Differential regulation of EGFR-MAPK signaling by deoxycholic acid (DCA) and ursodeoxycholic acid (UDCA) in colon cancer

Introduction

Bile acids are cholesterol derivatives that play an important role in fat metabolism. The two main functions of bile acids are to emulsify lipid aggregates, and to solubilize and transport lipids in hydrophilic environments. In addition, bile acids can serve as signaling molecules, capable of activating receptors as well as signaling pathways. Moreover, bile acids have been implicated in the pathogenesis of many diseases, particularly colon cancer.

The seven most common bile acids present in the human gut include: cholic acid (CA), chenodeoxycholic acid (CDCA), glycocholic acid, taurocholic acid (TCA), lithocholic acid (LCA), deoxycholic acid (DCA) and ursodeoxycholic acid (UDCA). These bile acids can be categorized into three groups, primary, secondary and tertiary. Primary bile acids are synthesized from cholesterol in the liver, and are the source from which secondary and tertiary bile acids are derived. Secondary bile acids are the product of intestinal bacterial metabolism of primary bile acids, whereas tertiary bile acids are formed via hepatocyte metabolism of reabsorbed primary bile acids, or primary bile acid intermediates [1–3].

After being synthesized in the liver, primary bile acids CA and CDCA can be conjugated to amino acids such as glycine or tarurine, and actively secreted into the gallbladder. Whether bile acids are conjugated or unconjugated may influence their biological activity. High fat intake stimulates the release of enteric hormones cholecystokinin (CCK) and secretin. The release of these hormones causes contractions of the gallbladder and common bile duct, resulting in delivery of bile into the small intestine, 67% of which is composed of bile acids. In the small intestine, bile acids perform the necessary task of emulsifying and transporting lipids and lipid soluble vitamins. Bile acids travel to the ileum of the small intestine, where 95% are actively transported into the portal vein. The remaining 5% travel to the colon where the glycine and taurine can be cleaved by hydrolysis, resulting in what are known as free bile acids. The free primary bile acids can then be converted to secondary bile acids such as DCA, and to a lesser extent UDCA through intestinal bacterial metabolism of cholic acid or chenodeoxycholic acid, respectively [4]. It is important to note that UDCA may be generated in the colon or in the liver, and therefore may be classified as a secondary or tertiary bile acid, respectively. From the colon both primary and secondary bile acids, along with intermediates can be passively transported in venous blood to the liver via the portal vein. Once in the liver, hepatocytes extract the bile acids from sinusoidal blood, and can metabolize them into tertiary bile acids. In some cases UDCA can be formed through the reduction of the primary bile acid intermediate, 7-Oxo-lithocholic acid by hepatocytes, and can therefore also be considered a tertiary bile acid [5]. Notwithstanding, the exact proportion of UDCA produced in this manner is unknown.

Bile acids have diverse biological functions. For example, secondary bile acids are implicated in the development of disease while tertiary bile acids, such as UDCA, may be protective agents. Secondary bile acids are found in several organs within the gastrointestinal system, such as the gall bladder, bile ducts, stomach, esophagus and colon and are associated with the development of neoplasias in all of these organs. DCA, in particular, has been suggested to be a significant contributor to the development of colon cancer. Another secondary bile acid that may contribute to colon cancer progression is LCA, but there is much less evidence supporting this. Conversely, the tertiary bile acid UDCA, has shown some promise in clinical trials as a chemopreventive agent. Interestingly, the chemical structures of DCA and UDCA are very similar, differing only by the position of 1 hydroxyl group which is either located in the C-7 or C-12 positions of the cholesterol nucleus in DCA or UDCA respectively (Figure 1A).

It is not completely understood why DCA and UDCA have different biological effects while maintaining such similar structures. One reason may be that the different position of their hydroxyl group contributes to the 3-dimentional conformation of the molecules. Figure 1B illustrates the conformational differences between DCA and UDCA. This variation in structure may affect the ability of DCA and UDCA to interact with other molecules. Another explanation is that the hydroxyl position influences the bile acids hydrophobicity [6]. DCA is strongly hydrophobic while UDCA is not. Relative hydrophobicty of these bile acids has been shown to correlate with their ability to induce apoptosis in colorectal cells[6]. Also, this property may allow for DCA to cause membrane effects, such as lipid rearrangement, and caveolin-1 distribution. These effects have been attributed to cholesterol disruption which is only caused by the strongly hydrophobic bile acids [7]. Interestingly, other bile acids that are hydrophobic, like CDCA behave similarly to DCA, causing apoptosis and membrane perturbations while the less hydrophobic bile acid CA exerts cytoprotective effects similar to UDCA [8–10]. Most likely a combination of these features determines DCA and UDCA’s contrasting effects on cellular processes and pathology of disease.

