Ion and Water Balance in the Ixodid Tick Dermacentor Andersoni: III. Influence of Monovalent Ions and Osmotic Pressure on Salivary Secretion

During feeding of the adult female tick, Dermacentor andersoni, salivation accounts for approximately 75 % of the total water excreted from the large blood meal (Kaufman & Phillips, 1973 a). Since the salivary gland is capable of secreting in vitro (and so in the absence of an external hydrostatic pressure), this organ apparently operates by a secretory rather than a filtration-resorption mechanism of fluid production. Isolated glands do not secrete in the presence of fresh haemolymph collected from salivating ticks (Kaufman & Phillips, 19736); this suggests that fluid secretion is unlikely to be controlled by a diuretic hormone analogous to those which activate the Malpighian tubules of some insects (Maddrell, 1964 a, b, c;Berridge, 1966; Pilcher, 1970) unless such a factor is very rapidly destroyed. On the other hand, in vitro glands are sensitive at low concentrations to adrenaline, noradrenaline and dopamine, but are insensitive to acetylcholine, pilocarpine and glutamate (Kaufman & Phillips, 1973b). Along with histological evidence (Meredith & W. R. Kaufman, in preparation) these observations suggest that secretion by the salivary glands is triggered by means of catecholaminergic nerve fibres.

Preliminary observations suggested that chloride ions might be actively transported by the salivary glands of Dermacentor (Kaufman & Phillips, 1973 a). This in turn might drive the secretion of water by local osmosis (Diamond & Bossert, 1967) or a similar process. The availability of a successful in vitro preparation (Kaufman & Phillips, 1973b) makes it possible to examine this hypothesis by observing how altering the ionic and osmotic composition of the bathing medium affects the rate of secretion.

The adult females used in this study were cold-diapaused for 3–6 months. The rearing procedure, analytical methods for Na, K, Cl and osmotic pressure, and method for setting up salivary glands in vitro are outlined in earlier papers (Kaufman & Phillips, 1973 a, b). Salivary glands were dissected from partially fed ticks (generally 40– 300 mg), and adrenaline (concentration indicated for individual experiments) was used to stimulate secretion.

Table 1 a indicates the components which were common to all the experimental and control media (but not the dissecting media). Table 1 b lists the additional compounds present in various experimental media. Paired media, designed to test the effect of substituting one ion species for another, were identical in all respects except for the two ions in question. Thus various mixtures of such paired media resulted in solutions with desired ionic ratios. Reference is made in this paper to ‘normal Ringer solution’ ; the composition of the latter is given in a previous paper (Kaufman & Phillips, 1973 b) ; it was prepared by mixing Na-Ringer and K-Ringer in the proportion, 95:5. Usually all the media were prepared in concentrated form, stored frozen, and then thawed and diluted to normal strength prior to use. The osmotic pressure of the ‘reference medium’ in some experiments was increased by the additions of sucrose, NaCl, KC1, or NaNO₃. In this way media of identical osmolality but varying solute composition were prepared.

The rate of secretion by in vitro salivary glands decayed steadily with time in normal Ringer solution even when the bathing medium was frequently renewed from the oxygenated stock. As a result, any decay of secretory rate arising from an experimental treatment was made up of two components - one due to the treatment itself and one due to the unavoidable deterioration of the gland with time. Similarly, any stimulation induced by an experimental treatment was partially masked. Therefore correction factors were applied in both protocols used in these experiments - (1) the maintained dosage method (glands continuously bathed in adrenaline), and (2) the pulse method (glands exposed only intermittently to adrenaline). In the first method glands were allowed to secrete for approximately 1 h in control medium (with replacement from the oxygenated stock every 15 min) and the rate of secretion was recorded over the last 15 min period. The glands were then exposed to the experimental medium and were observed until the effect of this treatment was apparent (approximately 15–30 min). The glands were then re-exposed to the control medium. The average secretory rate during the experimental period was compared with the average of the initial and final control periods. If the rate of secretion during the experimental period fell rapidly to zero and then recovered following return to the control medium, the experimental rate was taken to be ‘zero ‘. The pulse method (2) was more reliable for rate comparisons since the secretory rate in normal Ringer solution decayed linearly with time (see Kaufman & Phillips, 1973b). Thus of three pulses, the first and third were controls and the second was experimental. To estimate what the control rate would have been at the time when the gland reached maximum rate in the experimental medium, a straight line was drawn through the maximum secretory rate for the two control pulses. The actual maximum rate during the experimental pulse was compared with the estimated control rate.

