Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The isoprostanes are a group of PG-like compounds formed by the free radical-catalyzed peroxidation of AA independent of cyclooxygenase activity (Morrow et al., 1990b, Morrow and Roberts, 1996). They differ from the PGs by the cis orientation of their side chains (fig. 1). Formation proceeds through four positional peroxyl radical isomers of AA that undergo endocyclization to yield PGG₂-like bicyclic endoperoxides. The endoperoxides are then reduced to form four PGF₂-like regioisomers (F₂ isoprostanes), each of which may, in theory, be composed of a mixture of eight racemic diastereomers (Morrow et al., 1990b). Both D₂ and E₂ isoprostanes can also be formed following the rearrangement and subsequent reduction of the bicyclic endoperoxides (Morrow et al., 1994). The F₂ isoprostanes have been shown to be formed in situ on phospholipids and, unlike cyclooxygenase-derived PGs, are released preformed (Morrow et al., 1992a). It has, however, been suggested that 8-iso-PGF₂ may also be produced via a cyclooxygenase-dependent activity (Pratico et al., 1995).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Chemical structures of the principal agonists and antagonists used in the study.

The biological effects of the isoprostanes on diverse test systems appear to be inhibited at least partially by TP receptor antagonists. It has been debated whether this indicates activity at the classic TP receptors (Crankshaw, 1995; Ogletree, 1992; Yin et al., 1994) or on a unique isoprostane receptor that bears some similarity to the TP receptor (Banerjee et al., 1992; Fukunaga et al., 1993a, 1993b, 1995; Longmire et al., 1994; Morrow et al., 1990b, 1992a, 1992b).

We sought to answer the following questions: Do the isoprostanes, specifically 8-iso-PGE₂ and 8-iso-PGF₂, have biological activity on the canine colon? Are these effects exerted on a unique isoprostane receptor, or do they involve other prostanoid receptors? We anticipated that contrasting the effects of the isoprostanes on the canine colonic epithelium and muscularis mucosae would provide useful information. Our rationale was as follows: earlier studies indicated that the epithelium did not possess TP receptors because no responses had been obtained to the selective TP receptor agonist U46619. Therefore, any isoprostane effects on this preparation would involve receptors other than the TP receptor. In contrast, the canine colonic muscularis mucosae responded well to U46619, so isoprostane effects could well be mediated through TP receptors. Tissues were set up in Ussing chambers to record the responses of the colonic epithelium or in muscle baths to record force changes in the muscularis mucosae. In the absence of selective antagonists for certain prostanoid receptors, a number of experiments used desensitization procedures. When selective antagonists were available, pK_B values were found using selective agonists and the isoprostanes. Figure 1 shows the structures of the agonists and antagonists used in this investigation.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Tissue preparation. Two tissue preparations from the canine proximal colon were used in this study. The first, a functionally "nerve-free" colonic epithelial preparation was used to examine ion transport functions. This preparation has been previously described in detail (Keenan and Rangachari, 1989; Rangachari and Prior, 1994). A second preparation, used for the contractility studies, was obtained from strips of the canine colonic muscularis mucosae. These strips were prepared using a dissection similar to that previously described for the canine gastric muscularis mucosae (Muller et al., 1994). Both preparations were obtained from mongrel dogs of either sex that had been killed with sodium pentobarbital (100 mg/kg i.v.). The proximal colon was removed, opened along the mesenteric border and rinsed in warm, oxygenated PSS buffer of the following composition: 116 mM NaCl, 4.6 mM KCl, 1.5 mM CaCl₂, 1.2 mM MgCl₂, 22 mM NaHCO₃, 1.2 mM NaH₂PO₄ and 10 mM glucose. The preparation was then pinned, mucosal surface down, in a Petri dish containing PSS buffer. The longitudinal and circular smooth muscle layers were cut off in strips, leaving the submucosal and mucosal tissue layers. The subsequent steps were modified slightly to obtain tissues for the Ussing chamber and contractility experiments. These procedures are described separately below.

Ussing chamber experiments. Forceps were used to pinch and lift the muscularis mucosae so an opening could be made in it with sharp scissors without damaging the underlying epithelium. Fine forceps were put through this opening to carefully separate the muscularis mucosae from the epithelium, thereby allowing the muscularis mucosae to be removed by blunt scissors and leaving a circle of exposed epithelium. A surrounding ring of muscularis mucosae was left intact to provide structural support. From each dog, eight such tissues were prepared. The tissue was mounted in Lucite Ussing chambers providing a tissue surface area of 1.96 cm². The mucosal and serosal surfaces were bathed by separate baths kept at 37°C. The baths contained PSS buffer that was oxygenated (5% CO₂/95% O₂) and continuously circulated using a gas lift system. Electrical measurements were done using standard methods and equipment previously described in detail (Rangachari and Prior, 1994). I_scs were chosen as the indices of tissue responsiveness and recorded with a high-impedance multivoltmeter (WPI; Sarasota, FL) and recorded using the MP100 data collection hardware (BIOPAC Systems, Goleta, CA). The voltage clamp was generated by passing sufficient current across the tissue to reduce the potential difference to zero, with appropriate corrections for solution resistance.

