bind to serum response elements, a cis-element responsible for activation of immediate-early gene expression following mitogenic stimulation (34). It has been reported that serum, growth factors, and phorbol 12-myristate 13-acetate stimulate the phosphorylation of Elk-1 via the Raf-MEK-ERK pathway (35). Moreover, other seven transmembrane spanning G-protein-coupled receptor agonists have been shown to activate a Ras-MAPK signaling pathway (36). Because cytokines induce COX-2 expression by activating the ERK signal transduction pathway, we postulated that PGE₂ could up-regulate COX-2 through activation of Ras-MAPK signaling.

To investigate the mechanism responsible for the effect of PGE₂ to promote intestinal adenoma growth, we examined whether this lipid mediator stimulates intestinal epithelial cell proliferation and determined the downstream targets of PGE₂ using both in vivo and in vitro models. Here we show that PGE₂ enhances intestinal cell proliferation and induces COX-2 expression in Apc^min mouse adenomas. We further show that activation of Ras-MAPK pathway is required for PGE₂ to induce COX-2 expression and stimulate HCA-7 cell proliferation. Consistent with these findings, we also observed that PGE₂ enhances ERK and Elk-1 activity in intestinal adenomas and phospho-Elk-1 levels are dramatically elevated in human colorectal cancers compared with matched normal mucosa. To our knowledge, this is first report that COX-2 induction by PGE₂ is mediated by activation of the Ras-MAPK pathway.

Animals. C57BL/6J-Apc^min male mice were obtained from the Jackson Laboratory (Bar Harbor, ME) at 5 weeks of age. After 1 week, Apc^min mice (n = 6) were randomly assigned to either a control or treatment group (vehicle and 150 μg PGE₂ per mice). Vehicle or PGE₂ in 100 μl sterile PBS was given twice daily via gavage feeding. After treatment for 7 weeks, mice were injected with 0.25 mL of BrdUrd labeling reagent (Zymed Laboratories, Inc., South San Francisco, CA) and sacrificed after 3 hours. The intestines were collected, opened longitudinally, fixed, and embedded in paraffin as reported previously (23). Sections were then used for cell proliferation assays and immunohistochemical staining.

In vivo Cell Proliferation Assays and Immunohistochemical Staining. Proliferating cells in tissue sections were detected using with an anti-BrdUrd antibody according to the instructions for ZYMED BrdUrd kit (Zymed Laboratories). In addition, tissue sections also were stained with mouse monoclonal antibodies against phospho-Elk-1 (Ser383) or phospho-ERK1/2 (Tyr204) at a dilution of 1:250 (Santa Cruz Biotechnology, Santa Cruz, CA). The immunohistochemical staining was completed by using a Zymed-Histostain-SP Kit (Zymed Laboratories).

Cell Culture. HCA-7 cells were maintained in McCoy's 5A medium containing 10% fetal bovine serum and penicillin-streptomycin. To arrest cell growth, cells were cultured in the absence of serum for 24 hours. The MEK1 (PD98059), the Ras (Ftase inhibitor III), and PI3K (LY294002) inhibitors (Calbiochem-Novabiochem Co., San Diego, CA) were prepared as a stock in DMSO (50 μmol/L).

Cell Proliferation ELISA and Cell Growth. Cell proliferation was measured using a Cell Proliferation ELISA kit (Roche Molecular Biochemicals, Indianapolis, IN) following the manufacturer's instructions. Briefly, HCA-7 cells (1 × 10⁴ cells per well) grown on 96-well culture plates were pretreated with either DMSO or inhibitors for 1 hour and then treated with the indicated concentration of PGE₂ for 24 hours after serum starvation. The cells were then labeled with BrdUrd for an additional 6 hours. Incorporated BrdUrd was measured colorimetrically with an ELISA reader, Spectra MAX 340PC (Molecular Devices, Sunnyvale, CA). Cell vitality was measured using a trypan blue exclusion assay.

