Abstract

It is widely accepted that alterations to cyclooxygenase-2 (COX-2) expression and the abundance of its enzymatic product prostaglandin E 2 (PGE 2 ) have key roles in influencing the development of colorectal cancer. Deregulation of the COX-2/PGE 2 pathway appears to affect colorectal tumorigenesis via a number of distinct mechanisms: promoting tumour maintenance and progression, encouraging metastatic spread, and perhaps even participating in tumour initiation. Here, we review the role of COX-2/PGE 2 signalling in colorectal tumorigenesis and highlight its ability to influence the hallmarks of cancer—attributes defined by Hanahan and Weinberg as being requisite for tumorigenesis. In addition, we consider components of the COX–prostaglandin pathway emerging as important regulators of tumorigenesis; namely, the prostanoid (EP) receptors, 15-hydroxyprostaglandin dehydrogenase and the prostaglandin transporter. Finally, based on recent findings, we propose a model for the cellular adaptation to the hypoxic tumour microenvironment that encompasses the interplay between COX-2, hypoxia-inducible factor 1 and dynamic switches in β-catenin function that fine-tune signalling networks to meet the ever-changing demands of a tumour.

A plethora of data indicates that human cancer is a multistage genetic and epigenetic disease. Like all cancers, colorectal cancers are thought to originate from a single replication-competent cell (stem cell or proliferative progenitor cell) ( 1 ). In colorectal tumours, the initiating event is thought to be the acquisition of a genetic alteration that deregulates the ‘gatekeeping’ pathway—the adenomatous polyposis coli (APC)/β-catenin pathway (also known as the WNT-signalling pathway)—with tumour progression occurring through clonal selection via the protracted acquisition of multiple genetic lesions ( 1 , 2 ). These alterations, while having the potential to produce a wide array of different cancer cell genotypes, have been proposed to affect six principal aspects of cell physiology that are obligate for cancer development. These are termed the ‘hallmarks of cancer’ ( 3 ): acquired capabilities that represent breaches to the normal regulatory mechanisms controlling cell survival, proliferation, migration, invasion and the interactions with neighbouring cells and stroma.

The genetic changes that occur in colorectal cancer appear to more or less follow the histological progression from small, pre-malignant adenomas, to advanced metastatic tumours, with both early and late stages of the disease associated with the number of sequentially obtained genetic alterations (reviewed in ref. 4 ). This implies that each successive genetic ‘hit’ confers an advantageous characteristic upon the expanding tumour mass. While it may be true that numerous different genes can become altered during the development of a given tumour, it has also recently been proposed that all cancers arise and are maintained by the deregulation of a relatively small number of signalling pathways ( 1 , 5 ). This is an attractive premise from a therapeutic standpoint, given that there are far fewer cancer-related pathways than cancer-related genes ( 1 ).

While perturbation of the WNT-signalling pathway is believed to account for the initiation of colorectal tumours ( 6 ), the aberrant expression of cyclooxygenase-2 (COX-2) that occurs in the majority of colorectal tumours is thought to play a crucial role during colorectal cancer development ( 7 ). Deregulation of COX-2 expression leads to an increased abundance of its principal metabolic product, prostaglandin E 2 (PGE 2 ), the pleiotropic effects of which appear to affect most, if not all, of the hallmarks of cancer. This ability of the COX-2/PGE 2 pathway to affect multiple aspects of cell physiology required for tumour development and maintenance may also offer an explanation for the effectiveness of COX inhibitors and non-steroidal anti-inflammatory drugs (NSAIDs) at reducing both the incidence and progression of intestinal tumours in animal models of cancer, and more importantly in human cancer patients (reviewed in ref. 8 ).

In this review, we discuss the impact of the COX-2/PGE 2 pathway on colorectal cancer development and highlight its influence on each of the hallmarks of cancer. Furthermore, we consider the emerging regulators of COX–prostaglandin signalling and novel roles for COX-2/PGE 2 signalling in the tumour microenvironment.

Cyclooxygenases and prostaglandins

NSAIDs such as aspirin have been used traditionally as analgesics and anti-inflammatories for centuries and are some of the world's most widely used drugs. The seminal work by Sir John Vane delineated the mechanism by which aspirin exerts its anti-inflammatory, analgesic and antipyretic actions. In the early 1970s, the enzyme prostaglandin G/H synthase, now more often referred to as COX, was elucidated as the target for inhibition by aspirin and other NSAIDs ( 9 ). Since this discovery, there has been a great deal of interest in identifying the components of the COX pathway as well as its functions in both physiological and pathological conditions. COX enzymes play key roles in the biosynthesis of prostaglandins from arachidonic acid following its release from the plasma membrane by the action of phospholipase-A2 (reviewed in ref. 10 ). In the initial cyclooxygenase reaction, the COX enzymes catalyse the formation of the unstable intermediate prostaglandin G 2 from arachidonic acid, which is then converted into prostaglandin H 2 by the peroxidase activity of COX ( 11 ). Prostaglandin H 2 is the precursor for several structurally related prostaglandins, which are formed by the action of specialized prostaglandin synthases (reviewed in ref. 12 ). The prostaglandins synthesized by this pathway include the aforementioned PGE 2 , as well as prostaglandin D 2 , prostaglandin F (PGF ), prostaglandin I 2 (also known as prostacyclin) and thromboxane-A2. The actions of these prostanoid ligands are mediated by their engagement of specific cell-surface G-protein-coupled receptors designated EP1–4 for PGE 2 and prostaglandin F receptor, thromboxane A2 receptor, prostaglandin I2 receptor and thromboxane A2 receptor for PGF , prostaglandin D 2 , prostaglandin I 2 and thromboxane-A2, respectively. These prostaglandins are important for a large number of normal physiological processes in a broad range of tissues. These include the modulation of immune responses, protection of the gastrointestinal mucosa, maintenance of renal homeostasis and the regulation of blood clotting. Furthermore, prostaglandins also function in pathological conditions where they can promote inflammation, swelling, pain and fever (reviewed in ref. 10 ).

Later work on the biology of COX enzymes led to the distinction between two COX enzymes: COX-1 ( 13 ) and COX-2 ( 14 ). A third form (COX-3) has also been reported ( 15 ), although recent studies indicate that this represents a splice variant of COX-1 that encodes a truncated protein lacking enzymatic activity ( 16 ). In humans, COX-1 is found constitutively expressed in a wide range of tissues including the kidney, lung, stomach, small intestine and colon. Thus, COX-1 is considered a housekeeping enzyme responsible for maintaining basal prostaglandin levels important for tissue homeostasis. In contrast, most tissues do not normally express COX-2 constitutively, notable exceptions including the central nervous system ( 17 ), kidneys ( 18 ) and seminal vesicles ( 19 ). However, the stimulation of COX-2 expression in Src-transformed fibroblasts ( 14 ), endothelial cells and monocytes treated with the tumour promoter tetradecanoyl-phorbol-acetate ( 20 ) or lipopolysaccharide ( 21 ) led to the notion that COX-2 is an inducible enzyme that produces prostaglandins during inflammatory and tumorigenic settings. Because of this, there has been a fervent interest in studying the biology of COX-2 in relation to tumorigenesis, particularly with regard to colorectal tumorigenesis where its actions appear to have a major impact on tumour development.

Cyclooxygenases and colorectal cancer

The first indication that COX enzymes might play a role in colorectal tumorigenesis came from the observation that patients with Gardner's syndrome (which causes multiple intestinal polyposis) treated with NSAIDs displayed a reduction in adenoma number ( 22 ). This was the first clinical evidence that NSAIDs could be useful in the prevention of cancer. Subsequently, in 1991, Thun et al. ( 23 ) undertook an epidemiological study, which reported that regular aspirin intake at low doses reduces the risk of colorectal cancer, and suggested the link between COX enzymes and cancer development. A large body of evidence now indicates that elevated levels of COX-2 are present in the majority of colorectal carcinomas ( 24–28 ) and in a subset of adenomas ( 24 , 25 ). The observation that COX-2 expression is up-regulated in both pre-malignant as well as malignant colorectal tissue is regarded as particularly significant—having potential implications for both the prevention and treatment of colorectal cancer ( 29 ).

Several lines of evidence now support a critical role for COX-2 during colorectal tumorigenesis. Most convincingly, in a randomized double-blind placebo-controlled trial, administration of the COX-2-selective NSAID celecoxib significantly decreased the occurrence of sporadic colorectal adenomas ( 30 ). Notably, in this scenario, celecoxib might potentially act not only by suppressing the growth of existing adenomas but also by preventing the formation of new adenomas ( 30 ). Similarly, in a study of patients with familial adenomatous polyposis, administration of celecoxib was associated with regression of adenomas of both the colon and the rectum ( 31 ). By the same token, studies have revealed that genetic disruption or pharmacological inhibition of COX-2 results in a substantial decrease in adenoma size and number in murine models of intestinal tumorigenesis [( 32 , 33 ), respectively].

The pro-tumorigenic effects of COX-2 in the colorectum are largely thought to be attributed to its role in producing PGE 2 . Indeed, increased levels of PGE 2 have been reported both in human colorectal adenomas as well as carcinomas ( 34 , 35 ). PGE 2 levels have also been reported to increase in a size-dependent manner in the adenomas of familial adenomatous polyposis patients ( 36 ) as well as in the adenomas of Apc Min/+ mice ( 37 ). Furthermore, in vivo studies have revealed that prevention of adenoma development in familial adenomatous polyposis patients is more effective when tissue prostaglandin levels are reduced through NSAID treatment ( 38 ), and more specifically, sequestration of PGE 2 by administration of a PGE 2 monoclonal antibody in mice has been demonstrated sufficient to retard the growth of transplantable tumours in vivo ( 39 ). In addition, PGE 2 has been found to result in augmented incidence and multiplicity of carcinogen-induced colon tumours in rats ( 40 ) and enhances intestinal adenoma growth in Apc Min/+ mice ( 41 ). Accumulating evidence from in vitro studies has further substantiated the in vivo evidence and also provided insights into the mechanisms underlying the oncogenic role of PGE 2 (reviewed in ref. 42 ). Nevertheless, it is worth noting that while good evidence exists suggesting that COX-2/PGE 2 pathway inhibition may be useful in the prevention of colorectal tumours, this may not be the case under all circumstances. Indeed, COX-2/PGE 2 signalling is likely to act in concert with other important signalling pathways that become deregulated in cancer, meaning that such pathways would also need to be targeted in conjunction with the COX-2/PGE 2 pathway for efficient prevention (or treatment) of colorectal cancer.

While there is a large body of data indicating an important tumour-promoting role for COX-2 during tumorigenesis, it is important to note that some reports imply that the role of COX-2 and/or prostaglandins in cancer might not be as straightforward as initially proposed. For example, some NSAIDs produce antitumoral effects in vitro independently of their ability to inhibit COX-2 or to reduce PGE 2 synthesis ( 43–45 ), and furthermore, some NSAIDs actually induce COX-2 expression ( 46 , 47 ). In addition, the COX-2/PGE 2 pathway may even elicit what might be considered counterintuitive effects under certain circumstances, by acting in a tumour-suppressive manner ( 48–51 ). In one study, the authors investigated the effects of exogenous administration of the stable synthetic PGE 2 analogue 16,16-dimethyl-PGE 2 on tumour development in Apc Min/+ mice; unexpectedly, the authors observed reductions in both the number and size of intestinal tumours ( 48 ). Furthermore, transgenic mice engineered to over-express COX-2 via the keratin 14 promoter (resulting in COX-2 over-expression and elevated PGE 2 levels in the skin) were found to be more resistant, rather than more sensitive, to the development of skin tumours induced by an initiation/promotion protocol ( 49 ). These findings contrast with other similar studies reporting tumour-promoting effects of COX-2/PGE 2 signalling in mice ( 41 , 52 , 53 ), adding further layers of complexity to the role of the COX-2/PGE 2 pathway during intestinal tumour development.

