Prostaglandin E2-induced colonic secretion in patients with and without colorectal neoplasia

Colorectal cancer (CRC) is the third most common type of cancer and the second leading cause of death among cancers in the Western world [1]. Therapy is usually through surgery, which in severe cases is followed by chemotherapy [2]. There is a need for additional medical therapy and prevention of CRC, which necessitates further insight into the presently poorly understood mechanisms of colorectal mucosal defence, repair and carcinogenesis. In particular, the mechanisms and signal pathways of pre-neoplastic colorectal epithelial cells are of special interest as these could be target for pharmacotherapy in the prevention of colorectal neoplasia (CRN) and CRC.

By their inhibitory action on the cyclooxygenase enzyme (COX), non-steroidal anti-inflammatory drugs (NSAIDs) are partly chemopreventive against CRC, an effect maybe particularly due to attenuation of the enzyme isoform 2 (COX-2) [3‐6]. This protection is believed to be mediated, at least in part, through reduction of prostaglandin E₂ (PGE₂) levels [7, 8], as PGE₂ promotes cell growth, migration and angiogenesis and reduce apoptosis [4]. In general, the downstream cascades of the PGE₂ signaling are altered (due to gene mutations) in CRN [9‐11]. These mutations may both affect PGE₂ production per se via a regulation of COX enzymes in pro-inflammatory cells including epithelial cells as well as the PGE₂-dependent signaling pathways in target cells, Figure 1. Whether changes in levels of PGE₂ are primary or secondary causes of CRN remains unclear. With respect to PGE₂ receptors, CRC cells and their neighboring cells have augmented expression of receptors EP2 and EP4, while initially the EP3 receptor expression often is lowered [4, 8, 12‐15], Figure 1.

PGE₂ is the primary endogenous agonist for the EP receptors and stimulates all 4 EP receptor subtypes [16]. Butaprost is a selective EP2 agonist [16]. Sulprostone is mainly an EP3 agonist but it is also a weak agonist of EP1 [16]. PGE₁ (OH) is mainly an EP4 agonist but it is also a weak agonist of EP3 [16].

Studies of transport mechanisms in human intestinal epithelia in vivo require invasive bothersome procedures. In order to circumvent these problems, the Modified Ussing Air-Suction (MUAS) chamber has been developed for the study of duodenal and colonic epithelia in vitro [17, 18]. Fairly easily, this method enables us to study epithelial electrogenic ion transport in human biopsies obtained during endoscopy and has been proven useful for functional receptor studies [19].

In this study we sought to establish possible differences in functional expression/response of PGE₂ EP-receptor subtypes in colonic biopsies from CRN patients and control subjects by the use of indomethacin, PGE₂ and selective EP receptor agonists. Our findings indicate that normally appearing colon in CRN patients, including CRC patients, express higher COX enzyme activity than in control patients but with no difference in the functional expression of the four EP receptor subtypes with respect to transepithelial ion transport.

The present study provides evidence that electrogenic transport is altered in histological normal appearing colonic mucosa in CRN patients with respect to indomethacin-sensitive mechanisms. This finding supports that NSAID-sensitive mechanisms are activated not only in tumor tissue [22] but also in normal appearing tissue.

Levels of PGE₂ in the para-/auto-crine milieu are elevated in CRN [23‐25] and the clinical benefits of reducing PGE₂ levels for CRC has been documented [5, 22, 26‐30]. Accordingly, related therapeutic strategies are suggested [31, 32]. But what are the mechanisms for the elevated PGE₂ levels in CRN and how do we explain the effect of indomethacin in the present study?

Several enzymes and regulatory pathways control the level of PGE₂ both in colonic cells and in the intercellular environment of colonic tissue, Figure 1. In particular, it appears that the COX-2 enzyme expression and hence its activity is elevated in human colonic carcinogenic cells [27, 33‐35]. Activation of COX-2 enzymes leads to an immediate increase in the intracellular level of PGE₂, Figure 1. Regulation of the COX-2 expression is simultaneously controlled by a host of intracellular and extracellular signaling pathways [35‐37], Figure 1. In addition, the extracellular concentration of PGE₂ is controlled by several other means. Increasing the extracellular PGE₂ through prostaglandin secretion in ABC efflux transporters, MRP4 and MOAT, is probably augmented by elevated activity in the MRP4 transporter as found in CRC cells, Figure 1[38]. Removal of extracellular PGE₂ is partially performed by a specific prostaglandin transporter, PGT, and further degradation of PGE₂ is by the enzymes 15-prostaglandin-dehydrogenase, 15-PGDH, and carbonyl reductase, CBR, of which, both the activity by PGT and 15-PGDH are reduced in CRC cells, Figure 1[3, 38, 39]. Additional support for the involvement of altered15-PGDH enzyme activity in CRC development was recently documented in a mouse model of 15-PGDH -/-mouse demonstrating resistance to COX-2-inhibitor celecoxib' prevention of colon tumors [40]. Thus, the combined effects of increased synthesis and export together with reduced elimination and degradation of extracellular PGE₂ in CRC point to a maintained higher level of PGE₂ in the auto-/paracrine milieu in the CRC colonic epithelium, Figure 1. Interestingly, indomethacin down-regulates the expression of some of the ABC transporters, which are up regulated in CRC [38, 41]. Therefore several mechanisms could account for indomethacin in regulating the bioavailability of PGE₂.

