Introduction

Top Abstract Introduction Methods Results Discussion References

PAF (1-alkyl-2-acetyl-sn-glycero-3-phosphocholine) is a phospholipid inflammatory mediator that is synthesized by mesenchymal cells such as endothelial cells, macrophages, mast cells, neutrophils, eosinophils and possibly fibroblasts (Arnoux et al., 1980; Braquet et al., 1987; Casmussi et al., 1983; Chao and Olson, 1993; Lee et al., 1984; Oda et al., 1985). It is not stored preformed in these cells but rather is synthesized and released after activation of phospholipase A₂. The production of this bioactive substance is tightly regulated by a key enzyme, GPC acetyl transferase, and its degradation is regulated by 1-alkyl-2-acetyl-GPC-acetylhydrolase. Although its name is derived from its ability to activate platelets from certain species, many other proinflammatory actions have been attributed to it, including vasoconstriction (Hsueh et al., 1988), enhanced adhesion of neutrophils to endothelial cells (Kubes et al., 1990a, 1990b) and activation of the inducible prostaglandin synthase gene (Bazan et al., 1994). PAF is proposed to have a role in the inflammation of inflammatory bowel disease, ischemic colitis and necrotizing enterocolitis (Ferraris et al., 1993; Resnick et al., 1995; Sobhani et al., 1992; Travis and Jewell, 1994). Unlike lyso-PAF, PAF also has the ability to stimulate water and electrolyte secretion by the intestine (Bern et al., 1989; Hanglow et al., 1989; MacNaughton and Gall, 1991).

PAF stimulation of intestinal Cl secretion is due predominantly to the release of leukotrienes and prostaglandins in the intestine (Bern et al., 1989; Hanglow et al., 1989; MacNaughton and Gall, 1991), although there is some controversy regarding this point (Buckley and Hoult, 1989). In most laboratories in which this has been studied, the PAF-stimulated Cl secretory responses of rat intestine and simultaneously measured colonic PGE₂ and PGI₂ (measured as 6-keto-PGF₁) production are inhibited 75% to 90% by cyclooxygenase inhibitors but not by inhibitors of 5-lipoxygenase, histamine or serotonin (Bern et al., 1989; MacNaughton and Gall, 1991).

The intestinal secretory response to PAF is often inconsistent (Bern et al., 1989; Hanglow et al., 1989). The dose response may vary considerably, and on occasion, the response may be entirely absent. To accomplish the primary goal of this study (i.e., understanding whether PAF may play a role in the Cl secretion stimulated by immune cells), it was necessary to first understand the cause of this erratic secretory response to exogenous PAF. We postulated that it may be mediated by endogenous desensitization or by release of a PAF inhibitor by the intestine. After determining how to obtain consistent responses to exogenous PAF, specific stimuli were used to determine whether PAF is involved in the intestinal Cl transport response to immune system-mediated secretion. The following stimuli were used as models of immune cell-mediated secretions: chemotactic peptide FMLP, which is specific for phagocytes (Marasco et al., 1984); anti-rat IgE, which is a potent but not entirely specific stimulus of mast cells (Ishizaka and Ishizaka, 1984); and H₂O₂, which is released by phagocytes during the respiratory burst and is a potent stimulant of prostaglandin and PAF secretion by mesenchymal cells (Karayalcin et al., 1990; Lewis et al., 1986).