It is not clear what accounts for the apparent disparities between these studies, but a number of possibilities can be considered. For instance, it could be speculated that the precise level of PGE 2 cells are exposed to may result in different outcomes in a given experimental system. Indeed, we have previously reported that human colorectal adenoma cells are growth stimulated by low concentrations of PGE 2 , but growth inhibited by high PGE 2 concentrations ( 54 ). Conversely, colorectal carcinoma cells are growth stimulated by both low and high concentrations of PGE 2 ( 54 ). We hypothesize that during adenoma progression, further events—such as over-expression of the EP4 receptor—may be necessary for tumours to become growth stimulated by high concentrations of PGE 2 . Therefore, it is possible that the use of stable PGE 2 (16,16-dimethyl-PGE 2 ) ( 48 ) led to high levels and/or accumulation of PGE 2 , resulting in the inhibition of adenoma growth. Another interesting possibility, as noted earlier, is whether there could perhaps be an important role for other genetic lesions occurring in parallel to deregulation of the COX-2/PGE 2 pathway; that is, additional changes might be required for COX-2 to promote tumorigenesis efficiently. Thus, although much of data suggest an important role for COX-2/PGE 2 signalling in the promotion of tumorigenesis, further research is still required in order to elucidate its precise roles in specific cellular and experimental contexts.

COX-2, prostaglandins and their contribution to the hallmarks of cancer

As mentioned previously, COX-2 is frequently over-expressed in colorectal cancer, and PGE 2 has been identified as the principal prostanoid promoting cell growth and survival in colorectal tumours. PGE 2 is able to exert pleiotropic effects in colorectal tumours, promoting proliferation, survival, angiogenesis, migration and invasion. Here, we discuss the influence of the COX-2/PGE 2 pathway on the hallmarks of cancer ( 3 ).

Evasion of apoptosis

Apoptosis, the process of programmed cell death ( 55 ), is a critical mechanism by which metazoan organisms control cell number, where selective cell suicide enables the efficient removal of superfluous, damaged or infected cells (for reviews, see refs 56 , 57 ). Apoptosis plays an essential role during embryonic development and is required for maintaining tissue homeostasis throughout the lifespan of metazoan organisms. Disturbances to the apoptotic machinery that result in excessive cell survival or cell death underlie a number of pathological conditions, but most pertinent here, the failure to initiate cell death in response to apoptotic stimuli is thought to play a central role in tumorigenesis ( 58 ). Indeed, deregulated cell proliferation, coupled with an acquired resistance to apoptosis, has been proposed to constitute a platform both necessary and sufficient for tumour growth and malignant progression ( 59 , 60 ).

While the acquired ability of tumour cells to evade apoptosis can arise via a variety of mechanisms, the majority of such changes result in an impaired ability of a cell to engage the intrinsic cell death machinery (i.e. the mitochondrial pathway of apoptosis). This pathway is governed by the ratio of pro-apoptotic and pro-survival BCL-2 family members that set a threshold for activation of the apoptotic caspase cascade upstream of the mitochondria ( 61 ). Establishing the connection between the receipt of extracellular signals and the activation of pathways such as the phosphatidylinositol-3-OH kinase (PI3K)/AKT (also known as protein kinase B) and Ras-mitogen-activated protein kinase (MAPK)/extracellular signal-regulated protein kinase (ERK) cascades to the suppression of apoptosis via the intrinsic cell death machinery was a crucial finding of recent years ( 62 , 63 ). Indeed, this is likely to represent the mechanism by which the COX-2/PGE 2 pathway regulates apoptosis, via the autocrine and/or paracrine effects of PGE 2 acting at one or more of the four cell-surface EP receptors.

Although still not entirely clear, several mechanisms for the suppression of apoptosis by the COX-2/PGE 2 pathway have been suggested, and engagement of one or more of the signalling pathways downstream of EP receptor activation may be responsible for this pro-survival response. Moreover, the ability of the COX-2/PGE 2 pathway to control an apoptosis evasion programme in tumour cells may differ between cell types and depend on factors such as the tumour microenvironment (for example, in hypoxic conditions). The initial indication that COX-2 might modulate the intrinsic pathway of apoptosis came when its forced over-expression led to elevated levels of the pro-survival protein BCL-2 and conferred increased resistance to the dietary fibre fermentation product butyrate-induced apoptosis in rat intestinal epithelial cells ( 64 ). Interestingly, the first suggestion that chemopreventive agents might produce their effects via the induction of tumour cell apoptosis arose from studies with butyrate in colorectal tumour cells ( 65 ); thus, the ability of COX-2 expression to suppress butyrate-induced apoptosis suggests that the chemopreventive action of COX-inhibiting NSAIDs may act in part by re-sensitizing cells to apoptosis-inducing luminal factors. Further studies indicated that the mechanism by which COX-2/PGE 2 might suppress apoptosis and increase the expression of BCL-2 was via activation of the Ras-MAPK/ERK pathway ( 66 ). More recently, several studies indicate that the COX-2/PGE 2 pathway might alter apoptotic thresholds by engaging a number of different signalling pathways. Indeed, PGE 2 has been reported to activate pro-survival pathways including the PI3K/AKT pathway ( 67 , 68 ), ERK signalling ( 69 ), cyclic adenosine monophosphate (cAMP)/protein kinase A signalling ( 70 ) and activation of epidermal growth factor receptor (EGFR) signalling ( 71 , 72 ). PGE 2 has more recently been shown to promote cell survival in murine intestinal adenomas by indirectly transactivating the nuclear peroxisome proliferator-activated receptor (PPAR)-delta via a PI3K/AKT-dependent mechanism ( 52 ). However, while this and other evidence ( 73 ) point to an important role for PPARδ in intestinal tumour development, there remains an interesting debate as to the precise role PPARδ plays during intestinal tumorigenesis ( 74 ). It is also interesting to note that in addition to the apoptosis-suppressive effects of COX-2-derived PGE 2 , deregulated expression of the COX-2 protein itself might also alter the susceptibility of cells to undergo apoptosis by reducing the cellular pool of its substrate arachidonic acid, which can stimulate apoptosis by stimulating ceramide production ( 45 ).

While the exact mechanism by which the COX-2/PGE 2 pathway suppresses apoptosis remains to be fully elucidated, it is clear that the influence of the COX-2/PGE 2 pathway has on this hallmark of cancer plays an important role in aiding tumour progression. For instance, under conditions of hypoxia—a situation conducive to the induction of cell death—PGE 2 appears to promote cell survival in colorectal tumour cells by the stimulation of Ras-MAPK signalling ( 75 ). As well as promoting the survival of cancer cells under harsh microenvironmental conditions, PGE 2 might also play a crucial role in encouraging the transition from adenoma to carcinoma. We have previously reported that increased EP4 receptor expression occurs during the colorectal adenoma–carcinoma sequence ( 54 ) and that EP4 receptor over-expression confers an anchorage-independent phenotype to otherwise anchorage-dependent human colorectal adenoma cells ( 54 ). This suggests an important role for a COX-2/PGE 2 /EP4-axis in the suppression of apoptosis; indeed, in a recent complementary study, Hull and colleagues ( 76 ) reported that EP4 receptor over-expression results in the suppression of apoptosis and enhanced tumorigenic behaviour of the human colon cancer cell line HT29 ( 76 ).

Finally, it is important to note that while the ability of the COX-2/PGE 2 pathway to suppress intrinsic cell death mechanisms might facilitate cancer development, it may also confer a decreased sensitivity of colorectal tumour cells to cytotoxic cancer treatments. For example, it has previously been reported that PGE 2 attenuates radiation-induced apoptosis in colorectal cancer cells via activation of EGFR/AKT signalling and a mechanism that prevents the translocation of pro-apoptotic Bax to the mitochondria ( 68 ). Observations such as these merit further investigation into the potential of NSAIDs and/or COX-2-selective inhibitors as adjuvants to current chemotherapy/radiation regimens for patients with colorectal cancer.

Self-sufficiency in growth signals

The behaviour of normal cells in multicellular organisms is controlled by the coordinated regulation of complex signalling pathways, which transduce signals from growth factors and cytokines into decisions that direct cell fate. It is thought that the deregulation of these signalling networks underlies tumour initiation and progression, causing cells to grow uncontrollably and unregulated by environmental cues. Two major pathways frequently mutated in many human cancers, including colorectal, are the Ras-MAPK- and the PI3K/AKT-signalling pathways; activation of which stimulates cell growth, proliferation and survival (reviewed in ref. 77 , 78 ). Deregulation of these signalling pathways can occur by a variety of mechanisms, including alterations to both upstream and downstream components of the pathways. Most commonly, constitutive activation of the Ras-MAPK pathway occurs via mutations to KRAS [in ∼50% of colorectal tumours ( 79 )] or BRAF [in 10–20% of colorectal tumours ( 80 )] genes, whereas deregulation of the PI3K/AKT pathway occurs most commonly by PIK3CA mutation [in 32% of colorectal tumours ( 81 )], phosphatase and tensin homolog mutations [in up to 20% of colorectal tumours ( 82 )] or mutations in the pleckstrin homology domain of AKT [in 6% of colorectal tumours ( 83 )].

As already mentioned, COX-2-derived PGE 2 is able to signal via the PI3K/AKT- and Ras-MAPK/ERK-signalling pathways to enhance cell survival. An interesting possibility arising from such studies is the prospect that the COX-2/PGE 2 pathway represents an alternative mechanism by which tumours acquire growth factor autonomy—in the absence of PI3K pathway- and MAPK pathway-activating mutations. In other words, aberrant activation of the COX-2/PGE 2 pathway might phenocopy-activating mutations in the PI3K/AKT and/or Ras-MAPK pathways, which could play an important role in promoting tumour progression. This supports the notion that while the cohort of alterations to cancer-related genes and proteins might differ between individual tumours, alterations to seemingly disparate genes may in reality be functionally equivalent; bestowing one or more of the hallmarks of cancer on cells by deregulating the same molecular pathway ( 1 ). Indeed, deregulation of the COX-2/PGE 2 pathway has been shown to behave in a similar manner to constitutively active Ras in murine intestinal adenomas, resulting in a positive feedback loop that boosts COX-2 expression and further stimulation of tumour growth ( 41 ). This mechanism might also operate during hypoxic (as well as normoxic) conditions. In fact, we have previously reported that during hypoxia, the transcription factor hypoxia-inducible factor-1 (HIF-1) directly up-regulates COX-2 expression and increases PGE 2 production ( 75 ). This leads to activation of the Ras-MAPK pathway, which may act in a positive feedback loop to maintain an active pro-survival COX-2/PGE 2 pathway during hostile microenvironmental conditions ( 75 ).

Insensitivity to anti-growth signals

As noted earlier, disturbances to normal tissue homeostasis that shift the balance from a state of equilibrium towards increased cell growth will invariably lead to the appearance of a neoplastic population of cells ( 84 ). Under normal physiological conditions, in addition to the strict control of apoptosis and growth-stimulatory signals, a further crucial mechanism for maintaining tissue homeostasis is via the restraint of cell growth through the action of anti-proliferative signals ( 3 ). These include soluble growth inhibitors, as well as membrane-bound ligands, both of which act on cell-surface receptors that couple to intracellular pathways to repress cell growth. Hanahan and Weinberg ( 3 ) proposed two discrete mechanisms by which anti-growth signals can block proliferation to restrain inappropriate cell growth and maintain tissue homeostasis. Importantly, both of these mechanisms are frequently hijacked in colorectal cancers. Furthermore, deregulation of the COX-2/PGE 2 pathway may provide an additional and/or secondary mechanism by which tumour cells can attain this hallmark.

