INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ACTIVATION OF THE salt-secreting rectal gland of the spiny dogfish, Squalus acanthias, is thought to occur chiefly via a cardiac natriuretic peptide (26) native to the shark (sCNP; Refs. 17, 32) secreted into the bloodstream in response to an increase in central blood volume (28, 29). The mode of action of CNP on the rectal gland is not yet fully understood and appears to be complex. Both the related mammalian hormone, atrial natriuretic peptide (ANP), and extracts of native shark heart liberate vasoactive intestinal peptide (VIP) from nerve terminals within the gland (22), stimulating secretion via a cascade involving adenyl cyclase and cAMP (30). However, CNP also induces secretion in preparations of isolated tubules and cultured cells that are devoid of neural elements, probably by a pathway involving guanylyl cyclase (7). The present experiments were designed to clarify the mechanisms by which CNP induces salt secretion in the shark rectal gland.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dogfish of either sex were taken by gill nets from Frenchman's Bay, ME, and kept in marine live cars until used, usually within 3 days of capture. The dogfish were pithed, and their rectal glands were removed via an abdominal incision.

Rectal gland perfusions. The rectal gland artery, vein, and duct were catheterized with PE-90 tubing. The glands were placed in a glass perfusion chamber maintained at ~15°C with running seawater and perfused by gravity at a pressure of 40 mmHg. The composition of the perfusion media was (in mM) 280 Na, 290 Cl, 5 K, 8 bicarbonate, 1 phosphate, 2.5 Ca, 3 Mg, 0.5 sulfate, 350 urea, and 5 glucose; pH 7.6 when gassed with 99% O₂-1% CO₂. Rectal gland secretion was collected over 10-min intervals in 1.5-ml tared conical centrifuge tubes or calibrated glass pipettes when volume was <200 µl/10 min. In some experiments, peptides were infused directly into the arterial catheter through a rubber sealed port, in close proximity to the gland, over an interval of 1 min in an amount calculated to provide the desired final concentrations. When more than one concentration of peptide was used, they were added in order of increasing concentration and the secretion of chloride was allowed to return to basal levels before each subsequent addition. In other experiments, CNP or VIP was infused continuously for a period of 30 min after a 30-min control period. In perfusions where 10 mM procaine or 10 nM staurosporine were used, they were added to the perfusate solution before the start of the perfusion at the beginning of the experiment.

Confluent monolayers of rectal gland cells. Primary cultures of rectal gland cells were prepared by the protocol developed by Valentich and Forrest (33). Under sterile conditions, the gland was minced and then digested in a collagenase solution (0.2% in 30 ml of sterile shark Ringer) at room temperature (25°C) for 1 h under constant agitation. At the end of the incubation, the suspension was washed with sterile shark Ringer and the tubules were harvested and kept on ice. The process was repeated several times. The accumulated tubules were then washed and suspended in 20 ml of sterile shark Ringer. The tubules were then suspended in medium containing equal parts of DMEM and Ham's F-12, supplemented with 5% Nu-serum; 1% insulin, transferrin, and selenium; and Pen/Strep (100 U/ml-100 µg/ml). The tubules were seeded in 35-mm Petri dishes onto 18-mm disks made of 150-µm nylon mesh and coated with collagen (Collagen Biomedical). The tubules were allowed to grow to confluence, usually in 10-15 days, and the nylon mesh covered with cells was removed from the Petri dish and mounted in a Ussing chamber. Both the apical and basolateral sides of the monolayer are bathed with shark Ringer containing 5 mM glucose at pH 7.6. Measurements of short-circuit current (I_sc), transepithelial voltage, and transepithelial resistance are made using standard electrophysiological methods (33). The transepithelial voltage and I_sc were allowed to reach a steady state, usually within the first 30 min of the experiment. Human CNP (hCNP) was then added to the basolateral side of the chamber, and transepithelial voltage and I_sc were recorded until it reached its peak effect.