The electropotential differences across the salivary gland epithelium were measured by three methods :

(A) Glass microelectrodes were drawn out on a ‘Palmer’ microelectrode puller and were filled by boiling in 1·5 M-KCI. The microelectrode made contact with a silver-silver chloride recording electrode attached to a ‘Mediator ‘cathode follower. The indifferent electrode made contact with the Ringer solution bathing the specimen. Only electrodes having a resistance of 10–20 MΩ were subsequently used.

Each tick was dissected in a manner similar to that described for the in vitro preparation (Kaufman & Phillips, 1973b) up to and including evisceration; dissecting medium 1 (Table 1 b) was used. The salivary acini did not on their own offer enough resistance to the advancing microelectrode to permit piercing of the basal plasma membrane. Consequently a cradle was constructed from an e.m. copper grid (coated with collodion) glued with epoxy resin to a bent pin. The pin was secured to a glass rod which in turn was held in a micromanipulator. It was possible, by manoeuvring the cradle amidst the acini, to prevent a given acinus from rolling freely when it was confronted with the glass electrode. The whole preparation was observed through a dissecting microscope and the trans-epithelial potentials were displayed on a ‘Tectro- nix’ model D54 oscilloscope after the asymmetry potential had been backed off. The Ringer solution used to bath the glands in situ was similar to that of Rehacek & Brzostowski (1969), but lacked the vitamins, phenol red, and heat-treated haemolymph fisted in their table of composition.

(B) The second method involved recording the potential difference across the whole gland, while the latter was secreting in vitro in normal Ringer solution. Saturated KCl-1 % agar gel bridges were made up in PE 90 tubing fitted at one end with a glass capillary. One bridge made contact with the secreted droplet, and the other with the bathing medium. The free ends of the PE tubing were placed in separate vessels containing saturated KCl and calomel electrodes. A high-impedance millivoltmeter (Keithley model 602) completed the circuit. Prior to experiments, the two agar bridges were placed in the medium bathing the gland and the asymmetry potential was recorded.

(C) The third method was in most respects identical to the second except that a glass microelectrode (described under A) was used to measure trans-acinar potentials. However, since some groups of acini in vitro always rested at the oil-Ringer interface on the Parafilm covering the Petri dish, the cradle described under Method A was not required. Electrodes first pressed an acinus to the Parafilm and then entered the lumen when the cathode follower was advanced by means of a micromanipulator. Occasional readings across the whole gland (as distinct from the acinus itself) were also made with the microelectrodes to confirm the results obtained with the KCl- agar bridges of Method B.

The rate of fluid secretion by Dermacentor salivary glands declines over a 6 h period (Kaufman & Phillips, 1973b). Since the present study required the collection of saliva over a 6 h period, it was important to determine whether or not the composition of this fluid remained constant with time when the gland was bathed in any single medium. Fig. 1 shows that the ion concentration of the secreted fluid did not fluctuate significantly, even though the rate of fluid secretion had fallen to 20 % of the initial rate. This suggests that the same active processes generating fluid secretion were operative throughout the experimental period, albeit at a continuously declining rate. Consequently, any differences in saliva composition resulting from the exposure of the glands to different bathing media cannot be explained on the basis of a deteriorating preparation.

