Introduction

Some studies have speculated that the coelom of cartilaginous fish (Elasmobranchii) may be a part of these animals’ osmoregulatory organs along with the gills, kidneys and rectal gland (Bles 1897; Smith 1929; Hartman et al. 1941; Donnelly et al. 2019). These speculations were based on studies reporting the composition of the coelomic fluid in various cartilaginous fish species (Smith 1929; Hartman et al. 1941; Rodnan et al. 1962; Bernard et al. 1966; Maren 1967; Murdaugh and Robin 1967; Thorson et al. 1973; Donnelly et al. 2019). However, all of these studies were done in wild caught animals in the field and not under controlled laboratory conditions. Thus, the physiological implications for coelomic fluid’s possible role in osmoregulation are speculative at best.

A compelling argument that the coelomic cavity and the fluid it produces are part of the osmoregulatory organs was the observation that this fluid had an easy exit from the coelomic cavity to the external environment through paired abdominal pores (also called coelomic pores) near the cloaca (Bridge 1879; Turner 1879; Bles 1897, 1898; Daniel 1934; Weichert 1959; Romer and Parsons 1977; Donnelly et al. 2019). Although Bridge (1879) postulated that the abdominal pores were vestigial, Smith (1929) reported observing fluid exiting the abdominal pores in various species of skates and sharks supporting his contention that the coelom was part of the osmoregulatory system of elasmobranch fish. Confirmatory studies by Abel et al. (1994) showed that marker molecules injected into the cardio-coelomic cavity of various elasmobranch species exited through the abdominal pores under laboratory conditions. Furthermore, the patency of the abdominal pores between the coelomic cavity and the surrounding seawater was shown by Donnelly et al. (2019) who easily inserted catheters through them and withdrew coelomic fluid for diagnostic health analyses.

Supportive evidence for this pathway of coelomic fluid to the external environment through abdominal pores also came from phylogenetically related groups (e.g., class Teleostei), to the cartilaginous fishes (class Chondricthyes) which also have abdominal pores (Bles 1897, 1898). These observations have been extrapolated to elasmobranch fish to include speculations of the functions of the renal system, reproduction, pH regulation, and as a lymphatic fluid reservoir (Schweigger-Seidel and Dogiel 1867; Bridge 1879; Bles 1897, 1898; George et al. 1982; Dobbie 1988). For example, studies in teleost fish in which a marker fluid, India ink, and Gram-negative bacteria were injected into the coelom directly through the abdominal wall showed these substances were excreted through the abdominal pores (Dobbie 1988).

We undertook the current study to test the hypothesis that the composition of coelomic fluid in a representative euryhaline elasmobranch fish species changes with alterations in the composition of the animals’ external environment and is thus a part of the animal's osmoregulatory system. We used a widely euryhaline elasmobranch, the Atlantic stingray, Hypanus sabina. This species belongs to the family, Dasyatidae, which has worldwide distribution where they predominately inhabit the mouths of rivers, estuaries and habitats of reduced salinity (Smith 1931; Gunter 1945; Bigelow and Schroeder 1953; de Vlaming and Sage 1973; Taniuchi 1979; Piermarini and Evans 1998). In these nearshore environments, the stingrays are subjected to extreme variations in external salinity caused by the daily tides and influx of fresh water from rivers. In North America, during the late autumn and winter these elasmobranch fish migrate into deeper water offshore (de Vlaming and Sage 1973; Schwartz and Dahlberg 1978) where they adapt to full-strength seawater for many months. On the other hand, populations of Atlantic stingrays have been reported to complete their entire life cycle in fresh water (Johnson and Snelson 1996; Piermarini and Evans 1998).

The stingrays used in the present study were transferred under laboratory conditions from full-strength seawater to seawater diluted by approximately 50%. We compared the concentrations (mass/volume) and the total (absolute) amounts of hydrogen ions (pH), as well as osmolytes: Na+, Cl-, K+, urea and protein in plasma, coelomic fluid and the surrounding seawater under both environmental conditions.

