Ion and Water Balance in the Ixodid Tick Dermacentor Andersoni: I. Routes of Ion and Water Excretion

Terrestrial blood-sucking arthropods concentrate the nutrient portion of the blood meal by selective elimination of excess water. It has long been established that the Malpighian tubules are largely responsible for this water elimination in insects (Wigglesworth, 1931; Ramsay, 1955; Maddrell, 1964a,b). The argasid ticks excrete excess water via a pair of coxal glands (Boné, 1943), but until recently (Gregson, 1967; Tatchell, 1967b) the main excretory route employed by ixodid ticks was problematical.

A major contribution to our understanding of water-balance mechanisms in ixodid ticks comes from the work of Lees (1946, 1947). He suggested for Ixodes ricinus (1946) that excess fluid of the blood meal might be excreted by evaporation from the integument, although Lees himself recognized some shortcomings of this hypothesis (e.g. ticks feeding in a saturated or near-satiirated micro-environment were still able to regulate body water content). A possible answer to this problem was suggested by Gregson ( 1967), who proposed that the salivary glands may function in osmoregulation. Tatchell (1967b) substantiated Gregson’s suggestion by demonstrating: (1) that the estimated total water loss from engorging Boophilus microplus was far greater than that measured through the integument, and (2) that tritiated water injected into the haemocoele of engorging ticks could be recovered from the blood and urine of the host. Tatchell inferred from the latter that the tracer could only have entered the host via the salivary glands. Belozerov (1967) also implicated the salivary gland as an important route for water excretion in two ixodid species (Ixodes ricinus and Derma-centor marginatus) largely on the basis of cuticular water-loss studies and the observation that the salivary glands increase in size during the progression of the feeding cycle. He points out that precedent has been set for such an hypothesis, since bloodsucking gamasid mites excrete water from the salivary glands after feeding (Belozerov, 1958).

Although to date the ‘salivary gland hypothesis’ has been reasonably supported in one species of ixodid tick (Boophilus microplus-, Tatchell, 1967b, 1969), the latter species cannot be considered ‘typical’ of the family. First, it is a one-host tick (remaining on a single host throughout the larval, nymphal and adult feeds); secondly, and more important, it does not discharge fluid from the rectal sac throughout the feeding cycle (Seifert, Springell & Tatchell, 1968). Other ixodid species, such as Dermacentor andersoni, defaecate considerably during the feeding period, and so one must consider the possibility that a sizeable proportion of the total water loss may be excreted via this route. It was primarily for this reason that the relative importance of the integument, anus and salivary glands as sites of water loss was re-assessed in D. andersoni. We also hoped to estimate the relative importance of the salivary glands and the anus in eliminating monovalent ions.

All experiments were conducted on adult female Dermacentor andersoni taken from a laboratory culture.

About 2 weeks after hatching, larvae were confined to the ear and scalp region of a rabbit by means of a cloth sac taped to the head. The larvae became engorged within 4 or 5 days and moulted into nymphs about 2 weeks later when kept at room temperature (21–25 °C) over saturated KNO₃ (relative humidity (R.H.) = 98%; O’Brien, 1948).

Two-week-old nymphs were confined to the shaven back region of a rabbit by means of a foam rubber enclosure cemented to the skin. Nymphs engorged within 6 or 7 days and were stored at room temperature over saturated NaCl (R.H. = 88%) until they moulted into adults 3 or 4 weeks later.

The newly moulted adults were stored over saturated NaCl for 1 month, transferred to clean vials, and subsequently kept over saturated KNO₃ at 5 °C for 3–6 months. A few days prior to the adult feed the ticks were returned to room temperature and kept over saturated NaCl. As cautioned by Loomis (1961) only rabbits which had never before been infested with ticks were used for feeding adults.