The first mechanism by which anti-growth signals can block proliferation is by preventing cells from advancing through the G 1 phase of the cell cycle, which serves to maintain cells in a quiescent (G 0 ) state. The archetypal example of such an anti-proliferative signal is the soluble signalling factor transforming growth factor-beta (TGFβ), which blocks progression through the G 1 phase of the cell cycle via the suppression of c-Myc and activation of cyclin-dependent kinase inhibitors such as p15Ink4B and p21Cip1 (reviewed in ref. 85 ). Often, cancer cells acquire the inability to respond to the growth-suppressive effect of TGFβ, which can occur as a result of inactivating mutations in the downstream signalling effectors of TGFβ signalling (e.g. loss of SMAD function) or mutations to the receptors themselves (reviewed in ref. 85 ). The latter mechanism is particularly relevant in colorectal tumours with microsatellite instability, where TGFβ type II receptor mutations occur at a high frequency ( 86 ). However, not all cancer cells exhibit mutational inactivation of the TGFβ type II receptor, and a further mechanism by which cancer cells lose TGFβ responsiveness is via the down-regulation of the TGFβ receptors. Indeed, this could represent a mechanism by which the COX-2/PGE 2 pathway prevents the receipt of anti-growth signals since over-expression of COX-2 has been reported to cause down-regulation of the TGFβ type II receptor ( 64 ).

A second mechanism by which cells can be instructed to curb their proliferative potential involves the implementation of differentiation programmes that instigate an irreversible post-mitotic (or terminally differentiated) state ( 3 ). An example of such an anti-growth pathway operates in the colorectal epithelium and is often deregulated in neoplasia. Under normal physiological conditions in the colonic crypt, the long-lived stem cells continually give rise to rapidly proliferating progenitor cells (also known as transit-amplifying cells), which comprise the proliferative region in the lower two-thirds of the crypt. These progenitor cells are able to differentiate into the cell lineages that occupy the upper third of the crypt: absorptive enterocytes, enteroendocrine cells and mucin-secreting goblet cells ( 87 ). Upon reaching the top third of the crypt, committed progenitor cells undergo cell cycle arrest, become differentiated and are eventually shed in to the lumen. However, in colorectal tumours, aberrant activation of β-catenin (i.e. the WNT pathway) serves to block this normal differentiation programme and maintains cells in a progenitor-like state ( 88 ). Recent evidence indicates that the COX-2/PGE 2 pathway can activate the APC/β-catenin pathway ( 89 ). Thus, in the context of colorectal cancer, where mutations in the APC/β-catenin pathway are common, the COX-2/PGE 2 pathway might enhance the activation of the APC/β-catenin pathway to keep cells in a progenitor-like state. Furthermore, another attractive hypothesis would be that inappropriate activation of the COX-2/PGE 2 pathway discourages differentiation in the absence of APC/β-catenin mutations; in this scenario, the activation of the APC/β-catenin pathway by PGE 2 might even be able to contribute to tumour initiation. In addition, the ability of PGE 2 to switch on β-catenin signalling might also contribute to the acquisition of an immortalized phenotype, as discussed below.

Limitless replicative potential

The intrinsic capacities for self-renewal and limitless replicative potential are characteristics thought to be shared by stem cells and cancer cells ( 90 ). Because of these apparent similarities, it has been proposed that cancer arises from the deregulation of pathways that maintain the stem/progenitor cell phenotype in a given tissue ( 6 , 91 ). As mentioned earlier, colorectal tumours are thought to originate from such replication-competent cells (i.e. a proliferative progenitor cell or stem cell) initiated by activating mutations in the APC/β-catenin pathway. Perturbations to the APC/β-catenin pathway are detectable in the earliest lesions of the colorectal epithelium (aberrant crypt foci) and are thought to occur in virtually all colorectal tumours. This most commonly arises by loss of the APC tumour suppressor gene, occurring in ∼80% of all colorectal tumours ( 92 ). The tumour-suppressive action of APC is thought to be mediated in large part by its ability to regulate the levels of intracellular β-catenin, which transduces signalling by the WNT ligand. The mutations in either APC or indeed other components of the WNT pathway observed in colorectal tumours lead to the formation of a constitutively active β-catenin–T-cell factor (TCF) complex that mimics active WNT signalling. Evidence suggests that the proliferative compartments (stem cells and progenitor cells) in intestinal crypts are maintained through activation of the WNT pathway ( 88 ). Indeed, disruption of the WNT pathway by deletion of Tcf4 in mice results in loss of the stem cell and proliferative compartments in the small intestine ( 93 ). This suggests that the WNT pathway maintains the crypt stem/progenitor cell phenotype and that the same pathway is active in both colorectal cancer cells and the stem/progenitor cells of the intestinal crypt.

The recently discovered connection between the WNT pathway and the COX-2/PGE 2 pathway ( 89 ) indicates that PGE 2 signalling may be able to contribute to the crypt progenitor phenotype by activating β-catenin/TCF signalling in colorectal cancer cells. Gutkind and colleagues ( 89 ) reported that upon engagement of the EP2 receptor, PGE 2 stimulates a dual-signalling cascade involving activation of the PI3K/AKT pathway by the EP2-associated G-protein βγ subunits and the association of the G-protein α s subunit with Axin ( 89 ). This leads to the activation of β-catenin by a dual mechanism: first, the activation of AKT inhibits glycogen synthase kinase-3 beta (GSK3β), which reduces the inhibitory effect of GSK3β-mediated phosphorylation on β-catenin; and second, the complex between the EP2-associated G-protein α s subunit and Axin promotes the release of β-catenin from the Axin–GSK3β complex, freeing β-catenin and facilitating its nuclear accumulation ( 89 ). Thus, by activating the β-catenin/TCF pathway, it is tempting to speculate that COX-2/PGE 2 signalling might contribute to limitless replicative potential by promoting the acquisition of a progenitor or stem cell-like phenotype; although, at present, whether COX-2/PGE 2 signalling is able to promote such a phenotype is unknown.

In addition, it is also possible that the COX-2/PGE 2 pathway, by enhancing cell survival and growth, serves to prime cells for the acquisition of further cellular alterations that contribute to immortalization and the progression towards the full malignant phenotype.

Lastly, it is of interest to note that recent studies suggest a role for COX-2/PGE 2 signalling in the maintenance of haematopoietic stem cell homeostasis ( 94 ) and in the suppression of embryonic stem cell apoptosis ( 95 ). Whether the COX-2/PGE 2 pathway influences the behaviour of stem cells of the intestinal and colonic epithelium awaits investigation.

Sustained angiogenesis

In the course of solid tumour development, it is well recognized that the avascular tumour mass becomes dependent on angiogenesis for maintenance and progression, leading to the concept known as the ‘angiogenic switch’ ( 96 ). The initiation of this morphogenic process is controlled by the relative balance of pro-angiogenic and anti-angiogenic factors in a given tissue, such that a tipping of the balance towards pro-angiogenic factors results in the stimulation of blood vessel formation ( 96 ). In the case of colorectal cancer, over-expression of COX-2 in colon cancer cells induces the production of angiogenic factors such as vascular endothelial growth factor (VEGF) and basic fibroblast growth factor, which are instrumental in stimulating the formation of new blood vessels—a requirement for tumours should they wish to develop beyond a few millimetres in size. These pro-angiogenic effects of COX-2 can be inhibited by NSAIDs, resulting in the inhibition of endothelial cell proliferation, migration and vascular tube formation in vitro ( 97 ). The inhibitory effect of NSAIDs on angiogenesis in this context can be rescued by adding exogenous PGE 2 ( 98 ), suggesting that COX-2-derived PGE 2 is at least part responsible for the pro-angiogenic effects of COX-2 over-expression; although it is important to note that prostaglandin-independent mechanisms of NSAID action on angiogenesis were also reported ( 98 ). Consistent with this, PGE 2 has been reported to stimulate VEGF expression in colon cancer cells through the activation of HIF-1, a key regulator of VEGF expression ( 99 ). Furthermore, in vivo studies have also demonstrated an important role for the COX-2/PGE 2 pathway in angiogenesis. In mice, homozygous deletion of COX-2 (but not COX-1) results in impeded growth of tumour xenografts and is associated with a reduced tumour vascular density ( 100 ). Such observations are likely to be at least in part mediated by the resultant reduction in PGE 2 , since the intestinal adenomas present in APC Δ716/+ mice that are null for the PGE 2 receptor EP2 are defective in VEGF induction ( 101 ). One mechanism by which COX-2 might promote tumour vascularization is via the production of PGE 2 and prostaglandin I 2 , which have been shown to participate in inducing endothelial cell spreading and migration by integrin αVβ3-mediated activation of the small guanosine 5′-triphosphatases Cdc42 and Rac ( 102 ). More recently, PGE 2 has been shown to regulate angiogenesis though the modulation of chemokine receptor signalling: PGE 2 can enhance VEGF and basic fibroblast growth factor-induced chemokine receptor-4 that is important for microvessel assembly in vivo ( 103 ). In addition, PGE 2 can induce the expression of the pro-angiogenic chemokine CXCL-1 in vivo ( 104 ). PGE 2 may also work concomitantly with the hypoxic tumour microenvironment to orchestrate the process of angiogenesis, as discussed later ( 75 ).

Tissue invasion and metastasis

It is well established that the primary cause of cancer mortality is the formation of distant metastases, making the capacity of tumour cells to invade and metastasize one of the most pertinent hallmarks of cancer from a therapeutic perspective. In order to achieve metastases, cancer cells must exhibit a more motile, invasive phenotype, dissociate from neighbouring cells within the tumour, invade through extracellular matrix components and intravasate into local blood or lymphatics (reviewed in ref. 105 ). Having made their escape from the primary tumour, cancer cells must then extravasate from the blood or lymphatics into the surrounding tissue in order to colonize distant sites. Several lines of evidence indicate that COX-2 and the prostaglandins play important roles in aiding these processes—more specifically, PGE 2 is thought to promote a more metastatic phenotype in colorectal tumour cells. While inhibition of COX-2 in vivo can attenuate the metastatic potential of colorectal tumours in both humans ( 106 ) and mice ( 107 ), over-expression of COX-2 in intestinal cells can modulate their adhesive properties ( 64 ) and increase matrix metalloproteinase activity to promote invasion ( 108 ). Further studies have demonstrated that PGE 2 promotes cytoskeletal reorganization and increases colorectal cancer cell migration and invasion via PI3K signalling ( 67 ). The stimulation of invasion and motility by PGE 2 is dependent on the intracellular Src-mediated transactivation of the EGFR ( 72 ). Furthermore, the hepatocyte growth factor receptor, c-Met, is also transactivated by PGE 2 in an EGFR-dependent fashion in colorectal cancer cells ( 109 ). Hepatocyte growth factor/c-Met signalling is classically associated with loss of cell–cell contact (or scattering) and invasive growth ( 110 ). Transactivation of c-Met by PGE 2 leads to increased nuclear β-catenin accumulation, increased urokinase-type plasminogen activator receptor expression and invasion through matrigel ( 109 ). Furthermore, COX-2, c-Met and β-catenin are co-expressed at the invasive front of colorectal tumour specimens ( 109 ). Although PGE 2 is the most abundant prostaglandin found in colorectal cancer tissue ( 35 ), PGF has also been detected in colorectal tumours, and PGF can stimulate motility in both colorectal carcinoma and adenoma cells and invasion of colorectal carcinoma cells ( 111 ).