	Methods

Top Abstract Introduction Methods Results Discussion References

Transport studies. Male Sprague-Dawley rats weighing 350 to 480 g and New Zealand White rabbits weighing 2.5 to 3 kg were killed by cervical dislocation or intravenously administered pentobarbital sodium (60 mg/kg), respectively. The colon was removed, opened longitudinally and washed of contents with oxygenated Ringer's solution. The entire rat colon or a 5-cm segment of distal rabbit colon was stripped of its outer muscle layers through a combination of blunt and sharp dissection as previously described (Bern et al., 1989). Segments of colon were mounted in Lucite Ussing chambers with a 0.5-cm² aperture, with each side incubated with 10 ml of Ringer's solution maintained at 37°C and pH 7.4 when gassed with 95% O₂/5% CO₂. The Ringer's solution contained (in mM): Na⁺ 140, K⁺ 5.2, Ca⁺⁺ 1.2, Mg⁺⁺ 1.2, Cl 119.8, HCO₃ 25, H₂PO₄ 0.4, HPO₄ 2.4 and glucose 10. The bathing solutions were connected via agar bridges to calomel electrodes to measure the electrical PD across the tissue. The tissues were short-circuited to zero PD with an automatic voltage clamp (World Precision Instruments, Sarasota, FL) using Ag-AgCl electrodes connected to the bathing solution via agar bridges. Tissues were continuously short-circuited except for 5-sec intervals every 15 to 30 min when the open-circuit PD was read. Pilot studies showed no region-specific differences in rat colon in its response to PAF stimulation.

Studies of PAF desensitization or inhibition. To investigate the influence of hypoxia on intestinal desensitization to PAF or the possible release of endogenous PAF inhibitors, tissues were rendered hypoxic for 5- to 20-min intervals by bubbling argon in the Ringer's solution before mounting. In these hypoxia experiments, great care was taken to keep the time interval between the preparation and the mounting procedure as short and consistent as possible. Untreated tissues or washed hypoxic tissue was used as controls in experiments to study the response to exogenous PAF, to PAF antagonists and to theophylline, BK or VIP.

Washing procedure. In an attempt to remove PAF or any putative endogenous lipid PAF inhibitor(s), Ussing-chambered intestine was incubated with a wash solution consisting of 0.01% faf BSA. This wash solution was added to both sides of the tissue and drained after 10 min. Both sides were then rinsed twice with 50 ml of prewarmed, oxygenated Ringer's solution. During the rinsing procedure, which required 15 min, the fluid of both bathing solutions was carefully removed through continuous suction and replaced at the same time to maintain a constant fluid level with minimal hydrostatic or mechanical disturbances of the tissue. The washed and unwashed tissues were then allowed an additional incubation period of 30 min before the addition of agonists.

Extraction of lipids from wash solution. Wash fluid from the serosal compartment of 22 Ussing-chambered tissues, rendered hypoxic for 5 to 20 min, was extracted according to the Bligh and Deyer method using four volumes of chloroform/methanol/concentrated HCl (100:200:1) (Bligh and Dyer, 1959). The nonaqueous phase was then subjected to two extractions with 1:3 volumes chloroform/0.1 N HCl, and the extract was dried in a vacuum-centrifuge (Jouan Inc., Winchester, VA). The samples were dissolved in 2 ml of ethanol (95%) and chromatographed on a C18 Silica column (Sep-Pak, Waters, Milford, MA). Methanol (70%) was used to elute a prostaglandin-containing fraction from the column. The column was then eluted twice with 2 ml of 100% methanol (PAF-containing fractions 1 and 2). The methanol fractions were then dried as described above and redissolved in DMSO (100%). In mock experiments, 90% of ³H-PGE₂ was extracted by the 70% methanol elution, and 62% of ¹⁴C-PAF was eluted from the column with 100% methanol (data not shown). Both PAF-containing fractions were used separately for experiments in which ~10% of these fractions were added back to the serosal bathing solution of a single intestine mounted in 6 to 10 chambers.

Prostaglandin measurements. After the tissues were mounted in Ussing chambers, 1-ml samples were removed from the serosal bathing solution at varying intervals, put into plastic vials, gassed with argon and stored at 70°C until they were analyzed for prostanoids PGE₂ and 6-keto-PGF₁ by direct radioimmunoassay on 100- to 300-µl samples. Internal standards were assayed in the presence of the various inhibitors and agonists that were used. Prostaglandin production rate was calculated from concentrations measured at 15-min time periods. All values were expressed as ng/15 min/cm².