Materials and methods

Sexually mature (wingspan ~45 cm, 0.5–0.8 kg) Atlantic stingrays, Hypanus sabina (Schwartz and Dahlberg 1978), of both sexes were captured between May and December in close proximity to the Charleston Harbor, South Carolina, USA. Animals were captured by trammel net and transferred to a 15,000 l holding tank. The tank water, aerated and constantly filtered, was drawn from Charleston Harbor at high tide which was representative of that in which the animals were captured (886.0 mOsm/l H2O). Stingrays were fed shrimp twice per week ad libitum. Sixteen animals (eight males and eight females) were held for 30 days in this holding tank at ambient temperatures (21–23°C) and 12-h light/dark cycle.

To compare the effects of decreasing external environmental osmolality on the composition of coelomic fluid and plasma, eight sexually mature stingrays (four males and four females) were transferred to a 5,500 l tank filled with harbor water diluted with distilled water (v/v) to a final osmolality of 407.3 mOsm/l H2O. We refer to this as diluted seawater. The animals were subjected to diluted seawater conditions for 24 h. The remaining 8 animals (4 males and 4 females) from the holding tank were transferred to a separate tank (same size as the diluted seawater tank) filled with the same seawater as the holding tank. Conditions for temperature, light/dark cycle, aeration and filtration were the same in the 2 smaller tanks (full-strength and diluted seawater). After acclimation to the respective tank conditions, all animals in both salinity environments were prepared in vivo for coelomic fluid and blood collection. Individual animals were transferred to a container, ~50 l, containing the anesthetic MS-222 (3-aminobenzoic acid ethyl ester, 0.5 g/l, Sigma Chemical Co., St. Louis, MO, USA) in aerated water from the respective tanks from which they were adapted (Janech et al. 2006).

For coelomic fluid collection, animals were placed ventral side up on a slant board (approximately 45°) with the rostrum and gills elevated. We had determined from previous post-mortem dissections, that in this anatomical position, all the coelomic fluid pooled at the lower end of the peritoneal cavity on each side of the cloaca. In the distal end of the peritoneal cavity lies the internal opening of the abdominal pore, thus allowing access to the coelomic fluid from the ventral body surface (Donnelly et al. 2019). Aerated water from the respective tank was used to irrigate the gills during the coelomic fluid collection that took approximately 1–2 min. To harvest coelomic fluid, a polyethylene tubing (PE 320) was inserted into each of the two abdominal pores and the coelomic fluid gently aspirated into 10 ml syringes and transferred to a 50 ml conical Falcon tube. The total volume of coelomic fluid for each animal was measured using a 5 ml disposable pipette with increments of 0.1ml. The coelomic fluid was then immediately put on ice, followed by centrifugation at 2,000 x g for 10 min (Dynas II, Clay Adams, Inc. Parsippany, NJ, USA). The supernatant was used either fresh or frozen at -80°C for later analyses. Subsequently, a blood sample was taken from the lateral cutaneous tail vein (Janech et al. 2006). A small (~2 mm) superficial incision was made at the collection site and a PE 10 catheter inserted. Blood (0.5–1.0 ml) was withdrawn into a 3 ml heparinized syringe. The collected blood was rapidly placed into a cooled micro-centrifuge tube and centrifuged at 5,000 rpm for 5 min. The plasma was aspirated and stored at −80 °C until analysis. An antiseptic solution was swabbed on both the abdominal pore and the tail incision following collection.

Collection of both coelomic fluid and blood took less than 4 min, after which the animal was placed back in the respective tank water. Complete recovery, indicated by active, unassisted swimming, took approximately 3 min.

Chemical analyses. Analyses of coelomic fluid, plasma and tank water were achieved by the following: osmolality by vapor pressure osmometry (Wescor, Inc., Logan, Utah, USA); Na+, K+ by flame photometry (Instrumentation Laboratory, Lexington, MA) and Cl- by chloride titration (Radiometer, Copenhagen, The Netherlands); protein analysis by the Bradford technique (Pierce Chemical Co., Dallas, TX, USA), pH by electrode (Mettler-Toledo) and urea by quantitative colorimetric determination (Kit #535-A, Sigma Aldrich Chemical Co., St. Louis, MO, USA). The absolute amount of each osmolyte was calculated from the concentrations per unit volume in proportion to the actual volume collected for full-strength seawater and diluted seawater except for pH. The absolute amount of hydrogen ions in coelomic fluid was determined from a conversion table for pH (Fiorica 1968) expressed as nEq.