More recently, the significance of COX-2—not only as a critical player in tumour development but also as being necessary for dissemination of cancer cells to other organs—was demonstrated during an elegant series of experiments in an in vivo model of breast cancer cell metastasis to the lungs ( 112 ). In this study, using both genetic and pharmacological approaches, the authors identified COX-2 as one of four key ‘metastasis progression’ genes, which collectively synergize to mediate both tumour development and metastasis to other organs. Notably, the tumour maintenance and progression functions of COX-2 were mediated solely by the COX-2 expressed by the tumour cells themselves—as opposed to COX-2 expression and PGE 2 production by other cells (i.e. stromal cells) in the tumour microenvironment ( 100 , 113 ). Therefore, these findings contribute to the ongoing discussion concerning the roles of tumour-derived and stroma-derived COX-2 during the different stages of tumorigenesis and within specific cellular contexts ( 114 ).

Evasion of the antitumour immune response and its regulation by the COX-2/PGE 2 pathway

In addition to the six hallmarks of cancer originally proposed by Hanahan and Weinberg, additional characteristics of tumours have recently been proposed, one of which is the ability of tumours to evade attack by the immune system (reviewed in ref. 115 ). Tumour-specific antigens are capable of eliciting an antitumour immune response involving cytotoxic CD8+ T cells, resulting in tumour cell lysis. However, in cancer patients, these antitumour immune responses are generally ineffective and increasing evidence suggests that tumours evolve several strategies to escape tumour-specific immunity ( 116 , 117 ). In light of the known ability of PGE 2 to modulate multiple aspects of the immune response ( 118 ), this additional cancer hallmark may therefore be influenced by inappropriate expression of COX-2 and PGE 2 production that occurs commonly in colorectal tumours. Furthermore, COX-2 over-expression and the resulting increase in PGE 2 levels may represent a strategy adopted by tumours that contributes to the evasion of tumour-specific immune response ( 119 ). For example, PGE 2 has been reported to shift the production of cytokines by antigen-presenting dendritic cells, away from a Th1 (type 1 T cell) profile, leading to a reduced activation of antitumour cytotoxic CD8+ T cells ( 118–120 ). This ability of PGE 2 to suppress these immune responses may allow tumour cells to escape immunosurveillance, adding to the already countless roles of the COX-2/PGE 2 pathway during tumour development.

Emerging regulators of the COX-2/PGE 2 pathway: PGE 2 synthesis and signalling as alternative chemopreventive/therapeutic targets

Until relatively recently, COX expression was believed to be the major determinant of PGE 2 levels, but recent data indicate that multiple levels of control exist for the regulation of PGE 2 production and turnover. In addition to regulation via COX enzymes, PGE 2 synthesis can be regulated via expression of specific prostaglandin E synthases (PGES), which act on the COX product prostaglandin H 2 to produce PGE 2 (reviewed in ref. 42 , 121 ). Moreover, in addition to regulated synthesis, PGE 2 levels are also determined by the rate of its degradation, which can be regulated via the expression of 15-hydroxyprostaglandin dehydrogenase (15-PGDH) and the prostaglandin transporter (PGT) ( 122 ). The cellular response to PGE 2 can also be regulated via the differential expression of the four PGE 2 receptors (EP1–4), which activate distinct downstream signalling responses (reviewed in ref. 42 , 123 ). An understanding of these further levels of regulation controlling levels of (and response to) PGE 2 may reveal novel strategies for prevention and/or treatment of colorectal cancer.

There are currently three characterized PGES enzymes: cytosolic PGES, microsomal PGES-1 and microsomal PGES-2 (reviewed in ref. 42 ). Microsomal PGES-1 is, like COX-2, induced by pro-inflammatory stimuli and up-regulated in colorectal tumours ( 124 ). Moreover, loss of microsomal PGES-1 expression is reported to suppress intestinal neoplasia in APC-mutant mice ( 125 ), although the reverse has also been observed ( 126 ), possibly reflecting the knock-on enhancement of synthesis of other prostaglandins with pro-tumorigenic effects (such as PGF ). Further work will be required to ascertain the potential utility of PGES inhibition as an anticancer strategy in the colon.

As noted earlier, the four characterized receptors for PGE 2 (EP1–4) each show distinct downstream signalling effects. For example, whereas activation of EP2 and EP4 receptors can both stimulate cAMP production, activation of EP3 results in the inhibition of cAMP production (reviewed in ref. 42 ). Recent evidence suggests a role for the EP receptors in colorectal neoplasia. As previously discussed, we have recently showed that up-regulation of EP4 occurs during human colorectal tumorigenesis in vivo ( 54 ) and others have shown that genetic or pharmacological inactivation of EP4 inhibits tumour growth in a mouse model of intestinal neoplasia ( 127 ). Other PGE 2 receptors are also probably play a role in colorectal neoplasia; for example, homozygous deletion of the gene encoding the EP2 receptor reduces adenoma size and number in Apc Δ716/+ mice ( 101 ). In addition, EP3 has been shown to stimulate angiogenesis and growth of tumours arising from implanted sarcoma cells in mice ( 128 ). However, down-regulation of EP3 has been observed in colorectal neoplasia and azoxymethane-induced tumorigenesis is enhanced in EP3-null mice, suggesting a tumour-suppressive role for EP3 in the intestine ( 129 ). It will be of interest to determine the contribution of different PGE 2 receptors to colorectal neoplasia and determine the potential utility of specific EP antagonists in anticancer therapy in the intestine ( 130 , 131 ).

Recent evidence suggests that deregulated catabolism of PGE 2 also plays a significant role in colorectal neoplasia. Prostaglandins are essentially local-acting hormones with diverse effects in numerous different tissues. Therefore, it is crucial that their catabolism is strictly regulated in order to prevent them from reaching distant organs in the bloodstream (reviewed in ref. 132 ). 15-PGDH is a cytosolic enzyme that inactivates prostaglandins by oxidation to their 15-keto form (reviewed in ref. 133 ) and was recently shown to act as a colorectal tumour suppressor ( 134 , 135 ). In the normal human colon, 15-PGDH is expressed in the non-proliferative epithelium at the top of the crypts and loss of 15-PGDH expression and activity is seen in human colorectal tumours ( 134 , 135 ). Several pathways deregulated in colorectal neoplasia may contribute to this down-regulation as 15-PGDH is negatively regulated by growth-stimulatory epidermal growth factor signalling (up-regulated in colorectal neoplasia) and positively regulated by TGFβ (which colorectal cancer cells become resistant to) ( 134 , 136 ). Functional evidence for the tumour suppressor activity of 15-PGDH in the intestine has been obtained using mouse models. Loss of the gene encoding 15-PGDH in mice results in a doubling of colonic mucosal PGE 2 levels and promotes colorectal tumour growth in Apc Min/+ mice ( 137 ). Furthermore, restoring 15-PGDH expression in colon cancer cells inhibits their ability to form tumours in immune-deficient mice ( 134 ).

Prostaglandin synthesis and degradation are believed to be compartmentalized within the cell. This implies that following synthesis and release, prostaglandins must be re-imported into the cell for degradation by the cytoplasmic-localized 15-PGDH. Prostaglandins are believed to traverse membranes poorly and their cellular uptake is thought to occur mostly via PGT (also known as solute carrier organic ion transporter family, member 2A1 or SLCO2A1) (reviewed in ref. 132 ). Indeed, in HeLa cells, co-expression of both 15-PGDH and PGT is required for metabolism of exogenously added PGE 2 ( 122 ). Similarly to 15-PGDH, PGT was recently shown to be down-regulated in colorectal neoplasia in both humans and Apc Min/+ mice ( 138 ), suggesting that PGE 2 signalling can be further controlled via regulated uptake.

It would be therapeutically desirable to increase prostaglandin degradation in colorectal tumours but at present relatively little is known about the potential for therapeutic re-expression of 15-PGDH and PGT. However, PGT expression can be induced by the histone deacetylase inhibitor trichostatin A and the histone demethylating agent 5-azacytidine; two classes of drug relevant to cancer therapy ( 138 ). Moreover, histone deacetylase inhibitors and 5-azacytidine can also up-regulate 15-PGDH ( 139 ). Future advancements in our understanding of prostaglandin catabolism in the intestine should allow insight into the potential for targeting 15-PGDH and PGT in colorectal cancer prevention and therapy.

Regulation of COX-2 expression by the tumour microenvironment

COX-2 up-regulation has been reported in a number of different tumour types (reviewed in ref. 140 ), and accordingly there has been a great deal of interest in attempting to achieve a greater understanding of the regulatory networks that control COX-2 expression. Whereas activating mutations in the PTGS2 gene (the gene encoding COX-2 in humans) have not yet been described, there are several known mechanisms underlying enhanced COX-2 expression in tumour cells. Principally, these mechanisms include deregulated growth factor signalling and oncogene activation. Examples of these mechanisms include activation of the WNT pathway ( 141 , 142 ) and the Ras-MAPK pathway ( 41 , 143 ) signalling via growth factor receptors including EGFR ( 144 ), TGFβ receptors ( 145 ), c-Met ( 146 ) and gastrin receptors ( 147 ). More recently, our laboratory has shown that COX-2 can be induced by the hypoxic microenvironment in colorectal tumour cells derived from both adenomas and carcinomas ( 75 ). This up-regulation of COX-2 is transcriptional and is mediated by the master regulator of transcription during hypoxia, HIF-1. COX-2 has also been reported to be up-regulated in lung cancer cells via a mechanism that is dependent on HIF-1 ( 148 ). COX-2 up-regulation during hypoxia has also been described previously in endothelial cells ( 149 ), monocytes ( 150 ) and corneal epithelial cells ( 151 ); whereas in endothelial cells, COX-2 up-regulation appears to be mediated by nuclear factor-kappa B ( 149 ), in corneal epithelial cells COX-2 over-expression was linked to PPARα/β activation ( 151 ).

The role of COX-2 and PGE 2 during hypoxia

As noted above, the HIF-1-dependent COX-2 up-regulation during hypoxia is associated with an increase in PGE 2 levels ( 75 ). Given the pivotal role of PGE 2 in colorectal tumorigenesis, these findings build on the already known roles of PGE 2 by demonstrating a role for PGE 2 in the promotion of cell survival under hypoxic conditions—a property that tumour cells must acquire in order to propagate and progress in vivo ( 152 ). Although the mechanism for this pro-survival attribute of PGE 2 is not yet defined, it is likely to involve the alteration of apoptotic thresholds by modulating the intrinsic cell death machinery; for example, by increasing BCL-2 expression, which has been previously demonstrated to promote cell survival under hypoxic conditions ( 153 ). The ability of hypoxia ( 154 ) and PGE 2 ( 155 ) to activate nuclear factor-kappa B may also be important in enhancing cell survival under hypoxic conditions. PGE 2 may also confer alternative survival strategies to cells via the up-regulation of pro-survival factors such as inhibitor of apoptosis-2 ( 156 ), which has been shown to be up-regulated by hypoxia ( 157 ).