Of the many bile acids that are known, DCA and UDCA have been the most extensively studied, with regard to colon cancer development. Notwithstanding, the exact mechanisms for the tumor promoting effects of DCA and the anti-tumorgenic actions of UDCA have not been fully elucidated. One explanation for the profoundly different biological effects of these structurally similar molecules is the differential regulation of oncogenic signaling pathways. One such pathway that both DCA and UDCA have been shown to act on is the EGFR-MAPK pathway. Over-activation of this pathway in colon cancer leads to upregulation of a number of genes involved in proliferation and cell survival. Positive regulation of the EGFR-MAPK pathway by DCA and negative regulation by UDCA may be one explanation for the dual roles of these two bile acids in colon cancer progression.

This review will present what is currently known about the diverging biological activities of DCA and UDCA on in vivo colon cancer animal models, as well as clinical observations. Moreover, the influence of these bile acids on signaling pathways shown to be involved in colon cancer development will also be presented. Finally, a detailed examination of the effects of DCA and UDCA on the EGFR-MAPK pathway, as well as potential mechanisms of regulation will be summarized. These data should offer some insight as to how bile acids may influence colon cancer development and prevention.

Clinical and biological activity

In 1940 J. W. Cook provided the first in vivo evidence that subcutaneous injection of DCA into mice caused, what they described as “spindle-celled tumors” [11]. This result was obtained after the mice were subjected to 15 DCA injections containing a total of 70 milligrams of the compound within 300 days. Since then, many groups have examined the effects of DCA on carcinogenesis. The strongest evidence for a link between DCA and carcinogenesis comes from colon cancer studies.

In early studies, DCA was thought to act as a carcinogen, capable of inducing transformation of epithelial cells. Another early hypothesis was that DCA may be a mutagen, responsible for inducing direct or indirect DNA damage. However, in the 1970’s E.L. Wynder’s research group showed that taurodeoxycholic acid (TDCA) could promote the formation of both colorectal adenomas and adenocarcinomas in a rat model [12–14]. In these studies the relationship between bile acids and cancer progression was only observed in animals that were also treated with the potent carcinogen, N-methyl-N′-nitro-N-nitrosoguanidine (MNNG). This new evidence led to the classification of DCA as a tumor promoter.

Although, not direct evidence, several clinical observations support the notion that DCA is a strong tumor promoter. In general, a western diet (high in fat, low in fruits and vegetables), is strongly linked with both a higher incidence of colon cancer as well as higher levels of fecal bile acids, primarily DCA. DCA is present at 3–4 times higher levels in individuals who consume high-fat diet, compared to those who maintained a low-fat diet [15]. In healthy individuals, the intestinal bile acid pool is comprised of 28% DCA, 33.8% CDCA, 38.1% CA, and less than 1% of both LA and UDCA [16]. However, DCA has been shown to account for up to 50% of the total bile acid pool [17]. Moreover, higher levels of serum DCA have been shown to correlate with the presence of colon adenomas in male patients when compared to individuals with negative colonoscopies [18].

Epidemiological associations, analysis of primary human tissue has generated strong evidence that DCA clearly behaves as a tumor promoter. Assessment of human colon biopsies showed that high DCA serum levels positively correlate with elevated proliferation rates of colon epithelium [19]. Moreover, proliferation was stimulated when human colonic biopsies were treated in vitro with DCA [20], an observation reproduced in vivo in animal model(s) [21]. This causal relationship between DCA and colonic epithelium proliferation has also been recapitulated by in vitro studies of AA/C1 human colonic adenoma cells [22]. It is important to mention that the proliferative effect induced by DCA is only observed in normal colonic epithelium or colonic adenoma cells, and not in cancer cells.