Ticks were allowed to commence feeding and at daily intervals a few were removed from the rabbit. The anal plates and mouthparts were plugged with a mixture of beeswax and resin (tacky wax). The ticks were weighed, returned to the rabbit, and re-weighed at 24 and 48 h. Thus these ticks were exposed to conditions that were identical in terms of relative humidity and temperature to those experienced by feeding ticks, but intake of blood and elimination of faeces were prevented by the wax plugs. The rate of integumentary water loss was taken as the rate of weight loss over the first 2 days following forced detachment from the host. This estimate includes metabolic weight loss and loss from the spiracles, genital orifice, and Gené’s organ, none of which was covered with tacky wax.

Six polyethylene cylinders (3 cm diameter and 3 cm high) were glued to the back of a rabbit with epoxy resin. The area of skin surrounded by the cylinder was sprayed with a plastic surgical dressing (‘Aeroplast’, Parke Davis and Co.). This spray-on film prevented the tick’s faeces from coming into direct contact with the salts associated with the rabbit’s skin yet did not prevent the tick from attaching. Single females were placed in each of five capsules and five males were placed in the sixth. Dried faeces were collected from each capsule daily and were stored in separate polyethylene vials. On the sixth day of attachment the males were removed from their capsule and each was presented to a female for mating. The anal plate of the male was plugged with epoxy glue to prevent any of its faeces from being tallied with those of the female.

Haemoglobin concentration was calculated by determining the iron content of each sample by the method of Breuer & Militzer (1938), and by converting this value to haemoglobin content on the assumption that, by weight, 0·34% of haemoglobin is iron (Lemberg & Legge, 1949). Although Breuer & Militzer’s method does include a wet-ashing step in the procedure, this was not always sufficient to eliminate turbidity from the final solutions when using some of the tick samples. Therefore all samples were first dried in small platinum boats at 100 °C and then ashed at 460 °C for 5 h in a muffle furnace before carrying out the iron determinations. Although the recovery of iron from ashed inorganic standards was 98%, the recovery from ashed rabbit haemoglobin was only about 80%. Values from subsequent organic samples were corrected accordingly. Iron concentrations were read at 480 mμon a Unicam SP 500 spectrophotometer.

Haemolymph was taken from a severed leg segment under a stereomicroscope which was housed in a moist chamber kept near 100% R.H. This limited evaporation of haemolymph from the open wound during the time required for collection.

Saliva was collected in a glass capillary placed over the chelicerae and hypostome of the tick in the manner described by Howell (1966) and Tatchell (1967 a), except that no pharmacological stimulant was used to enhance the flow of saliva. Before placing the glass capillary tube over the mouthparts the oral end of the tube was dipped in liquid paraffin; in this way the column of saliva released into the tube was protected from evaporation. The total time for collection of saliva from a tick was 5–10 min.

The collected haemolymph and saliva were kept under liquid paraffin until analysed.

Sodium and potassium were determined by emission flame spectrophotometry with a Unicam SP 900 or a Techtron AA 120 flame spectrophotometer. Sodium samples were dissolved in distilled water and potassium samples in a 500 ppm Na (as NaCl) swamp solution. Homogenates of whole ticks or tick faeces were dry-ashed, but samples of rabbit blood, tick haemolymph and saliva were untreated before diluting in the distilled water or sodium swamp. Chloride concentration and osmotic pressure of haemolymph and saliva were determined by the first electrometric titration method of Ramsay, Brown & Croghan (1955), and the cryoscopic method of Ramsay (1949) respectively.