In addition to promoting cell survival under hypoxic conditions, the amplification of the COX-2/PGE 2 pathway in hypoxic colorectal cancer cells may have important implications for stimulating angiogenesis, a process that is considered critical for solid tumour growth and progression ( 96 ). While bearing in mind the important separate roles of both hypoxia and COX-2/PGE 2 signalling in the promotion of tumour angiogenesis, we suggest that during hypoxia PGE 2 signalling may be integrated to act concordantly in the regulation of angiogenesis. Indeed, increased levels of PGE 2 appear to enhance HIF-1 transcriptional activity and VEGF production in hypoxic colorectal cancer cells ( 75 ). Another possible level for the integration between PGE 2 and hypoxia signalling in the modulation of angiogenesis could occur at the level of Src; a key factor in hypoxia-induced VEGF and PGE 2 -mediated transactivation of EGFR that has implications for colorectal tumour progression ( 72 ). Furthermore, as mentioned previously, PGE 2 has been shown to enhance the expression of chemokine receptor-4 ( 103 ), whose expression has been shown to be regulated by HIF-1/hypoxia ( 158 , 159 ). The ability of PGE 2 to enhance HIF-1 transcriptional activity stands to be a potentially important observation. Accordingly, PGE 2 may promote the expression of several HIF-1 targets (in addition to VEGF) that are crucially involved in tumorigenesis. If it turns out to be the case, PGE 2 may then be regarded as a key positive modulator of the transcriptional response to hypoxia, which is believed to be critical for the progression of solid tumours including colorectal cancer ( 99 ). The HIF-1-mediated transcriptional response provides both short- and long-term adaptive strategies, including cell survival, metabolic shift, angiogenesis, invasion and metastasis. Therefore, further investigations into cross talk between HIF-1 and PGE 2 signalling may provide greater insight into the mechanisms that promote adaptation to hypoxia and hence tumour progression.

Interplay between the COX-2/PGE 2 pathway, HIF-1 and β-catenin signalling

One of the mechanisms by which COX-2/PGE 2 signalling influences the hallmarks of cancer is via the ability of PGE 2 to simulate the activation of β-catenin, which leads to the stimulation of cell growth and proliferation via its interaction with TCF-4 in the nucleus ( 89 ). However, while in one study we found that COX-2/PGE 2 levels are increased in response to hypoxia ( 75 ), we also found that hypoxia results in a transient cell cycle arrest that involves the inhibition of β-catenin/TCF-4 activity ( 160 ). In the latter study, we found that the cell cycle arrest induced by hypoxia in colorectal cancer cells results from the direct binding of HIF-1α to β-catenin; this results in the suppression of β-catenin/TCF-4 activity and a subsequent shutting down of cell proliferation via the c-Myc-p21-axis ( 160 ). Furthermore, we also revealed a novel role for β-catenin in the enhancement of HIF-1 transcriptional activity, resulting in cell survival and adaptation to the hypoxic microenvironment. Hypoxia, therefore, appears to shift the balance from β-catenin signalling and co-activation of TCF-4 to β-catenin interacting with and enhancing of HIF-1 activity—thereby favouring cell quiescence and survival over proliferation. This scenario illustrates an intriguing dynamicity in β-catenin function in response to conditions of low oxygen tension. The ability of β-catenin to interact with and enhance HIF-1 transcriptional activity represents a novel function for β-catenin in colorectal tumorigenesis that is applicable to hypoxic conditions and may have implications for other aspects of carcinogenesis including metabolic adaptation, invasion and metastasis.

Together, our findings ( 75 ), with those of Castellone et al. ( 89 ), suggest on the one hand, a selective role for PGE 2 in the stimulation of cell proliferation that is restricted to normoxic conditions. On the other hand, during hypoxia, the proliferative effect of PGE 2 may be overridden as a result of HIF-1 competing with TCF-4 for β-catenin, causing the inhibition of β-catenin/TCF-4 activity and cell cycle arrest ( 160 ). Therefore, we propose a model in which hypoxia favours the pro-survival (anti-apoptotic) and pro-angiogenic functions of PGE 2 , whereas hampering the pro-proliferative effect of PGE 2 . The net effect of this selective strategy is enhanced tumour cell survival due to a cell cycle arrest that enables the mitigation of hypoxia-induced cellular stresses, which would otherwise compromise cell survival if proliferation was allowed to continue during hypoxic conditions. This scenario may represent a spatial–temporal switching in PGE 2 signalling and function during colorectal tumorigenesis. This model is summarized in Figure 1 .

Fig. 1.

Under normoxic conditions, COX-2 expression can be induced through the activation of the oncogenic pathways such as the Ras-MAPK pathway. Furthermore, COX-2-derived PGE 2 promotes cell proliferation at least in part via the stimulation of β-catenin/TCF-4 activity, as well as via enhancement of Ras-MAPK signalling. During hypoxia, β-catenin is displaced from TCF-4 (which results in TCF-4 repression) via its interaction with HIF-1, and β-catenin enhances HIF-1 transcriptional activity. In addition, COX-2 is up-regulated during hypoxia via HIF-1 activation, resulting in increased levels of PGE 2 . Both enhanced HIF-1 transcriptional activity and increased PGE 2 levels promote cell survival leading to adaptation to hypoxia. Whereas β-catenin acts as essential activator of TCF-4 in normoxia, β-catenin results in the enhancement of HIF-1 transcriptional activity during hypoxia. These effects are likely to be transient and reversible (dynamic) depending on the tumour microenvironment and fluctuations in oxygen levels.

Fig. 1.

Under normoxic conditions, COX-2 expression can be induced through the activation of the oncogenic pathways such as the Ras-MAPK pathway. Furthermore, COX-2-derived PGE 2 promotes cell proliferation at least in part via the stimulation of β-catenin/TCF-4 activity, as well as via enhancement of Ras-MAPK signalling. During hypoxia, β-catenin is displaced from TCF-4 (which results in TCF-4 repression) via its interaction with HIF-1, and β-catenin enhances HIF-1 transcriptional activity. In addition, COX-2 is up-regulated during hypoxia via HIF-1 activation, resulting in increased levels of PGE 2 . Both enhanced HIF-1 transcriptional activity and increased PGE 2 levels promote cell survival leading to adaptation to hypoxia. Whereas β-catenin acts as essential activator of TCF-4 in normoxia, β-catenin results in the enhancement of HIF-1 transcriptional activity during hypoxia. These effects are likely to be transient and reversible (dynamic) depending on the tumour microenvironment and fluctuations in oxygen levels.

The role of PGE 2 in the activation of β-catenin/TCF signalling also suggests that PGE 2 might serve an important function at early stages of colorectal tumorigenesis. This highlights the possibility that basal PGE 2 levels may switch on the β-catenin-signalling cascade, which in turn would promote cell proliferation that may contribute to tumour initiation. Given that COX-2 is not induced at the very early stages of colorectal tumorigenesis, COX-1 might be the source of PGE 2 during tumour initiation, perhaps enhancing β-catenin/TCF-4 signalling upon the loss of APC. This possibility is consistent with the findings that loss of either COX-1 or COX-2 decreases tumour burden in mouse models of intestinal tumorigenesis ( 53 ). This may also explain why inhibition of both COX-1 and COX-2 enzymes (for example with aspirin or sulindac) is as effective—if not potentially more so—as selective COX-2 inhibitors as chemopreventive agents for colorectal cancer.

Concluding remarks

Compelling evidence gained from mechanistic studies with cancer cell lines, mouse models of intestinal tumorigenesis and a number of clinical trials with both non-selective and COX-2-selective NSAIDs support an important role for the COX–prostaglandin pathway in the development of colorectal tumours. Furthermore, these studies have validated the COX-2/PGE 2 pathway as a bona fide target for cancer chemoprevention and therapy. Evidence suggests that the effectiveness of suppressing the COX-2/PGE 2 pathway in cancer prevention and therapy stems from its role in driving the hallmarks of cancer, without which tumours cannot sustain their growth and development.

Recent clinical trials involving highly selective COX-2 inhibitors as chemopreventive agents for colorectal cancer have cast doubt on the suitability of such drugs for long-term use due to increased risks of adverse cardiovascular events ( 30 , 161 ). Notwithstanding, the COX-2 selective inhibitor celecoxib remains a clinically relevant agent for colorectal cancer prevention and therapy, and further studies will be required to establish the usefulness of NSAIDs in chemoprevention—particularly in patients at a high risk from colorectal cancer. Given the importance of COX-2/PGE 2 in influencing the hallmarks of cancer, it will also be important to determine the applicability of NSAIDs as adjuvants to current chemotherapy/radiation regimens, not only for colorectal cancer patients but also for patients with other tumour types displaying deregulated COX-2 expression.

Future studies into the emerging players within the COX-2/PGE 2 pathway may reveal novel approaches for more safely targeting this pathway for both cancer chemoprevention and therapy.

Funding

Cancer Research UK programme; The Citrina Foundation; The John James Bristol Foundation.

Abbreviations

    Abbreviations
  • APC

    adenomatous polyposis coli

  • COX

    cyclooxygenase

  • EGFR

    epidermal growth factor receptor

  • ERK

    extracellular signal-regulated protein kinase

  • GSK3β

    glycogen synthase kinase-3 beta

  • HIF-1

    hypoxia-inducible factor-1

  • MAPK

    mitogen-activated protein kinase

  • NSAID

    non-steroidal anti-inflammatory drug

  • PGE 2

    prostaglandin E 2

  • PGES

    prostaglandin E synthase

  • PGF

    prostaglandin F

  • PGT

    prostaglandin transporter

  • PI3K

    phosphatidylinositol-3-OH kinase

  • PPAR

    peroxisome proliferator-activated receptor

  • 15-PGDH

    15-hydroxyprostaglandin dehydrogenase

  • TCF

    T-cell factor

  • TGFβ

    transforming growth factor-beta

  • VEGF

    vascular endothelial growth factor

Conflict of Interest Statement: None declared.