PAF measurements. After the colon was mounted in Ussing chambers bathed in Ringer's solution containing 0.05% faf BSA, 1-ml samples of the sersosal bathing solution were obtained from control and agonist-treated tissues before and at 5- to 15-min intervals after the addition of the agonist. Preagonist and postagonist samples were extracted with chloroform/methanol before radioimmunoassay as recommended by the manufacturer. PAF concentrations were expressed as pg/100 µl of serosal bathing solution.

Materials. All chemicals were obtained from Sigma Chemical Co. (St. Louis, MO). Sheep anti-rat IgE was purchased from ICN Immunobiologicals (Lisle, IL). Radioimmunoassay reagents for PGE₂ and 6-keto-PGF₁ assays were obtained from Advanced Magnetics, Inc. (Cambridge, MA). The ¹²⁵I radioimmunoassay kit for PAF was purchased from New England Nuclear Research Products (Boston, MA). FMLP was dissolved in DMSO, divided into aliquots and stored at 20°C until use. Lyophilized anti-IgE sera were reconstituted with sterile water and stored at 4°C. PAF (L--lysophosphatidylcholine,-acetyl--o-hexadecyl) was obtained as dry powder, dissolved in 2.5% faf BSA in water and stored in aliquots at 20°C for single-time use. INDO was dissolved as a stock solution in DMSO. PAF antagonists WEB 2086, BN 52021, Ro 24-0238 and Ro 24-4736 were dissolved in DMSO; these were graciously supplied by Hoffmann-LaRoche, Inc. (Nutley, NJ).

Statistics. The n in these studies refers to the number of animals. Experiments were performed on paired tissue from the same animal. In some instances, 4 to 12 Ussing chambers were run simultaneously to compare a control tissue with other treatments on tissues from the same animal. In these instances of multiple comparisons with rat colon, it was not possible to obtain more than four to eight pieces of colon from a single animal; therefore, a control tissue was always paired with one to seven treatments and the data were normalized to percentage of control response. Statistical significances of differences were determined by paired t test when single paired comparisons were made. When multiple comparisons were made, the significance of differences were determined by parametric or nonparametric analysis of variance.

	Results

Top Abstract Introduction Methods Results Discussion References

Variability of rat colonic I_sc response to PAF. The additions of exogenous PAF to the serosal bathing solution of rat colon in the Ussing chamber elicits a biphasic I_sc response (Cl secretion), with a first peak at 3 to 5 min and a second peak at 15 to 30 min (fig. 1). This response has been noted previously, although the cause of its biphasic nature has not been determined (Bern et al., 1989). This secretory response requires a high concentration of PAF (10⁵ M) for a maximal effect. Furthermore, the magnitude of the response is variable; occasionally, there is no response to PAF. Because such erratic responses to PAF may be due to the release of endogenous PAF with resulting desensitization or to the presence of endogenous inhibitors of PAF, rat colon was washed with 0.01% faf BSA, as described above, and the dose response to PAF was determined in washed and unwashed tissue. As shown in figure 2, the ED₅₀ value for the I_sc response to PAF in washed colon (500 nM) was decreased by 2 logs compared with unwashed colon (50 µM).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   The maximum Isc to exogenous 105 M PAF in control (minimally hypoxic) tissues, hypoxic (10 min bubbling with argon) tissues and washed hypoxic (10 min bubbling with argon followed by faf BSA wash) tissues are shown. *, Significance of difference (P < .05) between washed hypoxic tissue and hypoxic unwashed tissue.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   The PAF-induced Isc of rat colon stripped of muscularis propria and mounted in Ussing chambers is biphasic with early (peak 1) and later (peak 2) increases in Isc. Washing rat colon with 0.01% fafBSA (-) shifts the dose response to exogenous PAF ~2 two logs to the left of that observed in unwashed (---) colon.

