ABSTRACT
A sandwich-type enzyme-linked immunosorbent assay (ELISA) was utilized to quantify crustacean hyperglycemic hormone (as Carcinus maenas equivalents) released by single X-organ–sinus gland systems of the crab Cardisoma carnifex during continuous perifusion. Basal rates of secretion (20–60 pg min−1) were stable for at least 4 h. Electrical stimulation (600 stimuli in 5 min) of the axon tract increased secretion two-to threefold, but only if it resulted in neural activity that was propagated to the terminals of the sinus gland. No difference was observable when stimuli were given repetitively or as a series of trains. Perifusion with saline having ten times the normal K+ concentration augmented secretion by as much as fivefold. Augmented secretion of crustacean hyperglycemic hormone evoked by either electrical or K+ stimulation appeared abruptly but declined slowly (over tens of minutes) after stimulation was stopped. K+-evoked secretion of crustacean hyperglycemic hormone was maintained without decrement for at least 1 h. Basal secretion increased in saline from which Ca2+ had been omitted, but decreased in saline containing Mn2+. Neither electrical stimulation nor high [K+] augmented secretion in Ca2+-deficient saline or if Mn2+ was present. Introduction of Mn2+ during K+-evoked secretion immediately reduced release to unstimulated levels; secretion resumed promptly upon removal of Mn2+. Tetrodotoxin reversibly blocked both electrical and secretory responses to axonal stimulation, but it did not block basal or K+-evoked secretion.
Release of crustacean hyperglycemic hormone by isolated axon terminals was augmented two-to threefold by perifusion with saline having ten times the normal K+ concentration. The responses were similar to those of the intact systems, having a rapid onset, well-maintained secretion and a long ‘tail’ of secretion after removal of the K+ stimulus.
Introduction
Our study attempts to explain the relationships among neuronal activity, depolarization of secretory terminals and the rate of secretion of a neurohormonal peptide, crustacean hyperglycemic hormone (CHH), that have not been previously examined in this way. It further asks whether these relationships, and the resulting pattern of peptide release, differ for this hormone compared with the relationships for other hormones in a manner reflecting its different physiological role.
In addition to the vertebrate hypothalamic-neurohypophyseal system, the crustacean X-organ–sinus gland (XOSG) system provided important early examples that led to acceptance of the concept of neurosecretion: integration of synaptic input to the neuron results in axonal impulse firing and subsequent release of a peptide hormone from its terminals into the circulation at a specialized neurohemal site to elicit responses at distant target tissues (for a review, see Cooke and Sullivan, 1982). Studies of excitation–secretion coupling require the collection of fluid for analysis of released hormone. Therefore, the time resolution of rates of release is governed by the sensitivity of the assay available, as this sets the minimum intervals at which samples of blood or perifusate can be taken and still include sufficient peptide for detection. Observations of the classical systems mentioned above have achieved a 15 s resolution for release of vasopressin (Bicknell et al. 1982a) and for radiolabelled peptides of the crab sinus gland (Stuenkel, 1985) but, more routinely, studies have reported observations at intervals of one or a few minutes. Although many studies have demonstrated secretion of hormones in response to electrical stimulation or perifusion of depolarizing levels of K+, relatively few studies provide any information about how rapidly secretion begins, how well it is sustained during continuing stimulation and how rapidly it ceases when stimulation ends. Observations from the few preparations for which there are time-resolved data suggest that there may be important differences in these variables. The differing time scales involved in the physiological effects of neurohormones would seem to call for widely differing patterns of secretion. In the case of the crustacean XOSG, rapid events, such as color change (Fingerman, 1985), may require a rapid brief elevation in the rate of secretion of, for example, red pigment-concentrating hormone (RPCH). However, release of crustacean hyperglycemic hormone, a metabolic regulator, may be continuous and/or may involve a diurnal rhythm, thus leading to elevated levels lasting for hours (Kallen et al. 1990; Keller and Orth, 1990). The pattern for secretion of molt-inhibiting hormone (MIH), which is believed to suppress steroid production over periods of months (Webster and Keller, 1988), is unknown.
Recent advances in immunocytochemical methods have greatly expanded the number of neurons shown to contain peptides. These include many that contain co-transmitters and which make synapses onto other cells. In some of these, recording of the postsynaptic electrical response has allowed resolution to within milliseconds or seconds of the relationships between neuronal activity and peptide secretion (Cropper et al. 1990; Whim and Lloyd, 1989; Lundberg et al. 1989; Turrigiano and Selverston, 1989; Peng and Horn, 1991; Peng and Zucker, 1993). A generalization emerging from such studies is that an increased rate of secretion of peptides usually requires prolonged repetitive impulse activity. This seems to be adaptive, given that the effects mediated by peptides are generally far more enduring than those of neurotransmitters and are often more widespread.
It has been shown that the electrical activity of peptidergic neurons associated with known or putative hormonal secretion can be so distinctive as to provide identification for neurons secreting different hormones. Examples include the oxytocin (long bouts of high-frequency firing) and vasopressin (patterned bursting) neurons of the hypothalamic–neurohypophyseal system of mammals (for a review, see Poulain and Theodosis, 1988) and the parabolic burster and bag cell neurons of the mollusc Aplysia californica. In general, patterned impulse activity, seemingly endogenously generated once initiated, is a characteristic feature of most neurosecretory cells (for a review, see Cooke and Stuenkel, 1985). The relevance of the patterning of impulses to the release of hormone has been clearly established in only a few cases. One of these is the secretion of vasopressin, in which facilitation by repetitive impulses within each burst is accompanied by rapid fatigue. Recovery occurs during the interburst interval (Dutton and Dyball, 1979; Cazalis et al. 1985).