References

1.
Vogelstein
B
, et al.  . 
Cancer genes and the pathways they control
Nat. Med.
 , 
2004
, vol. 
10
 (pg. 
789
-
799
)
2.
Fearon
ER
, et al.  . 
A genetic model for colorectal tumorigenesis
Cell
 , 
1990
, vol. 
61
 (pg. 
759
-
767
)
3.
Hanahan
D
Weinberg
RA
The hallmarks of cancer
Cell
 , 
2000
, vol. 
100
 (pg. 
57
-
70
)
4.
Fodde
R
The APC gene in colorectal cancer
Eur. J. Cancer
 , 
2002
, vol. 
38
 (pg. 
867
-
871
)
5.
Wood
LD
, et al.  . 
The genomic landscapes of human breast and colorectal cancers
Science
 , 
2007
, vol. 
318
 (pg. 
1108
-
1113
)
6.
Reya
T
, et al.  . 
Wnt signalling in stem cells and cancer
Nature
 , 
2005
, vol. 
434
 (pg. 
843
-
850
)
7.
Sinicrope
FA
, et al.  . 
Role of cyclooxygenase-2 in colorectal cancer
Cancer Metastasis Rev.
 , 
2004
, vol. 
23
 (pg. 
63
-
75
)
8.
Brown
JR
, et al.  . 
COX-2: a molecular target for colorectal cancer prevention
J. Clin. Oncol.
 , 
2005
, vol. 
23
 (pg. 
2840
-
2855
)
9.
Vane
JR
Inhibition of prostaglandin synthesis as a mechanism of action for aspirin-like drugs
Nat. New Biol.
 , 
1971
, vol. 
231
 (pg. 
232
-
235
)
10.
Funk
CD
Prostaglandins and leukotrienes: advances in eicosanoid biology
Science
 , 
2001
, vol. 
294
 (pg. 
1871
-
1875
)
11.
Smith
WL
, et al.  . 
Cyclooxygenases: structural, cellular, and molecular biology
Annu. Rev. Biochem.
 , 
2000
, vol. 
69
 (pg. 
145
-
182
)
12.
Cha
YI
, et al.  . 
Fishing for prostanoids: deciphering the developmental functions of cyclooxygenase-derived prostaglandins
Dev. Biol.
 , 
2006
, vol. 
289
 (pg. 
263
-
272
)
13.
Yokoyama
C
, et al.  . 
Primary structure of sheep prostaglandin endoperoxide synthase deduced from cDNA sequence
FEBS Lett.
 , 
1988
, vol. 
231
 (pg. 
347
-
351
)
14.
Xie
WL
, et al.  . 
Expression of a mitogen-responsive gene encoding prostaglandin synthase is regulated by mRNA splicing
Proc. Natl Acad. Sci. USA
 , 
1991
, vol. 
88
 (pg. 
2692
-
2696
)
15.
Chandrasekharan
NV
, et al.  . 
COX-3, a cyclooxygenase-1 variant inhibited by acetaminophen and other analgesic/antipyretic drugs: cloning, structure, and expression
Proc. Natl Acad. Sci. USA
 , 
2002
, vol. 
99
 (pg. 
13926
-
13931
)
16.
Snipes
JA
, et al.  . 
Cloning and characterization of cyclooxygenase-1b (putative cyclooxygenase-3) in rat
J. Pharmacol. Exp. Ther.
 , 
2005
, vol. 
313
 (pg. 
668
-
676
)
17.
Svensson
CI
, et al.  . 
The spinal phospholipase-cyclooxygenase-prostanoid cascade in nociceptive processing
Annu. Rev. Pharmacol. Toxicol.
 , 
2002
, vol. 
42
 (pg. 
553
-
583
)
18.
Adegboyega
PA
, et al.  . 
Immunohistochemical expression of cyclooxygenase-2 in normal kidneys
Appl. Immunohistochem. Mol. Morphol.
 , 
2004
, vol. 
12
 (pg. 
71
-
74
)
19.
Kirschenbaum
A
, et al.  . 
Immunohistochemical localization of cyclooxygenase-1 and cyclooxygenase-2 in the human fetal and adult male reproductive tracts
J. Clin. Endocrinol. Metab.
 , 
2000
, vol. 
85
 (pg. 
3436
-
3441
)
20.
Kujubu
DA
, et al.  . 
TIS10, a phorbol ester tumor promoter-inducible mRNA from Swiss 3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homologue
J. Biol. Chem.
 , 
1991
, vol. 
266
 (pg. 
12866
-
12872
)
21.
Hla
T
, et al.  . 
Human cyclooxygenase-2 cDNA
Proc. Natl Acad. Sci. USA
 , 
1992
, vol. 
89
 (pg. 
7384
-
7388
)
22.
Waddell
WR
, et al.  . 
Sulindac for polyposis of the colon
J. Surg. Oncol.
 , 
1983
, vol. 
24
 (pg. 
83
-
87
)
23.
Thun
MJ
, et al.  . 
Aspirin use and reduced risk of fatal colon cancer
N. Engl. J. Med.
 , 
1991
, vol. 
325
 (pg. 
1593
-
1596
)
24.
Eberhart
CE
, et al.  . 
Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas
Gastroenterology
 , 
1994
, vol. 
107
 (pg. 
1183
-
1188
)
25.
Elder
DJ
, et al.  . 
Human colorectal adenomas demonstrate a size-dependent increase in epithelial cyclooxygenase-2 expression
J. Pathol.
 , 
2002
, vol. 
198
 (pg. 
428
-
434
)
26.
Kargman
SL
, et al.  . 
Expression of prostaglandin G/H synthase-1 and -2 protein in human colon cancer
Cancer Res.
 , 
1995
, vol. 
55
 (pg. 
2556
-
2559
)
27.
Sano
H
, et al.  . 
Expression of cyclooxygenase-1 and -2 in human colorectal cancer
Cancer Res.
 , 
1995
, vol. 
55
 (pg. 
3785
-
3789
)
28.
Sheehan
KM
, et al.  . 
The relationship between cyclooxygenase-2 expression and colorectal cancer
JAMA
 , 
1999
, vol. 
282
 (pg. 
1254
-
1257
)
29.
Elder
DJ
, et al.  . 
COX-2 inhibitors for colorectal cancer
Nat. Med.
 , 
1998
, vol. 
4
 (pg. 
392
-
393
)
30.
Arber
N
, et al.  . 
Celecoxib for the prevention of colorectal adenomatous polyps
N. Engl. J. Med.
 , 
2006
, vol. 
355
 (pg. 
885
-
895
)
31.
Steinbach
G
, et al.  . 
The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis
N. Engl. J. Med.
 , 
2000
, vol. 
342
 (pg. 
1946
-
1952
)
32.
Oshima
M
, et al.  . 
Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2)
Cell
 , 
1996
, vol. 
87
 (pg. 
803
-
809
)
33.
Jacoby
RF
, et al.  . 
The cyclooxygenase-2 inhibitor celecoxib is a potent preventive and therapeutic agent in the min mouse model of adenomatous polyposis
Cancer Res.
 , 
2000
, vol. 
60
 (pg. 
5040
-
5044
)
34.
Pugh
S
, et al.  . 
Patients with adenomatous polyps and carcinomas have increased colonic mucosal prostaglandin E2
Gut
 , 
1994
, vol. 
35
 (pg. 
675
-
678
)
35.
Rigas
B
, et al.  . 
Altered eicosanoid levels in human colon cancer
J. Lab. Clin. Med.
 , 
1993
, vol. 
122
 (pg. 
518
-
523
)
36.
Yang
VW
, et al.  . 
Size-dependent increase in prostanoid levels in adenomas of patients with familial adenomatous polyposis
Cancer Res.
 , 
1998
, vol. 
58
 (pg. 
1750
-
1753
)
37.
Kettunen
HL
, et al.  . 
Intestinal immune responses in wild-type and Apcmin/+ mouse, a model for colon cancer
Cancer Res.
 , 
2003
, vol. 
63
 (pg. 
5136
-
5142
)
38.
Giardiello
FM
, et al.  . 
Prostanoids, ornithine decarboxylase, and polyamines in primary chemoprevention of familial adenomatous polyposis
Gastroenterology
 , 
2004
, vol. 
126
 (pg. 
425
-
431
)
39.
Stolina
M
, et al.  . 
Specific inhibition of cyclooxygenase 2 restores antitumor reactivity by altering the balance of IL-10 and IL-12 synthesis
J. Immunol.
 , 
2000
, vol. 
164
 (pg. 
361
-
370
)
40.
Kawamori
T
, et al.  . 
Enhancement of colon carcinogenesis by prostaglandin E2 administration
Carcinogenesis
 , 
2003
, vol. 
24
 (pg. 
985
-
990
)
41.
Wang
D
, et al.  . 
Prostaglandin E2 enhances intestinal adenoma growth via activation of the Ras-mitogen-activated protein kinase cascade
Cancer Res.
 , 
2005
, vol. 
65
 (pg. 
1822
-
1829
)
42.
Chell
S
, et al.  . 
Mediators of PGE2 synthesis and signalling downstream of COX-2 represent potential targets for the prevention/treatment of colorectal cancer
Biochim. Biophys. Acta
 , 
2006
, vol. 
1766
 (pg. 
104
-
119
)
43.
Elder
DJ
, et al.  . 
Induction of apoptotic cell death in human colorectal carcinoma cell lines by a cyclooxygenase-2 (COX-2)-selective nonsteroidal anti-inflammatory drug: independence from COX-2 protein expression
Clin. Cancer Res.
 , 
1997
, vol. 
3
 (pg. 
1679
-
1683
)
44.
Hanif
R
, et al.  . 
Effects of nonsteroidal anti-inflammatory drugs on proliferation and on induction of apoptosis in colon cancer cells by a prostaglandin-independent pathway
Biochem. Pharmacol.
 , 
1996
, vol. 
52
 (pg. 
237
-
245
)
45.
Chan
TA
, et al.  . 
Mechanisms underlying nonsteroidal antiinflammatory drug-mediated apoptosis
Proc. Natl Acad. Sci. USA
 , 
1998
, vol. 
95
 (pg. 
681
-
686
)
46.
Paik
JH
, et al.  . 
Two opposing effects of non-steroidal anti-inflammatory drugs on the expression of the inducible cyclooxygenase. Mediation through different signaling pathways
J. Biol. Chem.
 , 
2000
, vol. 
275
 (pg. 
28173
-
28179
)
47.
Pang
L
, et al.  . 
Cyclooxygenase-2 expression by nonsteroidal anti-inflammatory drugs in human airway smooth muscle cells: role of peroxisome proliferator-activated receptors
J. Immunol.
 , 
2003
, vol. 
170
 (pg. 
1043
-
1051
)
48.
Wilson
JW
, et al.  . 
The effect of exogenous prostaglandin administration on tumor size and yield in Min/+ mice
Cancer Res.
 , 
2000
, vol. 
60
 (pg. 
4645
-
4653
)
49.
Bol
DK
, et al.  . 
Cyclooxygenase-2 overexpression in the skin of transgenic mice results in suppression of tumor development
Cancer Res.
 , 
2002
, vol. 
62
 (pg. 
2516
-
2521
)
50.
Yegnasubramanian
S
, et al.  . 
Hypermethylation of CpG islands in primary and metastatic human prostate cancer
Cancer Res.
 , 
2004
, vol. 
64
 (pg. 
1975
-
1986
)
51.
Murata
H
, et al.  . 
Promoter hypermethylation silences cyclooxygenase-2 (Cox-2) and regulates growth of human hepatocellular carcinoma cells
Lab. Invest.
 , 
2004
, vol. 
84
 (pg. 
1050
-
1059
)
52.
Wang
D
, et al.  . 
Prostaglandin E(2) promotes colorectal adenoma growth via transactivation of the nuclear peroxisome proliferator-activated receptor delta
Cancer Cell
 , 
2004
, vol. 
6
 (pg. 
285
-
295
)
53.
Chulada
PC
, et al.  . 
Genetic disruption of Ptgs-1, as well as Ptgs-2, reduces intestinal tumorigenesis in Min mice
Cancer Res.
 , 
2000
, vol. 
60
 (pg. 
4705
-
4708
)
54.
Chell
SD
, et al.  . 
Increased EP4 receptor expression in colorectal cancer progression promotes cell growth and anchorage independence
Cancer Res.
 , 
2006
, vol. 
66
 (pg. 
3106
-
3113
)
55.
Kerr
JF
, et al.  . 
Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics
Br. J. Cancer
 , 
1972
, vol. 
26
 (pg. 
239
-
257
)
56.
Adams
JM
Ways of dying: multiple pathways to apoptosis
Genes Dev.
 , 
2003
, vol. 
17
 (pg. 
2481
-
2495
)
57.
Green
DR
Apoptotic pathways: the roads to ruin
Cell
 , 
1998
, vol. 
94
 (pg. 
695
-
698
)
58.
Thompson
CB
Apoptosis in the pathogenesis and treatment of disease
Science
 , 
1995
, vol. 
267
 (pg. 
1456
-
1462
)
59.
Green
DR
, et al.  . 
A matter of life and death
Cancer Cell
 , 
2002
, vol. 
1
 (pg. 
19
-
30
)
60.
Pelengaris
S
, et al.  . 
Suppression of Myc-induced apoptosis in beta cells exposes multiple oncogenic properties of Myc and triggers carcinogenic progression
Cell
 , 
2002
, vol. 
109
 (pg. 
321
-
334
)
61.
Danial
NN
, et al.  . 
Cell death: critical control points
Cell
 , 
2004
, vol. 
116
 (pg. 
205
-
219
)
62.
Datta
SR
, et al.  . 
Akt phosphorylation of BAD couples survival signals to the cell-intrinsic death machinery
Cell
 , 
1997
, vol. 
91
 (pg. 
231
-
241
)
63.
Bonni
A
, et al.  . 
Cell survival promoted by the Ras-MAPK signaling pathway by transcription-dependent and -independent mechanisms
Science
 , 
1999
, vol. 
286
 (pg. 
1358
-
1362
)
64.
Tsujii
M
, et al.  . 
Alterations in cellular adhesion and apoptosis in epithelial cells overexpressing prostaglandin endoperoxide synthase 2
Cell
 , 
1995
, vol. 
83
 (pg. 
493
-
501
)
65.
Hague
A
, et al.  . 
Sodium butyrate induces apoptosis in human colonic tumour cell lines in a p53-independent pathway: implications for the possible role of dietary fibre in the prevention of large-bowel cancer
Int. J. Cancer
 , 
1993
, vol. 
55
 (pg. 
498
-
505
)
66.
Sheng
H
, et al.  . 
Modulation of apoptosis and Bcl-2 expression by prostaglandin E2 in human colon cancer cells
Cancer Res.
 , 
1998
, vol. 
58
 (pg. 
362
-
366
)
67.
Sheng
H
, et al.  . 
Prostaglandin E2 increases growth and motility of colorectal carcinoma cells
J. Biol. Chem.
 , 
2001
, vol. 
276
 (pg. 
18075
-
18081
)
68.
Tessner
TG
, et al.  . 
Prostaglandin E2 reduces radiation-induced epithelial apoptosis through a mechanism involving AKT activation and bax translocation
J. Clin. Invest.
 , 
2004
, vol. 
114
 (pg. 
1676
-
1685
)
69.
Pozzi
A
, et al.  . 
Colon carcinoma cell growth is associated with prostaglandin E2/EP4 receptor-evoked ERK activation
J. Biol. Chem.
 , 
2004
, vol. 
279
 (pg. 
29797
-
29804
)
70.
Leone
V
, et al.  . 
PGE2 inhibits apoptosis in human adenocarcinoma Caco-2 cell line through Ras-PI3K association and cAMP-dependent kinase A activation
Am. J. Physiol. Gastrointest. Liver Physiol.
 , 
2007
, vol. 
293
 (pg. 
G673
-
G681
)
71.
Pai
R
, et al.  . 
Prostaglandin E2 transactivates EGF receptor: a novel mechanism for promoting colon cancer growth and gastrointestinal hypertrophy
Nat. Med.
 , 
2002
, vol. 
8
 (pg. 
289
-
293
)
72.
Buchanan
FG
, et al.  . 
Prostaglandin E2 regulates cell migration via the intracellular activation of the epidermal growth factor receptor
J. Biol. Chem.
 , 
2003
, vol. 
278
 (pg. 
35451
-
35457
)
73.
He
TC
, et al.  . 
PPARdelta is an APC-regulated target of nonsteroidal anti-inflammatory drugs
Cell
 , 
1999
, vol. 
99
 (pg. 
335
-
345
)
74.
Reed
KR
, et al.  . 
PPARdelta status and Apc-mediated tumourigenesis in the mouse intestine
Oncogene
 , 
2004
, vol. 
23
 (pg. 
8992
-
8996
)
75.
Kaidi
A
, et al.  . 
Direct transcriptional up-regulation of cyclooxygenase-2 by hypoxia-inducible factor (HIF)-1 promotes colorectal tumor cell survival and enhances HIF-1 transcriptional activity during hypoxia
Cancer Res.
 , 
2006
, vol. 
66
 (pg. 
6683
-
6691
)
76.
Hawcroft
G
, et al.  . 
Prostaglandin E2-EP4 receptor signalling promotes tumorigenic behaviour of HT-29 human colorectal cancer cells
Oncogene
 , 
2007
, vol. 
26
 (pg. 
3006
-
3019
)
77.
Downward
J
Targeting RAS signalling pathways in cancer therapy
Nat. Rev. Cancer
 , 
2003
, vol. 
3
 (pg. 
11
-
22
)
78.
Vivanco
I
, et al.  . 
The phosphatidylinositol 3-Kinase AKT pathway in human cancer
Nat. Rev. Cancer
 , 
2002
, vol. 
2
 (pg. 
489
-
501
)
79.
Bos
JL
, et al.  . 
Prevalence of ras gene mutations in human colorectal cancers
Nature
 , 
1987
, vol. 
327
 (pg. 
293
-
297
)
80.
Davies
H
, et al.  . 
Mutations of the BRAF gene in human cancer
Nature
 , 
2002
, vol. 
417
 (pg. 
949
-
954
)
81.
Samuels
Y
, et al.  . 
High frequency of mutations of the PIK3CA gene in human cancers
Science
 , 
2004
, vol. 
304
 pg. 
554
 