6

7

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Maximum Isc to 105 M PAF of control intestine (minimal hypoxia) or intestine rendered hypoxic by bubbling in argon for 5, 10, 15, and 20 min before mounting in the Ussing chamber. The peak 1 () and peak 2 () Isc values are shown as well as the subsequent response of each tissue to stimulation with 5 mM theophylline added to both mucosal and serosal bathing solutions. n is number of animals.

View larger version (37K): [in this window] [in a new window]   Fig. 4.   The Isc response of hypoxic colon (5 min) is shown as a percent of control (nonhypoxic). The response to PAF (two peaks) was significantly inhibited by prior hypoxia, whereas the single peak response to BK and VIP was not. *, P < .05, significantly different from control.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   The effect of addition of chloroform/methanol extract of wash solution back to washed rat colon (left arrow) on subsequent Isc to 105 M PAF (right arrow) is demonstrated. Both fraction 1 and fraction 2 had a significant inhibitory effect on the subsequent Isc response to PAF. Note also the lack of any stimulatory effect of the chloroform/methanol extract on basal Isc.

Inhibition of prostaglandin synthesis by PAF inhibitors. Previous studies have shown that 90% of the secretory effect of PAF on rat intestine is due to prostaglandin release by PAF (Guerrant et al., 1994; Guthrie et al., 1991; Hanglow et al., 1989). The remainder of the secretory response could be due to PAF stimulation of PAF receptors on colonic epithelial cells or perhaps on PAF receptors on other inflammatory cells in the lamina propria, causing release of other Cl secretagogues, such as histamine, serotonin or H₂O₂. To dissect the role of PAF in mast cell (anaphylactic)-mediated or phagocyte (inflammatory)-mediated secretion, stimulants of mast cells (anti-IgE) and phagocytes (the chemotactic peptide FMLP) were studied in conjunction with PAF antagonists.

1

5

6

View this table: [in this window] [in a new window]   TABLE 1 PGE2 and prostacyclin (measured as 6-keto-PGF1) production by rat colon in response to H2O2 (5 × 104 M) and effect of PAF antagonists and 1 µM INDO PG production rate (ng/15 min/cm2) in the serosal solution of Ussing-chambered rat colon measured 30 min before and 30 min after the addition of H2O2.

Effect of PAF antagonists on colonic Cl secretion. To choose the proper concentration of antagonist for studies of immune cell agonists, the response (I_sc) of unwashed but minimally hypoxic rat colon to PAF 10⁵ M was determined in the presence of varying concentrations of the PAF antagonists (fig. 6). A relative ED₅₀ value of 0.05 µM was calculated for Ro 24-4736, followed by Ro 24-0238 (0.3 µM), WEB 2086 (10 µM) and BN 52021 (50 µM). At 10⁴ M WEB 2086 and BN 52021 and at 10⁷ M Ro 24-4736, a 100-fold higher concentration of PAF (10³ to 10⁵ M) was necessary to obtain the same maximal I_sc response in colons without an antagonist present (data not shown). The inhibitory effect of compound Ro 24-0238 (10⁵ M) could not be completely overcome by increasing concentrations of PAF (10³ M), suggesting a partially noncompetitive inhibitory action of this compound, a finding that is in keeping with its cyclooxygenase inhibitory activity. In subsequent studies, a concentration of antagonists greater than these calculated ED₅₀ values was used.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Response curves to 105 M PAF in the presence of varying concentrations of four different PAF antagonists. Maximum inhibition of Isc (normalized to percentage) of the first () and second peak () is shown for each antagonist. The calculated, relative ED50 value for each antagonist is shown in the bottom right of each graph. These do not represent true ED50 values because the data are normalized and the degree of inhibition of peak 1 is different from peak 2.