A number of patterns of neuronal bursting, as well as steady or irregular repetitive firing, have been recorded intracellularly and extracellularly from neurosecretory somata of the crayfish X-organ (Iwasaki and Satow, 1969, 1971; Glanz et al. 1983; Onetti et al. 1990; Martínez et al. 1991) and from somata and terminals of the crab XOSG (Stuenkel, 1985; Cooke, 1985; Nagano and Cooke, 1987). However, it is not known what hormonal activity might be associated with any of these activity patterns.
Studies of the peptide composition of sinus gland extracts (Newcomb, 1983, 1987; Keller and Kegel, 1984) together with studies utilizing antisera to specific neuropeptides (Van Herp and Van Buggenum, 1979; Mangerich et al. 1986; Dircksen et al. 1988; for reviews, see Webster and Keller, 1988; Keller, 1992) have identified three distinct hormonal phenotypes of peptidergic neurons among the approximately 150–200 neurons of the XOSG system. More than half represent neurons producing CHH, and these are responsible for a large majority of the terminals in the sinus gland (Dircksen, 1992; Hanke et al. 1992). Fewer neurons give rise to MIH- and RPCH-containing terminals of the SG.
Among the time-resolved studies of peptide neurosecretion, the XOSG has provided two examples with a time resolution to 1 min or better. These are significant because they document patterns of secretion that differ sharply from each other. During perifusion with high-K+ saline, release of RPCH peaks rapidly (within 2 min) and equally rapidly declines (Cooke and Haylett, 1984). By contrast, release of radiolabelled peptides, predominantly CHH, rises rapidly, but is sustained, declining only slowly over 40 min (Stuenkel, 1985). These patterns are reminiscent of the rapid fatigue of vasopressin release and of the sustained secretion of oxytocin during sustained repetitive electrical stimulation (Bicknell et al. 1984).
The present report presents the first study, to our knowledge, of the secretion of specifically and quantitatively determined CHH from the crustacean sinus gland in response to axonal stimulation as well as to elevation of saline K+ concentration. The observations were made possible by the development of a sandwich-type ELISA with a sensitivity better than 2 pg (less than 0.25 fmol) of CHH. This immunoassay is based on antisera raised to CHH of the shore crab Carcinus maenas. Both the amino acid sequence and the prohormone structure of this neuropeptide have been determined (Weideman et al. 1989; Kegel et al. 1989).
The ELISA made it possible to follow secretion of CHH from single isolated XOSGs with a time resolution of 1 min. We show that axonal stimulation, if it results in propagated electrical responses observable from the secretory terminals, evokes a sustained two-to threefold or greater augmentation of the rate of CHH release, although even greater rates are observed in response to perifusion with saline having a 10-fold greater than normal [K+]o. The pattern of CHH secretion confirms the observations on secretion of radiolabel (Stuenkel, 1985) in being sustained and showing no susceptibility to depolarizing inactivation. A new feature that was observed is the very slow decline in the rate of CHH secretion after cessation of stimulation.
Materials and methods
Animals
Large (approximately 250 g), male semi-terrestrial crabs, Cardisoma carnifex (Herbst), were flown to Honolulu from Christmas Island, Line Island District, Republic of Kiribati. They were held for a few days to several weeks in outdoor screened cages provided with fresh water and sea water and fed rat pellets (Purina) supplemented with lettuce and vegetable scraps. Animals have been held for over a year under these conditions but have never molted.
Dissection
An eyestalk was removed and placed in chilled sea water. Its exoskeleton was removed, and muscle and connective tissue overlying the neuronal tissue were removed with the aid of a dissecting microscope. Under chilled sea water or crab saline, the X-organ, sinus gland tract and sinus gland were isolated. Some tissue surrounding the nerve tract was left to support it. The intact isolated system, measuring approximately 0.3 mm×4 mm, was transferred to the saline-filled experimental chamber in a fire-polished glass tube in order to avoid bringing it through the saline–air interface. In some preparations, it was then pinned with cactus spines through adhering connective tissue; in others, the electrodes were used to hold the preparation.
Stimulation and recording
For stimulation, the X-organ and a portion of the nerve tract, together with saline, were drawn into a fire-polished closely fitting capillary having internal and external silver wires (i.e. a suction electrode). Brief (0.3 ms), rectangular stimulus pulses were delivered from an isolation unit (Grass SIU 4A) controlled by a stimulator (Grass S48). The recording electrode was a fire-polished capillary having a tip diameter of less than 100 μm, in which saline made contact with an internal silver wire. This was led to one input of a high-input-impedance high-gain (1000X) differential a.c. preamplifier (Grass P15). The indifferent electrode was a silver wire placed in the chamber. Low-pass and high-pass filters were generally set at 0.3 Hz and 3 kHz (half amplitude) respectively. Outputs were led to a storage oscilloscope, an audio monitor and to an analog-to-digital converter (Sony PCM 701ES) for recording with a video tape recorder. A grounded silver electrode was placed between the stimulating and recording electrodes to reduce the stimulus artifact. Using a manipulator under guidance of the dissecting microscope, the opening of the recording electrode was pressed gently onto the surface of the sinus gland, usually near the most distal edge.
