Uses and Limitations of Inulin and Mannitol for Monitoring Gill Permeability Changes in the Rainbow Trout

Non-electrolytes, such as mannitol and urea, have sometimes been used as probes for describing the behaviour of passive channels in epithelia like isolated frog skin (Mandel & Curran, 1972). A few reports suggest that such compounds are transferred from blood to the external medium across the gills of fish (Lam, 1969; Masoni & Payan, 1974). The possibility that such compounds might be used to monitor gill permeability was suggested by our study of the action of poly-L-lysine (PLL) on the FW trout gill. Brief application of PLL resulted in a transient but massive leak of Na+ and Cl^-. This persisted for about 20 min after the PLL was removed ; then the ion permeabilities reverted to normal. In one experiment mannitol efflux behaved similarly; it was low prior to PLL application, increased markedly after PLL but returned to the control value about 20 min after the compound was removed (Greenwald & Kirschner, 1976).

This paper explores the efflux of inulin and mannitol across the trout gill under conditions known to modify ion fluxes, some of them by changing gill permeability. The results show that the non-electrolytes’ behaviour reflects that of Na+ in a number of experiments but not in all.

Rainbow trout (Salmo gairdneri), weighing 200-400 g were obtained from a commercial hatchery near Soap Lake, Washington. The fish were maintained in running, dechlorinated tap water at 12 ± 1 °C. Those to be adapted to sea water (SW) were transferred to 50% artificial SW (Instant Ocean, Aquarium Systems Inc., Eastlake, Ohio) for a week. The salinity was then increased with solid Instant Ocean to 75% SW (1 week), then to 100% SW (1 week) before animals were used. The SW system was closed and recirculating, but mortality was low for at least several additional weeks.

The experimental animals were anaesthetized in 0·03 % MS-222 (ethyl-m-amino-benzoate, methane sulphonate) and maintained with small increments of anaesthetic as needed. The gills were irrigated by recirculating a small volume (usually 200 ml) of medium at 12 + 1 °C as described by Kerstetter, Kirschner & Rafuse (1970). Radioactive isotopes (“Na, ¹⁴C- or ³H-labelled inulin and mannitol) were dissolved in Cortlahd’s solution (Wolf, 1963) and injected intraperitoneally. About 90 min was allowed for equilibration before measurements were made. Carrier-free “Na was purchased from New England Nuclear (NEN) and from ICN Chemical and Radioisotope Division. Inulin (³H(G) and methoxy-¹⁴C) and mannitol (1-³H) were obtained from NEN.

Samples of external medium were taken at 15 or 20 min intervals for an hour and added to 15 ml scintillation fluid (Aquasol, New England Nuclear). These served to establish the rate of isotope appearance in the medium. Blood samples were taken from the caudal vein at the beginning and end of each hour and were used to convert the isotope movement into an absolute flux or plasma clearance. Sample channels ratios were observed to correct for variable quenching; corrections were rarely necessary, because all samples were processed in the same fashion (e.g. identical volumes of water in the scintillation fluid). In some experiments the simultaneous efflux of ²²Na and either inulin or mannitol was monitored. The scintillation spectrometer easily resolved ²²Na and ³H which is the preferred pair in such runs; resolution of ²² Na and ¹⁴C was less good, although this pair was used in some early experiments. In several experiments simultaneous effluxes of [³H]mannitol and [¹⁴C]inulin were measured.

In experiments with FW fish concentrations of Na+ and Cl^- in the medium were measured, Na+ by atomic absorption spectrophotometry and Cl^- by silver titration. Plasma concentrations of these ions were measured in both FW and SW fish.