82.
Nassif
NT
, et al.  . 
PTEN mutations are common in sporadic microsatellite stable colorectal cancer
Oncogene
 , 
2004
, vol. 
23
 (pg. 
617
-
628
)
83.
Carpten
JD
, et al.  . 
A transforming mutation in the pleckstrin homology domain of AKT1 in cancer
Nature
 , 
2007
, vol. 
448
 (pg. 
439
-
444
)
84.
Siegel
PM
, et al.  . 
Cytostatic and apoptotic actions of TGF-beta in homeostasis and cancer
Nat. Rev. Cancer
 , 
2003
, vol. 
3
 (pg. 
807
-
821
)
85.
Massague
J
TGFbeta in Cancer
Cell
 , 
2008
, vol. 
134
 (pg. 
215
-
230
)
86.
Markowitz
S
, et al.  . 
Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability
Science
 , 
1995
, vol. 
268
 (pg. 
1336
-
1338
)
87.
Radtke
F
, et al.  . 
Self-renewal and cancer of the gut: two sides of a coin
Science
 , 
2005
, vol. 
307
 (pg. 
1904
-
1909
)
88.
van de Wetering
M
, et al.  . 
The beta-catenin/TCF-4 complex imposes a crypt progenitor phenotype on colorectal cancer cells
Cell
 , 
2002
, vol. 
111
 (pg. 
241
-
250
)
89.
Castellone
MD
, et al.  . 
Prostaglandin E2 promotes colon cancer cell growth through a Gs-axin-beta-catenin signaling axis
Science
 , 
2005
, vol. 
310
 (pg. 
1504
-
1510
)
90.
Reya
T
, et al.  . 
Stem cells, cancer, and cancer stem cells
Nature
 , 
2001
, vol. 
414
 (pg. 
105
-
111
)
91.
Feinberg
AP
, et al.  . 
The epigenetic progenitor origin of human cancer
Nat. Rev. Genet
 , 
2006
, vol. 
7
 (pg. 
21
-
33
)
92.
Kinzler
KW
, et al.  . 
Lessons from hereditary colorectal cancer
Cell
 , 
1996
, vol. 
87
 (pg. 
159
-
170
)
93.
Korinek
V
, et al.  . 
Depletion of epithelial stem-cell compartments in the small intestine of mice lacking Tcf-4
Nat. Genet.
 , 
1998
, vol. 
19
 (pg. 
379
-
383
)
94.
North
TE
, et al.  . 
Prostaglandin E2 regulates vertebrate haematopoietic stem cell homeostasis
Nature
 , 
2007
, vol. 
447
 (pg. 
1007
-
1011
)
95.
Liou
JY
, et al.  . 
Cyclooxygenase-2-derived prostaglandin e2 protects mouse embryonic stem cells from apoptosis
Stem Cells
 , 
2007
, vol. 
25
 (pg. 
1096
-
1103
)
96.
Hanahan
D
, et al.  . 
Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis
Cell
 , 
1996
, vol. 
86
 (pg. 
353
-
364
)
97.
Tsujii
M
, et al.  . 
Cyclooxygenase regulates angiogenesis induced by colon cancer cells
Cell
 , 
1998
, vol. 
93
 (pg. 
705
-
716
)
98.
Jones
MK
, et al.  . 
Inhibition of angiogenesis by nonsteroidal anti-inflammatory drugs: insight into mechanisms and implications for cancer growth and ulcer healing
Nat. Med.
 , 
1999
, vol. 
5
 (pg. 
1418
-
1423
)
99.
Fukuda
R
, et al.  . 
Vascular endothelial growth factor gene expression in colon cancer cells exposed to prostaglandin E2 is mediated by hypoxia-inducible factor 1
Cancer Res.
 , 
2003
, vol. 
63
 (pg. 
2330
-
2334
)
100.
Williams
CS
, et al.  . 
Host cyclooxygenase-2 modulates carcinoma growth
J. Clin. Invest.
 , 
2000
, vol. 
105
 (pg. 
1589
-
1594
)
101.
Sonoshita
M
, et al.  . 
Acceleration of intestinal polyposis through prostaglandin receptor EP2 in Apc(Delta 716) knockout mice
Nat. Med.
 , 
2001
, vol. 
7
 (pg. 
1048
-
1051
)
102.
Dormond
O
, et al.  . 
NSAIDs inhibit alpha V beta 3 integrin-mediated and Cdc42/Rac-dependent endothelial-cell spreading, migration and angiogenesis
Nat. Med.
 , 
2001
, vol. 
7
 (pg. 
1041
-
1047
)
103.
Salcedo
R
, et al.  . 
Angiogenic effects of prostaglandin E2 are mediated by up-regulation of CXCR4 on human microvascular endothelial cells
Blood
 , 
2003
, vol. 
102
 (pg. 
1966
-
1977
)
104.
Wang
D
, et al.  . 
CXCL1 induced by prostaglandin E2 promotes angiogenesis in colorectal cancer
J. Exp. Med.
 , 
2006
, vol. 
203
 (pg. 
941
-
951
)
105.
Weinberg
RA
Mechanisms of malignant progression
Carcinogenesis
 , 
2008
, vol. 
29
 (pg. 
1092
-
1095
)
106.
Fenwick
SW
, et al.  . 
The effect of the selective cyclooxygenase-2 inhibitor rofecoxib on human colorectal cancer liver metastases
Gastroenterology
 , 
2003
, vol. 
125
 (pg. 
716
-
729
)
107.
Yao
M
, et al.  . 
Inhibition of cyclooxygenase-2 by rofecoxib attenuates the growth and metastatic potential of colorectal carcinoma in mice
Cancer Res.
 , 
2003
, vol. 
63
 (pg. 
586
-
592
)
108.
Tsujii
M
, et al.  . 
Cyclooxygenase-2 expression in human colon cancer cells increases metastatic potential
Proc. Natl Acad. Sci. USA
 , 
1997
, vol. 
94
 (pg. 
3336
-
3340
)
109.
Pai
R
, et al.  . 
Prostaglandins promote colon cancer cell invasion; signaling by cross-talk between two distinct growth factor receptors
FASEB J.
 , 
2003
, vol. 
17
 (pg. 
1640
-
1647
)
110.
Birchmeier
C
, et al.  . 
Met, metastasis, motility and more
Nat. Rev. Mol. Cell Biol.
 , 
2003
, vol. 
4
 (pg. 
915
-
925
)
111.
Qualtrough
D
, et al.  . 
Prostaglandin F(2alpha) stimulates motility and invasion in colorectal tumor cells
Int. J. Cancer
 , 
2007
, vol. 
121
 (pg. 
734
-
740
)
112.
Gupta
GP
, et al.  . 
Mediators of vascular remodelling co-opted for sequential steps in lung metastasis
Nature
 , 
2007
, vol. 
446
 (pg. 
765
-
770
)
113.
Hull
MA
, et al.  . 
Regulation of stromal cell cyclooxygenase-2 in the ApcMin/+ mouse model of intestinal tumorigenesis
Carcinogenesis
 , 
2006
, vol. 
27
 (pg. 
382
-
391
)
114.
Christofori
G
Cancer: division of labour
Nature
 , 
2007
, vol. 
446
 (pg. 
735
-
736
)
115.
Kroemer
G
, et al.  . 
Tumor cell metabolism: cancer's Achilles’ heel
Cancer Cell
 , 
2008
, vol. 
13
 (pg. 
472
-
482
)
116.
Palucka
AK
, et al.  . 
Taming cancer by inducing immunity via dendritic cells
Immunol. Rev.
 , 
2007
, vol. 
220
 (pg. 
129
-
150
)
117.
Zou
W
Immunosuppressive networks in the tumour environment and their therapeutic relevance
Nat. Rev. Cancer
 , 
2005
, vol. 
5
 (pg. 
263
-
274
)
118.
Harris
SG
, et al.  . 
Prostaglandins as modulators of immunity
Trends Immunol.
 , 
2002
, vol. 
23
 (pg. 
144
-
150
)
119.
Harizi
H
, et al.  . 
The impact of eicosanoids on the crosstalk between innate and adaptive immunity: the key roles of dendritic cells
Tissue Antigens
 , 
2005
, vol. 
65
 (pg. 
507
-
514
)
120.
Ahmadi
M
, et al.  . 
Prevention of both direct and cross-priming of antitumor CD8+ T-cell responses following overproduction of prostaglandin E2 by tumor cells in vivo
Cancer Res.
 , 
2008
, vol. 
68
 (pg. 
7520
-
7529
)
121.
Park
JY
, et al.  . 
Prostaglandin E2 synthesis and secretion: the role of PGE2 synthases
Clin. Immunol.
 , 
2006
, vol. 
119
 (pg. 
229
-
240
)
122.
Nomura
T
, et al.  . 
The two-step model of prostaglandin signal termination: in vitro reconstitution with the prostaglandin transporter and prostaglandin 15 dehydrogenase
Mol. Pharmacol.
 , 
2004
, vol. 
65
 (pg. 
973
-
978
)
123.
Hull
MA
, et al.  . 
Prostaglandin EP receptors: targets for treatment and prevention of colorectal cancer?
Mol. Cancer Ther.
 , 
2004
, vol. 
3
 (pg. 
1031
-
1039
)
124.
Yoshimatsu
K
, et al.  . 
Inducible microsomal prostaglandin E synthase is overexpressed in colorectal adenomas and cancer