Experimental chamber and perifusion
The chamber was designed to minimize turbulence and mixing of perifusate during passage over the isolated preparation. A peristaltic pump fitted with small-bore tubing supplied saline at a rate of approximately 0.15 ml min−1. A combination of capillarity and siphoning was used to remove the perifusate to vials for later analysis. There was some fluctuation of the fluid level in the chamber corresponding to the drops falling from the siphon. A fraction collector using a drop counter (six drops per vial in approximately 1 min) moved the collecting vials under the siphon.
Studies with dye showed that a new solution arrived at the chamber as a well-defined front. Some dilution of dye occurred between the chamber and the arrival of dye in the collecting vial. The time from change of solution to its first appearance in a collecting vial was 2 min, corresponding to a tubing and chamber volume of about 0.2 ml. In Figs 2–4, the time of changing the perifusate at the pump is indicated.
Washout of a sample of 500 pg of a CHH standard in 10 μl of saline added near the sinus gland in the chamber was examined in one experiment. The level of CHH was less than half the peak in the next fraction (1 min) and had returned to the spontaneous baseline in the following fraction.
Salines
The normal perifusate was the saline devised for Cancer sp. by Pantin (1948) modified in that it was buffered with 2 mmol l−1 Hepes (pH 7.4). It contained (in mmol l−1): NaCl, 440; KCl, 11.3; CaCl2, 13.3; MgCl2, 26.3; Na2SO4, 23. ‘High-K+ saline’ had 113 mmol l−1 KCl; in ‘Ca2+-deleted saline’ the CaCl2 was omitted. Osmotic equivalency in these salines was maintained by a corresponding reduction or increase in [NaCl]. Osmolarity determinations for normal and for high-K+ salines were 958 and 954 mosmol l−1, respectively. In saline containing Mn2+, MnCl2 was substituted for MgCl2 and NaCl was substituted for Na2SO4. Tetrodotoxin (TTX, CalBiochem) was diluted just before use from a concentrated stock solution.
Dissociation and perifusion of sinus gland terminals
Sinus glands were isolated by dissection from 6–8 eyestalks and placed in 50 μl of saline in a semispherical depression in black acrylic resin. They were teased under a dissecting microscope with insect pins to separate the bleb-like terminals from their axons. Connective tissue and axonal tracts were removed with fine tweezers. The remaining material was transferred to a chamber with a long-tip micropipette along with additional saline used to rinse the dish and suspend the terminals. The chamber was designed to retain the terminals in a small volume in a stream of perifusate giving minimal turbulence or mixing. The height of a reservoir was adjusted to provide a flow rate of approximately 0.1 ml min−1. Fractions were collected at 5 min intervals. Because the time required to clear the deadspace in the perifusion system was approximately 1 min, switches to altered perifusate were made in advance of the change of collecting vial in the hope of avoiding mixing in the collected sample (see Fig. 5). In order to avoid possible mechanical artifacts, the chamber was held on a vibration-isolated table, separated from the stand on which perifusing solutions and the switching manifold were mounted and from the fraction collector. Changes between reservoirs of normal saline did not cause a change in the amount of CHH appearing in the fractions.
The isolation of terminals was necessarily incomplete and involved loss of peptides to the saline, so the total amount of CHH remaining in the chamber (i.e. held in terminals on the support) was evaluated at the end of the experiment by introducing distilled water and stopping flow for 20 min or more to lyse the terminals, then flushing the chamber contents, together with several rinses, into a vial for assay (as serial dilutions of up to 1/100 000). Thus, for example, the total remaining at the end of the experiment shown in Fig. 5 was 1.36 μg (of Carcinus CHH equivalents, see below) from six sinus glands, representing 11–13 % of the amount expected to have been in the undissociated sinus glands.
ELISA for CHH determination
Antisera
A heterologous assay of the sandwich type was developed, utilizing antisera that had been raised in rabbits against purified CHH of Carcinus maenas (Keller, 1988). IgG fractions were prepared by protein A/Sepharose chromatography according to standard procedures. Although Carcinus CHH could be measured by using IgG from single rabbits, both as first and second (biotinylated) antibodies, it proved necessary to use antibodies from two different animals to obtain cross-reactivity of Cardisoma CHH in this assay. Since the assay depends on the recognition of two different epitopes, we assume that one epitope in Cardisoma CHH is significantly different from the corresponding one in Carcinus CHH and that the population of antibodies in a single animal (either animal 1, T1, or animal 2, T2) did not contain antibodies to both epitopes. However, strong cross-reactivity was obtained either when IgG from T1 was used as first and that from T2 as second antibody, or when both were mixed in equal proportions and the mixture was used as both first and second antibody. It may be mentioned here that in a radioimmunoassay (RIA), which depended on single epitope recognition, Cardisoma CHH showed perfect cross-reactivity in the standard Carcinus CHH assay (R. Keller, unpublished results, see below).
Preparation of titer plates
Microtiter plates (Nunc Maxisorp F16, flat bottom) were activated with 0.25 % glutaraldehyde in 0.1 mol l−1 sodium phosphate buffer, pH 5 (100 μl per well) for 4 h at 37 °C and washed three times with the same buffer.
Coating with first antibody
Wells were filled with 100 μl of a mixture of equal amounts of IgG from T1 and T2 (diluted to 20 μg ml−1 with 0.1 mol l−1 sodium phosphate buffer, pH 8) and incubated for 6 h at 37 °C. This was followed by five washes with the same buffer. Wells were then filled with 400 μl of a 2 % bovine serum albumin (BSA, Sigma RIA grade, fraction V) solution in sodium phosphate buffer, pH 8.0, plus 0.02 % sodium azide and left for 1 h at room temperature (approximately 25 °C).