Haemolymph volume was estimated using the tracer-dilution method. Inulin-carboxyl-¹⁴C (50 μC/25 mg; supplied by New England Nuclear) was dissolved in 5 ml of the tissue culture medium of Rehacek & Brzostowski (1969). The inulin solution was injected through a severed hind leg using an ‘Agla’ micrometer syringe (Burroughs Wellcome and Co.). The needle of the syringe was fitted to a fine-tapered glass pipette via a sleeve of PE tubing. The pipette was sealed within the leg-stump with tacky wax before injecting the tracer, and was left in place for several minutes after the injection. After removal of the pipette the wound was sealed with the tacky wax. Leakage from the leg was prevented during the whole procedure by the use at appropriate times of a fine bulldog clamp. The dosage for unfed ticks (approximately 10 mg weight) was 1 μl, and this was increased up to 3 μl for ticks weighing over 100 mg. All haemolymph samples (usually 1 μl) were added directly to Bray’s solution (Bray, 1960) for determination of ¹⁴C-activity using a Nuclear Chicago ‘Mark I’ liquid scintillation counter and the channels-ratio method for quench correction. The radioactivity of samples ranged between 10 and 60 times background. Since it was shown in several ticks that the radioactivity per unit volume of haemolymph was the same 1, 2 and 3 h after injection, haemolymph was thereafter sampled about 2 h following injection.

Values for total intake and total excretion of four engorged females are present in Table 1.

Table 2 shows the integumentary water loss during a normal feeding period for 36 ticks separated into arbitrary weight ranges. Since the method used does not allow distinction between metabolic and evaporation losses, the estimates in Table 2 are probably maximal. Only about 40 mg of water were lost through the integument by the average adult female over a normal 7-day feeding period; this figure represents 2·5% at most of the total water loss (cf. Table 1).

Differentiating between the quantities of fluid lost via the anus and salivary glands proved to be difficult. Since the excreted saliva is injected back into the host tissues throughout most of the feeding period, it was not feasible to collect it and measure its volume directly. All attempts to collect wet faeces and measure their water content before evaporation occurred were also unsuccessful. It was finally decided to take advantage of information on salt concentrations in the fluids of the tick and host which permit one to calculate the probable fluid loss via the salivary glands. These calculations become possible because of the fortunate circumstance that once salivation has commenced the sodium concentration of saliva (161 ±3 m-equiv./l; mean±S.E.) collected in glass capillaries does not change with increasing size of tick (Fig. 1); the assumption was therefore made that saliva injected naturally into the host also showed little variation in sodium concentration with the phase of engorgement. The total amount of sodium ingested (Table 3 C) during the feeding period was calculated from the volume of imbibed blood (Table 1) and the measured sodium concentration of rabbit whole venous blood (101 ± 2 m-equiv./l; mean± S.E.). The total amount of sodium lost in the saliva by each tick during the feeding period was calculated from the following:

With values for total meal intake of sodium and potassium, and for the concentrations of these ions in the saliva and faeces (Table 4), it was possible to estimate the relative importance of the salivary glands and anus in the excretion of sodium and potassium. One arrives at the following conclusions from Table 4: of the total ingested sodium and potassium, 89 and 60 % were eliminated respectively. However, whereas 96 % of the excreted sodium was lost in the saliva, only 18 % of the excreted potassium was lost via this route. Conversely, 82% of the excreted potassium, but only 4% of the excreted sodium appeared in the faeces. The figures for chloride were not complete, but it appeared that this ion was excreted in a manner similar to that of sodium.

In order to gain some insight as to the mechanism of fluid secretion by the salivary gland (see Kaufman & Phillips, 1973 a, b), the ion concentrations of haemolymph and saliva were monitored throughout the feeding period (Fig. 1). Haemolymph of the unfed tick had a fairly high ionic content (280 m-equiv./l. Na), but this fell with the progression of feeding to about 160 m-equiv./l on the third day. This level was maintained until repletion. Likewise, the chloride concentration of haemolymph fell from 170 m-equiv./l initially, to 125 m-equiv./l after 5 days and then remained relatively constant. Similarly, potassium fell from about 20 to 7·5 m-equiv./l. Comparisons of sodium and potassium concentrations in saliva and haemolymph of individual ticks indicated no significant differences; the chloride concentration of the saliva, however, was about 10% higher than that of the haemolymph (P = 0·01; pairs t-test) over the whole time course of salivation. Osmotic pressure (Fig. 2) of the haemolymph fell from 527 ± 21 mOsm/1 (mean ± S.E.) on the first day of feeding and remained at 375 ± 6 mOsm/1 through the remainder of the feeding period. The saliva was slightly, but consistently, hypo-osmotic (P = 0·02) to the haemolymph by about 5 % (saliva = 356 + 4 mOsm/1).

