Ionic and Osmotic Influences on Prostaglandin Release from the Gill Tissue of A Marine Bivalve, Modiolus Demissus

Many types of invertebrate tissue contain (Nomura & Ogata, 1976) or synthesize (Christ & Van Dorp, 1972) prostaglandins. Recently a spectrophotometric assay, reported to be specific for prostaglandin endoperoxides, has revealed a wide distribution of these endoperoxides in coelenterates (Morse et al. 1978). Although the presence of prostaglandins in invertebrate tissue has been established, the physiological role of this compound has not been completely determined. Prostaglandins have been reported to function in the reproduction of molluscs (Morse et al. 1977). However, just as they do in mammals, prostaglandins probably have many other physiological actions in molluscs.

This study was undertaken to determine what effect various ionic and osmotic seawater solutions would have on prostaglandin synthesis in marine bivalves. The research was conducted on Modiolus demissus, a euryhaline osmoconformer (Pierce, 1970). The ability of these animals to regulate their volume is dependent on the presence of the divalent cations, calcium and magnesium (Pierce & Greenberg, 1973). Prostaglandins are released from vertebrate tissue following hyposmotic stress (Hall, O’Regan & Quigley, 1977) and are involved in the regulation of calcium flux (Greenberg et al. 1974). Therefore, the possibility existed that prostaglandins have a physiological role in the acclimation of Modiolus demissus to salinity stress. This research shows that Modiolus demissus contains and releases prostaglandins following an incubation in hyposmotic and magnesium-deficient sea water.

Modiolus demissus were collected from the salt marshes along the western shore of Assateague Island, Maryland. The animals were kept in aquaria containing aerated, 1010 m-osmol sea water (Instant Ocean, Aquarium Systems) at 15·0 °C. Hereafter, this sea water is referred to as tank sea water (TSW). The animals were used following a 2-week acclimation period.

Unless otherwise stated, the studies reported in this paper were performed on isolated gill tissues. Gill tissues were first dissected from the animal and sectioned to a length of 1·0 or 3·0 cm depending upon the experiment. To minimize mechanical stress on the tissue while transferring from one solution to another, gills were mounted on gill holders. Gill holders were made from 15 cm lengths of thin capillary tubing. One end was heat sealed, then two bends were placed in the tubing. One bed formed a 180⁰ angle 3 cm from the heat sealed end and the other bend formed a 30° angle 1 cm from the same end. The gill axis was positioned above the two bends, parallel to the gill holder. When the holder was lifted out of the sea water, the gill slid down and lodged at the 180⁰ angle. The tissue itself was never touched, blotted or squeezed with forceps. The tissue was held for 15 s to drain off excess water before being placed in an experimental solution.

Artificial sea water (ASW 1010 m-osmol; pH 8·1) was prepared according to the method of Wilkins (1972). Different cation-free sea-water solutions were also prepared according to the method of Wilkins (1972). For the prostaglandin inhibitor studies a 5 mm solution of aspirin, acetylsalicylic acid (Sigma) and a 1 mm solution of indomethacin (Sigma) were made in tank sea water (pH 7·8). To reduce pH drift tissues were transferred at the end of each hour to fresh inhibitor solutions with a pH of 7 ·8.

All organic solvents used in sample preparation for radioimmunoassay (RIA) were purchased as reagent grade solvents and then were freshly distilled before use. Test tubes were siliconized with Siliclad (Clay Adams). Labelled prostaglandins, used as internal standards to estimate recovery loss, were prostaglandin-A₁ [5,6-³H(N)] (87 Ci/mmol), prostaglandin E₁ [5,6-³H(N)] (89-5 Ci/mmol) and prostaglandin E₂ [5,6,11,12,14,15-³H(N)] (117 Ci/mmol; New England Nuclear).

Three different procedures were utilized for the preparation of samples for radio-immunoassay. One procedure was used to analyse tissue prostaglandin content and the other two procedures were used to analyse the amount of prostaglandins released into sea water. The two latter procedures were different in the presence or absence of a column chromatography separation step and an ethyl acetate extraction step. Although the three procedures differed mainly in sample preparations, they were treated as three distinct procedures and each one was validated independently. The following is a brief summary of the radioimmunoassay methods. A more complete and detailed procedure has been previously reported (Freas, 1978).