Conversely, there is strong evidence that UDCA may act to suppress colon tumor development. The first study which identified the chemopreventive effects of UDCA was conducted in a joint effort by the University of Arizona and the University of Chicago [23]. In this study, 0.4% (w/w) UDCA was able to decrease azoxymethane (AOM)-induced tumors in rats from a 47% to 22% incidence [23]. These data led to the development of several clinical trials [24–27]. In retrospective studies, there was a trend towards decreased adenoma prevalence, a significant decrease in polyp size, decreased probability of adenoma recurrence, and colonic epithelial cell proliferation in patients treated with UDCA [27]. In two other retrospective studies, UDCA treatment was given to patients with ulcerative colitis associated with primary sclerosing cholengitis, which significantly correlated with both an increase in the probability of survival without dysplasia [24] and decreased development of dysplasia [25]. A randomized clinical trial of UDCA’s chemopreventive properties was subsequently conducted at the University of Arizona Cancer Center. In this study, 1285 individuals who had a colorectal adenoma removed within 6 months prior to the trial were randomly given dietary supplementation with or without UDCA. This clinical trial exhibited a trend toward decreased adenoma recurrence, and a statistically significant decrease in the recurrence of colorectal adenomas with high-grade dysplasia in patients who received UDCA supplementation, [26]. Further evaluation of the clinical data presented above showed that gender may alter the response to UDCA [28]. UDCA caused an overall reduction in the development of advanced lesions in men, but led to a significantly higher risk of advanced lesion development in younger women, obese women and women that consumed high fat diets. These finding suggests that UDCA is acting in a complex manner that is not currently well understood.

It is evident that bile acids can have diverse effects on biological properties as well as cancer progression; however, if UDCA is to be used as a chemopreventive agent and the tumor promoting effects of DCA are going to be combated, the mechanisms driving the biological properties of these molecules must be examined further. In order to pursue these questions, researchers have begun investigating the influence of bile acids on signaling pathways, which play a prominent role in tumor development.

Bile acids and signaling

Bile acids were first discovered to act as signaling molecules via their ability to regulate cholesterol 7 alpha-hydroxylase (CYP7a) activity in the late 1960’s [29]. Soon after, bile acids were also shown to activate a number of different receptors, including the farnesoid X receptor-alpha (FXR-α), the preganane X receptor (PXR), the vitamin D receptor (VDR), G protein coupled receptor (TGR5) and EGFR in a variety of cell types [3,30–35]. Moreover, bile acids have been shown to activate signaling molecules directly. Some examples of the molecules activated include c-Jun N-terminal kinase (JNK), protein kinase B (AKT), and extracellular signal-regulated kinases 1 and 2 (ERK1/2), among others. The activation of many of these molecules in hepatocytes allow bile acids to regulate bile acid synthesis, conjugation, and transport, as well as cholesterol synthesis, uptake, and metabolism [3]. DCA and UDCA in particular, have been shown to effect specific signaling molecules in intestinal cells. The effects of bile acids on colon cells are varied from those shown in hepatocytes and other cell types. Table 1 includes a comprehensive account of all molecules shown to be modified by DCA, UDCA or both, and what their specific effects are in intestinal cells. DCA and UDCA have opposing effects on many of these molecules. Furthermore, in many cases UDCA has the propensity to block an effect directly induced by DCA. These instances are indicated by the superscripted symbol α.

Interestingly, a number of molecules influenced by bile acids are members of larger signaling cascades, many of which have been shown to play an integral role in colon cancer development. Figure 2 depicts major signaling pathways that are implicated in colon cancer development and progression. The specific molecules affected by bile acids are indicated by either an asterisk, in the case of UDCA, or highlighted in green if they have been shown to be affected by DCA. Major pathways implicated in colon cancer development include, WNT signaling, phosphatidylinositol 3-kinase (PI3K) signaling, MAPK signaling, p53 deregulation and more recently prostaglandin E2 (PGE2) signaling. The canonical WNT pathway is activated when any number of WNT ligands binds to Frizzled receptors. This interaction activates disheveled (DSH), which then inhibits the complex of proteins composed of axin, GSK-3β, and APC. This complex usually promotes the proteolytic degradation of β-catenin. By activating WNT signaling, this degradation is blocked and β-catenin is able to accumulate in the cytoplasm and then translocate to the nucleus where it is able to interact with TCF and LEF transcription factors, resulting in upregulation of specific genes involved in cellular proliferation [36]. In colon cells APC deficiency or β-catenin activating mutations are the most common early genetic modifications in the transition of normal epithelium to adenomatous polyps, precancerous lesions [37]. DCA has been shown to increase β-catenin activation and nuclear translocation in colon cancer cells [38,39], which may be a potential mechanism of its tumor promoting activity. However, DCA’s effects on APC and other members of this pathway have not been examined (Figure 2). Moreover, there is currently no data elucidating UDCA’s effects on WNT signaling or any member of this signaling cascade.