A major uncertainty in experiments such as these is whether the test compound is excreted unmodified or after being degraded. The possibility was examined for inulin as follows. A FW trout was injected with [³H]inulin (0·2 mg, 27µCi dissolved in 0·5 ml Cortland’s solution). The isotope was allowed to circulate for an hour, then the external medium was changed and 100 ml FW recirculated through the gills for an hour. The medium was then lyophilized and the dried residue redissolved in 1·9 ml H₂O. This was put on a Sepharose G-50 column (0·9 x 10 cm) and eluted with EDTA-Tris buffer, pH 7·7. Eluant flow was about 0·2ml/min and 1·6ml fractions were collected. A sample of the injection solution (about 0·5 µC) was also run through the column. Fig. 1 shows the elution patterns of ³H for the injection solution (open circles, solid line) and for the lyophilized bath sample (filled circles); there is no trace of smaller fragments in the bath sample. The behaviour of mannitol was not examined; since it is a non-metabolizable carbohydrate monomer it was assumed to be stable in the animal.

Fig. 2 shows the kinetics of branchial excretion of inulin and mannitol in a SW-adapted rainbow trout. Sampling did not commence until 90 min after an intraperitoneal injection of the two compounds, by which time the blood concentration was reasonably stable. In three successive 1 h periods, separated by 30 min intervals, the rate of inulin excretion was nearly constant. Extrarenal plasma clearance in this animal was 0·29 ml kg^-1 h^-1 during the first hour and 0·31 ml kg^-1 h^-1 during the last period. Mannitol clearance increased through the experiment, but this was not seen in all animals. The initial mannitol clearance was P24 ml kg^-1 h^-1; in the final period it was 1·94 ml kg^-1 h^-1.

Both compounds are also excreted across the gills of trout adapted to fresh water (FW). Table 1 shows mean plasma clearance values for both compounds in SW- and FW-adapted animals. In both groups of fish mannitol was excreted nearly 5 times faster than inulin. The excretion of each compound occurred at nearly the same rate in the two groups. Clearances are not significantly different for inulin (P > 0.3) and marginally so for mannitol (P < 0.1). However, the general correspondence of the values is more impressive than the small differences ; the point will be discussed below.

The effect of PLL on excretion of inulin and mannitol was also studied, and the results of a typical experiment are shown in Fig. 3. The upper panel shows inulin excretion during a 60 min control period, then for two 60 min periods after a 5 min application of PLL. Excretion during the time PLL was in the medium is not shown, but it can be seen that efflux is much higher during the first 20 min after its removal than during the control period. The rates during the last 40 min of this hour and during the next hour are considerably lower. The lower panel shows cumulative changes in Na+ in the external medium. During the control hour there is net Na⁺ uptake into the fish (decrease in the external medium), while immediately after PLL there is a large net loss of Na+. The animal appears to be a steady state for the last part of this period. During the last hour there is again some net Na+ uptake, but at a rate lower than during the control period. Table 2 shows the results of four experiments in which inulin and mannitol clearances were measured simultaneously. It is interesting that even when the gill is rendered leaky by PLL it is able to discriminate between the two compounds on the basis of molecular size.

Passive effluxes of both Na+ and mannitol in FW trout were also increased in response to elevated temperature. Table 3 shows data from five experiments in which fluxes were measured at 12⁰, then at 23°, and finally at 12 °C. Sodium and mannitol clearances increased when the temperature was raised, and the increments, between 3 and 4 times the control rates, were similar. Both fell when the temperature was returned to 12°, and again, the changes show good quantitative correspondence, although neither returned to the original (12°) value.

However, not all ionic flux changes are reflected by the behaviour of these probes. Gill permeability to Na+ is known to be decreased by elevated ambient Ca²+ in both FW (Maetz, 1974; McWilliams & Potts, 1978) and SW-adapted fish (Isaia & Masoni, 1976). We have examined the action of Ca²+ on simultaneous Na+ and non-electrolyte excretion in SW-adapted animals. Four SW-adapted trout were injected with ²²Na and [³H]inulin, arid efflux of both tracers was measured after a period of isotope equilibration. The fish were initially irrigated by deionized water, then by the same medium containing 2 mm-CaCl₂. The results are shown in Table 4, and it is clear that was markedly decreased without change in inulin movement.