After mounting, gill sections were acclimated in ASW for 2 h prior to the start of the experiment. Experimental incubations were performed in siliconized culture tubes, each containing a total of 3 ml ASW or experimental solution. Incubations were for 60 min unless stated otherwise. At the end of the incubation the tissue was removed, blotted and then dried at 68 °C for 48 h. The tissue was then weighed and the RIA results were expressed as immunoprostaglandin E₂ released/mg dry weight of tissue. Along with experimental samples, two additional sets of tubes were processed to monitor pH and loss of tritiated internal standards.

The following is a brief summary of the sample preparation for the radioimmuno-assay. Extraction of the sample was a modification of the technique developed by Jaffee & Behrman (1974). The sea-water sample (1·8 ml) was extracted with petroleum ether and the organic phase was discarded. The sample was then extracted with an extraction solution consisting of ethyl acetate, isopropyl alcohol and enough dilute citric acid to reduce the sample pH between 3·6 and 3·8. The organic phase was then removed, dried down and run through silicic acid columns to separate the prosta-glandins by type (PGA, PGE, and PGF). The column fractions were collected and evaporated by dryness.

To convert PGE and PGA to PGB, 1 ml of 0·1 M-KOH in a 50% methanol solution was added to each tube containing the dried extract. The tubes were then flushed with nitrogen gas, capped, and incubated for 30 min at 50 °C in a water bath.

At the end of the incubation, the pH was adjusted between 3·5 and 3·8. The sample was then extracted with 4 ml of chloroform followed by the addition of 2 ml of distilled H₂O. The upper aqueous layer was then discarded and the chloroform extract was taken to dryness under a stream of air at 45 °C.

The radioimmunoassay of PGB was performed with a radioimmunoassay kit (Clinical Assays, Inc.). Kit instructions (revised 1 October 1974) were used with the following modifications. As already stated, PGE was converted to PGB and then extracted, instead of simply neutralizing the aqueous reaction mixture. For quan-tifying unknown samples serial dilutions of PGE₁ (P-L Biochemicals Inc.) were run through the columns, extracted, converted to PGB₁ along with the experimental samples, and used to make the standard curve. Points on the standard curve were run in triplicate due to the added source of variation from the extraction and chromato-graphy steps. The kit PGB standards were not used. The binding data was trans-formed to logits (Rodbard, Rayford & Ross, 1970), where logit Y = log_e[y/(1− Y)] and Y = % bound/100. The slope and intercept of the standard curve were calculated by linear regression.

In this procedure the initial ethyl acetate extraction and column chromatography procedure were omitted. The rest of the procedure remained basically the same. Aliquots, of 0·2 and 0·4 ml, were taken from each sea-water sample. Aliquots were snap frozen in acetone and dry ice. The 9-ketoprostaglandins were then converted to PGB by adding 1 ml of 0·1 M-KOH in 50% methanol to the frozen sample. The samples were checked to ensure the pH was 12·5 or above. Following conversion the samples were extracted with chloroform. The dried chloroform extract was then analysed by the radioimmunoassay procedure previously described.

Minor modifications in the previous procedure were made for analysis of tissue prostaglandin content. In this procedure gill sections were trimmed to a length of approximately 1 cm. At the end of the experimental incubation the tissue was removed by its holder, and rapidly blotted on filter paper. The gill holder was then broken and the tissue dropped into a siliconized test tube containing 3·5 ml of chilled ethyl acetate.

Along with the tissue analysis, sea-water aliquots were simultaneously analysed to determine the quantity of prostaglandins released. Aliquots of 0·2 and 0·4 ml sea water were quickly frozen and assayed by the shortened RIA procedure. The tubes containing tissue and ethyl acetate, were acidified, flushed with nitrogen, capped and frozen at − 60 °C. The frozen tissue was thawed, and sonicated. After sonication, the mixture was vortexed lightly to avoid formation of an emulsion and then centrifuged at 650 g for 20 min. The ethyl acetate was then removed and the aqueous phase was re-extracted twice with 4 ml of ethyl acetate. The extract was dried down in a water bath at 45 °C under a stream of air. The aqueous phase remaining after extraction was used for protein determination. This phase was also evaporated down to near dryness under air and following solubilization in 2 NaOH was analysed by the method of Lowry et al. (1951) using a bovine serum albumin standard.

