Effects of Epinephrine on Branchial Non-Electrolyte Permeability in Rainbow Trout

Recent investigations on non-electrolyte permeation across various epithelia suggest that large polar solutes (e.g. mannitol, sucrose or inulin) pass through the tight junctions whereas small hydrophilic solutes (e.g. water or urea) and lipophilic molecules (e.g. ethanol or butanol) pass through both the apical and the baso-lateral membranes (Smulders & Wright, 1971 ; Smulders, Tormey & Wright, 1972; Wright & Pietras, 1974; Wright, Smulders & Tormey, 1972). Some study has been made of the effect of hormones on epithelial permeability. Antidiuretic hormone (ADH) has been observed to increase the permeability of toad bladder to small hydrophilic solutes, with a smaller augmentation of permeability to lipophilic substances, and to have no effect at all on permeability to large hydrophilic molecules (Pietras & Wright, 1975). This suggests that ADH increases the fluidity of plasma membranes and has no effect on the paracellular pathway.

In the present investigation we have investigated the effect and mode of action of epinephrine on the permeability of trout gill epithelium. Fish gill epithelium exhibits a low permeability to water (Evans, 1969; Motais & Isaia, 1972; Motais, Isaia, Rankin & Maetz, 1969) and to urea (Payan & Maetz, 1970), but has a relatively high permeability to very large molecules such as inulin (Lam, 1969; Masoni & Payan, 1974). Epinephrine increases branchial transfer of non-electrolytes. In particular, it increases the uptake of oxygen, a highly lipophilic solute. It also increases diffusional water fluxes across the gills of sea-water- and freshwater-adapted Mugil capito(Pic, Mayer-Gostan & Maetz, 1974), and freshwater-adapted trout (Haywood, Isaia & Maetz, 1977). Branchial influx and efflux of urea are also enhanced in freshwater-adapted trout (Bergman, Olson & Fromm, 1974; Haywood et al. 1977). Although this neurohormone produces alterations in the branchial irrigation which may entail changes in the surface available for diffusion (Bergman et al. 1974) it has been shown to produce a true change of permeability to water in the branchial epithelium, via β-adrenoceptors (Haywood et al. 1977).

Rainbow trout (Salmo gairdneri), obtained from a local trout farm, were kept in tanks of running aerated fresh water (Ca concentration: 1·5 M<SC>M</SC>), at 15 ± 3 °C. Both sexes, weighing from 130 to 200 g, were used.

The isolated perfused head preparation was carried out after Haywood et al. (1977) with similar experimental protocol. Labelled test substances were added to the perfusion fluid and effluxes were measured for two 30 min periods with the external medium in a closed circuit of 240 ml. These two periods were separated by a 5 min interval allowing for addition of epinephrine (L-epinephrine, Merck: 10⁻⁶ M), rinsing off the external medium and replacing it by fresh medium. The β-blocking agent propranolol (ICI: 10⁻⁶ M) was present during both experimental periods in one series of investigations. Throughout these experiments, the perfusion rate was monitored by a drop counter and by weighing collected perfusion fluid in each period.

Table 1 gives details of the various substances tested. In most of the experiments, THO was used simultaneously with another test-substance, which had either a ¹⁴C or a fluorescent label. Concentrations of ³H and ¹⁴C in the external and internal media were measured in an Intertechnique SL 40 scintillation counter with automatic quench correction. Fluorescence was measured in an Aminco SPF 125 spectrophotometer in samples made alkaline (pH 9·0) by the addition of NaOH.