Clin. Cancer Res.
 , 
2001
, vol. 
7
 (pg. 
3971
-
3976
)
125.
Nakanishi
M
, et al.  . 
Genetic deletion of mPGES-1 suppresses intestinal tumorigenesis
Cancer Res.
 , 
2008
, vol. 
68
 (pg. 
3251
-
3259
)
126.
Elander
N
, et al.  . 
Genetic deletion of mPGES-1 accelerates intestinal tumorigenesis in APC(Min/+) mice
Biochem. Biophys. Res. Commun.
 , 
2008
, vol. 
372
 (pg. 
249
-
253
)
127.
Mutoh
M
, et al.  . 
Involvement of prostaglandin E receptor subtype EP(4) in colon carcinogenesis
Cancer Res.
 , 
2002
, vol. 
62
 (pg. 
28
-
32
)
128.
Amano
H
, et al.  . 
Host prostaglandin E(2)-EP3 signaling regulates tumor-associated angiogenesis and tumor growth
J. Exp. Med.
 , 
2003
, vol. 
197
 (pg. 
221
-
232
)
129.
Shoji
Y
, et al.  . 
Downregulation of prostaglandin E receptor subtype EP3 during colon cancer development
Gut
 , 
2004
, vol. 
53
 (pg. 
1151
-
1158
)
130.
Fulton
AM
, et al.  . 
Targeting prostaglandin E EP receptors to inhibit metastasis
Cancer Res.
 , 
2006
, vol. 
66
 (pg. 
9794
-
9797
)
131.
Yang
L
, et al.  . 
Host and direct antitumor effects and profound reduction in tumor metastasis with selective EP4 receptor antagonism
Cancer Res.
 , 
2006
, vol. 
66
 (pg. 
9665
-
9672
)
132.
Schuster
VL
Prostaglandin transport
Prostaglandins Other Lipid Mediat.
 , 
2002
, vol. 
68–69
 (pg. 
633
-
647
)
133.
Ensor
CM
, et al.  . 
15-Hydroxyprostaglandin dehydrogenase
J. Lipid Mediat. Cell Signal.
 , 
1995
, vol. 
12
 (pg. 
313
-
319
)
134.
Yan
M
, et al.  . 
15-Hydroxyprostaglandin dehydrogenase, a COX-2 oncogene antagonist, is a TGF-beta-induced suppressor of human gastrointestinal cancers
Proc. Natl Acad. Sci. USA
 , 
2004
, vol. 
101
 (pg. 
17468
-
17473
)
135.
Backlund
MG
, et al.  . 
15-Hydroxyprostaglandin dehydrogenase is down-regulated in colorectal cancer
J. Biol. Chem.
 , 
2005
, vol. 
280
 (pg. 
3217
-
3223
)
136.
Mann
JR
, et al.  . 
Repression of prostaglandin dehydrogenase by epidermal growth factor and snail increases prostaglandin E2 and promotes cancer progression
Cancer Res.
 , 
2006
, vol. 
66
 (pg. 
6649
-
6656
)
137.
Myung
SJ
, et al.  . 
15-Hydroxyprostaglandin dehydrogenase is an in vivo suppressor of colon tumorigenesis
Proc. Natl Acad. Sci. USA
 , 
2006
, vol. 
103
 (pg. 
12098
-
12102
)
138.
Holla
V
, et al.  . 
Regulation of prostaglandin transporters in colorectal neoplasia
Cancer Prev. Res.
 , 
2008
, vol. 
1
 (pg. 
93
-
99
)
139.
Backlund
MG
, et al.  . 
Repression of 15-hydroxyprostaglandin dehydrogenase involves histone deacetylase 2 and snail in colorectal cancer
Cancer Res.
 , 
2008
, vol. 
68
 (pg. 
9331
-
9337
)
140.
Subbaramaiah
K
, et al.  . 
Cyclooxygenase 2: a molecular target for cancer prevention and treatment
Trends Pharmacol. Sci.
 , 
2003
, vol. 
24
 (pg. 
96
-
102
)
141.
Howe
LR
, et al.  . 
PEA3 is up-regulated in response to Wnt1 and activates the expression of cyclooxygenase-2
J. Biol. Chem.
 , 
2001
, vol. 
276
 (pg. 
20108
-
20115
)
142.
Howe
LR
, et al.  . 
Transcriptional activation of cyclooxygenase-2 in Wnt-1-transformed mouse mammary epithelial cells
Cancer Res.
 , 
1999
, vol. 
59
 (pg. 
1572
-
1577
)
143.
Araki
Y
, et al.  . 
Regulation of cyclooxygenase-2 expression by the Wnt and ras pathways
Cancer Res.
 , 
2003
, vol. 
63
 (pg. 
728
-
734
)
144.
Coffey
RJ
, et al.  . 
Epidermal growth factor receptor activation induces nuclear targeting of cyclooxygenase-2, basolateral release of prostaglandins, and mitogenesis in polarizing colon cancer cells
Proc. Natl Acad. Sci. USA
 , 
1997
, vol. 
94
 (pg. 
657
-
662
)
145.
Sheng
H
, et al.  . 
Cyclooxygenase-2 induction and transforming growth factor beta growth inhibition in rat intestinal epithelial cells
Cell Growth Differ.
 , 
1997
, vol. 
8
 (pg. 
463
-
470
)
146.
Jones
MK
, et al.  . 
HGF triggers activation of the COX-2 gene in rat gastric epithelial cells: action mediated through the ERK2 signaling pathway
FASEB J.
 , 
1999
, vol. 
13
 (pg. 
2186
-
2194
)
147.
Guo
YS
, et al.  . 
Gastrin stimulates cyclooxygenase-2 expression in intestinal epithelial cells through multiple signaling pathways. Evidence for involvement of ERK5 kinase and transactivation of the epidermal growth factor receptor
J. Biol. Chem.
 , 
2002
, vol. 
277
 (pg. 
48755
-
48763
)
148.
Csiki
I
, et al.  . 
Thioredoxin-1 modulates transcription of cyclooxygenase-2 via hypoxia-inducible factor-1alpha in non-small cell lung cancer
Cancer Res.
 , 
2006
, vol. 
66
 (pg. 
143
-
150
)
149.
Schmedtje
JF
Jr
, et al.  . 
Hypoxia induces cyclooxygenase-2 via the NF-kappaB p65 transcription factor in human vascular endothelial cells
J. Biol. Chem.
 , 
1997
, vol. 
272
 (pg. 
601
-
608
)
150.
Demasi
M
, et al.  . 
Effects of hypoxia on monocyte inflammatory mediator production: dissociation between changes in cyclooxygenase-2 expression and eicosanoid synthesis
J. Biol. Chem.
 , 
2003
, vol. 
278
 (pg. 
38607
-
38616
)
151.
Bonazzi
A
, et al.  . 
Regulation of cyclooxygenase-2 by hypoxia and peroxisome proliferators in the corneal epithelium
J. Biol. Chem.
 , 
2000
, vol. 
275
 (pg. 
2837
-
2844
)
152.
Vaupel
P
, et al.  . 
Tumor hypoxia and malignant progression
Methods Enzymol.
 , 
2004
, vol. 
381
 (pg. 
335
-
354
)
153.
Shimizu
S
, et al.  . 
Prevention of hypoxia-induced cell death by Bcl-2 and Bcl-xL
Nature
 , 
1995
, vol. 
374
 (pg. 
811
-
813
)
154.
Koong
AC
, et al.  . 
Hypoxic activation of nuclear factor-kappa B is mediated by a Ras and Raf signaling pathway and does not involve MAP kinase (ERK1 or ERK2)
Cancer Res.
 , 
1994
, vol. 
54
 (pg. 
5273
-
5279
)
155.
Poligone
B
, et al.  . 
Positive and negative regulation of NF-kappaB by COX-2: roles of different prostaglandins
J. Biol. Chem.
 , 
2001
, vol. 
276
 (pg. 
38658
-
38664
)
156.
Liston
P
, et al.  . 
Suppression of apoptosis in mammalian cells by NAIP and a related family of IAP genes
Nature
 , 
1996
, vol. 
379
 (pg. 
349
-
353
)
157.
Dong
Z
, et al.  . 
Up-regulation of apoptosis inhibitory protein IAP-2 by hypoxia. Hif-1-independent mechanisms
J. Biol. Chem.
 , 
2001
, vol. 
276
 (pg. 
18702
-
18709
)
158.
Staller
P
, et al.  . 
Chemokine receptor CXCR4 downregulated by von Hippel-Lindau tumour suppressor pVHL
Nature
 , 
2003
, vol. 
425
 (pg. 
307
-
311
)
159.
Ding
M
, et al.  . 
Loss of the tumor suppressor Vhlh leads to upregulation of Cxcr4 and rapidly progressive glomerulonephritis in mice
Nat. Med.
 , 
2006
, vol. 
12
 (pg. 
1081
-
1087
)
160.
Kaidi
A
, et al.  . 
Interaction between beta-catenin and HIF-1 promotes cellular adaptation to hypoxia
Nat. Cell Biol.
 , 
2007
, vol. 
9
 (pg. 
210
-
217
)
161.
Bresalier
RS
, et al.  . 
Cardiovascular events associated with rofecoxib in a colorectal adenoma chemoprevention trial
N. Engl. J. Med.
 , 
2005
, vol. 
352
 (pg. 
1092
-
1102
)