Addition of standards and samples
Both standards and samples were appropriately diluted with phosphate-buffered saline containing 0.1 % Tween 20 (PBS-T). Sample size was 100 μl per well. The fractions from the XOSG perifusion experiments (140–150 μl) were mixed with 200 μl of PBS-T, and two 100 μl samples were taken for duplicate assays. Dilutions of pure Carcinus CHH were used as standards and were assayed in quadruplicate. Plates were incubated for 24 h in a refrigerator (4–9 °C) and finally washed seven times with PBS-T.
Second (biotinylated) antibody
The mixture of T1 and T2 IgGs was biotinylated by use of a commercial kit (Amersham RPN 28) according to the manufacturer’s instructions, and diluted to 2 μg ml−1 with PBS-T. Each well received 100 μl of the diluted solution, was incubated for 6 h at 37 °C and was then washed eight times with PBS-T.
Addition of avidin–peroxidase complex
We used ready-made avidin and biotinylated horseradish peroxidase working solutions of the Vectastain-ABS kit (Vector Laboratories, Burlingame, CA). 25 μl fractions, containing 125 μg of avidin and 62.5 μg of enzyme, were added to 250 μl of PBS-T and allowed to react for 1 h at room temperature. After a 100-to 300-fold dilution with PBS-T, 100 μl samples (0.06–0.18 μg of peroxidase) were added to each well and incubated for 1 h at room temperature. Plates were finally washed 8–9 times with PBS-T.
Peroxidase substrates
A solution of 0.04 % 2, 2′ -azino-di-3-ethylbenzthiazolinesulfonate (6) (ABTS, Boehringer) in 0.5 mol l−1 sodium phosphate/citrate buffer, pH 4, containing 0.003 % H2O2, was added (100 μl per well) and absorption readings were taken 30–60 min later at 414 nm, using a Biorad model 2550 EIA reader.
Sensitivity
Fig. 1 presents an example of a standard curve such as was obtained for each plate. The graph shows the mean and standard deviations of the quadruplicate determinations of CHH standards plotted against the optical density (on logarithmic scales) with the calculated regression (r2>0.96) for one of the assay plates. Such a plot was obtained for each plate and was used to obtain the values of CHH in the experimental vials assayed on that plate. The lower limit of detection was 1 pg of Carcinus CHH equivalents; the usable range of the assay was between 1 and 50 pg. Since CHH from Carcinus maenas was used as the standard in the assay, the values reported here for the Cardisoma XOSG system are Carcinus CHH equivalents and cannot be taken as absolute amounts of Cardisoma CHH. When dilutions of extracts from whole sinus glands of Cardisoma were tested in the assay, a sample of 0.3×10−5 sinus gland equivalents was found to be close to the lower limit of detection, corresponding to approximately 2.3 pg of Carcinus CHH. This amounts to approximately 690 ng of CHH equivalents per sinus gland of Cardisoma. The mean amount of CHH in a Carcinus sinus gland is 1.2 μg (Keller et al. 1985). There is evidence, mostly from high performance liquid chromatography studies (Newcomb et al. 1985), that CHH is just as abundant in the sinus gland of Cardisoma as it is in that of Carcinus. Since the sinus gland of Cardisoma is bigger, the absolute amount may be even higher than 1.2 μg, perhaps in the range 1.8–2 μg. This suggests that Cardisoma CHH has a cross-reactivity of approximately 0.3 in the ELISA.
Statistical analyses
To test the statistical significance of responses to experimental changes, values were compared using a Student’s t-test with the level for significance considered to be P<0.05.
Results
CHH secretion in response to axonal stimulation
Dependence on propagated electrical responses
In seven of eight isolated X-organ–sinus gland systems prepared in the continuous perifusion chamber for electrical stimulation of the axonal tract and extracellular monitoring of responses from sinus gland terminals, brief suprathreshold stimuli (0.3 ms) were followed by complex waveforms lasting several milliseconds and having several discrete peaks as recorded by the pore electrode on the sinus gland (see Figs 2, 3). The appearance of a response occurred with a sharp threshold; with increasing stimulus amplitude, the delay between stimulus and response decreased to a constant value and additional complexities developed in the response. Beyond a limited range, further increase of the stimulus amplitude caused no further changes or, in some cases, a decrease in the responses. During repetitive stimulation, some fluctuation of the responses was observable, but in none of the preparations was there a marked decrement of the response levels during the periods of repetitive stimulation (generally 600 stimuli in 5 min at 2 Hz or in interrupted trains) given to examine evoked secretion of CHH. Unlike the recordings in a previous study (Cooke et al. 1977), very little spontaneous electrical activity was observed in this series of experiments. This is the result of using a pore electrode of smaller tip diameter which therefore sampled a smaller number of axons and terminals. The number may have been as few as three or four, as judged by the number of discrete peaks resolvable (e.g. Fig. 3).
Assays of the perifusate collected during and after bouts of axonal stimulation consistently showed an increase in the amount of CHH relative to that in samples collected before stimulation. In eight of nine preparations examined, the increase was greater than 50 %, and in six of these greater than 100 %. In the exceptional preparation, there was no observable propagated electrical response and the nerve appeared to be damaged (VI in Tables 1, 2). Table 1 summarizes data from six of the preparations in which the manner of perifusion and sample collection were comparable. It will be seen that all of the preparations showed unstimulated release of CHH (approximately 13–74 pg min−1). Rates of release during electrical stimulation often exceeded 100 pg min−1 (range 37–175 pg min−1). Pooling the averaged measurements (excluding those from VI), revealed that electrical stimulation caused a 145 % increase in rate of secretion over basal levels.