Statistical analyses. Basic descriptive statistical methods were used to calculate the mean and standard deviation of the concentration of each component in plasma, coelomic fluid and tank water, and the absolute amount of each component in coelomic fluid. Comparisons of the mean concentration values for each component across plasma, coelomic fluid and tank water, first in full-strength seawater, then in diluted seawater, were calculated using ANOVA analyses (experiment-wise alpha = 0.05) with post hoc correction for multiple comparisons made using the Bonferroni method. Independent t-test analyses (2-sided, alpha = 0.05) were used to compare the mean absolute amount of each component in coelomic fluid between full-strength seawater and diluted seawater.

Results

Coelomic fluid volume was 17.9 (± 5.4) ml in animals exposed to full-strength seawater and 28.1 (± 3.4) ml in animals exposed to diluted seawater. The measured amount of coelomic fluid in animals in diluted seawater included the residual amount of coelomic fluid each animal had from the full-strength seawater environment in addition to that they produced in the diluted seawater environment. The 56.9% increase in the volume of collected coelomic fluid in diluted seawater-acclimated animals was statistically significantly different from that in the animals in full-strength seawater.

Concentration of osmolytes (amount/per unit volume) for animals in full-strength seawater (Table 1). There were no significant differences in osmolality in the two body compartments (plasma and coelomic fluid) compared to each other and to that of the surrounding seawater. Under the same external environmental conditions, sodium, chloride and potassium values showed that the concentrations of these ions were significantly less in both plasma and coelomic fluid than in seawater, but statistically the same in plasma and coelomic fluid. As expected, there was no detectable protein or urea in tank water. However, there was 42 times less protein in the coelomic fluid than in the plasma, and urea constituted 38.4% and 36.4% of the coelomic fluid and plasma osmolality, respectively. The acidity, expressed as logarithmic units (pH), was slightly but significantly more basic in tank water (pH 7.7) than in plasma (7.3). However, the coelomic fluid was significantly more acidic at pH of 5.5 than the plasma and tank water.

Table 1 Concentrations of analyzed solutes of plasma, coelomic fluid and tank water for stingrays adapted to full strength (FSW) and diluted (DSW) seawater
Full size table

Changes in osmolyte concentrations when animals were transferred to diluted seawater (Table 1). When the tank water was diluted by 54.1% to produce a diluted external environment, the osmolality was 407.3 mOsm/kg. Concurrently, the plasma and coelomic fluid osmolality also dropped significantly from that of the tank water, but only 18.3% and 14.0%, respectively. These differences were not significant between the two body chambers. The pattern for changes in total osmolality was also recorded for sodium ions, which decreased in tank water from 370.4 mmol/l in full-strength seawater to 160.6 mmol/l in diluted seawater. Sodium ion concentrations decreased by 12.9% and 17.1% in plasma and coelomic fluid, respectively, during this shift. Chloride like sodium showed significant declines in concentrations in tank water by 57.6%. In both plasma and coelomic fluid, the decreases were 19.9% and 0.56%, respectively. As with sodium, there was significantly less chloride in the plasma of diluted seawater animals than those in full-strength seawater. There was statistically significant difference in Cl− concentration between plasma and coelomic fluid in the diluted seawater group. Potassium decreased significantly in the tank water in diluted seawater by 58.8% and in the plasma of animals (32.6%), but not in the coelomic fluid. These changes in plasma vs. coelomic fluid concentration were statistically significant. As in the full-strength seawater conditions, protein did not appear in the tank water. The concentration of protein decreased significantly in plasma (22.9%) from full-strength seawater to diluted seawater, but in coelomic fluid only 0.56% in animals during this shift. The protein concentrations were more than 35 times greater in plasma than in coelomic fluid. Urea was not detected in the tank water. This osmolyte remained elevated in the coelomic fluid and plasma of animals when they were acclimated to the diluted seawater conditions. There was no significant difference between the concentrations of this osmolyte in the plasma and coelomic fluid in the two environmental conditions (full-strength and diluted seawater). The pH was identical in diluted seawater conditions to that of full-strength seawater in both tank water (pH 7.7) and plasma (pH 7.3). However, the pH of coelomic fluid (pH 5.5) changed 0.3 log units to become slightly more basic in diluted seawater but still acidic at pH of 5.8. This change was not statistically significant.