Another pathway long associated with colon cancer progression is the EGFR pathway. Over-activation of this pathway in tumor cells is associated with tumor cell proliferation, survival, angiogenesis, invasion, and metastasis. Moreover, the expression of EGFR itself may be prognostic for many epithelial tumors including colorectal cancer [40]. EGFR has therefore become a drug target for colon cancer and at least one targeted therapy, Cetuximab, a monoclonal antibody that suppresses EGFR activity, has been approved for treatment of stage four colon cancer. Conventionally, the binding of a ligand such as EGF to EGFR promotes receptor dimerization and phosphorlyation. This results in the activation of downstream signaling cascades such as PI3K and MAPK signaling. Within the PI3K pathway both DCA and UDCA have been shown to phosphorylate AKT at serine 473, indicating that PI3K is being activated by both [39,41–43] (Figure 2). A conjugated form of DCA, TDCA has also been shown to activate the PI3K/AKT pathway, and promote the transactivation of EGFR leading to the promotion of colon cancer cell proliferation and survival [44,45]. Interestingly, this research team determined that matrix metalloproteinase-7 was responsible for TDCA’s ability to transactivate EGFR in H508 human colon cancer cells, a mechanism of activation which has not been evaluated for DCA [46]. In addition to the PI3K pathway, DCA and UDCA affect many of the molecules found within the MAPK branch of the EGFR pathway as well. In fact RAS, RAF, ERK1/2, protooncogene activator protein 1 (AP-1), as well as EGFR itself are all influenced by these bile acids and in an opposing fashion (Figure 2) [47–49].

As depicted in figure 2, another molecule with tumor promoting activities in colon cells is cyclooxygenase 2 (COX-2). Not surprisingly DCA and UDCA have different effects on this molecule. DCA leads to COX-2 activation and increased expression while UDCA is able to block these effects [47]. This may be explained by the fact that induction of COX-2 is associated with EGFR activation in colon cancer cells, as well as RAS activation status [47,50,51], other molecules known to be regulated differentially by DCA and UDCA. COX-2 expression can also be induced by the presence of proinflammatory cytokines. Stimulation of COX-2 leads to increased PGE2 levels, another molecule whose expression is enhanced by the presence of DCA. Once PGE2 production is induced in colon cells, it is able feedback in an autocrine fashion that stimulates a complex signaling cascade. This stimulation of signaling leads to increased cell survival, proliferation, angiogenesis and resistance to apoptosis.

DCA has also been documented to decrease the levels of p53, a tumor suppressor that regulates cell cycle progression [52]. The loss of p53 is an integral requirement for the development of many tumors including adenocarcinomas of the colon. In HCT116 cells the mechanism of p53 suppression by DCA was shown to be partially dependent on ERK1/2 activation, however UDCA had no effect [52]. Taken together it seems clear that the dichotomy between DCA and UDCA’s effects on these pathways may contribute to their dual roles in colon cancer progression. The most dramatic representation of this is demonstrated when examining the EGFR-MAPK pathway. DCA increases expression or the activation status of RAS, RAF, ERK and AP-1, while UDCA exhibits opposing effects on all of these molecules. Furthermore, the activation of MAPK signaling seems to play an integral role in the activation of additional pathways that are imperative to colon cancer development. Albeit, the precise mechanisms of bile acid regulation of the EGFR-MAPK pathway has yet to be fully elucidated. The remainder of this review will evaluate what is known and what is lacking in this particular field of research.