In order to provide the reader with an overview of a complete experiment, Fig. 2 presents all the assay data from one of the experiments (III in Tables 1, 2; not all samples were assayed) together with examples of the electrical recordings. Typically, an increase in response to electrical stimulation was first seen in the third sample following commencement of the stimulation, as, for example, in Fig. 2B (the early peak at A in Fig. 2 may be an artifact), reflecting the deadspace of the perifusion system. Unexpectedly, however, the rate of secretion failed to return promptly to background levels following cessation of stimulation. Decline of the augmented secretion to 50 % of maximum required a minimum of 3 min (e.g. Fig. 2A) and more typically 5–7 min (e.g. Fig. 3B). Observations of prompt inhibition of secretion by Mn2+ (to the spontaneous level within 2 min) to be presented below (Fig. 3A) rule out the possibility that the sustained tail of secretion resulted only from delayed diffusion from tissue spaces such as the internal hemolymph sinuses of the sinus gland.
The one preparation (VI in Tables 1, 2) in which electrical responses could not be recorded from the sinus gland in response to axonal stimulation, and in which the nerve was subsequently found to be damaged, provides reassurance that electrical stimulation was not causing release by some means other than propagated action potentials, such as by tissue breakdown. Bouts of electrical stimulation were delivered to this preparation in the same way as to other preparations, except that the stimulus amplitude was an order of magnitude higher. However, there was only a 23 % increase in the amount of CHH released to the perifusate (P=0.022). The recording electrode sampled a very small proportion of the population of terminals, so stimulation of some fibers distal to the break in the axon track (which was at the tip of the electrode) seems a plausible explanation for this small response.
Two preparations were tested for the effects of tetrodotoxin (TTX) (1 μmol l−1) on electrically elicited, as well as K+-evoked, secretion. In one preparation (Fig. 3; IV in Tables 1, 2), the onset of the TTX block of propagated action potentials was monitored by observing the response recorded from the sinus gland to a single axonal stimulus given every 5 s (Fig. 3A). Complete disappearance of the response occurred within 15 s after the first noticeable change, at the predicted time of arrival of the toxin in the chamber (during collection of sample 60). There was no change in the unstimulated level of CHH secretion during perifusion with saline containing TTX (P>0.1), therefore permitting the conclusion that spontaneous CHH release does not reflect existing propagated electrical activity. No secretory response to electrical stimulation of the axonal tract was observable during block of the electrical responses by TTX (not illustrated). In each of these preparations, recovery from TTX block of electrical responses had occurred 50 min after restoring normal perifusion and a secretory response to stimulation was again obtained (Fig. 3B).
Six of the electrically responsive preparations (including three of those in Table 1) were tested for possible differences in the effectiveness of patterned, compared with steady, axonal stimulation, in view of the demonstrated facilitation of vasopressin (Dutton and Dyball, 1979) and oxytocin (Bicknell et al. 1982b) secretion from the vertebrate neurohypophysis by patterned stimulation. The same number of stimuli (generally 600) was given over the same period (5 min), but stimuli were delivered continuously at 2 Hz or in trains of different frequencies (5, 10 or 20 Hz) containing 5, 10 or 20 stimuli per train, with intervening rest intervals (Fig. 2A,B). There was no significant difference between the rates of secretion achieved by a particular preparation in response to steady or patterned stimulation (P>0.05). In the only two preparations for which there were significant differences in response to successive bouts of electrical stimulation, the rate of CHH secretion increased, apparently because the bout occurred later during the experiment (Table 1, IV, V). Significantly decreased responses to second or third bouts of electrical stimulation were not seen.
Dependence of secretion on entry of extracellular Ca2+
Seven preparations were examined for the Ca2+-dependence of stimulated CHH secretion. The simplest test, omission of Ca2+ from the perifusate (isosmotically substituted with NaCl), was applied in two preparations. In both cases, introduction of this saline was followed by a gradual increase of spontaneous CHH secretion to approximately double its previous level. When electrical stimulation was tested after 20 min in the Ca2+-deleted saline, terminal responses were absent, and no augmentation of CHH secretion was observed. Upon return to standard perifusate, however, there was little decline in the spontaneous rate of secretion. Similar observations were obtained in a test in another preparation after the addition of 1 mmol l−1 Cd2+ to the perifusate.
Rapid, nearly complete and rapidly reversible block of CHH secretion was observed in three of four preparations perifused with a saline containing 26 mmol l−1 Mn2+ in place of MgCl2 (see discussion of K+-evoked secretion below). As shown in the experiment illustrated in Fig. 2, the introduction of Mn2+ saline (middle panel) resulted in a prompt and significant (P<0.001) decrease in the level of basal release. There was no secretory response when electrical stimulation was given during perifusion with saline containing Mn2+, nor was a propagated electrical response recorded in the sinus gland during perifusion with this saline (Fig. 2C). Electrical responses and secretory responses were again observable after normal perifusion had been resumed (Fig. 2D).