Assay for cGMP. For the measurement of cGMP in rectal gland tissue, the glands were frozen immediately in liquid nitrogen and subsequently stored at 80°C until assayed. The tissue was homogenized in 12% trichloroacetic acid (final concentration). The tissue samples were centrifuged at 2,300 g for 15 min, and the supernatant was removed, extracted with 10 vol of water-saturated ether four times, dried, and reconstituted in 50 mM sodium acetate buffer (pH 6.2) for measurement of cGMP content. For the assay of cGMP in venous effluent, the venous effluent was collected in test tubes on ice in 12% trichloroacetic acid (final concentration). The samples were extracted with 10 vol of water-saturated ether four times, dried, and reconstituted in 50 mM sodium acetate buffer (pH 6.2), as described for the tissue samples above. Samples with and without added peptide were assayed simultaneously. The concentration of cGMP was measured using a radioimmunoassay kit obtained from Biomedical Technologies (Stoughton, MA).

VIP, procaine, phorbol 12-myristate 3-acetate (TPA) and inactive TPA, oleyl acetyl glycerol, and zaprinast (MB22948) were purchased from Sigma. hCNP and human ANP (hANP) were purchased from Peninsula Laboratories. sCNP was provided by California Biotechnologies. Staurosporine was purchased from Calbiochem.

Chloride was measured by amperometric titration using a Buchler-Cotlove chloridometer. Chloride secretion is expressed as microequivalents per hour per gram wet weight. All values are means ± SE. The dose-response curves of chloride secretion to stimulation with CNP shown in Figs. 1 and 2 were analyzed using a mixed-models analysis of variance (9), which takes into account the relationships among multiple measurements in a gland. We used an autoregressive correlation matrix to analyze the effects of successive concentrations of CNP on the glands. CNP concentrations were tested on a log (base 10) scale; concentration 0 was coded one step below the lowest experimental concentration. The chloride secretion response was modeled in both the actual and log (base 10) scale; results were consistent and are presented in the actual scale. The models included factors for the type of CNP administered to the gland (human vs. shark), treatment condition (control or procaine 10 mM), linear and quadratic terms for CNP concentration, and all interactions. Nonsignificant terms were removed from the model. The post hoc comparisons between control and procaine experiments were done using contrasts for model parameters. All tests consider P < 0.05 as statistically significant. All statistical computations were done using SAS, Version 6.12 (SAS Institute, Cary, NC). Statistical significance of the data presented in Figs. 3, 5, 6, 7, and 9 was determined using analysis of variance with repeated measures. The post hoc tests used were the Bonferroni/Dunn test when the numbers were equal and the Student-Newman-Keuls test when there was a difference in the number of measurements. Paired t-test was used to determine the statistical significance of the data shown in Fig. 4 and throughout the report wherever appropriate.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of procaine on CNP stimulation of chloride secretion by isolated perfused rectal glands. Perfusion of isolated rectal glands with 10 mM procaine completely prevents chloride secretion induced by veratrine or ANP, but not by VIP (31). This concentration of procaine also prevents the release of VIP into the venous effluent of the gland, suggesting that both veratrine and ANP stimulate glandular secretion by depolarizing intrinsic VIPergic nerves (31). In contrast, chloride secretion stimulated by 1-min bolus infusions of sCNP and hCNP was reduced, but not eliminated, by exposure to procaine (Figs. 1 and 2).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effect of procaine on response of chloride secretion to 1-min bolus infusions of human C-type natriuretic peptide (hCNP). , Glands perfused with procaine (10 mM); , glands without procaine. hCNP stimulated secretion of chloride in a dose-response way both in presence and absence of procaine. Effect was significant in both conditions (P < 0.0001), and procaine reduced response to hCNP (P < 0.0001, by mixed-models analysis of variance). Values are means ± SE. Number of glands studied was 86: 48 without procaine and 38 with procaine. * P < 0.001 by post hoc comparison using contrasts for model parameters.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Effect of procaine on response of chloride secretion to 1-min bolus infusions of shark CNP (sCNP). , Glands perfused with procaine (10 mM); , glands without procaine. sCNP stimulated secretion of chloride in a dose-response way both in presence and absence of procaine. Effect was significant in both conditions (P < 0.0001), and procaine reduced response to hCNP (P < 0.0001, by mixed-models analysis of variance). Values are means ± SE. Number of glands studied was 60: 31 without procaine and 29 with procaine. * P < 0.001 by post hoc comparison using contrasts for model parameters.