Absolute (total) amounts of osmolytes in coelomic fluid (Table 2). The total amount of osmolytes in the 17.9 (± 5.4) ml of coelomic fluid collected in full-strength seawater conditions was 15.94 (± 0.75) mmol. Upon dilution of the tank water by 54.1%, the amount of total solutes in the coelomic fluid increased by 29.8% to 20.70 (± 0.74) mmol. The total amount of sodium increased from 4.96 (± 1.45) mmol in full-strength seawater to 6.46 (± 0.52) mmol in full-strength seawater. This 30.2% increase was the smallest increase of all the individually measured osmolytes in coelomic fluid after the environmental shift. Nevertheless, the increase was statistically significant. The total amount of chloride ions in coelomic fluid in the diluted seawater environment also increased by 56.2%, being nearly double the increase of the Na+ ions. This value was significant. Nevertheless, the values of chloride ions in full-strength seawater (4.79 ± 0.81 mmol) and diluted seawater (7.48 ± 0.33 mmol) were closer to the total amount of Na+ than to any other osmolyte.

Table 2 Absolute amounts of coelomic fluid solutes for stingrays adapted to full strength (FSW) and diluted (DSW) seawater
Full size table

Potassium ions, although the smallest actual amount of any of the measured osmolytes, had the greatest percentage increase of 68.8% going from 0.0938 (± 0.025) to 0.1583 (± 0.055) mmol upon environmental dilution shift, respectively. This increase was significant. The total amount of protein was small, being 6.68 (± 2.72) mg in full-strength seawater and increasing by 38.2% to 9.23 (± 2.11) mg in diluted seawater. This difference was not significant.

Urea, which makes up about 1/3 of the animals' plasma and coelomic fluid osmolality (Table 1), increased in total amount from 6.08 (± 1.04) mg to 9.42 (± 0.81) mg from full-strength seawater to diluted seawater, respectively. This accounted for a 54.9% increase, which was significant.

The acidity/alkalinity, measured as pH, was expressed as nEq to take into account that pH is expressed in a logarithmic scale. The amount of hydrogen ions in coelomic fluid in the diluted seawater conditions was less than that in full-strength seawater, but not significantly so. The values for full-strength seawater were 64.68 (± 50.23) and for diluted seawater 52.62 (± 32.91) nEq.

Discussion

The results of the current study show that under controlled laboratory conditions, the volume and composition of the coelomic fluid in the Atlantic stingray change rapidly in response to changes in the animals’ external salinity. Three notable results from these analyses were: (1) the coelomic fluid volume increased ~50% when the external osmolality decreased by approximately the same value, (2) the total (absolute) amount of all osmolytes (except protein and H+) in coelomic fluid increased significantly from ~30% to ~70% under diluted seawater conditions, (3) the pH of the coelomic fluid remained significantly acidic (7.5–7.7) which was greater than that of the plasma and tank water under both salinity environments.

We used the current protocol of changing this species from full-strength seawater to 50% seawater because: (1) previous studies indicated that this dilution in salinity was well within the normal physiological range of this species (de Vlaming and Sage 1973; Piermarini and Evans 1998) and (2) to compare our data with an identical study design on the same species, which measured changes in renal function and plasma osmolytes in the face of 50% seawater dilution (Janech et al. 2006).

The anatomical origin of the coelomic fluid in Hypanus sabina or any elasmobranch species is unknown, but would appear to be a combination of fluid surrounding the heart (pericardial) and that of the pleuroperitoneal cavity. In related elasmobranch fish species, studies have shown that the two cavities, pericardial and pleuroperitoneal, are anatomically connected via a pericardiaoperitoneal canal through which fluid travels from one compartment to the other (Smith 1929; Weichert 1959; Satchell 1971; Shabetai et al. 1985). We termed the fluid collected from the peritoneal cavity in the present study, “coelomic fluid” to include all possible origins anatomically.