Secretion of CHH in response to elevation of saline K+ concentration
Characteristics of the secretory response
When the perifusate was switched to a saline having 10-fold greater than the normal K+ concentration (113 mmol l−1, [NaCl] reduced), the amounts of CHH present in the perifusate fractions increased in all of the nine preparations tested. The smallest response was a 51 % increase over unstimulated levels; the other preparations showed, at least, a doubling of secretion, but the maximum response was undetermined as amounts exceeded the upper limits of the assay. Table 2 provides data from four of the preparations whose electrical responses were summarized in Table 1. The absolute rates of release during high-K+ stimulation exceeded 120 pg min−1 in all the preparations examined and, in several, saturated the assay at rates approaching 200 pg min−1 (Carcinus CHH equivalents). This rate of release represents approximately 0.02 % min−1 of the total CHH present in a sinus gland. In two preparations, rates were estimated to be greater than 500 pg min−1. When 1 min fractions were collected, the concentration of CHH in the vials began increasing in the third fraction (consistent with the 2 min deadspace of the system, e.g. Fig. 2, lower panel). The maximal amounts of CHH were observed 3–5 min after the beginning of the response.
Five of the preparations provided observations of both electrically evoked and K+-evoked secretion. Four of these are included among those for which data are given in Tables 1, 2. Although the absolute rates of release observed during stimulation were all greater during K+ stimulation than during electrical stimulation, in three of the preparations (including IV and V of Tables 1, 2) the differences were not statistically significant (P>0.05). The damaged nerve in VI and the use of a lower frequency of electrical stimulation in another preparation may account for these preparations showing larger differences.
The augmented secretion of CHH did not decline immediately following return to normal saline perifusion. Although we have not documented the decline in the rate of secretion, we estimate that the most rapid decline we observed (that for the first K+ stimulation, fraction 97, Fig. 2) took 10 min for the augmented secretion to fall to 50 % of maximum, whereas other examples (e.g. the last K+ stimulation in Fig. 2) took more than 20 min.
Eight preparations were challenged with a second perifusion of high-K+ saline, and all but one (preparation III, Table 2) showed responses that were not significantly different from that observed in the earlier challenge. In three of the preparations, the first perifusion of high-K+ saline lasted 45 min or more. During this period, the rates of secretion fluctuated, but none significantly declined. One of these preparations (VI, Table 2) showed increasing (although not statistically significant) rates of secretion when the averages of 10 min periods were compared, with the fractions from the last period having amounts of CHH that exceeded the limits of the assay. The durability of the preparations was demonstrated by one that was held under perifusion overnight and then challenged with high-K+ saline. CHH secretion increased threefold, compared with an approximately twofold increase the previous day, although the absolute rates of secretion were smaller the second day.
Tests for the calcium-dependence of K+-evoked CHH secretion
The unstimulated rate of CHH secretion increased during perifusion with Ca2+-deleted saline. Fig. 4 shows the results of assays from one preparation, commencing after 54 min in Ca2+-deleted saline. When high-K+, Ca2+-deleted saline was perifused, there was no increase in the rate of CHH secretion. When perifusion was switched to high-K+ saline containing the normal concentration of Ca2+, there was an approximately 30 % increase in the rate of CHH secretion.
In three of four preparations (Table 2), Mn2+ blocked K+-evoked secretion. Fig. 2 illustrates observations from one of the preparations (III, Table 2). As mentioned earlier, there was a significant (P<0.001) reduction in the unstimulated rate of release of CHH following the introduction of saline containing 26 mmol l−1 Mn2+. Introduction of high-K+ saline containing Mn2+ (at fraction 92, Fig. 2) produced a small increase in the rate of release, restoring the rate to a level not significantly different from the spontaneous release rate observed before the introduction of the Mn2+ saline. However, there was a rapid, very large (approximately 10-fold) further increase in CHH release (P<0.001) when high-K+ saline (without Mn2+) was perifused. One conclusion supported by this experiment is that the secretory process is not subject to inactivation by membrane depolarization resulting from the 10-fold increase in external K+ concentration.
In two preparations (IV, V, Table 2; see Fig. 3A), high-K+ saline containing Mn2+ was introduced 5 min after beginning perifusion with normal high-K+ saline, that is, at a time when the augmentation of CHH secretion was well under way. In both preparations, within 2 min (the deadspace time), the amount of CHH in the fractions decreased significantly (P<0.001), and in preparation IV (Fig. 3A), returned to the pre-stimulation level. Two minutes after return to high-K+ saline (without Mn2+), the amounts of CHH in the fractions were again at the augmented level observed just prior to the introduction of the Mn2+.
Mn2+ effectively blocks Ca2+-mediated action potentials in sinus gland terminals (Cooke, 1985), so these experiments permit the conclusion that secretion is dependent on entry of extracellular Ca2+. In addition, the observations support the conclusion that membrane depolarization caused by elevated extracellular K+ levels does not inactivate the secretory response. Finally, they demonstrate that the slow decline of augmented secretion following stimulation cannot be attributed to diffusional delays.
Failure of tetrodotoxin to inhibit K+-evoked secretion
If CHH secretion follows voltage-dependent entry of Ca2+, TTX would not be expected to interfere with release evoked by K+-induced depolarization. As reported above, there was no change in the unstimulated release of CHH when saline containing TTX was introduced. High-K+ saline containing TTX was tested on two preparations. It augmented CHH secretion by approximately twofold in one (Fig. 3) and by approximately threefold in the other. In both preparations, tests of the response to elevation of saline [K+] after recovery from TTX perifusion provided a higher rate of secretion (approximately 70 % in the preparation of Fig. 3, P=0.051; 46 % in the other preparation, P=0.067) than in the earlier challenge with high-K+ saline containing TTX (Fig. 3C).