6

2

Contractility experiments. As described under Tissue Preparation above, the colon was stripped of circular and longitudinal muscle layers, and a further dissection was performed to remove the intact muscularis mucosae layer from the epithelial layer. Tissue was used within 24 hr of being removed from the animal. If the tissue was not used immediately, it was kept at 4°C in oxygen-saturated PSS buffer until use. Preliminary experiments indicated that such tissues remained viable and responsive to agonists.

6

4

Drugs. PGE₂, 8-iso-PGE₂, 8-iso-PGF₂, U46619 {1R-[1,4,5- (Z),6(1E,3S*)]]-7-[6-(3-hydroxy-1-octenyl)-2-oxabicyclo[2.2.1]hept-5-yl]-5-heptenoic acid} and the TP receptor antagonist SQ29548 {[1S-[1,2(5Z),3,4]]-7-[3-[[2-[(phenylamino)carbonyl]hydrazino]methyl]-7-oxabicyclo-[2.2.1]hept-2-yl]-5-heptenoic acid} were purchased from Cayman Chemical (Ann Arbor, MI). GR32191 {[1R-[1(Z),2,3,5]]-(±)-7-[5-[[1,1-biphenyl)-4-yl]methoxy]-3-hydroxy-2-(1-piperidynl) cyclopentyl]-4-heptenoic acid hydrochloride} was obtained from Dr. D. J. Crankshaw (McMaster University) through the courtesy of Glaxo Wellcome (UK). L670596 [()-6,8-di-fluoro-9-p-methylsulfonyl benzyl-1,2,3,4-tetra-hydrocarbazol-1-yl-acetic acid] was supplied by Merck Frosst Centre for Therapeutic Research (Pointe Claire, Quebec, Canada) All other drugs were purchased from Sigma Chemical (St. Louis, MO). Indomethacin was dissolved in 22 mM NaHCO₃ at a concentration of 10³ M. All other compounds, except the following, were dissolved in 70% (v/v) ethanol at a concentration of 10¹ M and kept at 20°C until use. 8-Iso-PGF₂ was dissolved in ethanol at a final concentration of 2.5 to 3 mg/mL at 20°C, SQ29548 was prepared in ethanol at 10² M and kept at 4°C, L670596 was kept at 1 mg/mL in DMSO at 4°C and GR32191 was dissolved in ethanol at 10¹ M and stored at 4°C.

Statistics. All values are expressed as the arithmetic mean ± S.E.M. For statistical comparisons, either the unpaired or paired t test was used. Where there was more than one treatment group, a one-way analysis of variance was performed. Values of P < .05 were considered to be significant.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Studies on the Epithelial Preparation

Concentration-effect experiments. Serosal additions of 8-iso-PGE₂ elicited concentration-dependent increases in I_sc. Although the responses were similar to those seen with PGE₂, it was evident that there was a clear difference in potency. Thus, the estimated pEC₅₀ value for 8-iso-PGE₂ was 6.4 ± 0.1, whereas the corresponding value for PGE₂ was 7.4 ± 0.1. Both prostanoids were, however, equieffective as there were no significant differences in the maximal responses observed (fig. 2). The addition of 8-iso-PGF₂ produced no increase in epithelial I_sc, in either the presence or absence of indomethacin. The TP receptor antagonist SQ29548 (10⁶ M) had no effect on the increases in I_sc observed with 8-iso-PGE₂ and PGE₂.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   A comparison of the concentration-response curves to 8-iso-PGE2 () and PGE2 () on the epithelial preparation. All tissues were pretreated with indomethacin (106 M). The maximal responses (Tmax) obtained for 8-iso-PGE2 and PGE2 were not statistically different (167.0 ± 17.0 vs. 169.0 ± 19.0 µA·cm2); however, there was a significant difference in potency (pEC50) between 8-iso-PGE2 and PGE2 (6.4 ± 0.1 vs. 7.4 ± 0.1). The data shown are mean ± S.E.M. from 14 to 16 experiments. Inset (top left), typical trace of Isc taken from a representative experiment. Each arrow represents the sequential addition of increasing concentrations of agonist at the indicated intervals (ranging from 109 to 3 × 105 M final concentration).