Although in some poikilothermic vertebrates, the coelomic cavity is part of the reproductive, renal and lymphatic systems (Schweigger-Seidel and Dogiel 1867; Ewart 1876; Bridge 1879; Bles 1897, 1898; George et al. 1982; Kardong 2012), in elasmobranch fish it is not part of any of these organ systems.

Some published data on coelomic fluid composition are present in other species of elasmobranch fish from full-strength seawater environments (Smith 1929; Hartman et al. 1941; Rodnan et al. 1962; Bernard et al. 1966; Maren 1967; Murdaugh and Robin 1967; Thorson et al. 1973; Donnelly et al. 2019). Comparison of the data on coelomic fluid composition in the Atlantic stingray adapted to full-strength seawater conditions in the present study with the other species shows that our values are consistent with those reports. An exception is the report of protein and chloride ions. Smith (1929) and Donnely et al. (2019) reported coelomic fluid protein values in various species of elasmobranch fish, which were approximately 6 to 36 times greater than our measured values. Thorson et al (1973) reported a 43% greater concentration of chloride in coelomic fluid of the bull shark and Donnelly et al. (2019) reported 20% higher chloride values in the spotted eagle ray than that in the present study. Comparisons of concentrations of coelomic fluid components with those of plasma collected simultaneously in the same individuals are found in just a few early studies of elasmobranch fish by Smith (1929, 1931), Hartman et al. (1941), Rodnan et al. (1962) and Thorson et al. (1973). These comparisons of compositional differences in the two body cavities allow for speculation on the possible function of coelomic fluid. However, all previous studies were performed on animals in the wild at different geographical locations. Nevertheless, our values for coelomic fluid comparisons to plasma in the present study are roughly comparable to those from the five studies cited above, with the exception that Thorson et al. (1973) reported a 32% greater amount of potassium in plasma than reported here. Interestingly, in all published studies, urea made up a significant proportion of coelomic fluid and was not significantly different from its concentration in the plasma (Smith 1929; Hartman et al. 1941; Rodnan et al. 1962; Thorson et al. 1973).

Coelomic fluid analyses have been reported in two studies of euryhaline elasmobranch fish taken from freshwater environments in the wild. Smith (1931) showed in several species that the values of chloride in the coelomic fluid and blood were lower than in marine elasmobranch fish. Analysis of the environmental water in which the animals were caught was not reported in that study, so the extent of their freshwater adaptation cannot be compared with the results of the present study. Thorson et al (1973) compared individuals of the same species of shark, Carcharinus leucas, taken off the coast of Florida with those in the freshwater, Lake Nicaragua. That study also showed a reduction in serum and coelomic fluid Na+, Cl- and urea concentrations in animals from the freshwater lake. However, Thorson et al (1973) acknowledged that there was not a history of those animals' time in the lake, since they routinely swim from the marine environment of the Caribbean Sea up the San Juan River to Lake Nicaragua and back to the Caribbean Sea.

There was a relatively high concentration of urea in Atlantic stingray coelomic fluid in full-strength seawater (approximately 1/3 of total osmolality). These animals, like all marine elasmobranch fish, conserve urea through a renal mechanism in which plasma urea filtered through the renal glomeruli is subsequently reabsorbed (90–99%) back into the blood across the renal tubules to help maintain the hyper-osmolality of the animals' internal milieu (Shuttleworth 1988). However even in full-strength seawater conditions identical to those of the current study, renal function studies on Hypanus sabina showed that significant amounts of urea were also excreted in the urine (Janech et al. 2006).