Whole Cell Extracts and Western Blot Analysis. Whole cell extracts were prepared from cells treated with either vehicle or PGE₂ for the indicated times and after serum starvation for 24 hours. Western blots were done following protocols provided by Santa Cruz Biotechnology. The cells were lysed in 0.6 mL of radioimmunoprecipitation assay buffer with protease inhibitor cocktail tablets (Boehringer Mannheim Co., Indianapolis, IN) and 0.2 μmol/L sodium orthovanadate. Fifty micrograms of soluble protein were fractionated in a 10% SDS-PAGE reducing gel, and then blotted onto a 0.2-μm nitrocellulose membrane (Bio-Rad, Hercules, CA). The membrane was blocked with 5% dry milk in TBS-T buffer for 1 hour and then incubated for 12 to 16 hours at 4°C in a 1:1,000 dilution of a pan-Ras antibody (AB-3; Oncogene Research Products, Cambridge, MA), an anti-phospho-ERK1/2, or the anti-phospho-Elk-1, anti-Elk-1, anti-COX-2, and anti-ERK1/2 (Santa Cruz Biotechnology) in TBS-T buffer containing 5% dry milk. After three washings with TBS-T buffer, the membrane was incubated in a 1:3,000 dilution of the appropriate anti-mouse or anti-rabbit immunoglobulin conjugated with horseradish peroxidase (Boehringer Mannheim) in TBS-T buffer with 5% dry milk for 1 hour at room temperature. After three washings with TBS-T buffer, the protein bands were detected with the enhanced chemiluminescence Western blotting detection reagents (Amersham Pharmacia Biotech, Piscataway, NJ) according to the manufacturer's instructions. The blots were stripped and then reprobed with anti-ERK2 or anti-Elk-1.

Transient Transfection Assays. Cells (1.5 × 10⁵ in 12-well plates) were transiently cotransfected with 0.2 μg of COX-2 (−327 to +59) luciferase reporter gene and 5 ng of pRL-SV40 by the LipofectAMINE Plus reagent following manufacturer's protocol (Life Technologies, Inc., Rockville, MA). For cotransfection assays, the cells were transiently cotransfected with 0.3 μg COX-2 (−327 to +59) luciferase reporter gene and 5 ng of pRL-SV40 and 0.4 μg of empty vector or dominant-negative PI3K plasmids. Three hours later, the cells were placed in fresh serum-free media and incubated for another 4 hours. The cells were then treated with PGE₂ after pretreatment with vehicle or inhibitors for 1 hour. After 16 hours, cells were harvested in 1× luciferase lysis buffer. Relative light units from firefly luciferase activity was determined using a luminometer, Monolight 3010, (BD Biosciences/PharMingen, San Diego, CA) and normalized to the relative light units from Renilla luciferase using the Dual Luciferase kit (Promega, Madison, WI).

Ras Activation Assays. Ras activity was measured using a Ras Activation Assay Kit (Upstate Biotechnology, Inc., Lake Placid, NY) following the manufacturer's instructions. Briefly, quiescent cells were stimulated with PGE₂ at indicated concentrations and for indicated times. Cells were washed twice with ice-cold HBS and lysed in 1× Mg²⁺ lysis/washing buffer containing protease inhibitor cocktail tablets (Roche Molecular Biochemicals) for 15 minutes at 4°C. Cell lysates were centrifuged at 1,000 × g for 20 minutes. The supernatants were pretreated with glutathione-Sepharose-4B beads (Amersham Pharmacia Biotech) and the protein concentrations of the supernatants were then determined (Bio-Rad). Equal amounts of samples (400 μg) were immediately affinity-precipitated using 20 μg of recombinant glutathione S-transferase-c-Raf-1 ras binding domain (1149) fusion proteins conjugated glutathione-Sepharose beads (Amersham Pharmacia Biotech) for 1 hour at 4°C. The precipitates were washed thrice with 1× Mg²⁺ lysis/washing buffer and eluted by boiling in 1× SDS-PAGE sample buffer. The proteins were separated on a 12% SDS-polyacrylamide gel and then immunoblotted with pan-Ras antibody (AB-3; Oncogene Research Products). To normalize the amount of GTP-bound Ras to total amount of Ras, equal volumes of cell lysate were also subjected to Western blot analysis using the pan-Ras antibody.