The saliva to haemolymph (S/H) ratio for chloride in vivo is 1·1 (Kaufman & Phillips, 1973 a) and about 2 in vitro (Fig. 1). With these observations in mind, experiments were designed to test whether fluid secretion was dependent upon the presence of chloride in the bathing medium. Ticks were dissected in medium 1 (Table 1b). The control medium in each case was a chloride Ringer solution (Table 1b) and the experimental one was identical in all respects except that the chloride was replaced with the equivalent molarity of some other anion (nitrate, acetate, or bromide). In these experiments the maintained dosage protocol, using 10⁻⁴ M adrenaline, was followed.

When chloride in the bathing medium was replaced with increasing levels of nitrate, the rate of fluid secretion decreased. In pure nitrate Ringer solution the secretory rate was reduced to 5 % of the rate in pure chloride Ringer solution (Fig. 3). This inhibition could be partially reversed when the gland was offered chloride again (Fig. 2). It is unlikely that these results stem from possible toxicity of the nitrate ion, since the addition of 47 m-equiv. NaNO₃/l to medium 3 has the same effect on fluid secretion as adding the equivalent osmotic concentration of NaCl (Fig. 8). Thus inhibition was largely due to lack of Cl rather than presence of NO₃ ions.

The effect of replacing chloride with acetate was qualitatively similar to the effect of replacing chloride with nitrate. Pure acetate Ringer solution, however, was unable to support salivary secretion at all (Fig. 3). This inhibition was also reversible (Fig. 2) so the inhibition was unlikely to be merely one of permanent toxicity. Bromide was able to support fluid secretion equally as well as chloride. The rate of secretion of five glands in chloride Ringer solution was 23 ± 3 nl/min (mean ± S.E.) and that in bromide Ringer solution was 21 ± 3 nl/min.

The concentration of chloride in the saliva was linearly related to and was consistently higher (by approximately 80 m-equiv./l) than that of the bathing medium over a tenfold external concentration range (Fig. 4). This held true regardless of the rate of fluid secretion, or of whether the chloride concentration in the medium was arrived at by substitution of chloride with acetate or simply by the addition of NaCl. Fig. 5 shows that there is a direct relationship between the rates of chloride and fluid secretion; about 6 nl of fluid are secreted for each nano-equivalent of chloride transported. The direct dependence of fluid secretion on the chloride concentration of the medium and the high S/H ratio for chloride at all concentrations suggest that fluid movement is a consequence of chloride secretion.

The rate of salivation is linearly dependent upon rate of chloride secretion. Since chloride is probably transported against an electrochemical gradient (see potential difference measurements), this process is likely active. If so, by what mechanisms are the cations transported in order to achieve electroneutrality in the secreted fluid? If either sodium or potassium is actively co-transported with chloride, the rate of salivary secretion should be sensitive to the Na : K ratio ; but if the cations simply diffuse non- specifically through the glandular epithelium down an electrochemical gradient established by the active transport of chloride, then the salivary gland may be relatively indifferent to the presence of a specific monovalent cation. To test these hypotheses, glands were exposed to 10⁻⁵M adrenaline in media of varying Na:K ratios (total Na + K being constant). The experimental protocol was the pulse method.

Fig. 6 shows that the rate of fluid secretion was not directly proportional to the concentration of either cation. The control medium in this experiment (normal Ringer solution) contained 10 m-equiv./l K+ and 210 m-equiv./l Na⁺; the glands secreted most quickly in this medium. When the potassium in the control medium was completely replaced with sodium, the glands secreted at about 20 % of the maximum rate. However, when the potassium concentration was raised above 10 m-equiv./l the rate of fluid secretion also fell, approaching zero when the medium concentration of both cations was 110 m-equiv./l. To test whether this inhibition was caused by high potassium per se (rather than the concomitant low sodium), glands were exposed to 10⁻⁵ M adrenaline in media of constant NaCl concentration (100 m-equiv./l) but progressively higher KCl concentrations. As a control for the effect of elevated osmotic pressure resulting from the added KCl, the same glands were bathed in media to which an osmotically equivalent amount of sucrose was added. The experimental protocol was the pulse method. High concentrations of KCl had a much greater inhibitory effect on fluid secretion than did equivalent high concentrations of sucrose (Fig. 8).