Notably in all previous studies of coelomic fluid composition, the pH was significantly acidic ranging from 5.4 to 6.1 (Smith 1929; Robin 1962; Maren 1967; Murdaugh and Robin 1967; Thorson et al. 1973; Donnely et al. 2019) which we report in the present study as well. Experiments by Robin and colleagues (Robin 1962; Murdaugh and Robin 1967) showed that the pH of dogfish coelomic fluid did not change when the animal was intravenously administered hydrogen ion (sulfuric acid), but did become more alkaline when bicarbonate was administered via that route. Although the ability of the coelomic fluid to buffer administered bicarbonate was limited, these investigators noted that urine pH did not change, suggesting a limited or even nonexistent role for the kidney in this particular aspect of acid/base balance (Robin 1962; Murdaugh and Robin 1967). These reports led Heisler (1988) to suggest that the coelomic cavity may serve as an additional site to maintain acid/base homeostasis and that acidic coelomic fluid would be released to the environment via abdominal pores.

The current study is the only one to our knowledge in which the total (absolute) volume of coelomic fluid in elasmobranch fish was measured in animals in any environmental condition. By knowing the volume of coelomic fluid in animals in both environmental conditions (full-strength and diluted seawater) and the concentration of each measured component in that volume, we were able to calculate the absolute (total) amount of each solute in coelomic fluid. The results show (Table 2) that the absolute amounts of total solutes being secreted into the coelomic fluid during diluted seawater conditions increased by 29.8% simultaneous with fluid volume increase by 56.9%. The absolute amounts of sodium, chloride, potassium and urea increased significantly in diluted seawater conditions ranging from about 30% to almost 70%. It is interesting to note that as shown in Table 2, increases in absolute amounts of Cl- and urea are similar to the increase in volume, while the increase in Na+ was much lower compared to the increase in fluid volume. This may suggest active reabsorption of sodium from the coelomic fluid into the blood stream or other extracelular fluid space.

A previous study (Janech et al. 2006), using an identical study design of salinity dilution on the same species, showed that not only does urine volume increase significantly during external environmental dilution, but the absolute amounts of solutes excreted also increases significantly. This reflects the osmoregulatory strategy that elasmobranch fish use to remain hypertonic to the external environment even in the face of diluted external seawater. As the external environmental osmolality is decreased significantly, the animals not only increase the excretion of water which has osmotically come across the gills, but also decrease the amount of osmolytes in the internal milieu through renal secretion which thus decreases the osmotic gradient across the animals' body (Janech et al. 2006). We showed in the present study that the coelomic fluid follows this general pattern of increased volume output and increased solute excretion when animals are shifted to dilute environmental conditions, but to a lesser degree than the kidney. We currently cannot explain the differences in the individual solute concentrations or absolute amounts that differ from that of plasma and seawater. Clearly, the data in the present study suggest that there is some selectivity in the secretion of these solutes into coelomic fluid and that they are not just passively diluted by increases in water across the coelomic wall during environmental dilution.

To our knowledge, there are no reports on the microstructure of the epithelium lining the coelom of elasmobranch fish which might shed light on the mechanisms of coelomic fluid production. By comparison, published reports on the mammalian peritoneal epithelium (mesothelium) shows that it is very “leaky” from a structural perspective, since the paracellular pathway is not tightly sealed at the zonulae occuludentes (tight junctions) (Simionescu and Simionescu 1977), thus allowing large tracer molecules to easily pass through this transepithelial route (Cotran and Karnovsky 1968). By comparison, we would predict that the epithelium lining the coelomic cavity of elasmobranch fish would not be like that of the mammal, but be both structurally (multiple strands composing the zonulae occuludentes) and electrically (measured by transepithelial resistance) “tight” (Claude and Goodenough 1973) accounting for the concentration of hydrogen ions up to 100 times greater in the coelomic fluid than in the plasma as has been reported in every elasmobranch fish species thus far examined (Smith 1929; Robin 1962; Maren 1967; Murdaugh and Robin 1967; Thorson et al. 1973; Donnely et al. 2019). Other mechanisms of cellular secretion particularly in regard to water might include membrane aquaporins (Agre and Nielsen 1996).

Summary and conclusion. Taken together, the data of the current study show that the coelomic fluid of Atlantic stingrays, Hypanus sabina, rapidly changes in concert with that of the external environment. This provides suggestive evidence that the coelom of elasmobranch fish is an additional site for osmotic and pH regulation of their internal milieu. Additional studies are clearly needed to elucidate the sites and mechanisms of coelomic fluid production and regulation.