The conversion of tissue 9-ketoprostaglandins to PGB and the RIA procedure were essentially the same as that already described for sea-water samples. The same solutions were used; however, their concentration was slightly changed to ensure that the pH was between 12·5 and 13·0 for the conversion of PGB, and between 2·7 and 3·0 for the chloroform extraction. Solubilization of the lipid extract required extra time and repeated vortexing. Just prior to the RIA, the PGB residue was dissolved in 1 ml Isogel Tris buffer, reagent B of the kit RIA, and then divided into two 0-5 ml portions for duplicate RIA analysis.

The radioimmunoassay was validated for the three sample preparation procedures. Validation for accuracy was performed by demonstrating parallelism between spiked sample curves and PGE₁–PGB₁ standard curves. The sensitivity for the sea-water RIA procedure was 2-6 pg and for the column chromatography procedure was 1-9 pg. Precision for the tissue RIA averaged 11·4%. Precision for spiked sea-water samples was calculated at the various concentrations of added PGE_X (Table 1). Specificity was demonstrated by showing parallelism between dilutions of the PGB converted sample and the PGB₁ converted PGE₁ standards. More detailed RIA methods and assay validation results have been previously reported (Freas, 1978).

Prostaglandins from whole animals were extracted and purified by the following modifications of the method of Horton (1972). Animals were shucked, weighed, placed in 50 ml of 0·05 M Tris, 0·5 M-NaCl buffer, pH 7·8 and homogenized (Sorvall, Omni-Mixer, speed). The homogenate was allowed to stand at room temperature for 15 min. Then the incubate was acidified to pH 3·5 with HC1 and extracted three times with 100 ml of ethyl acetate. Samples were frozen to remove the emulsion. After thawing the ethyl acetate was removed and evaporated to near dryness. The residue was then extracted four times with 1·4 ml of 0·1 M phosphate buffer, pH 7·8. The ethyl acetate-lipid residue was discarded and the phosphate buffer was then extracted with 3 ml of petroleum ether. The petroleum ether was then discarded and the pH of the phosphate buffer was lowered to pH 3·4. The sample was then extracted three times with 4 ml chloroform. The phosphate buffer was then discarded and the chloro-form was evaporated to dryness. The residue was dissolved in 500 ;l of distilled ethanol and spotted on thin layer chromatography (TLC) plates (Silica Gel F-254, Brinkman). The plates were developed with a solvent consisting of 90 ml chloroform, 5 ml methanol, and 5 ml acetic acid (Ramwell & Daniels, 1969). Zones from the plates were removed by vacuum and collected in a trap made of glass wool. The silica gel trapped by the glass wool was then eluted three times with 3 ml of distilled MeOH. The methanol was evaporated to dryness and the sample fractions were assayed for biological activity and for immunological activity.

Immunoactivity was assayed with and without conversion to PGB. Tritiated PGE standards were used to assess the extent of the PGE to PGB conversion with the potassium hydroxide procedure already stated. To ensure that the conversion pro-cedure did not interfere with the radioimmunoassay, tubes were treated in the following manner. Samples converted to PGB received KOH and were incubated for 30 min at 50 °C. Samples not converted were incubated for 30 min at 50 °C and then KOH was added along with the next step in the extraction procedure. The tubes contained the same reagents added at different times, thus permitting one group of tubes to contain PGB, while the other group of tubes contained the original 9-keto-prostaglandins. For each TLC fraction the sample was split and analysed by radio-immunoassay with and without the PGB conversion.

Biological activity of the thin layer chromatography fractions was also assayed. The fractions were dissolved in Tyrodes solution (Coceani et al. 1967) and assayed on the rat stomach strip according to the method of Coceani & Wolfe (1966).

To establish whether PGE₁ or PGE₂ was the major prostaglandin released, sample extracts were passed through silicic acid columns and then applied to silver nitrate impregnated TLC plates. Following precipitation of the silver and extraction of the prostaglandins according to the procedure of Willis (1970) the sample was then dried and converted to PGB, re-extracted and assayed by RIA. For all thin layer work the location of PGE zones was first performed by visualization with phosphomolybdic acid or iodine on simultaneously run TLC plates containing prostaglandin standards. To prevent loss of biological activity, plates containing samples and standards were never visualized until after the sample zones had been removed.