During the early part of these investigations, control experiments were performed to verify that the gill is indeed permeable to inulin and dextran 20 and that the ¹⁴C radioactivity appearing in the external medium corresponded effectively to inulin and dextran 20 and not to small breakdown products, in particular to D-fructose or D-glucose (mol. wt: 180). Diaflo R ultrafiltration was used to check for inulin and dextran 20 appearing into the external medium. Inulin was also checked by thin-layer chromatography. Aminco membranes (UM 05 for inulin and UM 10 for dextran 20) were employed to filter both internal and external media at the end of a 30 min efflux period. These filters, permeable to molecules of less than 500 mol. wt (UM 05) or 10000 mol. wt (UM 10) was used under 7 kg/cm² and 4 kg/cm² nitrogen pressure respectively. Kieselgel (Merck F 254), 250 μm thick, was used for thin layer twodimensional chromatography of the external medium. Two solvents were used successively: ethanol/water (30/70), and ethylacetate/isopropanol/water (65/20/15). The first produced a parallel migration of D-fructose and inulin to R_f 0 ·5, while the second produced migration of D-fructose along to R_f 0·3,. The medium was filtered through a millipore filter (0 ·22 μ) and vacuum concentrated to about one-hundredth of its original volume at room temperature. The sugars were revealed by spraying on anisaldehyde reagent after oven-heating the plate to 100 °C for 5 min (Stahl & Kalten-bach, 1964). The radioactivities associated with the inulin and fructose spots were compared to that of a 100 μ laliquot of the concentrated external medium identical to the aliquot placed on the silica gel.

Permeability coefficients were computed from the fluxes of the various test substances. Successive 5 min samples were taken from the external medium and the corresponding regression line was used to calculate the appearance rate of radioactivity. For the most permeant solutes (butanol and THO) the mean radioactivities of the afferent and efferent fluids was used to calculate the specific radioactivity. With the present high perfusion rates, the difference between afferent and efferent fluid radioactivities amounted to no more than 10% of the input radioactivity. Fluxes (mol/100 g h) were calculated by dividing the rate of appearance of the label (fluorescent or radioactive) in the external medium by the specific activity of the test-substance in the perfusion fluid (Haywood et al. 1977). Permeability coefficients in cm/s were obtained by dividing the flux in mol/100 g s by the gill surface in cm² per 100 g body weight (i.e. 400 cm²) and by the concentration of the test-substance in the perfusion fluid in mol/cm³.

Permeability coefficients measured in the epinephrine-free preparation, are listed in decreasing order in Table 2. Three groups may be recognized. Butanol and water, with almost identical coefficients, head the list. A second group of substances, about 7 −12 times less permeant, includes small hydrophilic solutes, such as urea, and hydrophilic macromolecules with an 18-fold variation in molar radii, but with similar permeability coefficients. Finally, mannitol, about 60 −70 times less permeant than water or butanol, has a permeability coefficient significantly lower than that of urea. There is no significant difference between dextran 3, dextran 20 and inulin values.

Inulin itself, and not breakdown products, penetrates through the epithelium. Ultrafiltration showed that in the inulin used in the perfusion fluid there was 5 ± 1 % (n = 3) of metabolite with molecular weights less than 500, compared with 6 ±2% (n = 3) in the external medium. After thin-layer chromatography, 86 ±6% (n = 3) of the external medium radioactivity corresponds to the R_f of inulin.

For the dextrans, molecular weight distributions examined by gel filtration (samples and data by Pharmacia) indicated no overlap in the molecular weights but we have found similar permeability coefficients of the two populations. Concerning dextran 20 the ultrafiltration showed 6% ±2% (n = 3) of metabolite with molecular weights smaller than 10000 in the perfusion fluid compared to 4 ·0% ± 0 ·6% (n = 3) of metabolite in the external medium. The permeability of the gills to dextran 20 was measured by means of a fluorimetric technique and using ¹⁴C-dextran 20. The values are respectively: P₀ = 1 ·2±0 ·7 10⁻⁶ cm/s (n = 8) and P₀= 1 ·1±0 ·7 10⁻⁶ cm/s (n = 4) in the absence of epinephrine and P_ep = 1 ·7±0 ·6 10⁻⁶cm/s (n = 8) and P_ep = 2 ·0± 1 ·0 10⁻⁶cm/s (n = 4) when 10⁻⁶ M of epinephrine were added to the perfusion fluid. The values obtained by the two experimental techniques were not significantly different. These results are depicted in Tables 2 and 3. Clearly dextrans of very high molecular weight cross the gill.