Desensitization experiments. Because the TP receptor antagonist was not effective at blocking responses to 8-iso-PGE₂, we sought to determine whether the prostanoid acted at an EP receptor. In the absence of readily available, selective EP receptor antagonists, we used desensitization experiments. After tissues were set up, they were treated repeatedly with near-maximal concentrations of 8-iso-PGE₂ (10⁶ M). Usually by the third addition of the agonist, no further increases in I_sc were noted. The tissues were then treated with a high concentration of PGE₂ (10⁶ M). In companion pieces, the order of addition of agonists was altered, so tissues were first treated with PGE₂ and subsequently with 8-iso-PGE₂. All tissues were also tested with a nonprostanoid agonist (histamine; 10⁴ M) at the end of the experiment as an independent index of tissue responsiveness. Our earlier studies have shown that there was no involvement of prostanoids in the responses of the colonic epithelium to histamine (Rangachari and Prior, 1994).

6

6

6

6

View larger version (13K): [in this window] [in a new window]   Fig. 3.   The cumulative concentration-response curves to 8-iso-PGE2 on epithelial tissues desensitized to PGE2. All tissues received three consecutive applications of PGE2 (106 M) followed by increasing concentrations of 8-iso-PGE2. Under these conditions, a concentration-dependent decrease in Isc was observed. The maximal response obtained with 8-iso-PGE2 in PGE2-treated tissues was (71.0 ± 6.0 µA·cm2) with a pEC50 of 6.8 ± 0.1. The data are mean ± S.E.M. from 4 experiments. Inset (top left), typical trace of Isc taken from a representative experiment. Each arrow represents the sequential addition of increasing concentration of agonist at the indicated intervals (ranging from 108 to 3 × 105 M final concentration).

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Representative traces of Isc from two experiments demonstrating the inhibitory responses to U46619 (105 M) and 8-iso-PGE2 (105 M) on tissues desensitized to PGE2. A, Tissue was exposed to PGE2 (106 M). After the third addition, no further increases were noted, although the Isc remained elevated. The addition of a single concentration of U46619 (105 M) elicited a sharp decrease in Isc. After the responses had attained a plateau, a single addition of 105 M DK elicited a further decrease in Isc. B, Same experiment is shown but with the addition of 8-iso-PGE2 rather than U46619.

6

6

5

5

View larger version (21K): [in this window] [in a new window]   Fig. 5.   The effects of TP receptor antagonists on the inhibitory responses to U46619 (105 M), 8-iso-PGE2 (105 M) and DK-PGD2 (105 M) on tissues previously desensitized to PGE2. For these experiments, the following protocol was employed. Tissues from the same animal were set up in pairs and pretreated with indomethacin (106 M). Then, one member of each pair was treated with a fixed concentration of a TP receptor antagonist, either L670596 (106 M), SQ29548 (106 M) or GR 32195 (106 M). All tissues were then treated with PGE2 to desensitize the EP receptor as in the earlier experiments. After the desensitization had been achieved, each pair was treated with either 8-iso-PGE2 or U46619. All tissues were finally tested with the PGD2 metabolite DK. The data shown are mean ± S.E.M. for six experiments. Solid bars, control tissues (no antagonist); open bars, treated tissues (TP receptor antagonist treated).

Studies on Canine Muscularis Mucosae Strips

U46619 as well as the two isoprostanes (8-iso-PGE₂ and 8-iso-PGF₂) produced concentration-dependent increases in tension of the muscularis mucosae strips (fig. 6). All tension data were normalized to the maximal tension produced by the addition of carbachol (10⁴ M) to each strip. The maximal tensions generated by each of the three agonists were significantly different. Compared with carbachol, U46619 functioned as a full agonist, whereas both of the isoprostanes were partial agonists. The estimated pEC₅₀ values (the concentration required to produce half-maximal response) were as follows: U46619, 7.7 ± 0.1; 8-iso-PGE₂, 7.4 ± 0.1; and 8-iso-PGF₂, 6.9 ± 0.1.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Cumulative concentration-response curves to U46619 (; Tmax = 112.8.0 ± 9.8, pEC50 = 7.7 ± 0.1), 8-iso PGE2 (; Tmax = 84.0 ± 8.0, pEC50 = 7.4 ± 0.1) and 8-iso-PGF2 (; Tmax = 53.0 ± 6.0, pEC50 = 6.9 ± 0.1) in canine colonic muscularis mucosae strips. The values are mean ± S.E.M. of 13 or 14 experiments. These values represent changes in force expressed as a percentage of the maximum force generated in response to carbachol (104 M).