The S/H ratios for sodium and potassium in vitro are slightly greater than unity over the complete range of concentrations tested (Fig. 7), whereas in vivo they are not significantly different from unity (Kaufman & Phillips, 1973 a).

The experiments performed by method A (see Methods) were carried out before it was discovered that topical application of a catecholamine is required before the glands of dissected ticks become active; the transacinar P.D. measurements obtained by this method therefore reflect the situation in resting glands. Before P.D. measurements were made by methods B and C, it was first ascertained that the glands were secreting fluid. The results are presented in Table 2. The lumen was consistently negative with respect to the bathing medium whether or not adrenaline was present, but the secretory potentials (35 mV ; method C) were on the average twice as great as the resting potentials (16 mV; method A). ^* However, when measured between the secreted droplet and the bathing medium (method B), the potentials were always zero or at most a few millivolts positive. It is possible that, with the latter method, the current was short-circuited through a surface film of electrolyte solution along the outside of the duct - even though such a film does not induce the secreted droplet to flow back into the medium bathing the gland. An alternative explanation of these low P.D. values arises from histological considerations (Meredith & W. R. Kaufman, in preparation). The salivary gland is composed of three types of acini, each possessing several cell-types, all interconnected by a system of ducts. There is no reason to assume that the P.D.S are similar in all regions. However, the transacinar P.D.S are probably more relevant to the process of ion secretion, since there is reasonably good ultrastructural evidence that the site of primary secretion is one of the acinar cell types.

The Na and K requirements for salivary secretion are highly specific (Fig. 6), Na being required at relatively high and K at relatively low concentrations. The relationship shown in Fig. 6 is similar to that widely observed for ‘pump ATPases ‘. As ‘pump ATPases’ are inhibited by the cardiac glycoside, ouabain, it was of interest to test the effect of the latter on salivation in vitro. Ticks were dissected in medium 2 and then subjected to 10⁻⁵ M adrenaline in normal Ringer solution containing 10⁻³ M, 10⁻⁵ M or 10⁻⁶ M ouabain (Strophanthin G, Calbiochem Corp.) after the secretory rates in medium lacking ouabain had been established. The protocol employed was the pulse method. Ouabain completely inhibited salivary secretion at a concentration of 10⁻⁶ M (Table 3). In one case (gland 5) the inhibition was partially reversed after incubation in normal Ringer solution for 85 min; the rate of secretion after recovery was 5 nl/min. The results suggest that fluid secretion is dependent on the activity of a ouabain-sensitive ‘pump ATPase ‘.

All experiments were conducted according to the pulse method with 10⁻⁵ M adrenaline in the bathing media. Each gland was tested with a control medium before and after exposure to an experimental one. Thus for comparing the effects of additional NaCl, sucrose or NaNO₃, the first medium served as the control and one of the latter two as the experimental. Likewise, sucrose media served as the control when testing the effect of additional KC1. The rates of secretion resulting from all these treatments were compared to those of a separate series of glands pulsed three times in medium 3, the basic medium containing no added solute.

Increasing the osmotic pressure from 300 to 400 mOsm/1 with NaCl or sucrose resulted in a slight acceleration of secretory rate (Fig. 8 a); the effect of KC1 was strongly inhibitory (50%). At 475 mOsm/l NaCl accelerated but sucrose and KC1 diminished the rate of secretion. Beyond 475 mOsm/1 the rate of secretion was diminished by both NaCl and sucrose, though glands could still secrete in NaCl Ringer solution reasonably well as 920 mOsm/1. Since the NaCl curve paralleled that for sucrose above 475 mOs/ml (Fig. 8 a), this suggests that after an initial stimulation by the added NaCl, the osmotic effect (inhibitory) of the latter was dominant at higher concentrations. It should also be recalled that very high Na: K ratios (Fig. 6) result in considerable inhibition of the secretory rate; therefore the reduced rate seen in Fig. 8a caused by addition of NaCl, may also be partially due to a concomitant increase of the Na:K ratio.