The data were analysed by analysis of variance. Following analysis of variance, means were compared by Student-Newman-Keuls test. For all analyses, a significance level of 0·05 was used in the determination of differences between treatment groups

To determine if Modiolus demissus gill tissue could release prostaglandins, gill sections were incubated in the presence of various combinations of substrate, 2 × 10⁻⁸ M arachidonic acid, and cofactors, 1 mm 5-hydroxytryptamine, and 1 mm glutathione. When gill tissue was incubated in the presence of substrate alone, there was a significant (P < 0·05) increase in release of immunoreactive prostaglandin A and E. When cofactors were added, immunoreactive prostaglandin A was not detected, and immunoreactive prostaglandin E concentration did not significantly change (Fig. 1). Various salinities affected the release of immunoreactive prosta-glandins from gill tissues. There was a significant (P < 0·001) increase of immuno-reactive prostaglandins released as a result of hyposmotic stress. When artificial sea water was diluted to 25% of its normal strength, the release of immunoreactive Prostaglandins was significantly greater than at other salinities tested (Fig. 2).

Ionic, as well as osmotic, experiments were conducted on gill tissue. Incubation in magnesium-deficient sea water, 1010 m-osmol, resulted in significantly (P < 0·05) greater amounts of immunoreactive prostaglandins released than did incubations in either calcium free, sodium free, potassium free or complete artificial sea water (Fig. 3). The effects of magnesium and sodium were examined in separate experiments, using a blocked statistical design to remove variance due to animals. Sodium did not have a significant (P > 0·25) effect on prostaglandin release, either when sodium was replaced by Tris (hydroxymethyl) aminomethane, or when the sodium concentration was double that found in sea water (Fig. 4A). On the other hand, magnesium-free sea water significantly increased immunoreactive prostaglandin release as compared to either normal sea water, sea water with twice the normal magnesium concentration, or magnesium-free sea water to which the normal level of magnesium had been restored (Fig. 4B).

Another experiment was designed to determine if the increased release of immuno-reactive prostaglandins due to hyposmotic sea water was due solely to the reduce magnesium level in that sea water. The release of immunoreactive prostaglandins was measured after incubating gill tissues in ASW (53 mm-Mg²⁺), 20% ASW (10·6 mm-Mg²⁺), 20% ASW plus Mg²⁺ (53 mm-Mg²⁺), and 20% ASW plus an osmotic equivalent of Na⁺ (10·6 mm-Mg²⁺). Hyposmotic sea water with normal magnesium levels, 53 mm, produced a significantly (P < 0·05) greater release of prostaglandins than isosmotic ASW with the same magnesium concentration (Fig. 5).

Two methods of tissue sample preparation were examined; one in which the tissue was homogenized in ethyl acetate, the other in which the tissue was homogenized in 0·05 M Tris, 0·5 M-NaCl buffer, followed by a 20 min incubation period before ethyl acetate extraction. A comparative study was made of both procedures, using gill sections from the same animal. The value of the immunoreactive prostaglandin concentration obtained from the aqueous buffer homogenization procedure was four times greater than the value obtained from the ethyl acetate homogenization pro-cedure. Since both sets of tubes contained the same amounts of ethyl acetate and buffer at the time of extraction, the higher concentration of prostaglandins obtained from tissues homogenized in aqueous buffer was probably due to synthesis from endogenous substrate during homogenization. An examination of prostaglandin content in extracts of aqueous homogenates from various tissues of Modiolus demissus revealed that the lower visceral mass and mantle contained significantly less (P < 0·05) immunoreactive prostaglandins than extracts of aqueous homogenates from gill tissue, upper visceral mass, the tissue surrounding the siphon, or the posterior adductor muscle (Fig. 6). In this particular experiment, the values indicate a difference in prostaglandin metabolism between various tissues, not necessarily tissue prostaglandin content. In all other experiments the ethyl acetate homogenization procedure was used and, therefore, the values obtained were direct measurements of tissue pro-staglandin concentrations.