Table 3 summarizes the effects of epinephrine. The permeabilities of the test substances measured after the addition of epinephrine, P_ep, are given. Epinephrine significantly and uniformly increased the permeabilities to butanol, water and urea but had no effect on the permeability to inulin, dextrans 3 and 20 and mannitol. Simultaneous measurement showed that water transfer increased, so the lack of effect on mannitol, inulin and dextrans is not due to tissue insensitivity. For instance, in the experiments measuring the water and mannitol permeabilities with and without epinephrine, the ratio P_ep/P₀ for water was 1 ·71 ±0 ·17.

The ratio p^s/p^THO calculated for the periods under epinephrine stimulation significantly increased in the experiments with butanol (P < 0 ·01 using the paired ratio differences). Such changes did not occur with the other test molecules, including mannitol. In the butanol experiments, increased water permeability (P_ep/P₀ = 1 ·35 ±0 ·10), although significant, is definitely lower than that in all the other experiments (P_ep/P₀ = 2 ·01 ±0 ·15, n = 51) (P < 0 ·01). Butanol may interact with membrane lipids to affect P^THO.

The perfusion flow rate, 113 ±7 ·4 ml/100 g h under control conditions, increased by 10% (+ 14 ·5 ± 6 ·5 ml/100 g h) after addition of epinephrine.

Haywood et al. (1977) showed that the increased branchial water permeability produced by epinephrine involved a β-adrenergic receptor. Table 4 summarizes data from experiments in which propranolol was present throughout the flux periods. The effectiveness of this blocker is verified by the considerably decreased perfusion rate observed after addition of epinephrine, from 122 ± 12 · 2 ml/100 g h to 48 ± 6 ·7 ml/ 100 g h (n = 12; P < 0 · 01). Such an effect is to be expected from stimulation of α-adrenoceptors, causing vasoconstriction in this preparation (Payan & Girard, 1977).

The mean permeabilities observed from the various test-substances during the control period with propranolol, are not statistically different from the comparable values shown in Table 2. In the presence of the β-blocker, epinephrine no longer lugments effluxes of the various test-substances, and indeed permeability to urea decreases significantly.

The perfused head preparation offers two methods for measuring the efflux rate of test-substances. First, clearance (difference in concentration of the molecule in the afferent and efferent perfusion fluid) and the rate of perfusion may be used, providing the difference between afferent and efferent concentrations exceeds 3 % of the input. To increase the difference, it is necessary to reduce the perfusion rate. Steen & Stray-Pedersen (1975) used perfusion rates attaining only 2 ·5 −10% of the cardiac output normally observed in eels. At such low perfusion rates, the effective exchange area may be drastically reduced. For these reasons, the clearance method, most adequate for measuring the permeability of the most permeant solutes (lipophilic and small hydrophilic molecules) is almost certainly inadequate for evaluating the permeabilities of larger hydrophilic solutes (sugars, dextran or inulin).

The method used during the present investigation, however, consists of the direct measurement of the appearance rate of the solutes in the external medium, and thus avoids these difficulties. In addition, the present perfusion rate of about 2 ml/ 100 g min, is of a physiological order.