The saliva was consistently hypo-osmotic to the bathing medium by 30–35 mOsm/1 when the latter was varied between 300 and 675 mOsm/1, and was hypo-osmotic by 95 mOsm/1 when the bathing medium was 920 mOsm/1 (Fig. 8 b).

The present study suggests that salivation in D. andersoni involves passive movement of water coupled to the active transport of solute. The following observations are consistent with the hypothesis that the glandular epithelium contains a specific chloride pump and that the latter may be the principal driving force for fluid secretion :

1. The S/H ratio for chloride in vivo and in vitro is significantly greater than unity at all concentrations and net movement of this ion probably occurs in the acini against a P.D. of 35 mV.

2. Either chloride or bromide can support fluid secretion, but acetate or nitrate cannot.

3. The rate of fluid secretion is linearly related to both the concentration of chloride in the medium and the rate of chloride secretion.

The concentration difference of Na and K across the salivary epithelium could be supported by the observed transacinar P.D. Other evidence, however, speaks against a passive process such as simple diffusion for explaining the distribution of Na and K. K cannot replace Na in the nutrient solution even though the hydrated radius of K+ is smaller than that of Na⁺. Moreover, the curve relating K concentration to the rate of fluid secretion (Fig. 6) exhibits an optimum concentration in the vicinity of 10 mM/1 ; the general shape of this curve is reminiscent of the stimulating and competitive effects of Na and K on ‘pump ATPase’ systems (Skou, 1965). For the latter, maximum ATPase activity occurs when Mg²+ is present, and when Na and K are present in a specific ratio characteristic of the tissue in question. Concentrations of potassium which are high relative to sodium cause displacement of sodium from its specific site by competitive inhibition thus decreasing the ATPase activity. Cardiac glycosides such as ouabain interfere directly with the ‘pump ATPase ‘. The cessation of salivary secretion in the presence of 10⁻⁶ M ouabain (Table 3) lends further support to the suggestion that the role of Na and K in the secretory process is bound up with the activation of a ‘pump ATPase ‘. Nevertheless, our observations do not represent rigorous evidence either for the existence of a Na pump or for coupling of water movement to Na transport. Perhaps a ouabain-sensitive Na-K exchange pump exists and must function in order for general metabolic processes, or perhaps more specifically a Cl-pump, to operate. Nor can one exclude the possibility that ouabain interferes directly with the Cl-pump, although we have seen no report in the literature of such an effect on Cl-pumps in other tissues.

The saliva secreted by the glands in vitro was slightly, but significantly, hypo- osmotic to the bathing medium over the range 300–920 mOsm/1. This would appear to conflict with the earlier proposal that water movement was a consequence of the transport of ions. For such a proposal to be valid, the secreted fluid should be hyperosmotic or iso-osmotic but not hypo-osmotic to the bathing medium. The simplest explanation to account for a hypo-osmotic saliva is to postulate that the primary fluid secreted into the acinar lumen is hyper- or iso- osmotic, but that solute is subsequently reabsorbed back into the haemolymph in excess of water. In the rat parotid gland, reabsorption of Na occurs primarily in the striated ducts (Mangos, Braun & Hamann 1966) and in the rate submaxillary gland, reabsorption of Na occurs primarily in the sublobular and main excretory ducts (Young & Schôgel, 1966; Martinez, Holzgreve & Frick 1966). Oschman & Berridge (1970) have shown that the proximal clear region of the blowfly (Calliphora) salivary gland is the site of K reabsorption and that this region is responsible for rendering the final saliva hypo- osmotic. In the salivary gland of Dermacentor solute reabsorption could conceivably occur either in the acinus itself by cell-types other than the one which secretes the primary fluid, or along the duct system as occurs in mammals. Further discussion of this matter will be delayed until the morphology and ultrastructure of the salivary gland has been described (Meredith and W. R. Kaufman, in preparation).

The authors wish to thank Dr O. H. Petersen for reading the MS. The National Research Council of Canada supports the research carried out in this laboratory and generously provided a scholarship to W.R.K.