Prostaglandin tissue content was examined together with the release of prosta-glandins into the sea water under isosmotic and hyposmotic conditions (Fig. 7). There was a significant (P < 0·05) increase in immunoreactive prostaglandins released into sea water with time for both the hyposmotic and isosmotic treated tissues. However, the tissue concentration of immunoreactive prostaglandins did not significantly change as a result of incubation time. For both tissue and sea-water samples there was a significant treatment effect due to the hyposmotic sea water for all time periods examined.

The effect of treatment of gill tissue with various prostaglandin synthetase in-hibitors prior to incubation of the tissue in hyposmotic sea water was examined. Pretreatment with 5 mm aspirin and 1 mm indomethacin significantly inhibited the release of immunoreactive prostaglandins from gill tissue subjected to hyposmotic ASW. There was not a significant difference between pretreatment with mercapto-ethanol and non-treated control value (Fig. 8). Ten ;l of each inhibitor solution was added to 200 ;l of 25% ASW and then processed through the conversion and radio-immunoassay analysis to ensure that the inhibitors themselves were not interferring with the RIA antibody binding.

Organic extracts of tissue homogenates were purified and separated into different fractions by thin layer chromatography. Zones of the thin layer chromatography plate were then analysed for biological and immunological activity (Fig. 9). Radioimmuno-assay was performed on sample aliquots that either had or had not been converted to PGB. All peaks of immunoactivity occurred on samples converted into PGB. Further-more, the major immunoactive peak occurred in a TLC zone with the same R_F value as the prostaglandin E standard. A smaller peak occurred at the TLC zone that had an R_F value similar to the PGA standard. The bioassay revealed two peaks of bio-logical activity which had the same R_F values as the PGE and PGF standards.

Thin layer chromatography was also utilized to establish whether PGE_X or PGE₂ was the major prostaglandin released. This procedure produced two TLC spots which separated the prostaglandins on the basis of saturation. The PGE₁ zone had an R_F of 0·68–0·85 and the PGE₂ zone had an R_F of 0·40–0·57. In the PGE₁ zone 35·5 ± 1·9% of the PGE₂ and 0·4 ± 0·% of the PGE₂ tritiated internal standards were recovered. In the E₂ zone 2·6 ± 0·9% of the PGE₁ and 26·7 ± 2·1% of the PGE₂ tritiated standards were recovered. Analysis of the TLC zones for radioimmunological activity revealed that 4·2% of the activity was in the PGE₂ zone and 95·8% of the activity was in the PGE₂ zone, indicating that the major prostaglandin was PGE₂.

The presence of prostaglandins in Modiolus demissus gill tissue was established by various specific purification and detection techniques. Prostaglandin-like material was isolated by solvent extraction, thin layer and chromatographic procedures and was detected by radioimmunoassay and bioassay. The radioimmunoassay had sufficient sensitivity to quantify the release of prostaglandins from isolated gill tissue. Release of immunoreactive prostaglandin E from gill tissue was detected in the presence of arachidonic acid, serotonin, and glutathione. This release or possible conversion of arachidonic acid into immunoreactive prostaglandin E was very small. To verify that the radioimmunoassay was measuring prostaglandin like substance, and not inter-ference due to the presence of the substrate and cofactors, different items of the incubation mixture were systematically deleted (Fig. 1). There was negligible production of immunoreactive PGE in the absence of gill tissue, thus eliminating the possibility of cross-reactivity from cofactors or substrate.

In mammalian tissues, glutathione (Lands, Lee & Smith, 1971) and serotonin (Sih, Takeguchi & Foss, 1970) have increased prostaglandin E production at the expense of prostaglandins F or D. Glutathione also increases PGF synthesis at the expense of PGE in insects (Destephano, Brady & Woodall, 1976). Since PGE is the putative intermediate in the biosynthetic formation of PGA, and the cofactors act on the pro-staglandin synthetase enzyme complex, an increase in PGE should also result in an increase of PGA. In Modiolus demissus gill tissue this was not the case. The cofactors did not significantly increase PGE formation; they did, however, inhibit PGA formation. Therefore, it is possible that these cofactors either act on different enzymes involved in prostaglandin synthesis or they have different effects on gill tissue than they have in mammalian tissue.

Experiments discussed thus far give evidence that gill tissue is capable of producing prostaglandins, but they reveal nothing about the biological role of prostaglandins in Modiolus demissus. Therefore, prostaglandin release was examined under physio-logical conditions. Gill tissues were incubated in sea-water solutions of various osmotic concentrations. When the osmotic pressure was reduced to 25 % of its normal level, immunoreactive prostaglandins were released. A similar osmotic effect was demon-strated with frog intestine (Vogt, 1967).