6

4

5

View this table: [in this window] [in a new window]   TABLE 2 Inhibition of rat or rabbit colonic Cl secretory response [Isc (µA/cm2)] induced by H2O2, anti-IgE or FMLP (rabbit) with PAF antagonists, INDO alone or combinations of PAF antagonist, INDO and diphenhydramine

Release of PAF by immune cell agonists. To obtain direct evidence that PAF is released by H₂O₂ and anti-IgE in rat colon, as suggested by the antagonist studies above, PAF concentrations were measured in the serosal bathing solution before and after stimulation with H₂O₂ or anti-IgE. As shown in figure 7, both immune cell agonists significantly increased PAF production by rat colon.

View larger version (49K): [in this window] [in a new window]   Fig. 7.   PAF production by rat colon in response to 500 µM H2O2 and 120 µg/ml anti-IgE. The measurements were made on pooled samples of serosal bathing solution collected 5 to 15 min after agonist addition in the Ussing chamber. *, P < .05 compared with control.

	Discussion

Top Abstract Introduction Methods Results Discussion References

PAF is one of the many inflammatory mediators synthesized and released by activated immune cells that are located in a subepithelial location in the lamina propria of intestine. PAF is known to stimulate intestinal Cl secretion by rat small and large intestine, although the ED₅₀ value for this response (20 µM) is several logs higher than is necessary to stimulate chemotaxis and degranulation of phagocytes in vitro (0.01-100 nM) (Bern et al., 1989; Hanglow et al., 1989; MacNaughton and Gall, 1991). Furthermore, as observed by our laboratory and others, the I_sc response of intestinal tissues to exogenous PAF may be inconsistent (Bern et al., 1989; Buckley and Hoult, 1989; Hanglow et al., 1989; MacNaughton and Gall, 1991). This suggests prior desensitization due to endogenous PAF release or the presence of an endogenous PAF antagonist. There is evidence of endogenous PAF inhibitors coexisting in situ with PAF in different organ systems (Miwa et al., 1987; Nakayama et al., 1987). Among the inhibitors discovered are polyunsaturated free fatty acids, such as oleic acid, as well as lyso-glycero-phospholipids and sphingophospholipids, which have chemical structures similar to PAF (Miwa et al., 1987; Nakayama et al., 1987; Nunez et al., 1990; Smiley et al., 1991; Tokumura et al., 1989). PAF and PAF-like substances bind avidly to albumin (Ludwig et al., 1985); therefore, we washed Ussing-chambered rat colon with fafBSA and demonstrated a significant decrease of 2 logs in the ED₅₀ value of the I_sc response to exogenous PAF. Because PAF is released by the hypoxic intestine (Caplan et al., 1990; Kubes et al., 1990a), the role of hypoxia in generating the putative PAF antagonist or antagonists was investigated. A hypoxia time-dependent reduction in the I_sc response was observed to exogenous PAF but not to BK or VIP. To further explore the hypothesis that PAF or an endogenous inhibitor was present and down-regulating the PAF response, a chloroform/methanol/HCl extract of the wash fluid was added back to the serosal bathing solution. The extract had little effect on the I_sc but recreated the blunted response to exogenous PAF. Thus, we hypothesize that PAF or a lipid inhibitor of PAF is released from the rat large intestine in response to hypoxia and that this putative substance desensitizes PAF receptors. The fact that the lipid extract failed to stimulate the I_sc when added back to the Ussing chamber is a point against PAF-induced desensitization as the mechanisms of the blunted response. However, we have not determined whether the inhibitor substance is PAF or one of the reported inhibitor substances with a similar structure. A similar desensitization of PAF receptors has been demonstrated in guinea pig ileal smooth muscle in response to inflammation (Jeanneton et al., 1995).