To determine whether the effects of the two isoprostanes were exerted on TP receptors, we tested the inhibitory effects of three different TP receptor antagonists. All three antagonists tested (SQ29548, L670596 and GR32191) inhibited responses to 8-iso-PGF₂, 8-iso-PGE₂ and U46619. From the data obtained, we estimated the pK_B values for the three antagonists (table 2). The pK_B values obtained for SQ29548 using each of the three agonists were identical. Similarly, those seen with GR32191 were also not significantly different. However, with L670596, the pK_B value noted with U46619 was significantly higher than those noted with the two isoprostanes; the pK_B value for the antagonist did not differ between 8-iso-PGF₂ and 8-iso-PGE₂.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We sought answers to two questions: Do the isoprostanes 8-iso-PGE₂ and 8-iso-PGF₂ have biological activities on the canine proximal colon? Are these effects exerted on a TP receptor, a unique isoprostane receptor or another prostanoid receptors?

8-Iso-PGE₂ acted as a stimulant on both the epithelium and muscularis mucosae, whereas 8-iso-PGF₂ had biological activity only on the muscularis mucosae. The stimulant responses elicited by 8-iso-PGE₂ appeared to involved EP receptors, desensitization of which revealed the presence of inhibitory TP receptors. On the smooth muscle preparation, the effects of isoprostanes appeared to be mediated by TP receptors; however, these receptors may differ subtly from the classic TP receptor.

Concentration-dependent increases in I_sc observed with 8-iso-PGE₂ were not blocked by SQ29548, a selective TP receptor antagonist that also antagonizes the putative isoprostane receptor. Thus, another prostanoid receptor was involved. The most likely candidate in this tissue would be an EP receptor based on the results of the desensitization experiments. Thus, pretreatment with the isoprostane reduced subsequent responses to added PGE₂. Pretreatment with PGE₂, on the other hand, abolished the stimulant responses to 8-iso-PGE₂ and revealed an inhibitory component, so the stimulant effects of the isoprostanes appear to involve an EP receptor although the particular subtype involved has yet to be defined (Coleman et al., 1994). Results of such experiments must, however, be interpreted cautiously.

Desensitization with PGE₂ revealed an unexpected inhibitory response to 8-iso-PGE₂. Although the inhibitory component showed a clear concentration dependence, the magnitude of the response was significantly lower than the stimulant effects noted with the same agonist (fig. 2 vs. fig. 3). That these inhibitory responses involved a TP receptor was shown both by the desensitization experiments and results with the TP receptor antagonists.

It is not easy to define pA₂ values in epithelial preparations, and given the magnitude and variability of the responses, we resorted to determining whether fixed concentrations of the antagonists (1 order of magnitude higher than the reported pK_B values) would produce significant inhibition of the responses to near-maximal concentrations of the agonists. The results clearly show that all three antagonists significantly reduce the decreases in I_sc produced by 8-iso-PGE₂. However, with U46619, the degree to which TP antagonists reduced responses attained statistical significance only with SQ29548 and L670596. The inhibition obtained with the third antagonist GR32191 was ~41%, but the variability precluded clear establishment of statistical significance at the chosen level. That these antagonistic effects were selective was shown by the persistent responses to another inhibitory prostanoid, the PGD₂ metabolite DK. Because the three antagonists were structurally different, it is likely that the effects involved a TP receptor.

The involvement of an EP receptor in the effects of 8-iso-PGE₂ is not particularly surprising because it is a stereoisomer of PGE₂ and many naturally occurring PGs are known to have actions at more than one prostanoid receptor (Coleman et al., 1994). Fukunaga et al. (1993a) noted the lower potency of 8-iso-PGE₂ in comparison with 8-iso-PGF₂ on vascular smooth muscle. They suggested that this could arise in part from the effect of 8-iso-PGE₂ on classic EP receptors, which could have opposing biological effects to those seen on the isoprostane receptor. In the case of the colonic epithelium, the opposing EP and TP effects are clearly demonstrable. Another unusual aspect worthy of mention is the suggestion that TP receptors are linked to inhibitory effects. The locus of these effects is unclear; however, PGD₂, which has marked inhibitory effects on this tissue, reverses the increases in Cl secretion seen with other stimulants (Keenan and Rangachari, 1989). The effects of isoprostanes could be exerted by a similar mechanism.