The interference of unstirred layers within the epithelium cannot be overlooked, and the relative thickness of the epithelia and of their corresponding connective tissue is a good indicator of the importance of unstirred layers in impeding solute movement. Thus for the rabbit gall bladder and the amphibian choroid plexus and urinary bladder values of 200 −900 μm have been considered (Smulders & Wright, 1971 ; Wright & Pietras, 1974). From morphological studies of the gill (Conte, 1964; Hughes & Morgan, 1973; Morgan & Tovell, 1973; Steen, 1971), unstirred layers are of minor importance in view of the absence or extreme thinness of the connective tissue (< 1μm), of the respiratory epithelium (< 5 μm), or even of the mitochondria-rich cells (< 20 μm). Moreover, the importance of the unstirred layers is inversely related to the magnitude of the fluxes crossing the epithelium (Smulders et al. 1972). The gill is relatively impermeable to lipophilic and small hydrophilic solutes for which corrections are important in other epithelia. An important factor, however, is rapid equilibration of the branchial blood supply. Under various circumstances, changes in perfusion rate, pressure or vascular resistance may cause blood to be shunted from the lamellae and channelled elsewhere (Girard & Payan, 1976, 1977; Haywood et al. 1977; Hughes, 1972; Hughes & Morgan, 1973; Payan & Girard, 1977; Steen & Kruysse, 1964; Vogel et al. 1973, 1974). Whether these changes correspond to all-or-none effects will be discussed later in relation to the effects of epinephrine.

Except for the human skin, the gills are among the most impermeable epithelia, especially with respect to lipophilic and small hydrophilic solutes. There is good agreement for water and urea permeabilities in the gills of the eel and trout under the influence of epinephrine. The value found for water in the perfused head of the freshwater eel is however somewhat lower than that reported previously for in vivo studies at 19 or 15 °C, respectively 18 and 25 × 10⁻⁶ cm s⁻¹.

To characterize the gill permeability to lipophilic solutes, a small range of solutes would have to be examined to obtain the slope of the regression line between log P and log K₀₁₁ or to study the effect of increasing the chain length of the solute on P. We have pooled our data for butanol with the data given by Steen & Stray-Pedersen (1975) for ethanol and antipyrine and find that the slope of the line relating P and K₀₁₁(about 0 ·29) indicates that the gills are much less hydrophilic than the urinary bladder of the toad and about as hydrophilic as frog choroid plexus or rabbit gall bladder (Wright & Pietras, 1974). (With high branchial permeability molecules, the clearance technique is available.) This suggestion is substantiated by the increase of P by 1 ·7, observed when butanol and ethanol permeabilities are compared. An increase in chain length of two methylene groups (butanediol hexanediol) produces a P increase of 1 ·9 in the gall bladder and of 4 ·5 in the urinary bladder (Wright & Pietras, 1974).

The branchial permeability to small polar solutes (water, urea) resembles permeability of other epithelia and single-cell membranes insofar as it is obviously greater than is to be predicted on the basis of partition coefficients. It may be noted that the p^urea/p^THO ratio is much greater for gills than for the toad urinary bladder or for red cells, and is similar to that observed in the choroid plexus. It may be that pores, membrane carriers or a highly ordered membrane lipid configuration explain the high permeability observed for the small polar solutes.

For large polar solutes, permeation is thought to occur via a few large pores in the tight junction as shown for the rabbit gall bladder (Smulders & Wright, 1971). Small polar solutes exhibit lower apparent activation energies than larger solutes. Moreover, the ratio of inulin to sucrose branchial permeabilities is similar to that of their free-solution diffusion coefficients and the apparent activation energy for sucrose is not distinguishable from that reported for diffusion in aqueous solution. The branchial penetration of large polar solutes has paradoxical properties, in that the permeability to mannitol is of an order of magnitude smaller than in the rather leaky gall bladder and choroid plexus and identical to that reported for the urinary bladder which is considered to be a tight membrane. On the other hand the trout gill is twice as permeable to inulin as is the rabbit gall bladder. Permeability to mannitol is, in fact, ten times smaller than for inulin, whereas the free-diffusion coefficient of mannitol is about five times higher. Two permeation routes for large polar solutes across the branchial epithelium must be considered : a paracellular route for mannitol and possibly sucrose, and an intracellular ‘bulk-transport’ route for macromolecules like inulin and dextrans. Histochemical and light microscopic autoradiography reveal that various organic solutes including mannitol and inulin are accumulated in the mitochondria-rich cells (Masoni & Garcia-Romeu, 1972; Masoni & Isaia, 1973). The tubular system on the baso-lateral membrane is probably the accumulation site of these substances. Whether this tubular system makes contact with the external medium by means of clefts or infoldings of the apical membrane (Bierther, 1970; Philpott & Copeland, 1963; Shirai & Utida, 1970) or by a process of vesicular transport and exocytosis (Isaia & Masoni, 1976; Shirai, 1972) remains to be elucidated. Maetz & Pic (1977) discussed the possible role of vesicular transport in salt excretion by the gill in sea-water. The presence of numerous microfilaments and microtubules in the apex of the mitochondria-rich cells suggests a role of cell motility in the permeation processes of these cells. Addition of colchicine (an inhibitor of microtubular polymerization) to the external medium rapidly blocks salt secretion in the sea-water-adapted mullet. Water permeability, however, remains unchanged (Maetz & Pic, 1977). It would be of interest to study permeability to large polar solutes in animals treated with colchicine, and to examine why permeability to mannitol is smaller than that for inulin or dextran. Obviously, with non-specific bulk transport, the vesicles should handle mannitol and dextran indiscriminately. It may be that the walls of the vesicles are relatively permeable to the smallest of these molecules and these would leak out during transport across the cell. If the apical cell membrane were impermeable to mannitol, its only exit route would be the tight junction.