Of the 146 glands studied in these experiments, 86 (59%) were stimulated with hCNP and 60 (41%) with sCNP; a total of 67 (46%) glands was perfused with procaine, 38 (57%) with hCNP, and 29 (43%) with sCNP; 13 of the 146 glands had three measurements (9 hCNP/procaine and 4 sCNP/control) and the rest had two measurements. The statistical model examining the effect of procaine on the response of chloride secretion to stimulation with hCNP includes the effect of treatment condition, linear effect of CNP concentration and its interaction with treatment, and a quadratic effect of CNP concentration (each P < 0.0001). The slope of the concentration-response curve was reduced when the glands were perfused with procaine (P < 0.0001). The statistical model examining the effect of procaine on the response of chloride secretion to stimulation with sCNP also includes the effect of treatment condition, linear effect of CNP concentration and its interaction with treatment (each P < 0.0001), and a quadratic effect of CNP concentration (P < 0.03). The slope of the concentration-response curve was reduced when the glands were perfused with procaine (P < 0.0001). The combined model includes a quadratic relationship between the concentration and response that is the same both for the types of CNP and treatment conditions. The linear term differs between both treatments and CNP types: it is larger for control vs. procaine measurements, but this difference is less pronounced with hCNP. The slopes of the concentration-response curves when the glands are perfused with procaine were reduced with sCNP and with hCNP (each P < 0.0001). As previously reported (31), the stimulatory action of ANP was completely prevented by procaine (data not shown).

To ascertain that procaine did not adversely affect the rectal glands, in eight of the experiments with hCNP, after the secretion of chloride had stabilized at a new constant level, a bolus of 0.15 µM VIP was infused into the rectal gland arteries in the same manner as the infusion of hCNP while the glands were continuously perfused with procaine. In these glands VIP increased the secretion of chloride from 343 ± 64 to 1,697 ± 98 meq · h¹ · g¹ (P < 0.0001 by paired t-test). This response to VIP is comparable to that evoked in the absence of procaine (30). This experiment shows that procaine does not alter the capacity of the gland to respond to VIP in contrast to its inhibition of the response to CNP.

To make certain that the difference in susceptibility to procaine between CNP and ANP was not merely a consequence of more powerful stimulation by CNP, additional experiments were carried out in which the dose of CNP was adjusted to achieve lower rates of chloride secretion equivalent to those achieved by ANP. Again, procaine completely inhibited the stimulatory action of ANP from 427 ± 93, n = 6, without procaine, to 1.6 ± 16.6 µeq · h¹ · g¹, n = 10, with procaine, while only partially reducing stimulation by CNP, from 329 ± 114, n = 7, without procaine, to 156 ± 34 µeq · h¹ · g¹, n = 11, with procaine (Fig. 3; P < 0.01 by analysis of variance). These results suggested a non-neural action of CNP different from that of ANP and presumably directly on the rectal gland cells.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Comparison of effect of 10 mM procaine on stimulation of chloride secretion by hANP and hCNP. Left bars (labeled hANP), procaine completely suppressed stimulatory effect of hANP (* P < 0.001). Right bars (labeled hCNP), procaine reduced but did not prevent stimulatory effect of hCNP (** P < 0.01). In presence of procaine, hCNP is able to stimulate chloride secretion (# P < 0.001). Columns are means ± SE. Number of observations for hANP experiments is 6 without procaine and 10 with procaine; and for hCNP, 7 and 11, respectively. Statistical analysis was performed by analysis of variance.