In the present study we found no difference in basal electrogenic transport values or in the response to stimulation with PGE₂ or EP receptor subtype specific agonists. Thus, the observed altered expression of EP receptor subtypes in CRC patients compared to non-CRC subjects does not seem to be manifest in the functionality of the receptors in epithelial transport. We speculate that the observed increased sensitivity to indomethacin in the N-patients is due to a higher baseline production of prostaglandins in N-patients following an increased expression/activity of the COX-2 enzyme and export of PGE₂ [27, 38] and/or a lowered removal and degradation of PGE₂, Figure 1[38, 39]. This would be consistent with the lack of differences between specific agonists of the EP receptors as well as still explaining the chemopreventive effect of the NSAIDs. To prove the involvement of the isoform COX-2 as a direct link to our findings, experiments with a specific COX-2 inhibitor would be required.

Although there is no significant difference in the baseline secretion measured as SCC between N-patients and C-patients, the basal current is nearly three times higher than the indomethacin-induced lowering of the current. The significantly higher (5 to 6 μA·cm^-2) differential inhibition of the basal current by indomethacin in N-patients compared to C-patients is not discernable as a significant difference in the basal SCC as such. Biopsies from N-patients had a mere non-significant (3.3 μA·cm^-2) higher value of SCC in the total basal current compared to C-patients. A straight forward explanation for this lack of significance is that the total basal current is most likely a sum of several different electrogenic transport processes including amiloride-sensitive sodium absorption and Ca²⁺-induced anion secretion, not accounted for in the indometacin-sensitive transport. This is substantiated by a partial inhibition of the current by bumetanide and an even higher inhibition of the total current by ouabain compared to indomethacin inhibition, Figure 5. Thus a significant difference in a single transport type, as for indomethacin, is possibly smeared to insignificance by its inclusion in the total basal current with a higher variability.

Of the four employed EP receptor agonists, PGE₂ induced the largest increase in SCC (i.e. electrogenic secretion), which is not surprising as PGE₂ stimulates all four EP receptor subtypes. EP4 seems to be the receptor subtype that is responsible for the largest proportion of this secretion, as stimulation with PGE₁ (OH) brought about the largest increase in secretion. This fits with findings in the human duodenum using the same experimental technique [19]. Similar mechanisms appear to be present in rat colon, but in contrast to rats, where EP4 is the major mediator of the PGE₂ response [21], the human colon PGE₂-induced secretion seems to be induced by all four receptor subtypes, EP4, EP2 and EP1/EP3, Figure 4. This underlines the notable difference of functional EP receptors in some animals as compared to man. However, in agreement with findings in rat colon [21], PGE₂ was found to be only a partial agonist of secretion, as forskolin almost doubled the electrogenic secretory response of PGE₂. It is not possible to make any final conclusions regarding the observed differences between human and rat colon, although species or regional characteristics might explain the observed differences.

The PGE₁-derivative Lubiprostone is used clinically as a laxative. Besides its prokinetic effect, Lubiprostone is also believed to be a chloride channel, ClC-2, opener and now further a stimulant of EP4 receptor subtypes in human ileum and colon epithelium chloride secretion, dependent on a well-functioning cystic fibrosis transmembrane conductance regulator protein, CFTR [42], as well as of the EP4 receptor subtype in human and rat duodenal epithelial bicarbonate secretion [19, 43]. This information corroborates our use of PGE₁(OH) as a rather specific EP4 receptor subtype agonist and that this receptor subtype is the major pathway for PGE₂-induced colonic anion secretion in man, Figure 4.

Epithelial electrogenic transport is altered in histological normal appearing colonic mucosal biopsies from patients with previous or present colorectal neoplastic disease with respect to indomethacin-sensitive mechanisms. The mode of action for the observed effect of indomethacin does not seem related to the functionality of expressed prostaglandin EP receptor subtypes, EP1-4, when compared with non-neoplastic patient tissues. Therefore, our results point directly to a change in the activity of COX enzymes in the normal colonic tissue from patients with neoplasia, although at present we cannot differentiate between the isoform of COX enzymes possibly being involved. To do this selective COX-1 and COX-2 inhibitors would have to be tested.