Both neurohypophysial hormones and epinephrine increase water permeability in frog skin (Maetz, 1968; Rajerison et al. 1972) by processes involving the adenyl cyclase system. Epinephrine’s effect is mediated through a β-adrenoceptor and the mechanism appears similar for increased gill permeability to water, since the response is blocked by propranolol (Pic et al. 1974). As a β-receptor is similarly involved in decreased branchial resistance to blood flow, variations in regional blood distribution may indirectly change the flux, either by more efficiently mixing the internal medium or by increasing available exchange surface via lamellar recruitment. By ruling out changes in surface area, Haywood et al. (1977) showed that epinephrine induces a true increase in gill permeability to water, although it is appreciated that in most efflux experiments, changes in the available surface often cannot be completely ruled out. Present results show that epinephrine increases the apparent permeability to butanol, water and urea in a rather non-specific manner, by about 100%, suggesting a possible increase in the available surface. This is not supported, however, by the absence of effect on permeability to inulin, dextrans and mannitol. Antidiuretic hormone increases the permeability of amphibian bladder to lipophilic solutes in a rather nonspecific manner, by about 30% (Pietras & Wright, 1975). Transfer of small hydrophilic solutes is greatly increased, however, by a factor of 3 ·5 for urea, and more than 10 for water. Permeability to larger hydrophilic solutes such as mannitol remains unchanged (Pietras & Wright, 1975). ADH possibly has no effect on the paracellular pathway but increases membrane lipid fluidity, which is the preferred pathway for lipophilic and small hydrophilic molecules. This may also apply to epinephrine action on the branchial epithelium, although with some differences. In the urinary bladder the abnormally high permeability of small polar solutes, as compared with lipophilic solutes, is considerably accentuated by ADH. This is not observed in the gill under epinephrine. The hypothetical paracellular route of mannitol permeation remains insensitive to epinephrine in the gill and to ADH in the urinary bladder.

Our observations with propranolol confirm observations (Haywood et al. 1977) that β-adrenergic receptors mediate increases in permeability under epinephrine, irrespective of whether small hydrophilic or lipophilic solutes are considered. This again suggests a common mechanism such as increased membrane lipid mobility.

In simultaneous studies, epinephrine has been shown to increase branchial per meability more for butanol than for water. If transfer of lipophilic solutes is preferentially stimulated, it is pertinent to examine the actions of epinephrine on the permeability to the lipophilic respiratory gases. To date increased oxygen uptake due to epinephrine has been interpreted purely in terms of vascular effects. Recently P. Payan, J. P. Girard, C. Peyraud and M. Peyraud-Waitzenegger (unpublished observations) have suggested that epinephrine actually increased membrane permeability to oxygen in perfused trout heads.

The authors thank the Royal Society, for financial support to G. P. Haywood by means of a European Post-Doctoral Fellowship.