Effect of CNP on a rectal gland preparation lacking neural elements. Additional evidence for a direct effect of CNP on the rectal gland was obtained from studies of confluent monolayers of primary cultures of rectal gland cells, which are devoid of neural elements (33). In this preparation, CNP also increased transport. CNP was added at a final concentration of 10⁸ M to the basolateral side of the preparation. I_sc, an indirect marker of chloride transport in this preparation, increased almost sixfold from 4.8 ± 1.6 to 27.0 ± 7.8 µA/cm² (P < 0.01 by paired t-test) in 12 monolayers from six different cell culture preparations (Fig. 4). The rise in I_sc started ~2 min after the addition of CNP, reached its maximum in 10 min, and thereafter declined very slowly without returning to baseline values. Similar results with killifish CNP (11) and sCNP (10) have been obtained by Karnaky and co-workers.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Effect of CNP on short-circuit current (Isc) in cultured rectal gland cells. CNP induced a 6-fold increase in Isc in confluent monolayers of cultured rectal gland cells. Values are means ± SE, n = 12. * P < 0.01 by paired t-test.

Effect of staurosporine on continuous CNP stimulation of chloride secretion by isolated perfused rectal glands. To examine a possible role of the inositol phosphate pathway in mediating the direct action of CNP, fresh isolated glands were perfused with 10 nM staurosporine, 10 mM procaine, neither, or both. This concentration of staurosporine is one order of magnitude greater than the I₅₀ for protein kinase C (2). Thirty minutes after the perfusion was begun, a continuous infusion of 10 nM CNP was begun. In glands not perfused with procaine, staurosporine reduced stimulation by CNP by ~50%, from 1,617 ± 298 (n = 8) to 839 ± 207 µeq · h¹ · g¹ (n = 6; P < 0.0001 by analysis of variance). The combination of staurosporine and procaine blocked >90% of chloride secretion induced by CNP, reducing the secretion of chloride to 237 ± 61 µeq · h¹ · g¹ (n = 6; P < 0.0001 by analysis of variance). Thus procaine-insensitive chloride transport in perfused rectal glands was virtually completely inhibited by staurosporine.

As expected, 10 nM staurosporine did not inhibit stimulation of the rectal gland by VIP (1.5 nM), which is known to be mediated by the adenyl cyclase cascade (Fig. 5). Separate control experiments with VIP in the absence of staurosporine produced an identical response in chloride secretion (data not shown). To ascertain that the glands perfused with procaine and staurosporine were able to increase their secretion of chloride in response to secretagogues, a bolus of VIP, calculated to give a final concentration of 0.15 µM, was infused into the rectal gland artery at the 60-min point after CNP had failed to increase the secretion of chloride while the glands were still perfused with procaine and staurosporine. After the infusion of VIP, the secretion of chloride rose from 207 ± 50 to 1,505 ± 235 µeq · h¹ · g¹ (n = 6; P < 0.001 by paired t-test).

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Effect of staurosporine and procaine on CNP-stimulated chloride secretion. Either 108 M staurosporine or 102 M procaine (infused from start of the perfusion) reduced, by ~50%, stimulation of chloride secretion induced by CNP. A CNP infusion was started at time indicated by arrow. Combination of staurosporine and procaine blocked >90% of CNP-induced stimulation. Thus staurosporine reduced completely stimulation of chloride secretion attributed to direct effects of CNP. By contrast, staurosporine (108 M) failed to inhibit vasoactive intestinal peptide (VIP)-induced chloride secretion, which is mediated by adenyl cyclase, cAMP, and protein kinase A. Values are means ± SE. Number of experiments of each type is shown in parentheses. All comparisons were by analysis of variance with repeated measures. * P < 0.01 for comparisons to CNP alone. ** P < 0.01 for comparisons to CNP plus procaine. *** P < 0.01 for comparisons to CNP plus staurosporine.

These results suggest that protein kinase C mediates, at least in part, the direct action of CNP, which is insensitive to procaine, on shark rectal glands.

Effect of protein kinase C activators on chloride secretion. To stimulate protein kinase C, we added TPA (1 µM) to perfusions of isolated rectal glands either alone or in combination with ionomycin (100 nM) in 10 experiments. In four additional experiments, perfusions were carried out with oleyl acetyl glycerol (100 µM) as an alternative way to activate protein kinase C. No increase in chloride secretion was observed under these circumstances.