PGE₂ Induces Epithelial Cell Proliferation and COX-2 Expression in Intestinal Adenomas. Because epithelial cell proliferation may be one of the key mechanisms by which PGE₂ promotes intestinal adenoma growth, we investigated whether PGE₂ would affect the adenoma epithelial cell proliferation in the Apc^min mice. We treated 6-week-old Apc^min mice with vehicle (PBS) or PGE₂ for 7 weeks. Before sacrifice, mice were given an injection of BrdUrd to label the proliferating cell fraction. Significant BrdUrd uptake was seen in epithelial cells found in adenomas from PGE₂-treated mice as compared with controls (Fig. 1A). To further understand the molecular mechanism by which PGE₂ induces cell proliferation, we examined COX-2 expression in these adenomas because of our previous data indicating that elevated COX-2 expression was correlated with enhanced cell proliferation (37). Immunohistochemical staining and Western blot analyses were done to assess COX-2 expression levels. As shown in Fig. 1B, PGE₂ treatment resulted in a dramatic increase in COX-2 expression in intestinal adenomas. These results show that PGE₂ enhances COX-2 expression and stimulates epithelial cell proliferation in adenomas.

PGE₂-Up-Regulated COX-2 Expression Is Dependent on Ras-MAPK Activation. To further elucidate the signaling pathways responsible for PGE₂-stimulated cell proliferation and COX-2 expression, we evaluated the effects of PGE₂ in HCA-7 cells that are known to exhibit COX-2-dependent proliferation (37, 38). HCA-7 cells have been carefully evaluated as an in vitro model to investigate COX-2-dependent cell growth (37, 38). These cells express PGE₂ receptors (EP1, EP3, and EP4) and contain wild type Ras. As a first step in identifying the signaling pathways involved, we examined the effect of PGE₂ on HCA-7 cell proliferation and COX-2 expression. A cell proliferation assay was used to directly measure the PGE₂-induced cell proliferation based on the measurement of BrdUrd incorporation during DNA synthesis. As shown in Fig. 2A, cell proliferation was stimulated following 24 hours of PGE₂ treatment in a dose-dependent manner. In addition, compared with untreated cells, PGE₂ treatment for 5 days increased cell number by 2-, 3.2-, 3.6-, 4.2-, and 5.2-fold at 0.001, 0.01, 0.1, 1, and 10 μmol/L, respectively (Fig. 2B). These results show that PGE₂ can stimulate HCA-7 cell proliferation in a dose-dependent manner.

Next, we determined the ability of PGE₂ to affect COX-2 promoter activity in HCA-7 cells. As shown in Fig. 3A, a dose-dependent increase in luciferase activity (representing COX-2 promoter activity) was seen with PGE₂, but not with PGD₂, cPGI₂, or PGF₂α. Since our previous results showed that overexpression of constitutively active Ras up-regulates COX-2 expression (29), we determined whether the Ras-MAPK cascade mediates PGE₂ induction of COX-2 expression in colorectal cancer cells. Our results show that either a highly selective Ras inhibitor (Ftase inhibitor III) or a MEK inhibitor (PD98059) blocks the PGE₂-enhanced COX-2 promoter activity in a dose dependent manner, but not PI3k inhibitor (Ly294002; Fig. 3B). Overexpression of a dominant-negative Ras also inhibited PGE₂-up-regulated COX-2 promoter activity, whereas expression of a dominant negative PI3k failed to inhibit PGE₂-mediated COX-2 promoter activity (Fig. 3C). Similarly, both inhibitors of Ras and MEK inhibit PGE₂-enhanced COX-2 protein expression, but not PI3k inhibitor (Fig. 3D). Furthermore, overexpression of a dominant negative Ras also blocked both basal and PGE₂-induced COX-2 expression (COX-2 levels were barely detected; Fig. 3D , bottom). These results show that COX-2 is up-regulated by its own downstream product, PGE₂, and that this autoinduction is dependent on activation of the Ras-MAPK pathway.

PGE₂ Activates Ras-MAPK. To further show that PGE₂ can induce endogenous Ras (wild type) activity, HCA-7 cells were treated with PGE₂ for various times following 24 hours of serum starvation. Ras activation assays were done by selective affinity precipitation of GTP-bound Ras with immobilized glutathione S-transferase-c-Raf-1 ras binding domain and detection of Ras-GTP by Western blotting with a pan-Ras antibody. We found that PGE₂ increases activation of Ras in a biphasic manner (Fig. 4A). The early phase of Ras activation is dose dependent (Fig. 4B,, lanes 1, 2, 4, and 6), whereas the late phase of Ras activation is independent of PGE₂ concentration (lanes 1, 3, 5, and 7). Ras activation was evident as early as 1 minute, reached a maximum level at 5 minutes followed by return to the basal level by 20 minutes after 0.1 μmol/L PGE₂ treatment. However, Ras activation again increased at 180 minutes and persisted for a prolonged time (data not shown). To confirm that equivalent amounts of protein were loaded on the gel, the level of total Ras protein from all samples was examined by Western blotting. All the samples exhibited similar levels of total Ras (Fig. 4A -B, bottom). These results show that PGE₂ stimulates Ras activation in a time-dependent manner.