It is important to note that not all studies show that UDCA and DCA have opposing effects. For example activation of the PI3K/AKT pathway results from both DCA and UDCA treatments [42,43,53]. This is not surprising, because the chemical structures of these molecules are so similar; they are likely to have some overlapping functions. This notion is supported by genetic studies which showed that selection for HCT116 cells resistant to the growth inhibitory effects of UDCA also showed resistance to DCA-induced apoptosis [54]. Activation of the PI3K/AKT pathway by UDCA suggests tumor promoting effects like DCA; however, there is a vast array of literature to the contrary. This may be explained by the fact that the cellular response to UDCA stimuli is influenced by its effects on a number of pathways leading to different biochemical properties. DCA and UDCA behave differently with regard to regulation of apoptosis, proliferation, and tumor promotion. These outcomes may be due to the directly opposing effects of these bile acids on other signaling molecules which participate in these functions. Not all studies summarized in Table 1 have conducted side by side comparisons of DCA and UDCA in the same cells lines or using the same concentrations, which allows for some variability when comparing the two; but those that have show that at the same concentration and duration of treatment DCA activates molecules such as PKC, p53, RAS-GTP, AP-1, P38, and ERK1/2 while UDCA suppresses or does not change the activity of these molecules [47,52,55,56]. In other studies UDCA treatments were performed several hours prior DCA treatment. These papers show that UDCA can block DCA-induced caspase activation and PARP cleavage [42]. In addition, UDCA was also able to directly prevent DCA-mediated activation of several MAPK related molecules including P38, Raf-1, ERK1/2, EGFR, and AP-1 when conducted as a pretreatment [47–49]. UDCA is also able to block DCA induced CEBPβ, COX2, and NF-κB [47,49,57]. It is also likely that other factors or signaling molecules which have not yet been examined may also contribute to the differences observed between DCA and UDCA.

Mechanisms of bile acid induced EGFR-MAPK signaling

DCA was first documented to activate the MAPK pathway via direct activation of EGFR in hepatocytes in 2001 by two separate groups Qiao et al. and Rao et al. [58,59]. Both groups showed that treatment of primary rat hepatocytes with 50μM DCA directly induced the rapid phosphorylation of EGFR. Moreover, Qiao et al. identified the mechanism as being a ligand independent process, leading to prolonged ERK1/2 phosphorylation (>4 hours). The ability of DCA to induce MAPK signaling, shown by ERK1/2 phosphorylation, was blocked by dominant negative EGFR, dominant negative RAS, dominant negative MEK1, MEK1/2 inhibitors, and inhibitors of PI3K. Findings by Rao et al. also showed that both TDCA and DCA could induce RAF-1 activity, MEK and ERK phosphorylation, and that DCA was the best activator of ERK, when compared to the other bile acids at the same concentration. The bile acids that were investigated in this comparative analysis were cholic acid (CA), taurocholic acid (TCA), taurodeoxycholic acid (TDCA), chenodeoxycholic acid (CDCA), Ursodeoxycholic acid (UDCA), tauroursodeoxycholate (TUDCA) and glycoursodeoxycholate (GUDCA) [59]. Dominant negative RAS and EGFR were both able to completely block DCA-induced ERK activation, suggesting that DCA induces MAPK activation through the activation of the EGFR in primary rat hepatocytes, and that this is not dependent on the presence of EGFR ligands, such as EGF or TNF-α. However, the mechanism by which DCA activates EGFR was not examined. Notably, inhibition of the EGFR-MAPK pathway enhanced the DCA’s ability to induce apoptosis, supporting the hypothesis that DCA’s ability to induce apoptosis is compensated by its activation of mitogenic signaling pathways [58]. Activation of the MAPK pathway by bile acids is not limited to hepatocytes. In fact, a number of groups have shown the DCA is capable of activating the MAPK pathway in a variety of cells within the gastrointestinal system, where bile acids are found.

Groups who investigate the role of DCA in cholangiocarcinomas (cancer of bile ducts) have also documented activation of the EGFR-MAPK pathway [60–62]. In KMBC cells, a human cholangiocarcinoma cell line, DCA at 200μM was able to induce the phosphorylation of EGFR. This induction of phosphorylation in comparison to taurochenodeoxycholate (TCDC) and glycochenodeoxycholate (GCDC), showed that DCA was the strongest activator of EGFR [60]. Activation of EGFR by DCA was associated with MAPK activation as well as other MAPK related molecules. Moreover, the activation of MAPK and p38 were responsible for a DCA-mediated COX-2 induction. Another group showed that DCA (200μM) could induce EGFR phosphorylation in both KMBC and H-69 human cholangiocarcinoma cell lines [61]. This induction was found to be dependent on the EGFR-activating ligand, transforming growth factor alpha (TGF-α). This data is contradictory to the results observed in hepatocytes, which showed that DCA induces EGFR activation in a ligand-independent manner [58]. In addition, DCA has also been shown to activate EGFR in human esophageal cell lines [63,64], as well as in gastric carcinoma cells [65].