Effect of CNP on cGMP efflux. To determine whether CNP stimulated the production of cGMP in intact cells, we perfused isolated rectal glands with CNP and measured cGMP in the venous effluent. The effect of CNP (10 nM) on the release of cGMP into the venous effluent and on chloride secretion is shown in Fig. 6. There is an initial rapid rise in the release of cGMP into the venous effluent during the first 10 min of perfusion with CNP from 12 ± 3.7 to 40 ± 14 pmol/100 µl (P < 0.01 by paired t-test). This rise is followed by an equally rapid decline, with levels returning to 13 ± 4 10 min later and 20 ± 1 at 30 min. The secretion of chloride did not peak until after 20 min of perfusion with CNP, when it rose from a baseline of 127.0 ± 34 to a peak of 1,617 ± 279 µeq · h¹ · g¹ (P < 0.0001 by paired t-test) in a pattern different from that of the release of cGMP.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Effect of hCNP on release of cGMP in venous effluent and chloride secretion in isolated perfused rectal glands. Duration of each collection period was 10 min. First collection period is control without hCNP. hCNP (108 M) was infused during periods 2, 3, and 4. Initial rapid rise in release of cGMP is followed by an equally rapid decline with levels returning to or slightly above control values, a pattern different from that of chloride secretion, which peaks 10 min later. Values are means ± SE, n = 8. * P < 0.05, ** P < 0.0001, by analysis of variance.

Failure of 8-bromoadenosine-cGMP alone to stimulate rectal gland secretion. sCNP releases cGMP into the venous effluent of the gland, presumably because it activates guanylyl cyclase in plasma membranes of rectal gland cells (7). To determine whether cGMP caused an increase in the secretion of chloride, rectal glands were perfused with the permeant and phosphodiesterase-resistant analog of cGMP, 8-bromoadenosine-cGMP (8-BrcGMP; n = 10). In none of these experiments did 8-BrcGMP have any effect on chloride secretion (Fig. 7). A bolus of 1 µg of VIP, final concentration 1.6 × 10⁷ M, was injected into the arteries of the glands to make certain that they were able to respond to this secretagogue. All glands responded with increased chloride secretion to the bolus of VIP.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   A phosphodiesterase resistant analog of cGMP, 8-bromoadenosine-cGMP (8-BrcGMP), does not stimulate secretion of chloride. After 30 min of perfusion, 10 isolated rectal glands received 8-BrcGMP (104 M; arrow). Perfusion with 8-BrcGMP was continued for remainder of experiment (). 8-BrcGMP did not stimulate chloride secretion. , Control perfusions; , effect of hCNP (108 M) added at 30 min. CNP was added at time indicated by arrow. Values are means ± SE. * P < 0.0001 by analysis of variance.

To confirm that intracellular levels of cGMP were indeed elevated by the exogenous administration of 8-BrcGMP, we measured the cGMP content of the perfused gland 10 min after the infusion of 8-BrcGMP was interrupted, when the cGMP content of the venous effluent was negligible and the extracellular concentration of cGMP within the gland could reasonably be neglected. Tissue levels of cGMP found after perfusion with 8-BrcGMP (which failed to stimulate secretion) exceeded those found after administration of CNP (which greatly stimulated secretion; Fig. 8)

View larger version (15K): [in this window] [in a new window]   Fig. 8.   Effect of 8-BrcGMP and hCNP on intracellular levels of cGMP. Tissue levels of cGMP were measured by radioimmunoassay in isolated rectal glands perfused with either 104M 8-BrcGMP or 108M hCNP. Both 8-BrcGMP and hCNP increased dramatically intracellular levels of cGMP; control value of 99 ± 40 fmol/mg protein does not show in figure. 8-BrcGMP caused a 4-fold larger increase in tissue levels than hCNP. Values are means ± SE. * P < 0.01 for control vs. 8-BrcGMP. ** P < 0.01 for control vs. hCNP; n = 13 for control, 7 for 8-BrcGMP, and 8 for hCNP

In additional experiments, an inhibitor of cGMP phosphodiesterase, zaprinast (100 µM), was added to the solution perfusing isolated rectal glands before stimulation with 100 µM 8-BrcGMP was attempted to diminish intracellular hydrolysis of cGMP. Again, no stimulation by 8-BrcGMP was detected in any experiment. Thus, although glandular guanylyl cyclase is directly stimulated by CNP, a high intracellular level of cGMP alone did not appear sufficient to induce chloride secretion in perfused glands.