Because the MAPK cascade is one of the Ras downstream signaling pathways, we next examined whether PGE₂ can activate MAPK signaling. HCA-7 cells were treated with PGE₂ for 5 or 180 minutes. Western blotting was then done with an antibody which recognizes the phosphorylated (activated) form of ERK1/2 and its downstream target ELK-1. As shown in Fig. 4C, PGE₂ treatment enhanced phosphorylation of both ERK1/2 and ELK-1 without affecting the levels of total ERK1/2 and ELK-1 proteins. However, there is evidence that the activation of ERK by other G-protein coupled receptor agonists involves both Ras-Raf1-MEK dependent and independent effects. Thus, we examined whether Ras and MEK are required for PGE₂-enhanced ERK activation. HCA-7 cells were pretreated with the selective Ras and MEK inhibitors for 1 hour and then treated with PGE₂ for 5 minutes. Both of the inhibitors significantly blocked the PGE₂-stimulated ERK activation as well as its downstream target Elk-1 (Fig. 4D) without showing apparent signs of cytotoxicity at the concentrations and incubation times studied. Furthermore, the late phase of ERK and Elk-1 activation by PGE₂ (180 minutes) was also sensitive to both inhibitors (data not shown). These results show that a Ras-Raf-MEK cascade mediates the PGE₂-induced ERK and Elk-1 activation.

In determining the biological significance of a signaling pathway, results obtained in vivo are helpful to verify the data obtained from cultured cells. Thus, we examined whether PGE₂ can activate the MAPK cascade in Apc^min mouse adenomas in vivo. As shown in Fig. 5A, PGE₂ treatment resulted in dramatic increases in both phosphorylated ERK1/2 and Elk-1 in intestinal adenomas as determined by immunostaining. Moreover, because ERK and c-jun-NH₂-kinase activities increase modestly in a subset of human colorectal carcinomas (39), we examined whether the Elk-1 activation is also increased in human colorectal cancers. Western blotting was done to examine Elk-1 activation in human colon cancers and matched normal tissues. Among the 15 pairs of human colon cancer and the matched normal mucosal samples, eight pairs showed high levels of Elk-1 activation (phosphorylation) in the cancer specimens (Fig. 5B). To our knowledge, this is the first report that Elk-1 activity is increased in colorectal cancer. Taken together, these results show that PGE₂ up-regulates COX-2 expression through activation of the Ras-MAPK pathway.

A Ras-MAPK Cascade Is Required for PGE₂-Stimulated Cell Proliferation. Finally, we examined whether the Ras-MAPK pathway is required for PGE₂-induced cell proliferation. HCA-7 cells were pretreated with either the Ras or MEK inhibitor for 1 hour, and then treated with indicated concentration of PGE₂. Both cell proliferation assay by ELISA and cell counting were used to measure cell proliferation and cell growth. We observed that PGE₂-induced cell proliferation was totally blocked by the Ras inhibitor at 20 μmol/L or MEK inhibitor at 25 μmol/L, respectively (Fig. 6). These data show that the Ras-MAPK cascade is required for PGE₂-stimulated cell proliferation and growth.

There is considerable preclinical evidence indicating that COX-2 plays a role in the development of various types of cancers, including colorectal, breast, bladder, skin and lung cancer. COX-2 derived PGE₂ was found to promote colorectal adenoma growth in Apc^min mice. The aim of our current study was to understand the molecular mechanism(s) by which PGE₂ promotes intestinal tumor growth.

Our laboratory was the first to report significant elevation of COX-2 expression in 85% of human colorectal carcinomas and in ∼50% of colorectal adenomas (1). However, the mechanism(s) by which COX-2 is highly expressed in a number of solid malignancies is not yet completely understood. COX-2 mRNA and protein are normally very low or undetectable in normal intestinal tissues, but are rapidly induced in response to inflammation, cytokines, growth factors, oncogenes, endotoxins, and other chemicals. Here we show that PGE₂ can amplify the expression of COX-2, a key regulatory enzyme in the PG biosynthetic pathway, in colorectal carcinoma cells through a positive feedback loop. This self-amplifying loop may help explain why COX-2 is constitutively overexpressed in majority of colorectal cancers.