In this study it was important to determine whether inulin could be secreted by the salivary glands or Malpighian tubules, or whether it could diffuse into a large compartment such as the gut and thus result in an overestimate of haemolymph volume. The following experiments showed that errors resulting from the latter possibilities were probably quite small. Six partially fed ticks were injected with large doses of ¹⁴C-inulin and then returned to the rabbit to recommence feeding, salivation, and defaecation. The next day freshly extruded faeces was collected, and haemolymph and saliva were sampled. There was no detectable ¹⁴C-activity in any of the six saliva samples, although the haemolymph contained 5000–40000 cpm/μl of radioactivity. Clearly inulin is not secreted by the salivary gland. The activity of the faeces at the same time was 310 cpm/mg dry weight (mean value for faecal samples pooled from a number of injected ticks); thus inulin is only slowly eliminated from the haemolymph.

Haemolymph volume increased linearly with increasing weight of the tick, and always comprised about 23% of the body weight (Fig. 3). That the haemolymph volume is maintained at a constant proportion of the body weight (as the tick becomes engorged to 75 times its unfed weight) suggests that the volume of this compartment is under some control. We supposed that it is the act of salivation which ultimately regulates the volume of haemolymph, and therefore we tested whether artificially increasing haemolymph volume could stimulate or induce salivation into glass capillary tubes. However, compared to non-injected controls, salivation was inhibited when the haemolymph volume was increased by 25 % or 50 % with iso-osmotic saline (1·2% NaCl) (Fig. 4). One can coax a tick forcibly removed from a rabbit to secrete into a capillary tube and then restore its haemolymph volume by injections of saline; however, this treatment failed to induce salivation (Fig. 4).

A pictorial summary of the essential findings is presented in Fig. 5. The female tick excretes ions via both the anus and the salivary glands; however, most of the excreted sodium and probably chloride are lost in the saliva and most of the excreted potassium in the faeces. One explanation for the differing routes of excretion is that the major portion of the faeces is derived from host blood passing directly into the rectal sac from the midgut. If the midgut epithelium can transport sodium, chloride and water into the haemolymph but, is relatively impermeable to potassium, this would result in the observed high potassium and low sodium in the remains of the meal entering the rectal sac. There is some evidence for this hypothesis. Once the haemolymph has attained a stable composition 3 or 4 days after feeding has commenced, the sodium concentration in the haemolymph (160 m-equiv./l) is somewhat higher than the sodium concentration in the meal (100 m-equiv./l). Similarly, the chloride concentration in the haemolymph (125 m-equiv./l) is also higher than that in the meal (90 m-equiv./l). However, haemolymph potassium (7·5 m-equiv./l) is considerably less than that of the meal (42 m-equiv./l). These figures are consistent with the transport of sodium and chloride across the gut epithelium in excess of potassium. The relative impermeability of the gut epithelium to potassium has been suggested for the argasid tick, Ornithodorus moubata (S. E. Kaufman, 1971). Alternatively, one cannot rule out another plausible explanation. Since potassium is the major cation in the Malpighian tubule secretion of several insects (Ramsay, 1953; Berridge, 1968; Irvine, 1969; Maddrell, 1969; Pilcher, 1970) it might be absorbed from the midgut and then rapidly secreted by the Malpighian tubules of the tick. Provided that reabsorption of potassium in the rectal sac were lower than in insects studied to date, the net result would be a low potassium concentration in the haemolymph and a high concentration in the faeces. Although there is no clear-cut evidence which opposes the latter mechanism, there are some facts available which make it less attractive than the first explanation. On the basis of histological investigation, both Balashov (1958) and Till (1961) report that the Malpighian tubules do not become very active until after the tick detaches from the host. Most of the accumulation of guanine in the Malpighian tubules and rectal sac occurs at that time and is probably due to metabolism associated with egg development. This appears to be the situation in D. ander-soni as well (unpublished observations). However, more direct evidence on the quantity of fluid and potassium secreted by the Malpighian tubules during the feeding period is clearly desirable.