CHH secretion from dissociated sinus gland terminals
Response to elevation of saline K+ concentration
Seven preparations in which dissociated terminals from 6–8 sinus glands were perifused showed two-to threefold increases in the rate of CHH release when the perifusate was switched to high-K+ saline. The assay results from one such experiment are presented in Fig. 5. The absolute amounts of CCH released per sinus gland were less than 50 % of those from the intact preparations, as might be expected given the non-quantitative dissociation of terminals by mechanical teasing. Unlike the previous figures, note that samples were collected every 5 min and that the change of perifusate was performed in advance of changing the collection tube so that the arrival of the new perifusate corresponded to the beginning of fraction collection. The augmented CHH level was thus found in the first fraction to contain high-K+ saline. By contrast, after returning to normal saline, one or more of the 5 min fraction collections exhibited augmented CHH levels. Therefore, isolated terminals stimulated by a high concentration of K+ show a period of continued augmented secretion comparable to that seen following axonal or K+ stimulation of the intact system.
In all three of the preparations in which a second high-K+ challenge was given, there was a clear, although reduced, response to the second stimulus. In one of the experiments, the first exposure to high-K+ saline caused a 150 % increase in CHH secretion that was maintained throughout the 45 min stimulation. The second challenge, after 75 min of perifusion with normal saline, produced a smaller but distinct, response (a 73 % increase).
Effects of omission of Ca2+ from the perifusate
Two preparations of isolated terminals were perifused with Ca2+-deleted saline. Within 30 min of the change from normal to Ca2+-deleted saline, unstimulated CHH release had increased by four-to fivefold. When Ca2+-deleted high-K+ saline was subsequently introduced, there was no marked change in the existing elevated rate of release. However, when a normal Ca2+ concentration was restored during continuing perifusion with high-K+ saline, there was an immediate further increase in the rate of secretion (20 % and 57 % respectively). In one of the two preparations, unstimulated secretion returned to levels observed before testing with Ca2+-deleted salines and, when later challenged with perifusion of high-K+ saline containing the normal concentration of Ca2+, a fivefold increase in the rate of secretion was observed. In the other preparation, unstimulated secretion remained at the level prevailing in Ca2+-deleted saline (i.e. that measured prior to testing with high-K+ saline containing the normal concentration of Ca2+). The subsequent high-K+ challenge resulted in a 16 % increase in the rate of secretion. These observations on the effects of Ca2+-deleted saline on the isolated terminals closely correspond to those made on the intact X-organ–sinus glands reported above.
Discussion
A sensitive and quantitative ELISA has been employed to study the physiologically evoked neurosecretion of crustacean hyperglycemic hormone (CHH), the predominant neuropeptide of the X-organ–sinus gland (XOSG) system of decapods. One of the most significant observations to emerge is that the pattern of CHH secretion, and presumably some of the underlying mechanisms, differs markedly from those previously described for the secretion of another neurohormone of the crab X-organ–sinus gland system, red pigment-concentrating hormone (RPCH) (Cooke et al. 1977; Cooke and Haylett, 1984).
We have now demonstrated, as we had previously for RPCH, that CHH is secreted in response to axonal stimulation that results in action potentials propagated to the terminals of the sinus gland. Both are secreted at high rates in response to depolarization by elevation of saline K+ concentration and both show dependence on external Ca2+. The striking difference in the secretory responses is that, once initiated, whether by axonal stimulation or by elevated [K+], CHH secretion continues for tens of minutes following cessation of stimulation. By contrast, RPCH secretion ceases immediately on stopping electrical stimulation and occurs as a brief peak that declines within 1 min during perifusion of high-K+ saline.
The defining physiological effect of CHH is to maintain and to increase hemolymph glucose levels. Although the response to injected hormone occurs within minutes, examination of CHH levels in the hemolymph of crayfish has suggested a diurnal rhythm in CHH levels, with hormone levels elevated over periods of 2 h or more (Kallen et al. 1990; Keller and Orth, 1990). Increased levels of CHH and hyperglycemia occur when animals are subjected to stressors such as rapid salinity or temperature changes (for a review, see Webster and Keller, 1988). Large amounts of CHH are stored in the sinus gland: up to 10 % of the total protein of the sinus gland in Carcinus maenas (Keller and Wunderer, 1978) and Cardisoma carnifex (Newcomb, 1983). Thus, the role of CHH in elevating glucose levels during periods of stress or high activity and its abundance are consistent with a pattern of sustained release. Hamann (1974) showed that maintenance of resting levels of glucose is dependent on the presence of the sinus gland in crayfish and observed an overall decrease in glucose levels after selective sinus gland ablation. The observation that the glucose level does not fluctuate much under normal conditions, combined with the finding of a relatively short half-life of CHH injected into the hemolymph (R. Keller, unpublished results), also appears to argue for a continuous release of CHH under physiological conditions.
One explanation for the differences between CHH and RPCH secretion may lie in possible differences in the sensitivity of voltage-dependent Ca2+ channels to inactivation by membrane depolarization: those of CHH terminals being resistant, and those of RPCH terminals being susceptible, to such inactivation. Such differences are implied by the observations made in the presence of divalent ions that block Ca2+ channels. In high-K+ saline (which depolarizes the cells), CHH secretion resumes promptly upon removal of the Ca2+ channel blocker (Mn2+ in these experiments). In contrast, secretion of RPCH does not occur upon removal of a blocker, even if the blocker has prevented any previous secretion (Cooke and Haylett, 1984). Ca2+ channels showing little or no voltage inactivation appear to be the only type of Ca2+ channel present in CHH neurons isolated from the Cardisoma X-organs (Meyers et al. 1992; Meyers, 1993). Stuenkel (1985) has previously reported the sustained release and the lack of depolarizing inactivation of secretion of radiolabelled peptide, shown to be predominantly CHH. He also observed the contrasting brief peak and immediate decline of RPCH secretion (as measured by bioassay).