Synergistic effect of cGMP and protein kinase C activation in isolated perfused rectal glands. When glands were perfused with 1 µM TPA in combination with 100 µM cGMP and 100 µM zaprinast, chloride secretion increased substantially from 93.8 ± 30.3 to 505.9 ± 61.1 µeq · h¹ · g¹ (n = 4, P < 0.01), a level similar to that obtained with CNP in the presence of procaine. The substitution of an inactive phorbol ester in these experiments eliminated the synergistic effect so that no increase in secretion occurred (Fig. 9).

View larger version (25K): [in this window] [in a new window]   Fig. 9.   Effect of phorbol ester, 8-BrcGMP, and zaprinast on chloride secretion. After 30 min of control collections, rectal glands were perfused with 106 M phorbol 12-myristate, 13-acetate (TPA), 104 M 8-BrcGMP, and 104 M zaprinast (, n = 4) or with an inactive analog of TPA substituted for active form (, n = 4). Combination of TPA, 8-BrcGMP, and zaprinast stimulated chloride secretion. Magnitude of stimulation is ~40% of that seen with CNP alone (compare to peak values at 50 min in Fig. 4) and similar to that observed during perfusion with CNP and procaine (, n = 6). Perfusion with 8-BrcGMP, zaprinast, and inactive form of phorbol ester failed to stimulate chloride secretion. TPA, inactive TPA, and CNP were added at time indicated by arrow. Values are means ± SE. * P < 0.05 by analysis of variance.

These results suggest that the direct effect of CNP on intact rectal glands requires the simultaneous activation of guanylyl cyclase and protein kinase C.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Control of secretion by the salt gland of the shark appears to be extraordinarily complex, involving several types of receptors and second messengers. The gland can be stimulated via adenyl cyclase by VIP and high concentrations of adenosine (5, 6, 30), by cardiac peptides that activate guanylyl cyclase (26, 27), and by genistein, a nonspecific inhibitor of tyrosine kinases (15). Chloride secretion is inhibited by several peptide hormones native to the shark, including somatostatin (21), bombesin (19), NPY (18), and epidermal growth factor (unpublished observations), as well as by purine nucleotides (20) and low concentrations of adenosine (<10 µM; Ref. 13). Because the fine adjustment of salt secretion by the rectal gland is critical to homeostasis of the volume and composition of the shark's body fluids, it is not surprising that its modulation should be complicated and include parallel and overlapping pathways. In considering the relative importance of these influences, a case can be made for placing primary reliance on the behavior of the intact organ rather than of isolated or cultured cells, in which the number, location, or reactivity of receptors may be altered.

The present experiments illustrate some of these principles. Earlier reports employing mammalian ANP suggested that induction of rectal gland secretion by this cardiac natriuretic hormone could be entirely accounted for by stimulation of rectal gland nerves to release their neurotransmitter, VIP, because the effect of ANP was completely inhibited by the local anesthetic procaine and other maneuvers that block neurotransmitter release (22). Although ANP evoked an I_sc when applied to monolayer cultures of rectal gland cells (12), no stimulation could be observed in isolated perfused tubules (22) or dispersed tubules freshly prepared by collagenase treatment of whole glands (22). In the present experiments, on the other hand, when C-type cardiac peptides resembling those native to the shark were infused into isolated glands, only about half of the stimulatory effect was blocked by procaine. The remaining moiety of chloride secretion, insensitive to procaine, presumably represents a direct effect of CNP on rectal gland cells, and such an effect can be demonstrated in both freshly isolated tubules (23-25) and, as in the present experiments, cultured cells (10, 11). It is interesting that in fresh whole perfused glands, equimolar concentrations of sCNP and VIP have roughly similar effects on chloride secretion (26), whereas, in dispersed tubules, sCNP produces less than half of the respiratory stimulation evoked by equimolar amounts of VIP (23-25). This suggests that in isolated tubules devoid of neural elements, a portion of the action of sCNP apparent in the intact gland is no longer present.