In this study, BN 52021, an alkaloid derived from the Gingko tree, and three synthetic compounds, WEB 2086 (a triazolodiazepine) (Casals-Stenzel et al., 1987; Dent et al., 1989), Ro 24-0238 (a pentadienylamide) (Guthrie et al., 1991) and Ro 24-4736 (thienodiazepine) (Crowley et al., 1991), were tested for their ability to block PAF-induced intestinal Cl secretion stimulated by diverse agonists. WEB 2086 (Dent et al., 1989) and Ro 24-4736 (Crowley et al., 1991) are specific PAF antagonists devoid of cyclooxygenase inhibitory activity (table 1). In dose-response studies to exogenous PAF, all four antagonists showed significant inhibition of the I_sc response. Furthermore, these PAF antagonists inhibited the Cl secretion induced by H₂O₂ and anti-IgE but not the Cl secretion induced by FMLP.

H₂O₂, which may be created in vivo from the respiratory burst of primed phagocytes (Nathan, 1987), is capable of stimulating PAF and eicosanoid production by endothelial cells (Lewis et al., 1986). We have shown here that H₂O₂ releases PAF from rat colon, and previously our laboratory demonstrated that H₂O₂ stimulates colonic prostaglandin production (Karayalcin et al., 1990). Both PAF and H₂O₂ stimulate intestinal Cl secretion by releasing prostaglandins, which stimulate the enteric nervous system (Bern et al., 1989; Hanglow et al., 1989; Karayalcin et al., 1990; MacNaughton and Gall, 1991). The experiments reported in table 2 suggest that both PAF and eicosanoid production is stimulated by H₂O₂ and that both contribute to the I_sc response induced by H₂O₂. PGE₂ and PGI₂ production was not affected by the selective PAF antagonists, yet these antagonists were capable of inhibiting the I_sc response to H₂O₂, indicating that PAF accounts for part of the response.

We have also shown here that PAF is released by the colon stimulated with anti-IgE and that it contributes to the I_sc response of anti-IgE. Mast cell degranulation may result in Cl secretion through release of mediators such as histamine, serotonin, adenosine and eicosanoids. Because exogenous PAF releases PGE₂ and PGI₂ in rat colon and 50% of the anti IgE-mediated I_sc response is inhibited by INDO (table 2; Bern et al., 1989), the lack of an additive effect of PAF antagonists to that of INDO suggests that most of the I_sc response to anti-IgE is due to release of prostaglandins by PAF. PAF may therefore mediate mast cell-induced Cl secretion through stimulation of prostaglandin production by intermediate targets in the lamina propria such as other immune cells or mesenchymal cells. The combination of a PAF antagonist, cyclooxygenase inhibitor and H₁ antagonist was additive, suggesting that all three agonists play some role in mast cell-mediated secretion. Others have presented evidence for PAF involvement in the electrolyte secretion stimulated during mast cell-mediated intestinal anaphylaxis, and our findings are consistent with those data (MacNaughton et al., 1992).

The chemotactic peptide FMLP was used as a phagocyte stimulant in rabbit colon. It has been shown previously that almost 90% of the FMLP-mediated I_sc response of rabbit colon was inhibited by INDO (Bern et al., 1989). Of the four PAF inhibitors used, only compound Ro 24-0238 inhibited the response to FMLP, and at the concentrations used in our study, this compound was found to significantly inhibit prostaglandin synthesis by intestinal tissue. Therefore, we conclude that PAF does not contribute to the FMLP-stimulated phagocyte Cl secretory response in rabbit colon.

In summary, these experiments provide evidence that hypoxia releases PAF or, more likely, an endogenous lipophilic inhibitor of PAF action that desensitizes the in vitro rat colon. PAF also appears to contribute to some, but not all, immune cell-mediated intestinal water and electrolyte secretion. The Cl secretory responses to exogenous H₂O₂ and anti-IgE are partially PAF dependent, but FMLP-stimulated secretion is not. Also of interest is the recent report implicating PAF in intestinal prostaglandin production and secretory response to cholera toxin (Guerrant et al., 1994). PAF has Cl secretory effects in part by releasing eicosanoids from lamina propria mesenchymal cells, and it may also directly stimulate epithelial cells and enteric nerves (Berschneider and Powell, 1992; Willard, 1992).