Lack of depolarization-mediated inactivation of voltage-dependent Ca2+ current could explain, and would predict, a longer period of continuing augmented secretion following removal of the high-K+ saline (as contrasted with electrical stimulation), for direct recordings from terminals under similar perifusion changes have shown that, although depolarization occurs in less than 1 min, repolarization of the terminals takes 5–10 min (Cooke and Haylett, 1984; Stuenkel, 1985). The rapidity with which secretion is blocked by the addition of Mn2+ or resumes upon switching to Mn2+-free perifusate provides additional evidence that excludes the possibility that diffusional barriers explain the slow decline in CHH secretion. It also points to an obligatory link between existing entry of extracellular Ca2+ and secretion. It does not support explanations that invoke release of internal Ca2+ following priming by extracellular Ca2+ admission (for a review, see Tsein and Tsein, 1990) or slow restoration of stimulus-augmented internal Ca2+ concentrations to non-secretory levels.
The continuing secretion of CHH after the cessation of axonal stimulation may involve continued entry of Ca2+ through voltage-dependent channels. Relevant to this are observations from intracellular recordings of terminals (Cooke, 1985; Stuenkel, 1985; Nagano and Cooke, 1987), which reveal the initiation or augmentation of ‘spontaneous’ firing in some terminals, often involving depolarized plateau potentials with superimposed impulse bursts, that persisted for tens of minutes following ‘priming’ bouts of axonal stimulation. We did not record spontaneous electrical activity during our experiments, but our small-diameter extracellular pore electrode sampled a very limited population of terminals, possibly as few as 3–4, as judged by the number of discrete deflections observable in response to changes of stimulus intensity. The depolarized plateaux may be more relevant to admission of Ca2+ than the impulse bursts in these terminals. However, there also remains the possibility that axonal stimulation activates a process not directly linked to action potentials.
Our observations of the effect of elevated [K+] on CHH secretion from the isolated terminals are important because they show that the terminals, like the intact system, exhibit a slow decline of secretion following their return to normal saline. Therefore, it is necessary to look at the physiology of the terminals themselves for explanations. The preparation of dissociated terminals, by contrast with the intact sinus gland in which there are hemolymph sinuses, lacks any tissue compartmentalization. The dissociated terminal preparation thus excludes diffusional delays as an explanation for the continuing elevated secretion rates.
The only peptidergic secretory system (that we are aware of), other than the XOSG, in which the decline of peptide levels in the perifusate following the cessation of stimulation has been examined is the neurohypophysis. The analysis in this case indicated a prompt cessation of peptide secretion, because the decline followed the same time course as diffusion of an extracellular marker (Ingram et al. 1982; Bicknell et al. 1984). It is worth noting that, when the amount of peptide in the perifusate was corrected for diffusion, it was found that, under sustained repetitive electrical stimulation, secretion of vasopressin fatigued within 3 min, whereas release of oxytocin was sustained. Secretion of vasopressin from isolated neurohypophyseal endings was found to decline by half within 45 s, despite the continued presence of an elevated K+ concentration (Stuenkel and Nordmann, 1993). Although there may be relatively sustained postsynaptic responses or changes in the behavior of a central pattern generator involving peptidergic transmission, these seem to be attributable to the mechanisms underlying the responses rather than to a continuing tail of peptide secretion, because they can occur in response to brief applications of the appropriate peptide (e.g. Nusbaum and Marder, 1988; Turrigiano and Selverston, 1989; Peng and Horn, 1991).
The independence of the secretory mechanism from transmembrane Na+ currents, except for their role in depolarizing the membrane and thus activating Ca2+ currents, has been accepted since the demonstration of secretion from isolated neurohypophyses in Na+-deficient saline (Douglas and Sorimachi, 1971). Thus, it was expected that the presence of TTX would not inhibit K+-evoked secretion of CHH, as was observed. TTX did not reduce the unstimulated rate of secretion of CHH, indicating that this release is not the result of ‘spontaneous’ action potentials propagating to the terminals. It does not exclude the possibility of spontaneous Ca2+-mediated regenerative potentials, which the terminals have been shown to support in saline containing TTX (Cooke, 1985; Nagano and Cooke, 1987). By contrast with TTX, addition of saline containing Mn2+, which blocks Ca2+-mediated terminal responses, reduced the unstimulated release rate. Our data on K+-stimulated secretion of CHH in the presence of TTX contrast with those on RPCH release in that they show no evidence of the enhancement of peak secretion rates seen in the presence of TTX for release of RPCH, an effect also seen in saline having a reduced external [Na+] (Cooke and Haylett, 1984). In each of the CHH preparations, a subsequent test of K+-evoked secretion in the absence of TTX showed a higher secretion rate (although not quite statistically significant) than in the earlier test in its presence. These observations again point to possible differences in the mechanisms of secretion of the different peptide hormones.
ACKNOWLEDGEMENTS
We thank Susan Grau and Jana Watanabe for technical assistance and Carol Kosaki for preparation of the manuscript. This work was supported by NSF grant BNS 89-10432 to I.C., by the Ida Russell Cades Fund of the University of Hawaii Foundation and by a travel grant from the Deutsche Forschungsgemeinschaft (Ke 206/7-8) to R.K.