The effect of natriuretic peptides in most cells is thought to be mediated by cGMP, because these peptides, including CNP, stimulate guanylyl cyclase (7, 12, 14). In plasma membranes of shark rectal gland both sCNP and hCNP activate guanylyl cyclase in a dose-dependent fashion and are equipotent with a Michaelis constant of ~10 nM (7). When applied to primary cultures of rectal gland cells, 1 mM 8-BrcGMP elicits a slow increase in I_sc of small magnitude (12). However, high external concentrations (0.1 and 1 mM) of 8-BrcGMP failed in the present experiments to stimulate perfused rectal glands. Furthermore, ANP has been noted to stimulate the production of cGMP in isolated rectal gland tubules without increasing oxygen consumption (14). Similar dissociation between the physiological effects of natriuretic peptides and those of cGMP have been observed in other tissues. In the adrenal, the effect of ANP is not reproduced by 8-BrcGMP even though ANP stimulates the production of cGMP in adrenal cells (1). In aortic smooth muscle, the vasorelaxant potency of ANP and analogs can be dissociated from their capacity to stimulate the production of cGMP (3). These considerations raised the possibility that other cell transduction systems such as the inositol phosphate pathway (4) might be responsible in part for the effect of CNP on shark rectal gland. Accordingly we tested the effect of staurosporine (10 nM) on CNP stimulation in perfused glands. At this concentration, staurosporine is said to exert a strong inhibitory effect on protein kinase C but not on protein kinase A, protein kinase G, and other serine/threonine kinases (16). Staurosporine completely inhibited the procaine-resistant moiety of CNP-induced stimulation but did not affect stimulation by VIP (presumably mediated by protein kinase A). These results strongly suggest that the direct action of CNP on intact rectal glands is mediated, at least in part, via protein kinase C.

Activation of protein kinase C appears to be necessary but not sufficient for CNP stimulation of salt secretion in this system, and the same may be true for guanylyl cyclase. Phorbol ester, a well-established activator of protein kinase C, did not stimulate salt secretion when infused, nor did high concentrations (1 mM) of 8-BrcGMP, even when combined with an inhibitor of cGMP phosphodiesterases. The combination of protein kinase C activation and cGMP infusion, however, proved synergistic, stimulating chloride secretion to a level similar to that observed when CNP was infused into glands blocked by procaine. It seems reasonable to hypothesize, therefore, that the direct effect of CNP on rectal gland secretion requires and entrains activation of both protein kinase C and guanylyl cyclase. The complexity of this pattern is reminiscent of that which apparently characterizes the action of natriuretic peptides in other systems, e.g., proximal tubular cells of the human kidneys (8).

Perspectives

View larger version (17K): [in this window] [in a new window]   Fig. 10.   Schema of effect of CNP on secretion of chloride in dogfish rectal gland. Volume expansion causes release of CNP from heart. CNP binds to a guanylate cyclase-linked B-type receptor and activates guanylyl cyclase with the production of cGMP. Increase in cGMP is not itself sufficient to produce an increase in chloride transport. A parallel stimulation of protein kinase C (PKc) is probably mediated by phospholipase C and the phosphoinositide pathway producing a synergistic effect on chloride transport. CNP also has an indirect effect of stimulating chloride secretion by eliciting release of VIP from nerves within rectal gland. VIP stimulates adenyl cyclase, thus increasing cAMP, which activates apical chloride channel, homologous to human cystic fibrosis transmembrane conductance regulator (CFTR). PKa, PKg, protein kinase A and G, respectively.