In order to detect a significant release of immunoreactive prostaglandins from gill tissue a large hyposmotic stress was required. A 25% water solution has a salinity of 8· 6‰. Pierce (1970) has shown that Modiolus demissus may survive salinities as low as 3‰ for a few days and may survive salinities from 8‰ to 24%, for 21 days or more. To ensure that the hyposmotic treatment was not simply lysing the cells of the tissue, the gills were examined microscopically at the end of the experiment to observe the activity of the beating cilia.

Two stimuli, hyposmotic stress and low magnesium levels, both produced a signifi-cant increase in prostaglandin release. The effect of hyposmotic stress was not due solely to the reduced magnesium concentration in the hyposmotic sea water. When gill tissues were placed in hyposmotic sea water with normal magnesium concentration, 53 mm, there was still an increase in release of prostaglandins. Therefore, there appeared to be two distinct stimuli that evoked prostaglandin release by gill tissue, one ionic and one osmotic.

Immunoreactive prostaglandins were not only released by gill tissue, they were also synthesized by the tissue. As seen in Fig. 7, isolated gill tissue in hyposmotic sea water released more and contained more immunoreactive prostaglandins than isolated gill tissue in isosmotic sea water. Although the release of immunoreactive prosta-glandins peaked at 60 min, most of the synthesis occurred within the first 5 min. In both hyposmotic treated and control tissues, the initial peak in tissue immunoreactive prostaglandin concentration is probably due to manipulation of the tissues. According to Flower (1974) even mild physical manipulation of tissue is sufficient to provoke synthesis of prostaglandins.

In mammalian tissue such as the spleen, lungs, adrenals, stomach, and intestine the amount of prostaglandins released when stimulated is always more than the tissue contained (Piper & Vane, 1971). In these tissues prostaglandins are believed to be synthesized and released but not stored. It appears that in Modiolus demissus tissue approximately one third of stored prostaglandins can be released.

Inhibitors, aspirin and indomethacin, inhibited the release of immunoreactive prostaglandins from the gill tissue (Fig. 8). Therefore, indomethacin appears to be an inhibitor of prostaglandin synthetase in mammals (Flower, 1974) and Modiolus demissus, but not in HaHotis rufescens (Morse et al. 1977), Plexaura homomalla (Corey, Wash-burn & Chen, 1973) or Acheta domesticus (Destephano et al. 1976). Aspirin appears to be a more universal prostaglandin synthetase inhibitor, since it is effective in mammalian systems, Modiolus and Haliotis.

Supporting evidence that the RIA was measuring prostaglandins is displayed in Fig. 9. Thin layer fractions of the prostaglandin extracts possessed immunological activity only if they were first incubated with KOH at 50 °C for 30 min, a process that converted the 9-ketoprostaglandins into PGB. Furthermore, this immunological activity had the same thin layer chromatography R_F factor as PGE standards. There also was a smaller peak of immunological activity which ran on the thin layer chromato-graphy plate with the PGA standard. The immunological activity in this area could have been due to endogenous PGA, but most of it was probably formed as a result of dehydration of PGE during the extraction process (Lee et al. 1967). In this particular experiment, tritiated internal standards were not run to check the conversion of PGE into PGA.

Each thin layer chromatography fraction was analysed for biological activity. Two peaks of biological activity were found. One peak was associated with the PGE standards, and the other was associated with the PGF standard. Since the first antibody used in the radioimmunoassay had no cross reactivity with PGF, the radioimmuno-assay could not detect the presence of PGF. However, both PGE and PGF are biologically active and capable of contracting the rat fundus stomach strip. Therefore, the bioassay further established the presence of PGE and furnished some evidence for the existence of PGF in the prostaglandin extract.

Demonstrating the existence, quantifying the concentration, and determining the stimuli that alter synthesis were the first steps in establishing the physiological role for prostaglandins in marine bivalves. These data reveal that Modiolus demissus contains and releases prostaglandins following hyposmotic stress. Therefore, the pos-sibility exists that prostaglandins may play a role in acclimation of the animal to this stress.