The ability of DCA to induce the activation of the MAPK pathway in colon cancer cells has also been evaluated [41,48,52,55,66,67]. In early studies, Qiao et al. clearly showed that physiologically relevant concentrations of DCA (250μM) found in the colon could activate strong and prolonged (>4 hours) ERK1/2 signaling in the human colon adenocarcinoma cell line, HCT116 [55,66]. Notably, only ERK1/2 and protein kinase C (PKC) activation were responsible for activation of AP-1, composed of JunD, Fra-1 and c-Fos [55,66]. Interestingly, the mechanism of DCA-induced ERK1/2 activation in these cells was not evaluated, however, other groups have suggested that PKC activation by DCA is due to membrane perturbations or interactions [68–70]. The ability of DCA to modify membrane organization and cause activation of membrane-associated molecules is another potential mechanism by which DCA could lead to EGFR activation, which has yet to be examined. In concert with earlier work conducted in hepatocytes [58], Qiao et al. also showed that blocking the MAPK pathway using the specific MEK inhibitor, PD98059, resulted in an increase in DCA-induced apoptosis [66]. These data support the aforementioned hypothesis that the apoptotic effects of DCA are partially countered by the simultaneous activation of the MAPK pathway.

Other groups have also documented the mitogenic effects of DCA on colon cancer cells. In HM3 colon adenocarcinoma cells, DCA (200μM) was able to activate the EGFR/Ras/Raf-1/ERK pathway, as well as others and lead to the upregulation of mucin 2 (MUC2) [41]. In this case, AG1478, a specific EGFR inhibitor was able to block DCA-induced MUC2 upregulation. These data suggest that DCA may be capable of acting directly on EGFR. This is supported by work conducted in T-84 and HCT116 human adenocarcinoma cells DCA was able to directly activate EGFR with concentrations ranging from 50μM to 500μM, respectively [48,67]. In the HCT116 cells, phosphorylation of EGFR by 500μM DCA led to an increase in Raf-1 kinase activity, phosphorylation of EGFR, and increased AP-1 DNA binding activity. It is important to note that the concentrations of DCA used in the studies described in this review vary, and may influence which pathways are activated. Collectively, these studies support the hypothesis that DCA mediates EGFR activation in a ligand-independent manner, whether it is by membrane modification or direct interaction with EGFR remains to be elucidated.

These data taken together suggest four potential mechanisms of DCA-induced EGFR-MAPK induction (Figure 3). The first is a ligand-dependent mechanism, meaning that natural ligands for EGFR, such as EGF or TNF-α must be present in order for DCA to activate this pathway. Other groups suggest that DCA is capable of inducing MAPK signaling independent of ligand presence. This may happen through lipid peroxidation or structural modification of the membrane, which may indirectly activate EGFR. In addition, DCA may be able to directly activate EGFR, leading to receptor dimerization and phosphorylation, however this has not been specifically evaluated. Lastly, MAPK activation may be independent of EGFR in its entirety. DCA may be able to directly activate signaling molecules. In hepatocytes, esophageal cells, and colon adenocarcinoma cells, DCA has been shown to induce the production of reactive oxygen species (ROS) [71–74]. ROS generation in BCS-TC2 colon adenocarcinoma cells was mediated through the activation of plasma membrane-associated enzymes such as NAD(P)H oxidase and phospholipase A2 (PLA2) [74]. These molecules, along with ROS, have been associated with the induction of receptor tyrosin kinases as well as pro-survival signaling. Through the production of ROS, NAD(P)H oxidase, or PLA2 or other mechanisms that have yet to be explicated, DCA may be capable of directly inducing MAPK signaling. Moreover ROS are known to block the activity of receptor tyrosin phosphatases, which may allow for the prolonged activation of receptor tyrosine kinases such as EGFR. Further research is necessary to fully elucidate the mechanisms behind DCA’s ability to regulate EGFR-MAPK signaling, and to determine its contribution in colon cancer development.