On dry-weight basis the sodium concentration in male faeces was about six times that in female faeces (217 m-equiv. Na/kg dry weight and 38 m-equiv. Na/kg dry weight, respectively). This finding suggests that males eliminate salt and water differently from females. Although there is little doubt now that the salivary glands are important for water regulation in the female of two ixodid species, for a variety of reasons it has not yet been necessary to postulate the same for the male. First, growth of the salivary glands during feeding is not as marked in the male as in the female (Till, 1961; Chinery, 1965). Secondly, the male imbibes only a modest amount of blood, and so it is not faced with the task of excreting large volumes of excess fluid. Finally, the paralytic factor (most probably carried in the saliva) is only rarely transmitted by the male (Gregson, 1943). One would suspect that anal and integumentary water loss might suffice to account for osmoregulation in the male.

When one compares the ionic composition of haemolymph and saliva from Derma-centor andersoni and that from Boophilus microplus (Tatchell, 1969), some differences emerge. In Dermacentor the saliva to haemolymph (S/H) ratio for sodium and potas sium is insignificantly different from one: in Boophilus, however, the S/H ratio for sodium is greater than one, and that for potassium less than one, although Tatchell does not state whether the differences are significant. In both species the S/H ratio for chloride is I-I. The saliva of Boophilus is hyperosmotic to the haemolymph (S/H = 1·23), whereas in Dermacentor it is slightly (but significantly) hypo-osmotic to the haemolymph (S/H = 0·94). With the assumption that the primary secretion in Dermacentor is iso-osmotic or hyperosmotic to the haemolymph (i.e. that flow of fluid is driven by a local osmotic gradient), then reabsorption of solute relative to water may occur somewhere between the salivary acini and the oral cavity; micropuncture studies on vertebrate salivary glands demonstrate clearly that the ducts are responsible for solute reabsorption and hence the elaboration of a hypotonic saliva (Martinez, Holzgreve & Frick, 1966: Mangos, Braun & Hamann, 1966; Young & Schögel, 1966). Such may also be the case for Dermacentor. Since in Boophilus the saliva is hyperosmotic (as expected for a secretory system), this would suggest that in the latter species, either the ducts serve merely as a delivery system for the saliva, or that they possibly secrete solute as well. With this in mind it would be interesting to compare the ultrastructure of the ducts in these two species.

Despite the large net flux of ions and water through the haemolymph compartment (in all, 9 to 12 times the haemolymph volume measured at repletion), the ratio of extracellular fluid to body weight remains constant throughout the feeding cycle (Fig. 3). This suggests that the rate of salivary secretion is correlated with fluid intake. Maddrell (1964c) concluded that the release of diuretic hormone (i.e. the stimulus to urine secretion by the Malpighian tubules) in Rhodnius is linked to fluid intake through abdominal distention via receptors in the tergo-sternal muscles. We have not performed the experiments necessary to reveal stretch-receptors controlling salivary gland activity in Dermacentor; but at least salivary secretion does not appear to be related in a straightforward way to haemolymph volume (Fig. 4).

This paper provides evidence that the salivary gland is the major route whereby excess NaCl and water are excreted in Dermacentor, and implies that as a result of this process control over the volume of haemolymph may be exercised. The mechanism of salivation (whether fluid is produced by a secretory or a filtration - resorption process) and the control (whether by nerves or hormones) are examined in a subsequent paper (Kaufman & Phillips, 1973 a).

The tick culture in this laboratory was established from wild specimens kindly provided by Mr J. D. Gregson, formerly at the Canada Department of Agriculture, Kamloops, British Columbia. The authors wish to thank Dr M. J. Berridge for his most helpful comments on a draft of the manuscript. The National Research Council of Canada provided research funds and a scholarship to W. R. K. which are gratefully acknowledged.