Another system of interest is the gill of SW-adapted fish which shows a marked diminution of when transferred from SW to FW (Motais, 1967). Na⁺ efflux can then be stimulated by adding K⁺ to the dilute external medium (Maetz, 1969). Much, if not all of the increased Na+ flux has been ascribed to voltage-dependent, passive movement (Potts & Eddy, 1973; Kirschner, Greenwald & Sanders, 1974) or to stimulation of active extrusion (Evans, Carrier & Bogan, 1974). But gill permeability has not been assessed in such experiments, and a permeability change, as well as changes in active or passive driving forces, could modify ion fluxes. A group of animals was double-labelled with [³H]mannitol and ²²Na. Effluxes were measured simultaneously in each animal, first in SW, then in FW before and after the addition of K⁺. The data, expressed as plasma clearances, are shown in Table 5. Sodium efflux was suppressed in FW and stimulated when K+ was added, but mannitol efflux was the same in all three media. In a few experiments [³H]inulin showed substantially the same behaviour; its clearance was neither reduced in FW nor stimulated by K+.

Although branchial excretion of non-electrolytes has been reported previously (Lam, 1969; Masoni & Payan, 1974) the phenomenon has received little attention. Our demonstration that the trout gill clears inulin without modifying it rules out the possibility that the results are due to an isotope artifact following degradation of the parent compound. The mechanism of permeation is of some interest. At least three possibilities suggest themselves. A membrane transport system seems unlikely because of the lack of specificity and the nature of the test compounds which are foreign to the animal. A pinocytotic system might excrete these compounds, and such a mechanism would be more compatible with the movement of a polymer like inulin than would a membrane transport system. However, pinocytosis might be expected to show little discrimination based on molecular size, whereas the clearance of mannitol is more than 5 times faster than that of inulin. The quantitative relationship between effuxes of the two compounds is similar to that between their diffusion coefficients in aqueous solution; 0.22 x 10^-5 cm²/s for inulin (Bunim, Smith & Smith, 1937), 0.61 x 10^-5 cm²/s for mannitol (Ellerton & Dunlop, 1967); the ratio is 0.36. The ratio of fluxes (from Table 1) is about 0.15; the two ratios are different, and it is apparent that the larger molecule moves more slowly relative to the smaller that can be accounted for by the diffusivities in aqueous solution. Such discrimination suggests that the gill is permeable in the classical, uncomplicated sense; that it contains pathways through which non-electrolytes can move in the direction of their chemical gradients, though a substantial permeability to a molecule as large as inulin is surprising. This, in turn, suggests that compounds such as mannitol or inulin can be used to monitor at least some types of permeability changes.

The efflux of inulin was an order of magnitude slower in our trout than has been reported for the eel (Masoni & Payan, 1974). This might be a species difference; Na⁺ and Cl^- fluxes are also nearly an order of magnitude less in the trout. An isotope artifact is also possible. [³H]Inulin was used in the eel, and the chemical form of the excreted isotope was not determined. We have used [³H]inulin in some experiments on trout and fluxes comparable to those reported for eels were sometimes observed, though these were exceptions and are not reported here. The possibility that this labelled form is unstable has been discussed previously (Beyenbach & Kirschner, 1976), and the methoxy-¹⁴C compound was used for most experiments ([³H]inulin, used in the experiments shown in Fig. 1 and Table 4, gave clearances similar to the ¹⁴C compound). Whatever the basis for this discrepancy, inulin fluxes in eel and trout are similar in an important regard; the rates are essentially the same in animals adapted to FW and to SW. And the same is true for mannitol fluxes in trout. This observation was unexpected, because the transport characteristics and ionic permeabilities of FW and SW gills differ greatly (Kirschner, 1979). It suggests that these ‘leaks’ occur across the respiratory epithelium rather than through or between the transport cells, since the extent and structure of the respiratory surfaces are probably similar in both environments. Such a proposition is also supported by the fact that inulin efflux across the skin of intact frogs is negligible, although many of its ionic permeation characteristics are nearly identical to those of FW fish (Kirschner, 1970; Greenwald, 1971). It does not seem to be adaptive for the respiratory epithelium to be leaky to non electrolytes, since this must engender a constant loss of nutrients into the medium. However, such losses are probably small ; if plasma glucose were cleared at about the same rate as mannitol a trout would lose only a few per cent of the circulating sugar per day. This may simply be the price necessary for optimizing gas exchange across an extremely thin, highly vascular surface.