In contrast to DCA, UDCA is considered a cytostatic agent, suppressing cell growth while not promoting apoptosis. It is predominantly used in the clinic for the treatment of primary biliary cirrhosis, because of its ability to protect from cell death. In HCT116 colorectal adencarcinoma cells, UDCA was found to protect against DCA-induced cell death [48]. Interestingly, UDCA pretreatment was also able to block DCA-induced EGFR phosphorylation, ERK1/2 activation, and AP-1 DNA binding. Feldman and Martinez showed that UDCA treatment to HT-29 colorectal adencarcinoma cells led to the ubiquitanation and degradation of EGFR, in a process dependent on c-Cbl and caveolin-1 [75]. Moreover, UDCA was also able to reduce EGF-induced ERK1/2 activity, an observation enhanced in cells stably transfected with caveolin-1. In addition, UDCA was also able to decrease the half-life of EGFR by 30% or five fold in HT-29 parent or caveolin-1 transfected HT-29 cells, respectively. UDCA may have direct negative regulatory effects on the EGFR-MAPK pathway, while other membrane proteins may be able to enhance these effects.

Discussion

DCA and UDCA seem to have opposing effects on colon cancer progression. One of the explanations for this difference is the diverging effects on oncogenic signaling pathways, specifically on the EGFR-MAPK pathway. The EGFR-MAPK pathway has been shown to play a role in colon cancer progression, where EGFR expression may even be a prognostic factor. DCA can activate the EGFR-MAPK pathway in hepatocytes and in a number of gastrointestinal cells, including Barrets-associated esophageal cells, gastric cancer cells, cholangiocarcinoma cells and colon cancer cells. In colon cancer cells, EGFR-MAPK activation is associated with increased survival, and when MAPK signaling is blocked, DCA-induced apoptosis increases significantly. Moreover, DCA or EGF activation of MAPK signaling can be blocked by UDCA pre-treatment.

DCA activation of MAPK signaling may lead to the activation of numerous downstream targets, which may play a role in cancer development. In colon cancer, DCA-induced activation of MAPK signaling leads to activation of the protooncogene AP-1 and the suppression of the tumor suppressor gene p53. In addition, MUC2, a mucin aberrantly expressed in colon cancer, may also be upregulated by DCA activation of ERK signaling [41]. In other gastrointestinal cells, DCA-induced MAPK activation is associated with COX-2 and MCl-1, markers of inflammation and metaplasia, respectively. COX-2 has also been shown to play a role in the development of colon cancer, however, the effects of DCA on COX-2 expression in colon cells has not been assessed.

In clinical samples, DCA may lead to the proliferation of normal colonic epithelium, however, the role of the EGFR-MAPK pathway in these tissues have not been evaluated. Additionally, the influence of DCA on immortalized normal colonic cells has also not been evaluated. This important evaluation will help lead to a thorough understanding of DCA and UDCAs respective roles in the promotion or suppression of tumor development. Many of the colon cancer cell lines used to evaluate the effects of DCA on the EGFR-MAPK pathway have RAS mutations, or other mutations within this pathway that may have altered the results. The effects of DCA on cells that do not contain mutations in the EGFR-MAPK pathway, or converging pathways, have yet to be fully evaluated.

DCA has also been shown to regulate additional converging pathways such as P38, PKC and PI3K. The role of these pathways in the ability of DCA or UDCA to induce or suppress MAPK signaling has also not been fully elucidated. Nonetheless, DCA and UDCA both have the ability to interact with membrane proteins that may induce membrane perturbations. The contribution of membrane modifications induced by DCA and UDCA has also not been evaluated with respect to EGFR activation. It is also unclear whether DCA’s effects on EGFR activation is solely ligand independent, or if the bile acid itself may bind to any part of the receptor. The variations in the regulation of EGFR-MAPK signaling by DCA and UDCA may contribute to the diverging effects of these bile acids on colon cancer progression. Furthermore, there may be a link between bile acid regulation of EGFR-MAPK signaling and membrane proteins. However, these questions must be address in order to fully understand and prevent the tumor promoting effects of DCA. In addition, elucidating the mechanisms involved in UDCA’s chemopreventive activity is necessary for the development of the therapy or may even lead to the identification of a more targeted therapy for the prevention of colorectal cancer.