The responses of the muscularis mucosae to the isoprostanes appear to involve only TP receptors because the responses to both were antagonized by TP antagonists. The pK_B values for three structurally unrelated TP receptor antagonists (fig. 1) were used to compare the effects of the isoprostanes (8-iso-PGF₂, 8-iso-PGE₂) with a standard TP receptor agonist, U46619 (table 2). Of the three antagonists tested, both SQ29548 and GR32191 produced similar pK_B values when tested against each of the three agonists. The pK_B values obtained with the third antagonist, L670596, were significantly different. The pK_B value determined with U46619 as the agonist was significantly different from the values obtained for the two isoprostanes. The pK_B values of 8-iso-PGE₂ and 8-iso PGF₂ were not different from each other, raising the possibility that the isoprostanes exert their effects on a receptor that is homologous with but distinct from the TP receptor.

Although tempting, this interpretation must be made cautiously. There are no reported pK_B values for TP receptor antagonists at canine TP receptors. Of the three antagonists studied, both SQ29548 and GR32191 have been evaluated in numerous studies, whereas more limited information is available for L670596. Ford-Hutchison et al. (1989) reported a pA₂ value of 9.0 (8.9-9.1) for L670596 in the guinea pig trachea; however, in another study, the pK_B value has been reported to be 8.6 (Senchyna and Crankshaw, 1996). In our experiments, we obtained a pK_B value of 8.93 with U46619 and values of 8.66 and 8.62 for 8-iso-PGE₂ and 8-iso PGF₂, respectively. Due to the lack of available literature on the effects of L670596, a cautious interpretation would be that although a statistically significant difference was noted in our study, the biological significance of this observation may not be great. Had there been clear differences in the pK_B values noted with at least one of the better studied antagonists, our confidence in assessing the biological significance would have been much greater.

As mentioned above, 8-iso-PGF₂ had no effect on the epithelium, whereas both isoprostanes stimulated the muscularis mucosae strips. 8-Iso-PGE₂ and 8-iso-PGF₂ are stereoisomers of PGE₂ and PGF₂ and vary only in the cis orientation of their side chains (the side chains of PGE₂ and PGF₂ are oriented trans). In the muscularis mucosae, the trans orientation of the side chains of 8-iso-PGE₂ and 8-iso-PGF₂ appears to confer selectivity for TP receptors. 8-Iso-PGE₂, in particular, has no significant action at EP receptors in the muscularis mucosae, as demonstrated by the antagonism of 8-iso-PGE₂ by TP antagonists with pK_B values similar to those found with 8-iso-PGF₂. The TP selectivity of 8-iso-PGE₂ is also apparent in the renal vasculature of the rat, in which 8-iso-PGE₂ acts as a vasoconstrictor and PGE₂ acts as a vasodilator (Longmire et al., 1994). In fact, in all other systems investigated to date, 8-iso-PGE₂ and 8-iso-PGF₂ have only been reported to act at TP (or "TP-like") receptors. No accounts of other prostanoid receptor action exist. However, this study has shown that 8-iso-PGE₂ can act on EP receptors as well. In the present study, the EP effects predominate. The reasons for this are unclear; we could speculate that these differences could be related to receptor numbers and/or more efficient coupling to the signal transduction systems.

The isoprostanes have an undefined role in colonic abnormalities; however, other eicosanoids have been implicated in inflammatory bowel disease (Halm and Frizzel, 1990; Rachmilewitz et al., 1989). Isoprostanes could play a significant role because their concentrations are 1 to 2 orders of magnitude greater than those of cyclooxygenase-derived eicosanoids, even in normal plasma (Morrow et al., 1990a). In the context of inflammatory states, the free radicals produced could enhance the production of isoprostanes, and this in turn could affect both the epithelium and muscularis mucosae. The observation that 8-iso-PGE₂ has both stimulatory and inhibitory effects and that the latter are revealed only after desensitization of the EP receptors may be interesting in another context. Alternating diarrhea and constipation form part of the clinical picture of inflammatory bowel diseases. It is tempting to speculate that varying levels of PGE₂ and the corresponding isoprostane may contribute to this picture.

Although such arguments are speculative, some experimental data exist to suggest that the possibilities are real. A colitis-like inflammation can be induced in rabbits by the intrarectal administration of trinitrobenzenesulfonic acid. Responses to exogenous PGE₂ of both the muscularis mucosae and epithelium from such animals were attenuated (Goldhill et al., 1993a, 1993b; Percy et al., 1993). These effects appeared to involve a specific desensitization to the prostanoid because responses to vasoactive intestinal peptide were unaffected. The increased production of PGs in these tissues could have played a significant role in this regard. An exploration of the production and effects of isoprostanes in such models may prove interesting.