It is likely that an appreciable fraction of the ‘basal’ NaCl efflux also takes place across the respiratory surfaces in FW fish. Total Na+ clearances are not dissimilar to those of mannitol, perhaps somewhat lower (cf. the control values in Table 3). It is reasonable that a pathway permeable to mannitol and inulin would also permit some loss of sodium, since the diffusion gradient for the latter is appreciable (Kerstetter et al. 1970). If so, much of the ion loss would be another consequence of specialization of the respiratory epithelium to promote rapid gas exchange. Such a possibility is supported by the observation that the Na+ efflux in FW trout is faster when the animals are active than when quiet (Randall, Baumgarten & Malyusz, 1972). Presumably, more respiratory lamellae are perfused by blood during activity, and the increase in perfused area could account for the augmented flux.

The proposition that compounds such as inulin and mannitol might be useful in estimating permeability changes that also modify ion fluxes appears to receive support in the PLL experiments. Polycations were shown earlier to induce a massive but reversible leak of NaCl across FW trout gills (Greenwald & Kirschner, 1976). If application is brief the permeability reverts to normal within 30 min. The data here show that PLL also causes a large, reversible increase in efflux of both non-electrolytes, and the magnitude and time course is similar to changes in ion clearances.

The same parallel is evident in comparing the effect of environmental temperature on Na+ and mannitol effluxes. An abrupt rise, from 12° to 23°, was accompanied by flux increases of 3.9 times for Na+ and 3.6 times for mannitol. When the temperature was returned to 12° both fluxes dropped to about half of their peak values.

It is also requisite for a permeability that its behaviour be unaltered by treatments that may modify.ion fluxes without changing gill permeability. Here, too, inulin and mannitol pass at least one test. In SW-adapted trout Na⁺ efflux is low in a dilute medium and is markedly stimulated when K+ (10 mm) is added to the medium. Flux changes in such experiments have been ascribed to changes in voltage across the gill or to modified active extrusion, with no suggestion that permeability is altered. As seen in Table 5 these solution changes had no effect on non electrolyte fluxes.

However, the parallels are not perfect. In SW-adapted trout Ca²⁺ in the external medium lowers Na+ effluxes, an observation that agrees with results on SW eels (Isaia & Masoni, 1976). But the data in Table 4 show that mannitol clearance is unaffected by Ca²⁺. The result is not hard to understand. Cation fluxes in SW teleosts are between one and three orders of magnitude faster than in FW fish, a difference seen even when the same (euryhaline) animal is tested in the two media (Motais, 1967; Kirschner, 1979). Electrical measurements show that this is due to rapid permeation through cation-selective ‘channels’ (Potts & Eddy, 1973; Kirschner et al. 1974). This pathway is probably associated with the system of interlamellar chloride cells that develop in SW (Bierther, 1970; Sardet, Pisam & Maetz, 1979). If fluxes of mannitol and inulin pass across respiratory lamellae, changes in cation channels in a different location should leave them unchanged. While the lack of correlation can be rationalized,it points to an unfortunate limitation in the use of mannitol or inulin as monitors of permeability changes. They can provide no information about the cation pathway in SW fish gills. For these another probe must be sought.