Our results show that PGE₂ activates Ras in an early (5 minutes) and a late responsive manner (3-6 hours). A likely explanation for this observation is that there is a negative feedback loop that inactivates EP receptor function following ligand binding, which in turn inhibits Ras activation. In general, ligand binding to G-protein coupled receptors results in receptor phosphorylation, desensitization, and sequestration. For example, chemokine receptor CXCR2 desensitization occurs within 1 minute, whereas receptor sequestration is a much later event (30-60 minutes) (40). We postulate that PGE₂ binding to its cognate receptors also leads to the desensitization and sequestration of EP receptors, which provides negative feedback after 5 minutes of PGE₂ stimulation (Fig. 4A). Another possible explanation is that PGE₂ is metabolized to an inactive cyclopentenone PGA₂ in cultures similar to that has been observed in other biological fluids (41). Therefore, Ras activation in the late activation phase may depend primarily upon newly synthesized PGE₂ through an autocrine loop.

Our results also show that Ras activation during the early phase results in increased COX-2 expression via a MAPK-dependent pathway. We have previously shown that COX-2 expression is elevated within 30 minutes after transforming growth factorα or 12-O-tetradecanoylphorbol-13-acetate treatment in rat intestinal epithelial (RIE-1) cells (42). In addition, PGE₂ increases COX-2 mRNA expression in human prostatic carcinoma PC-3 cells with the highest levels of stimulation seen at 3 hours (43). These and our present results suggest that the constitutive (late phase) Ras activation is dependent on the newly synthesized PGE₂ by COX-2. Alternatively, increased levels of PGE synthase in addition to COX-2 may drive new synthesis of PGE₂. This is consistent with the observation that PGE₂ synthase is up-regulated after activation of the MAPK pathway (44). Our present findings suggest a novel mechanism by which COX-2-derived PGE₂ constitutively activates Ras in a positive autocrine feedback loop. Thus, PGE₂ may mimic the effects of an aberrant Ras function even in the absence of an actual Ras gene mutation. This implies that the influence of wild-type Ras in human colorectal cancer is even greater than what is expected based on the known frequency of Ras mutations.

Activated ERK can directly phosphorylate transcription factors, such as Elk-1, which binds to the serum-responsive element of the c-fos promoter (45, 46). In this respect, PGE₂ can induce the expression of early response genes, such as c-fos, c-jun, jun B, and egr-1 (47) and heterodimers or homodimers of the Jun and Fos family members form the activator protein transcription factor complex. A putative activator protein cis-activating consensus element is present in the COX-2 promoter and activator protein activation mediates the induction of COX-2 in intestinal epithelial cells in response to the G-protein-coupled receptor ligand, Bombesin (48). Thus, it is likely that PGE₂ signals through a Ras-MAPK-dependent pathway to activate activator protein, which then regulates COX-2 gene expression, leading to increased cell proliferation. However, further investigation is needed to confirm this speculation. Nonetheless, analysis of signaling pathways affected by PGE₂ reveals that a Ras-ERK-Elk-1 pathway mediates PGE₂ downstream signaling leading to increased cell proliferation and up-regulation of COX-2 expression. In conclusion, we show for the first time that a key oncogenic signaling pathway mediate the regulation of COX-2 by PGE₂ in colorectal carcinoma cells. Our data support a novel mechanism by which COX-2-derived PGE₂ promotes human cancer cell growth by autoregulation of COX-2 expression, which depends primarily on PGE₂-induced Ras-MAPK pathway. An understanding of PGE₂ downstream signaling transduction pathways critical for tumor growth control may lead to the development of novel therapeutic and chemoprotective options for colorectal cancer patients.

Grant support: USPHSs grants RO1DK 62112 (R.N. DuBois), P01-CA-77839 (R.N. DuBois), R37-DK47297 (R.N. DuBois), R37 HD12304-26 (S.K. Dey) and HD33994 (S.K. Dey), T.J. Martell Foundation (R.N. DuBois), and National Colorectal Cancer Research Alliance (R.N. DuBois).

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